f Hy mong

iy
i
'

BULLETIN

OF THE

MUSEUM OF COMPARATIVE ZOOLOGY

AT

HARVARD COLLEGE, IN CAMBRIDGE

VOLES 122

CAMBRIDGE, MASS., U.S. A.
1959 - 1960

Tue Cosmos Press, INC.
CamBripGE, Mass., U.S.A.

CONTENTS

PAGE

No. 1—SKELETON AND MuscULATURE OF THE HEAD OF
GELASTOCORIS OCULATUS (FABRICIUS) (HEMIPTERA-
HETEROPTERA). By Margaret C. Parsons. Decem-
ber, 1959 . ; : : . : : : 1

No. 2.—A PRELIMINARY REVIEW OF THE FAMILY GONOSTO-
MATIDAE, WITH A KEY TO THE GENERA AND THE
DESCRIPTION OF A NEW SPECIES FROM THE TROPICAL
Paciric. By Marion Grey. February, 1960 . 2 oo

No. 3.—TRICHECODON HUXLEY! (MAMMALIA: ODOBENIDAE)
IN THE PLEISTOCENE OF SOUTHEASTERN UNITED
States. By Clayton E. Ray. (2 plates.) March,
1960 : : : : : 5) PAT

No. 4.—CoNTRIBUTIONS TOWARD A RECLASSIFICATION OF THE
Formiciwage. III. Trrse AMBLYOPONINI (HYMEN-
opTERA). By William L. Brown, Jr. March, 1960 143

No. 5.—LAND SHELLS or Navassa IsLAND, WEST INDIES.
By Ruth D. Turner. (7 plates.) March, 1960 . = 2a

No. 6.—A TRANSCRIPTION OF DARWIN’S First NOTEBOOK ON
‘“TRANSMUTATION OF SPEcIES’’. By Paul H. Bar-
Retin Aprile L960 2 : ; ; : ; 7 245

No. 7.—SKELETON AND MUSCULATURE OF THE THORAX OF
GELASTOCORIS OCULATUS (FABRICIUS) (HEMIPTERA-
HetTeROPTERA). By Margaret C. Parsons. May, 1960 297

No. 8.—THE PALATINE PROCESS OF THE PREMAXILLA IN THE
PassERES. A STUDY OF THE VARIATION, FUNCTION,
EVOLUTION AND TAXONOMIC VALUE OF A SINGLE CHAR-
ACTER THROUGHOUT AN AVIAN ORDER. By Walter J.
Bock. June, 1960. : ‘ : ; 3 o09

No. 9.—THE SNAKES OF Ecuapor. A CHECK LIST AND KEY.
By James A. Peters, June, 1960 . : : . 489

ky allt jeere e |
~koert EEE) Brieeenics
“eed, | Othe >>
7 Ne
fh rom ‘ +
, thes yaa
aes'h’ aye

ie ay

Py)

Bulletin of the Museum of Comparative Zoology
AGT) HVACR VirAeReD) 7 CO ihiGeH

MOE, s L220... NO.s 1

SKELETON AND MUSCULATURE OF THE HEAD OF
GELASTOCORIS OCULATUS (FABRICIUS)
(HEMIPTERA-HETEROPTERA)

By Maracaret C. PARSONS
Harvard Biological Laboratories

CAMBRIDGE, MASS., U.S.A.
PA ND Oe) eS NMinwas: Hew
DECEMBER, 1959

PUBLICATIONS ISSUED BY OR IN CONNECTION
WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

BULLETIN (octavo) 1863 — The current volume is Vol. 121.

BREVIORA (octavo) 1952 — No. 115 is current.

Memoirs (quarto) 1864-1938 — Publication was terminated with
Vol. 5p:

JOHNSONIA (quarto) 1941 — A publcation of the Department of
Mollusks. Vol. 3, no. 39 is current.

OccASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo)
1945 — Vol. 2, no. 23 is current.

PROCEEDINGS OF THE NEW ENGLAND ZooLoGicaAL CLUB (octavo)
1899-1948 — Published in connection with the Museum. Publication
terminated with Vol. 24.

The continuing publications are issued at irregular interva!s in num-
bers which may be purchased separately. Prices and lists may be
obtained on application to the Director of the Museum of Comparative
Zoology, Cambridge 38, Massachusetts.

Of the Peters ‘‘Cheek List of Birds of the World,’’ volumes 1-3 are
out of print; volumes 4 and 6 may be obtained from the Harvard Uni-
versity Press; volumes 5 and 7 are sold by the Museum, and future
volumes will be published under Museum auspices.

Bulletin of the Museum of Comparative Zoology
ACT SHAR VAR De COL bGr
Vou; -122.No:.1

SKELETON AND MUSCULATURE OF THE HEAD OF
GELASTOCORIS OCULATUS (FABRICIUS)
(HEMIPTERA-HETEROPTERA)

By MARGARET C. PARSONS
Harvard Biological Laboratories

CAMBRIDGE, MASS., U.S.A.
PRN DED HOR THE MUSEUM
DECEMBER, 1959

A)? ye

m ear

tj il i] ’
t na NB , iA Au Nae =D im ii itt i
oA Wea ih i ii

Bin Te

AO aa
1 L Wie

eA)

ial opie!

Sree pn hy iY 28
i i

7

1
0
ie

ig
paE

tal
PP hit

No. 1 — Skeleton and Musculature of the Head of Gelastocoris
oculatus (Fabricius) (Hemiptera-Heteroptera)

By MARGARET C. PARSONS

Harvard Biological Laboratories

TABLE OF CONTENTS

Page
MELO GUCELO MY PEAT oN airs MERI MCR rR TURN DE cei Tiny ey ena T ECD an tla CRW Arua MIC Ie esti eM AnaCae 3
Materialssand Methods icerccineg ira shes Acne | Sede epee ea eee 5
extern ale rAm a tomiye ile yeueaielemsbys euler mural ccmslitich 5 cate Ua yeep ON) ce Me) ATU ala ge 6
Internal Anatomy
Generalustructre suisse cseeie ee se ae ee ae PVA DE apa eh pre tt 11
Food pump, lora, and epipharyngeal plate ........................ 13
Salivaryspuim py yeoman caer ies cue Uinta Cael Nl vet me cr ube Ae Aaa eR OE 19
Hypopharyngeal wings, suspensory plate, and genal sacs ........... 19
Mandibular lever and) stylet 0s. ee ee 94
Maxillary wevervandastyleti smc sa toe) satis isis) ante RUN tlw ease ti ue Vast: 26
Babium and el aorumyye ose ye Nyy eRe ERO eons waa eel elt 27
Muscles}ofithe thea discs Sith. es Mne spay fal MEUM OP Sa aay ian ae Pee ee ERLE 31
DTS CUSSTO TN Ea Ae rane pened Gee aioe PRUNREER SEN COE SME RUE UE) CVO RMUULe apt Mei cob Be Me. 42
rteratunTed C1tedw ial tau Coes an se eel IMMA AGHA R Go IE NAAN se legap Ua Ane iy RUA AMO Me det 49
Abbreviations, Uisedane iouregi i. 9.024 h io Wiad aay ieee eet vn lens aesele 52

INTRODUCTION

The family Gelastocoridae contains two genera, Gelastocoris
and Nerthra. Its representatives are littoral, living in swamps or
along the shores of ponds and streams, where they frequently
burrow into the damp sand or mud. They are predaceous, feed-
ing mainly upon other insects. Their prominent, laterally pro-
jecting compound eyes, their jerky, hopping means of locomo-
tion, and their dark-colored, roughened exoskeleton have earned
them the common name of ‘‘toad bugs.’’

Very little is known about the anatomy of gelastocorids. Earlier
studies on the systematics of the Heteroptera have mentioned
some of the more obvious features of their external anatomy,
such as their possession of ocelli and short antennae. Hungerford
(1922) studied the life history of Gelastocoris oculatus and made
a few ecological observations on this species. In his extensive

4 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

study of the head capsule of the Hemiptera, Spooner (1938)
discussed briefly the external appearance of the head of Gelasto-
coris sp. Todd (1955) made a taxonomic study of the family,
describing the genitalia and some of the features of the exo-
skeleton. The presence of cephalic glands in G. oculatus has
been reported (Parsons, in press), but, to the author’s knowl-
edge, almost nothing else is known of the internal anatomy of
the Gelastocoridae.

The phylogentie position of the family is somewhat uncertain.
It is generally believed that the aquatic Heteroptera arose from
terrestrial forms, and most workers consider the shore-dwelling
families to represent an intermediate stage in the progression
from land to water. But whether the gelastocorids are more
closely related to the totally aquatic Hydrocorisae or to the
semi-aquatie Amphibicorisae is not generally agreed upon.
Spooner (1938) placed them in the Amphibicorisae; China
(1955b), however, considers them to be early offshoots of the
ancestral Hydrocorisae. Since most of the theories presented by
earlier workers have been based upon external anatomy alone, it
appears that a study of the internal structure of a gelastocorid
would be of value. The chief purpose of the present study is to
present additional morphological evidence which might link the
velastocorids more definitely with either the Hydrocorisae or the
Amphibicorisae. Although this paper is limited to the muscula-
ture and sclerotized parts of the gelastocorid head, the author
plans in the future to extend the study to other systems and other
parts of the body.

The literature on the hemipterous head is both extensive and
confusing. One source of confusion is the variety of different
terms which have been used for many of the structures. Another
problem is that very few investigators have adequately described
the complex relationships between the various endoskeletal ele-
ments. In the present study, an attempt has been made to bring
together, wherever possible, the different terms which have been
used for each structure, and to show how the various parts of the
endoskeleton are interrelated.

I am grateful to the members of the C. V. Riley Entomological
Society, of Columbia, Missouri, who supplied the live and pre-
served specimens used in this study. I also wish to thank Mr.

oD

PARSONS: HEAD OF GELASTOCORIS

Edwin P. Marks of Washburn University for his helpful sug-
gestions with regard to the gelastocorid food pump, and my
husband, Dr. Thomas S. Parsons of Harvard University, for his
advice and help with the preparation of the manuscript. This
study was carried out during the tenure of the Ellen C. Sabin
Fellowship, awarded by the American Association of University
Women.

MATERIALS AND METHODS

Most of the observations were made on insects preserved in
alcoholic Bouin’s solution and stored in 70% alcohol. In order
to have fresh material available when needed, live Gelastocoris
were kept in the laboratory. They were placed in several large
aquaria, allowing approximately 8 to 10 square inches of surface
area per insect. The bottoms of the aquaria were filled with fairly
coarse sand, which was kept moist at all times; the insects died
if left without moisture for more than an hour. The bugs were
fed vestigial-winged Drosophila, which were supplied in large
numbers in order to discourage cannibalism. Cheesecloth cover-
ings were placed over the tops of the aquaria to prevent the
escape of the Drosophila; the gelastocorids were never observed
to fly, or to climb more than a few inches up the sides of the
aquaria.

In the microdissections of the heads, the techniques of Marks
(1958 and 1959) were employed. The sclerotized structures were
studied in heads which had been left in a hot concentrated solu-
tion of potassium hydroxide for approximately half an hour;
this dissolved away most of the soft tissues. They were then dis-
sected in glycerine or in 70% alcohol, under a stereoscopic
microscope. The position of the frontal ganglion was determined
by dissection (in 70% alcohol) of whole heads or of ones which
had been cut parasagittally with a razor blade. In the larger of
the two pieces thus obtained, the brain and frontal ganglion
connectives were first located, and the latter were then traced to
the frontal ganglion. The nervous and muscular tissues were
elarified by introducing, with a fine pipette, a drop of Dela-
field’s hematoxylin onto the surface of the dissection; this tech-
nique was suggested by Mr. R. B. Willey, of Harvard University.

Transverse, sagittal, and frontal serial sections were made
through the heads of two adults, seven fifth instar nymphs, and

6 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

one fourth instar nymph. One of the nymphs was fixed in
Carnoy’s fluid, while the rest were preserved in alcoholic Bouin’s
solution. The heads were prepared for sectioning according to
the Peterfi technique, embedded in 60-62° Tissuemat, and sec-
tioned at 7 uw. The slides were stained in either Mallory’s triple
connective-tissue stain or Delafield’s hematoxylin and eosin.

EXTERNAL ANATOMY

Figure 1 shows the dorsal surface of the gelastocorid head.
It is triangular in shape, with prominent, laterally-projecting
compound eyes, and with two ocelli. Between the ocelli is a
broad, U-shaped indentation, and two shorter grooves lie lateral
to each ocellus. These indentations produce, on the inner surface
of the cranium, ridges for muscle attachments. Anterior to them

Figure 1. Dorsal view of the head. 14 X.

Figure 2. Ventral view of the head.

he a pair of small pits, which mark the points at which the
clypeal connectives (to be discussed later) contact the exo-
skeleton.

Although Spooner’s (1938) diagram of the head of a Gelasto-
coris nymph shows an epicranial suture which forks in approxi-
mately the same position as the U-shaped indentation, neither
the fifth instar nymphs nor the adults examined in the present
study showed such a suture. Thus the boundaries of the facial
selerites are difficult to determine. According to Snodgrass

PARSONS: HEAD OF GELASTOCORIS

(1947), the medial facial sclerites can be distinguished by their
muscle attachments; cibarial muscles originate on the elypeus
and pharyngeal muscles on the frons. He further states that the
forking of the eedysial cleavage line (epicranial suture), which
divides the frons from the vertex, occurs immediately above or
behind the fronto-clypeal muscles. If Snodgrass’ eriteria, which
have been employed by many workers, are followed here, the
region of the head from the base of the labium to the U-shaped
indentation may be termed the clypeus (C). As will be seen later,
all the muscles originating in this area are anterior to the frontal
ganglion, and are therefore cibarial. There is nothing to divide
this region into an anteclypeus and a postelypeus, and lateral
paraclypeal sutures are also absent.

The frons is much more difficult to define on the basis of muscle
attachments. The anterior pharyngeal dilator muscles are re-
duced, in Gelastocoris, to only a few narrow strands (Fig. 27,
17 A) which run between the brain and the frontal ganglion to
attach to the U-shaped indentation. Also originating on this
indentation, however, is a well-developed group of cibarial
muscles (Mig. 27, 16, 17). Strictly speaking, therefore, the frons
is limited to only a few small spots on the U-shaped indentation,
while the rest of the indentation, like the region anterior to it,
must be considered clypeal in nature. Rawat (1939) found a
similar situation in Naucoris. Snodgrass’ definition of the frons
will be followed in the present study, although with some reser-
vations. DuPorte (1946) has criticized Snodgrass’ use of muscle
attachments for the identification of sclerites, and it may be that
the indentation and the region anterior to it should be termed
the ‘‘fronto-clypeus.’’

The vertex (VX) is the region behind the U-shaped indenta-
tion. Spooner (1958) termed this the ‘‘frons’’ in adult gelasto-
corids; however, since none of the dilator muscles of the food
pump attach to it, his term is incorrect.

A ventral view of the head (Fig. 2) shows the short, four-
segmented antennae which le beneath the compound eyes. The
exoskeleton at the base of each eye is indented to receive the
knob-shaped terminal segment of the antenna. Just anterior to
the antennal bases, and situated on slight protuberances, lie the
external orifices of the cephalic glands (CO); these have been

o/2)

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

deseribed in an earlier paper (Parsons, in press). The large
occipital foramen (OF) is encircled by a collar-like rim of exo-
skeleton, which shall be termed, for convenience, the postocciput
(P), although there is no definite suture separating it from the
rest of the cranium. Ventrolaterally, it bears a pair of occipital
condyles (OC) which project into the prothorax, and just dorsal
to these is a pair of shorter lateral apodemes (LA). The ventral
exoskeletal plate of the head is usually termed the gula (@) in
the Heteroptera, and that term shall be used here. in Gelasto-
coris, the anterior edge of the gula is notched to allow the labium
to swing downwards.

Figure 3. Lateral view of the head.

The only definite sutures on the cranium cecur at its sides
(Fig. 3). Extending anteriorly from the anterior border of the
cephale gland orifice to the lateral corner of the labrum is a
fairly well defined suture. At the corner of the labrum it curves
dorsally and then turns posteriorly to run to the posterior border
of the glandular orifice. The dorsal part of this suture shall be
termed the dorsal suture (D), and the ventral part the ventral
suture (V). In adults, the dorsal suture is usually visible for
only the anterior third of its length, but it may be seen in its
entirety in fifth instar nymphs. The side of the head is quite
flat, and forms rather sharp angles with the dorsal and ventral
surfaces of the cranium. The sutures divide this flattened lateral

PARSONS: HEAD OF GELASTOCORIS 9

region into three general areas, a central one which is surrounded
by the sutures (Fig. 3, MXP), a dorsal one (Fig. 3, C), and a
ventral one (Fig. 3, GL).

The region above the dorsal suture appears to be a continua-
tion of the clypeus, since, as will be seen later, the epipharynx
is an inflection of its margins along the dorsal suture. Of the
two remaining regions, Spooner (1938) has termed the central
element the ‘‘paraclypeus’’ and the ventral one the ‘‘ maxillary
plate.’’ The external topography of this region makes Spooner’s
interpretation seem quite plausible. In the majority of Heter-
optera studied by that author, two sclerites are found in the
lateral region of the head, a dorsal (or posterior) paraclypeus
and a ventral (or anterior) maxillary plate. Typically the latter
fuses with the gular area without a suture, thus resembling the
region labelled GZ in Figure 3.

Spooner’s interpretation, which is based upon external ap-
pearances alone, is shown to be incorrect after examination of the
internal structure of the head. One argument against his idea
concerns the position of the mandibular lever. Other workers
appear to agree that this lever is articulated to the cranium at
the genal suture, which divides the more dorsal sclerite (the
paraclypeus of Spooner, 1938; ‘‘lorum’’ of Hamilton, 1931,
Butt, 1943, and Sprague, 1956; ‘‘jugum’’ of Snodgrass, 1935;
‘‘mandibular plate’’ of Griffith, 1945, and Akbar, 1957; ‘‘lamina
mandibularis’’ of Benwitz, 1956; and ‘‘fulerum’’ of Barth, 1952)
from the more ventral sclerite (the maxillary plate of Hamilton,
1931, Spooner, 1938, Butt, 1943, Griffith, 1945, Sprague, 1956,
and Akbar, 1957; ‘‘processus maxillaris’’ of Becker, 1929;
‘‘maxillary sclerite’’ of Rawat, 1939; and ‘‘lamina maxillaris”’
of Barth, 1952, and Benwitz, 1956). In Gelastocoris the man-
dibular lever joins the exoskeleton at the dorsal suture, at the
point where this suture becomes indistinct in adults (Figs. 4, 5,
12). It seems, therefore, that the dorsal suture represents the
genal suture of other authors, and that the sclerite lying below
it (between it and the ventral suture) is not the paraclypeus but
the maxillary plate. As will be discussed later, the paraclypeus
of Gelastocoris (termed the lorum in this study) is wholly in-
flected within the head, and, as in some aquatie Heteroptera, is
not visible externally.

10 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

A second argument against Spooner’s use of the term para-
clypeus for the sclerite here termed the maxillary plate is offered
by the point of origin of the maxillary protractor muscles. Pre-
vious investigations all seem to indicate that the maxillary pro-
tractors originate on the maxillary plate or on an internal
extension of this sclerite. In Gelastocoris these muscles originate
on the anterior wall of the genal sae (to be discussed later), a
pouch formed by marginal inflections of the sclerite which is
enclosed by the dorsal and ventral sutures. This sclerite must,
therefore, represent the maxillary plate. The area ventral to it,
which Spooner termed the maxillary plate, is not separated from
the eula in any way, and seems to be merely an anterolateral
triangular extension of that exoskeletal zone. For convenience,
it will be termed the gular lobe (GL).

At the base of the compound eye, lying just dorsal to the
second segment of the antenna, is a peculiar curved ridge com-
posed of numerous small tubercles (Fig. 3). A second, tooth-like
protuberance is found at the ventromedial margin of the com-
pound eye, just beyond the apical segment of the antenna (Figs.
2, 3). Whether these two protuberances have any special function
is not known. The surface of the cranial exoskeleton is fairly
smooth on the sides of the head, but the dorsal and ventral
regions are covered with small tubercles giving them a rough-
ened texture. Along the ventrolateral edge, the anterior dorso-
lateral edge, and the margins of the compound eyes are numerous
long hairs (not shown in the figures), and finer, shorter hairs
are scattered over the entire surface of the head, particularly on
the antennae. Grains of sand cling to these hairs, and, along
with the gray or brown color of the insect, produce an excellent
camouflaging effect against sandy or muddy backgrounds.

The labium (Figs. 2, 3, L) is four-segmented and normally
projects ventrally from the anteriormost part of the head. The
dorsal part of each segment is invaginated to form the stylet
groove (SG), which contains the apical parts of the four stylets
(the stylet bundle). Lying across the posterior part of the basal
segment is a broad, short labrum (ZB) which is attached by a
membrane to the anterior margin of the clypeus. It extends
laterally as far as the edges of the maxillary plates.

PARSONS: HEAD OF GELASTOCORIS ial

INTERNAL ANATOMY
GENERAL STRUCTURE

Figure 4 shows a ventral view of the head after the removal of
most of the exoskeleton. Anteriorly, the most ventral endoskele-
tal element is the broad, flat suspensory plate (SU). This extends
anteriorly to the base of the labium; laterally it attaches to the
dorsal border of the gular lobe, along the ventral suture. Along
its posterior margin it is continuous with the right and left
ventral walls of the genal sacs (VG). The genal sacs are paired
pouches lying immediately dorsal to the suspensory plate; they
are formed from inflections of the dorsal, ventral, and anterior
margins of the maxillary plate. Dorsal to the genal sacs and
extending posteriorly along the midline of the head is the food
pump (F'). On either side of the food pump are the two trough-
shaped hypopharyngeal wings (DH and VH), the ventral surfaces
of which are visible in Figure 4. Ventrally these are continuous
with the united posterior portions of the suspensory plate and
ventral walls of the genal sacs. Along the posterior border of
the suspensory plate, lying between the bases of the two hypo-
pharyngeal wings, is the salivary pump (S). Finally, attached
to the exoskeleton at the dorsal sutures on either side of the head,
lie the triangular mandibular levers (MDL), whose medial
corners articulate with the bases of the mandibular stylets (MD).

The more dorsal structures are shown in Figure 5. Here it
may be seen that the anterior walls of the food pump diverge and
pass laterally to join the exoskeleton at the dorsal suture, just
anterior to the point of articulation of the mandibular lever.
Of this lateral diversion, the limited portion which arises from
the floor or hypopharyngeal part of the food pump is here termed
the lorum (LO). Fused with the lorum and extending anteriorly
from it is a broad plate which reaches to the anteroventral margin
of the labrum. This is the epipharyngeal plate (PL), and forms
the anterior portion of the roof or epipharyngeal part of the food
pump. Just ventral to the epipharyngeal plate lie the dorsal walls
of the genal sacs (DG). Posteriorly these dorsal walls are con-
tinuous with the anterior edge of the mandibular lever; along
their anterior margins they are continuous with the ventral walls
of the genal sacs (not shown in Fig. 5). The dorsal surface of

\
!
'
1 1
| |
! |
\ \

MX MD MDL DH

Figure 4. Ventral view of the endoskeletal structures after removal of
most of the cranial exoskeleton. The right side has been tilted ventrally.
Part of the suspensory plate has been cut away to reveal the genal sacs, and
the posterior part of the medial wall of the genal sae (left side) has been
cut off to reveal the maxillary lever. On the right side, the maxillary stylet
and the medial wall of the genal sae are not shown. The labium and the
stylet bundle are omitted.

Figure 5. Dorsal view of the endoskeletal structures after removal of most
of the cranial exoskeleton. The right side has been tilted ventrally. The
medial wall of the genal sac on the right side, along with the labium and the
stylet bundle, are omitted.

PARSONS: HEAD OF GELASTOCORIS 13

the hypopharyngeal wing extends anteriorly between the epiph-
aryngeal plate and the dorsal wall of the genal sac, and becomes
continuous with the lorum and with the side of the anterior
portion of the food pump (concealed, in Figure 5, by the
epipharyngeal plate). Between the hypopharyngeal wine and
the mandibular stylet lies the mazillary stylet (MX). A short,
slender maxillary lever (MXLI) connects its base with the tip of
the hypopharyngeal wing. Extending between the maxillary
stylet and the hypopharyngeal wing is a very delicate, thin
double membrane, the medial wall of the genal sac (MG). In the
anterior part of the head, this medial wall surrounds the stylets
and the mandibular lever (Fig. 4, left side) ; posteriorly, its two
lamellae enclose the maxillary lever.

Foop Pump, Lora, AND EPIPHARYNGEAL PLATE

When the food pump is cut down the midline, as in Figure 6,
its two-layered nature is revealed. The hypopharyna (H) (in-
correctly termed the ‘‘anterior arms of the tentorium’’ by Ham-
ilton, 1931), which forms the floor, and the epipharynax (E), or
roof, lie close together, one upon the other. Transverse sections
through the pump show it to be U- or V-shaped (Fig. 13, A-C).
Posteriorly, the heavily sclerotized hypopharynx and the mem-
branous epipharynx are joined along their dorsolateral margins,
and a narrow cavity, the lumen of the pump, is thus enclosed
between them. Anterior to the lora, however, the hypopharynx
and epipharynx are not joined, although they are contiguous in
the lateral part of the head (see Fig. 13D, right side).

The walls of the food pump are not, as in many Heteroptera,
fused with the clypeus. Instead, the structure is held firmly in
place by a few short, sclerotized connections, the clypeal con-
nectives (CN), which run from the elypeus to the dorsolateral
margins of the pump (Fig. 7). Anterior to these connectives, the
pump is relatively broad, while posteriorly it gradually tapers
to a narrow structure (Figs. 5, 9). A lateral view of the food
pump (Figs. 6, 7) shows it to be somewhat arched, the region
from the clypeal connectives to the lorum being raised above the
anterior and posterior extremities.

14 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

6

G3 SP LO Gt

!
|
| !
|

| !
! 1
| ne
!

|

!
1
|
I
!
I
|
l
{ —
t
{
;

Figure 6. Lateral view of the inner surfaces of the left halves of the food
and salivary pumps, after removal of the right halves. The anterior portion
of the epipharyngeal plate is shown somewhat raised from its normal vo-
sition.

Figure 7. Lateral view of the outer surfaces of the food and salivary
pumps. The lateral portions of the lorum, epipharyngeal plate, and hypo-
pharyngeal wing (dorsal surface) have been cut off.

PARSONS: HEAD OF GELASTOCORIS 15

The hypopharyngeal part of the pump is relatively simple.
Just ventral to the clypeal connectives there is a slight trans-
verse thickening in the hypopharyngeal wall. From this thick-
ening a narrow, sclerotized hypopharyngeal flap (HF) extends
anteriorly on the outer surface of the pump (Figs. 6, 7). Most
of the hypopharyngeal wall directly in front of this flap is trans-
parent, and thus there appears to be a break at this point. The
hypopharynx is widest between this flap and the posterior edges
of the lora; anterior to the latter, it tapers sharply (Fig. 8),
forming a narrow trough (AF) which ends between the anterior
borders of the genal sacs. Merged with either side of this narrow
anterior trough he the anteriormost extensions of the dorsal
surfaces of the hypopharyngeal wings (Fig. 8); these are also
continuous with the ventral surfaces of the lora of either side.
They form a support for the anterior tip of the hypopharynx,
and also, as will be discussed later, help to direct the path of
the stylets in this region.

In most Heteroptera, those internal sclerotized structures
which are here termed the lora are also visible externally. The
external part of each lorum is a sclerite which, as was discussed
earlier, has received many different names (‘‘lorum,’’ ‘‘para-
elypeus,’’ ete.). The internal extension of this sclerite, which is
continuous with the hypopharynx of the food pump, has been
termed the lorum (Snodgrass, 1938; Butt, 1943; and Akbar,
1957), the ‘‘mandibular plate’’ (Griffith, 1945), or the ‘‘lamina
mandibularis’’ (Benwitz, 1956). Unfortunately, many authors
have used the same term for both the external sclerite and the
internal element, a practice which makes their descriptions
somewhat confusing.

The lora of Gelastocoris appear to be contained entirely within
the head, and there is no external sclerite present. It is difficult
to determine the exact boundaries of the lora, since there appears
to be considerable fusion between them and other structures.
Laterally, they are partially fused with the underlying genal
sacs. Dorsally, they join the posterior margin of the epipharyn-
geal plate, the line of fusion continuing the dorsolateral margins
of the food pump (Fig. 8). Medially the epipharyngeal plate
extends dorsal to the lora, while laterally it appears to come off
from the anterior margins of the lora. According to Snodgrass

16 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

(1938), the lora are hypopharyngeal in nature. If this definition
is used, the lora should be the parts which remain after the
epipharyngeal plate is carefully torn off and after the walls of
the genal sacs are pulled away from the hypopharynx. When
this is done, the lora appear as in Figure 8; laterally they are
very narrow, but they become somewhat broader medially where
they join the hypopharyngeal walls of the food pump and the
dorsal surfaces of the hypopharyngeal wings. The latter, which
merge anteriorly with the sides of the food pump’s floor, repre-
sent the anteriormost extensions of the lora. The anterior groups
of mandibular protractor muscles (Fig. 26, 8) originate on the
posteromedial loral surfaces. In most specimens, the posterior

Figure 8. Anterolateral view of the food pump, lora, and genal sacs. The
labium, labrum, eclypeus, and most of the gula have been removed. The
epipharynx has been removed from the hypopharynx, leaving a torn edge in
those regions in which it was fused to the hypopharyngeal elements. The
suspensory plate is bent ventrally, and the stylets are not shown.

edges of the lora are connected to the clypeus by a few short,
sclerotized connectives, similar to the elypeal connectives of the
food pump.

The epipharyngeal part of the food pump is somewhat more
complex than the hypopharyngeal portion. Since the epipharynx
is an inflection of the margins of the clypeus, it can be raised
from the floor of the pump by separating the clypeus from the
maxillary plate along the dorsal suture. Figure 9 shows the

PARSONS: HEAD OF GELASTOCORIS 17

ventral surface of the epipharynx after removal from the hypo-
pharynx. Its posterior part is similar in shape to that of the floor
of the pump. Most of its broad middle region, however, is modi-
fied to form a pair of roughly triangular, thickened plates (SP)
which are joined posteriorly. The surfaces of these structures
bear a series of regular, parallel striations, and they are, there-
fore, termed striated plates in the present study. Examination
of a striated plate under a compound microscope reveals count-
less minute secondary ridges running at roughly 40 degree angles
to these striations. At regular intervals along these secondary

Figure 9. Ventral surface of the epipharynx after removal from the
hypopharyngeal part of the food pump. The labrum and a small part of the
clypeus are shown in place.

Figure 10. A marginal portion of a striated plate, as seen under the com-
pound microscope. The thickenings in the narrow secondary ridges produce
the striations which are visible in the dissections. Approx. 2000 X.

ridges are thickenings which lie side by side along adjacent
ridges, thus forming the major striations (Fig. 10).

The margins of the posterior part of the epipharynx are
strengthened by a well sclerotized, pigmented border. The latter
extends anteriorly to a point approximately midway between
the striated plates and the epipharyngeal plate, and then turns
medially, disappearing along the midline of the epipharynx.

The epipharyngeal plate is continuous, along the dorsal suture,
with the ventral margin of the clypeus (Figs. 5, 9, 12, 13E).

18 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Anteriorly it is fused with the ventral surface of the labrum
(Fig. 9). An anteromedial tongue-like projection of the epiph-
arynx extends anteriorly a little beyond the labrum; this
epipharyngeal projection (EP) helps to hold in place the stylet
bundle, which lies directly ventral to it in the stylet groove of
the first labial segment. For most of its length, the projection is
somewhat concave ventrally. Its margins diverge laterally, just
posterior to the base of the labrum, forming a transverse epiph-
aryngeal ridge (Fig. 9, ER). This ridge probably represents the
‘*Gaumenfalte’’ of Becker (1929); it lies just anterior to the
anterior margins of the genal sacs, which underlie the epi-
pharyngeal plate (Fig. 11). The stylets pass through the medial
gap in the epipharyngeal ridge, which helps to direct them into
the stylet groove of the labium.

Just posterior to the epipharyngeal ridge is an oval, sclerotized,
raised area on the midventral surface of the epipharyngeal plate.
This raised area hes just posterior to the point at which the ef-
ferent duct from the salivary pump opens onto the tip of the
hypopharynx; its function may be to prevent the backflow of
saliva into the food pump.

The muscles which operate the food pump insert by slender,
sclerotized tendons projecting from the dorsal surface of the
epipharynx along the midline. As shown in Figure 6, there are
four groups of these processes, which shall here be termed Groups
1 through 4 respectively. Group 1 (G@1) extends from just
posterior to the oval, sclerotized area to a point somewhat
anterior to the striated plates. Group 2 (@2) is located between
the striated plates, while Group 3 (G3) consists of only four or
five very lone processes attached to the joined posterior edges of
the striated plates. Moderately long tendons make up Group 4
(G4), which extends posteriorly from the striated plates, ending
a short distance anterior to the posterior limit of the epipharynx.
The muscles inserting by the tendons of Groups 1, 2, and 4 raise
the roof of the pump, which snaps back into place elastically
when the muscles relax; these movements create the pumping
action of the food pump. The striated plates are moved back and
forth in an anterior-posterior plane by the action of the muscles
attaching onto Group 3; they are also raised and lowered by the

PARSONS: HEAD OF GELASTOCORIS 19

muscles of Group 2. These motions cause the highly ridged sur-
faces of the plates to rasp against the contents of the lumen of
the pump. The author has noted no particulate matter in the
eut of Gelastocoris, and the insect appears to ingest only fluids.
It may be that the plates serve as straining or filtering devices,
possibly breaking up small clots of material in the ingested fluid.
Marks (1958 and 1959) has described somewhat similar epipha-
ryngeal specializations in the food pumps of several fluid-feeding
aquatic bugs, and interprets these as straining devices.

SALIVARY Pump

As seen in Figures 6 and 7, the salivary pump lies just ventral
to the anterior half of the food pump. It is a large, bell-shaped
structure, with a short, narrow efferent duct (ED) leading from
its anterior end to open through the ventral surface of the
hypopharyngeal tip of the food pump. Two slender afferent
ducts (AD), which run from the salivary glands to the pump’s
ventral surface, convey the saliva to the lumen of the pump.
The expanded posterior end of the salivary pump is invaginated,
its walls becoming quite thickened. To the center of this thick-
ened invaginated portion is attached a thin, plate-like tendon
for the insertion of the salivary pump muscles. The latter enable
the invaginated wall to act as a piston, which is pulled out by
museular contraction and sprines back into place elastically when
the muscles relax.

HYPOPHARYNGEAL WINGS, SUSPENSORY
PLATE, AND GENAL Sacs

The hypopharyngeal wings (Hamilton, 1931; Butt, 1943;
Qadri, 1951; and Benwitz, 1956) have been variously termed
the ‘‘posterior plates of the hypopharynx’’ (Snodgrass, 1938),
the ‘‘sclerotized plates of the genal apodeme’’ (Rawat, 1939),
the ‘‘maxillary sheaths’’ (Sprague, 1956), or the ‘‘ventral
processes of the hypopharynx’’ (Akbar, 1957). Several authors,
such as Ekblom (1926) and Becker (1929) mistook them for part
of a tentorium. Whether or not a tentorium is present in the
Heteroptera appears to be a disputed point; although Spooner
(1938) and Butt (1943) deny its existence in heteropterans,

20 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Akbar (1957) has described a remnant of a tentorium in the
eoreid Leptocorisa. The tentorium appears to be completely
lacking in Gelastocoris.

As shown in Figure 5, the hypopharyngeal wings are plate-
like processes on either side of the food pump. Transverse

Figure 11. Ventral view of genal sacs, salivary pump, and associated
structures. Most of the gula has been removed, and the gular lobes have
been bent medially to show the ventral suture. The anterior part of the
suspensory plate and the posterior portions of the hypopharyngeal wings
have been removed. The stylets and the labium are not shown.

Figure 12. Dorsal view of the medial wall of the genal sac on the left side
of the head. The anterior parts of all structures are cut off; the most medial
portions of the epipharyngeal plate and lorum are cut away a little to the
left of the midline. Posteriorly, part of one membrane of the medial wall
has been removed to show the maxillary lever.

sections through the head (Fig. 13B) show that fine fibrillae
(indicated by dotted lines) connect them with the walls of the
food pump. The wings are folded along their longitudinal axes
so that each is trough-shaped, divided by the medial fold into a
dorsal surface (DH) and a ventral surface (VH). Dissections

PARSONS: HEAD OF GELASTOCORIS 21

of the hypopharyngeal wings reveal that they are actually two-
layered; these layers can be separated anteriorly where the
ventral surface contacts the exoskeleton, but posteriorly, in the
trough-like portions, the two lamellae seem to be fused together,
behaving as a single layer.

The anterior part of the dorsal surface of the wing has already
been described. This surface extends directly anterior, becoming
continuous with the ventral surface of the lorum (Figs. 12, 13D)
and with the tip of the hypopharyngeal part of the food pump
(Fig. 8). Anteriorly, the hypopharyngeal wing develops a series
of longitudinal ridges and grooves (Fig. 13D). These act as
tracks for the mandibular and maxillary stylets, and hold them
in position.

The anterior modifications of the ventral surface of the hypo-
pharyngeal wings are equally complex. Figure 4 shows that the
lateral margin of the ventral surface diverges laterally to meet
the exoskeleton along the ventral suture. If, in an undissected,
potassium hydroxide-treated head, the gular lobe and the maxil-
lary plate are separated from each other along the anterior part
of the ventral suture, it may be seen that from this part of the
suture an invagination arises. The invagination extends pos-
teriorly to a point just below the place where the mandibular
lever articulates with the dorsal suture. It produces two layers,
a dorsal one and a ventral one, which extend medially into the
head. Posteriorly, these lamellae are continuous with each other,
and the line of fusion between them (‘‘Querbalken’’ of Becker,
1929) is a continuation of the lateral edge of the ventral surface
of the hypopharyngeal wing. Anteriorly, however, the two layers
are entirely separate. The more ventral one, the suspensory plate,
is continuous with the dorsal edge of the gular lobe, while the
more dorsal layer, the ventral wall of the genal sac, is an inflee-
tion of the ventral margin of the maxillary plate (Fig. 13E).

The suspensory plate (‘‘Aufhangeblatt’’ of Becker, 1929)
extends from one side of the head to the other as a broad, single
lamella. The posteromedial margin of this plate forms a bridge
between the two hypopharyngeal wings, and extends across the
ventral wall of the salivary pump, with which it is fused (Fig.
11). The suspensory plate extends anteriorly as far as the base
of the first segment of the labium, with which it is partially or

22 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

completely fused. Laterally, it is bordered by the exoskeleton of
the sides of the head. The medial part of its anterior edge forms
a trough which is a posterior continuation of the stylet groove
of the labium (Fig. 8).

The ventral walls of the genal sacs (‘‘unterer Blatter der
lamina maxillaris’’ of Becker, 1929) from the two sides of the
head do not join medially, and do not project as far anteriorly
as does the suspensory plate. They extend medially to the base
of the salivary pump, but do not fuse with it. Instead, their
medial margins run from the base of the salivary pump anterior-
ly as far as the tip of the food pump, to which they are lateral
and somewhat ventral (Fig. 11). The efferent duct from the
salivary pump lies between their two parallel medial margins.

The anterior margin of each ventral wall extends laterally
to the anterior edge of the maxillary plate; along this margin
it is continuous with the dorsal wall of the genal sac. The latter
arises from an invagination along the anterior part of the dorsal
suture, and extends posteriorly as far as the articulation of the
mandibular lever. The dorsal layer of this invagination, alone
the ventral edge of the clypeus, gives rise to the epipharyngeal
plate, which has already been discussed; the ventral layer of the
invagination, along the dorsal edge of the maxillary plate, forms
the dorsal wall of the genal sae (Fig. 13E). Thus both the dorsal
and ventral walls of the genal sac are internal inflections of the
anterior margins of the maxillary plate (Fig. 8) ; Becker (1929)
termed these internal extensions the “‘laminae maxillares.’’ The
genal sacs are open only posteriorly; laterally they are closed
by the maxillary plate, and medially by the medial wall of the
genal sac, to be discussed next.

The medial wall of the genal sac is a rather complicated struc-
ture which is as difficult to describe and diagram as to dissect.
It can best be understood by a study of transverse sections
through the head (Fig. 13; indicated by broken lines). It is a
delicate, thin-walled, double membrane attached to the hypo-
pharyngeal wing and associated with the stylets and the levers
of the stylets. In the posterior part of the head, it runs dorsally
from the hypopharyngeal wing to the dorsal surface of the
maxillary stylet (Fig. 13A). The maxillary lever lies between
the two membranes near their posterior ends. More anteriorly,

Figure 13. Transverse sections through the head of a fifth instar nymph.
The medial walls of the genal sacs are indicated by broken lines, and the
muscles are stippled. The right side of each section is slightly anterior to
the left. Only the sclerotized structures and the muscles are shown. <A,
section at the level of the occipital foramen and the bases of the antennae.
30 X. B, section at the level of the striated plates; the fine fibrillae between
the food pump and the hypopharyngeal wings are indicated by dotted lines.
30 X. C, section at the level at which the mandibular lever articulates with
the mandibular stylet. 40 X. D, section at the level at which the lorum of
the right side contacts the cranial exoskeleton. 40 X. E, section through
the anterior part of the genal sacs, anterior to the salivary pump. 50 X.
F, oblique section through the first two labial segments, showing the stylet
bundle in the stylet groove. 50 X.

24 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

the medial wall comes to surround the base of the maxillary
stylet, as shown in Figure 13B. At the point where the man-
dibular lever articulates with the mandibular stylet, the double
membrane passes dorsolaterally to surround both the mandibular
stylet and lever (Fig. 13C). At this point the medial wall of the
genal sac resembles a sort of sling enclosing the stylets, the lever,
and the dorsal surface of the hypopharyngeal wing. It probably
helps to hold the stylets in position and to guide them onto the
track-like ridges and grooves of the hypopharynx.

At the level of the lora (Fig. 13D, right side), the medial wall
of the genal sac appears to consist of only one layer. Either the
more dorsal membrane has ended at this point, or else the two
membranes have fused together. The lateral part of the remain-
ing one becomes fused here with the ventral surface of the lorum.
Its medial part lies lateral to the stylets, holding them against
the ridges of the dorsal surface of the hypopharyngeal wing
(here merged with the lorum), and separating the stylets from
the cavity of the genal sac. In Figure 13E, a section through
the anterior part of the genal sacs, the membrane forms a short
medial connection between the dorsal and the ventral walls of
the sacs. It disappears at the anterior margin of the sacs, where
the dorsal and ventral walls join. The general appearance of
the medial wall of the genal sae is shown in Figures 4, 5, and
12. For simplicity it is omitted in most of the figures.

The two-layered medial wall of the genal sacs has been termed
the ‘‘Wangensack’’ (Becker, 1929), the ‘‘membranous sling”’
(Hamilton, 1931), the ‘‘bristle pouch’’ (Snodgrass, 1935), the
‘<visceraler Ast des Tentoriums’’ (Barth, 1952), and the ‘‘Stech-
borstenfalte’’ (Benwitz, 1956). Butt (1943) termed it the ‘‘max-
illary sac’’ where it surrounds the maxilla, and the ‘‘mandibular
sae’’ where it encircles the mandible; the two sacs together he
termed the ‘‘bristle sac.”’

MANDIBULAR LEVER AND STYLET

The mandibular lever (Figs. 4, 5, 12, 14), upon which the
mandibular protractor muscles insert, is triangular in shape.
One corner of the triangle is attached to the cranium at the
dorsal suture, a little posterior to the point where the lorum
contacts the cranium. A second corner articulates medially with

bo
On

PARSONS: HEAD OF GELASTOCORIS

the thickened base of the mandibular stylet, while the third is
unattached, projecting posteriorly in the head. The lateral and
medial margins of the lever are heavily sclerotized, while the
central part is membranous, continuing anteriorly directly into
the dorsal wall of the genal sae without a break. The lateral
margin of the lever is slightly longer than the medial margin,

Figure 14. Dorsal view of the right mandibular stylet (middle part omit-
ted) and lever.

Figure 15. Transverse section through the stylet bundle in the stylet
groove of the labium (taken from the section shown in Figure 13F).

Figure 16. Detail of the tip of the right mandibular stylet, showing the
teeth.

Figure 17. Dorsomedial view of the right maxillary stylet (middle part
omitted) and lever.

the shape of the lever being somewhat diferent from that figured
by Spooner (1938) for G@elastocoris sp.

Figure 14 shows the general form of the long, slender man-
dibular stylet. It is a hollow structure, approximately 2.5 mm.
in length. From the medial side of the thickened base a very
slender, sclerotized tendon extends posteriorly, providing an

26 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

insertion for one of the two groups of retractor muscles. A cross-
section through the apical half of the mandibular stylet (Fig. 15)
reveals that it is much wider ventrally than dorsally, allowing
it to conform to the contours of the maxillary stylet within the
stylet bundle. Along the lateral side of its tip are from five to
seven tooth-like projections (Fig. 16), which extend posteriorly.
Similar mandibular teeth have been reported in a great many
Heteroptera, and appear to be of common oceurrence. Since the
mandibular stylets he on the outside of the stylet bundle, these
teeth are able to rasp against the tissue of the prey when the
gelastocorid is feeding, and also help to anchor the tip of the
stylet bundle during ingestion.

MAXILLARY LEVER AND STYLET

The maxillary lever (the ‘‘maxillary guide-rod’’ of Ekblom,
1929) is a slender, curved bar attached to the dorsal side of the
base of the maxillary stylet (Figs, 4, 5, 12, 17). It curves ven-
trally and passes medial to the stylet to articulate with the lateral
surface of the hypopharyngeal wing. As has been previously
mentioned, the maxillary lever lies between the two membranous
layers of the medial wall of the genal sae (Fig. 13A). Unlike
the mandibular lever, it does not serve as a place for muscle
attachments, and therefore the term lever is used here with some
reservations. Neither Becker (1929; Naucoris) nor Ekblom
(1929; several genera of Heteroptera) noted any muscular at-
tachments on the maxillary lever.

The hollow maxillary stylet is shaped much like the mandibular
one, although its base is somewhat thicker, since it provides a
surface for the attachment of the protractor muscles as well as
the retractors (Fig. 17). It is sh¢htly longer than the mandibu-
lar stylet, being approximately 2.75 mm. in length, and lies
ventral and medial to the latter within the cranium. Running
alone the medial surface of the stylet are two longitudinal
grooves, separated by a ridge. Within the stylet bundle (Fig.
15), the two maxillary stylets lie together in the middle, and the
opposed grooves form a dorsal food canal (FC) and a ventral
salivary canal (SC); the ridge separating these hes ventral to
the center, so that the salivary canal is narrower than the food
canal. The stylet is broadest along its dorsal edge.

€

PARSONS: HEAD OF GELASTOCORIS 27

Along the medial surfaces of the tips of each stylet are longi-
tudinal rows of anteriorly-directed bristles. Examination of the
maxillae of two individuals revealed that the tip of the right
maxilla differs from that of the left. The right maxilla bears
four bristle rows, as shown in Figure 18B; the two outer ones,
located on the dorsal and ventral margins, consist of rather fine
hairs projecting outwards, while the two inner ones are com-
posed of stiffer bristles which extend medially. The latter are
located along the ridge dividing the food and salivary canals
and along the dorsal margin of the dorsal groove. The left

eee IBA
==

Figure 18. Details of the medial surfaces of the tips of the maxillary
stylets, showing the rows of bristles. A, left maxillary stylet. B, right
maxillary stylet.

maxilla bears a tuft of fine hairs on the dorsal margin of the
dorsal groove, along with only two rows of bristles, one on the
ventral margin and one on the separating ridge (Fig. 18A).
These rows are shorter than those of the right maxilla. The
bristles of the more dorsal row are stiffened and project medially.
Longitudinal sections through the stylet bundle reveal that the
bristles of the opposed separating ridges on the two stylets inter-
lock, thus holding the right and left maxillae together. It is
possible that the maxillary hairs and bristles help to sift out
particulate matter from the ingested food. Many authors have
reported similar rows of maxillary bristles in other Heteroptera,
and some asymmetry in their arrangement seems to be common.

LABIUM AND LABRUM

The main body of the gelastocorid beak is formed by the
labium, the labrum being only a dorsal flap overlapping its base.
The labrum (Figs. 3, 5, 9) is shaped like a broad triangle, with
a wide posterior base and curved sides ending in an anteriorly

28 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

projecting apex. Its base is attached by a membrane to the an-
terior border of the clypeus, as far as the margins of the maxil-
lary plates. Fused to its ventral surface is the anterior portion
of the epipharyngeal plate, and the epipharyngeal projection
extends a short distance anterior to it (Fig. 9).

The labium is attached to the cranium by a membrane extend-
ing from the ventral part of its posterior border to the anterior
edge of the gula. This membrane ends at the anterior margins
of the maxillary plate. Dorsally, the labium does not articulate
with the exoskeleton; here the most basal part of the first (or
basal) labial segment is membranous, and is partially or wholly
fused with the anterior edge of the suspensory plate. The labium
is composed of four segments, of which the second and fourth

Figure 19. Anterolateral view of the first two segments of the labium,
showing the dorsal and ventral intersegmental sclerites.

are the shortest and the third the longest. Each is eylindrieal,
its dorsal part being membranous and invaginated to form the
stylet groove which contains the apical portions of the four stylets
GRigs, loe20) 21).

The most basal labial segment is much longer dorsally than
ventrally (Figs. 19, 20). Its ventral part is often concealed by
the anterior edge of the gula. The posterior margin of the dorsal
side is not straight, but is somewhat V-shaped, the apex of the V
projecting anteriorly and the two arms extending posteriorly.
The basal portion of the dorsal side is membranous, while the
anterior portion is heavily sclerotized; the labrum overlaps the
membranous part.

In the connecting membrane between the first and second seg-
ments are two pairs of intersegmental sclerites (Fig. 19), an

PARSONS: HEAD OF GELASTOCORIS 29

elongate, dorsally located one, and a much smaller, ventral one,
which is roughly oval in shape. These provide insertions for parts
of the main levator and depressor muscles of the labium.

The second labial segment is the most complex and irregularly
shaped of the four. The midventral part of its basal margin is
bent inwards, forming a short projection, and a second, larger

SG IS

0

Figure 20. Lateral view of the inner surface of the left half of the labium.
The labium has been cut down the midline, and the right half has been
removed. The broken lines indicate the pigmented, sclerotized structures
supporting the stylet groove.

Figure 21. View of the inner surface of the dorsal half of the labium.
The ventral half has been removed. The unstippled areas anterior to the
oblique plate represent the pigmented, sclerotized structures supporting the
stylet groove.

30 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

apodeme originates from its basal margin dorsally (Figs. 20, 21)
The latter is a posterior continuation of the entire dorsal half
of that margin, extending into the cavity of the first labial seg-
ment in the form of a broad, flat oblique plate (OP). It serves
as a point of attachment for two labial muscles. Apodemes iden-
tical to the oblique plate in both form and position have been
deseribed by Becker (1929) in Naucoris, and by Butt (1943)
in Notonecta. The author has also observed them in Nepa,
Ranatra, Pelocoris, and Lethocerus. Griffith (1945) reported a
structure in corixids which he considered to be homologous with
this apodeme, although Benwitz (1956) disagreed with his in-
terpretation. In Gelastocoris, the floor of the stylet groove ap-
pears to be reinforced, in the second labial segment, by an under-
lying sclerotized shelf which is continuous with the base of the
oblique plate (Fig. 21). This shelf, which is heavily pigmented,
extends into the posterior part of the stylet groove of the third
labial segment. Similar supporting structures of the stylet groove
have been described in Naucoris by Becker (1929).

The third and fourth segments are relatively simple and eylin-
drical. The stylet groove is located more centrally in these than
in the two preceding segments. Between the two terminal seg-
ments are a pair of sclerites which resemble short rods on the
dorsal outer surface (Fig. 20, ZS). They lie side by side directly
above the stylet groove, and the dorsal margins of the third and
fourth segments are somewhat notched to accommodate them.
The anterior and posterior ends of each sclerite extend ventrally
to encircle the stylet groove beneath them (Figs. 20, 21). Of the
two rings thus formed, the posterior one is incomplete ventrally
while the anterior one is complete. These structures apparently
support the stylet groove. Similar sclerites have been observed
between the two terminal selerites of Benacus, Gerris, and Velia
(‘‘palpus labialis’’ of Leon, 1897), Belostoma (‘‘labiappendices’’
of Crampton, 1921), Naucoris (‘‘Lappchen’’ of Becker, 1929),
and Nepa (‘‘labial appendages’’ of Hamilton, 1931). The author
has also observed them in Notonecta. The stylet groove of the
terminal segment is further supported by a pair of pigmented,
sclerotized bars lying in the walls of the sides of the groove, and
by a third sclerotized ring which encircles the groove at the tip
of the segment (Fig. 21). Similar supporting structures have

PARSONS: HEAD OF GELASTOCORIS oil

been reported in the terminal labial segments of Sphaerodema
and Anisops (Qadri, 1951) and of Leptocorisa (Akbar, 1957).

MuSCLEs oF THE HEAD

In the following section, an attempt has been made to list, for
each muscle, similar muscles which have been reported in other
Heteroptera. Those which are listed are included because both
their origins and their insertions, as described in the literature,
are the same or very similar to those of the corresponding muscle
in Gelastocoris. Whether or not they are actually homologous
to the gelastocorid muscle which they resemble cannot, in most
cases, be definitely stated.

In Figures 22 through 28, the muscles are designated by the
numbers given below.

Labial muscles
1. M. LEVATOR LABII (Figs. 13F, 22, 23, 24, 25)

A very well developed, paired muscle.

Origin: Ventral surface of the suspensory plate, anterior to the
genal sacs.

Insertion: Posterior margin of the second labial segment, both
dorsally (on either side of the stylet groove) and laterally, as
far as the ventral intersegmental sclerites. Some strands also
insert onto the dorsal and ventral intersegmental sclerites.

Action: Raises the labium; probably also allows some lateral move-
ment.

Similar muscles: Second adductors of the labium (Notonecta; Butt,
1943): abductor of the labium (Leptocorisa; Akbar, 1957).

2. M. DEPRESSOR LABIT PRIMUS (Figs. 13C-E, 22, 24, 25)

A well developed, paired muscle.

Origin: Medial surface of the hypopharyngeal wing, just posterior
to the salivary pump.

Insertion: Posterior margin of the second labial segment, on either
side of the ventromedial apodeme; extends along the ventral
side as far as the ventral intersegmental sclerites. A few
strands appear to insert on the latter.

Action: Lowers the labium.

Similar muscles: First adductors of the labium (Notonecta and
Oncopeltus; Butt, 1943); extrinsic adductor of the labium
(Leptocorisa ; Akbar, 1957).

32 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

3. M. TRANSVERSALIS LABII PRIMUS (Figs. 13F, 22, 24)

A short, thick, unpaired muscle.

Origin: Dorsal surface of the oblique plate of the labium.

Insertion: Floor of the stylet groove of the first labial segment.

Action: Depresses the floor of the stylet groove, possibly foreing
together the dorsal lips of the groove. Possibly lowers the
labium by raising the oblique plate (?).

Similar muscles: Median transverse muscles of the bristle trough
(Notonecta; Butt, 1943).

Figure 22. Lateral view of the labium, with the exoskeleton removed on
the right side in order to show the muscles. The oblique plate and the sup-
porting structures of the stylet groove are heavily stippled. Part of muscle
no. 1 has been removed in order to show muscle no. 3. Muscles nos. 1 and 5
are omitted on the left side; muscle no. 7 and part of muscle no. 4 are
omitted on the right side.

Figure 23. View of the inner surface of the dorsal half of the labium,
showing the muscles. The supporting structures of the stylet groove are
heavily stippled.

Figure 24. View of the inner surface of the ventral half of the labium,
showing the muscles.

4. M. DEPRESSOR LABII SECUNDUS (Figs. 13F, 22, 23, 24).

A well developed, unpaired muscle.

Origin: Ventral surface of the oblique plate of the labium.
Insertion: Posteroventral margin of the third labial segment.
Action: Lowers the third and fourth labial segments.

PARSONS: HEAD OF GELASTOCORIS 33

Similar muscles: M. depressor labii (Naucoris; Becker, 1929) ;
third adduetors of the labium (Notonecta; Butt, 1943) ; muscle
group B (?) (Rhamphocorizxa; Griffith, 1945).

5. M. TRANSVERSALIS LABIIT SECUNDUS (Figs. 22, 23, 24).

A slender, paired muscle.

Origin: Posterolateral margin of the third labial segment.

Insertion: Lateral margin of the anterior portion of the sclerotized
pigmented shelf supporting the stylet groove in the third labial
segment.

Action: Moves the third and fourth labial segments laterally (?).

Similar muscles: Muskel der Endoskelettplatte (?) (Naucoris ;
Becker, 1929); M. transversalis 1 (Cimex; Kemper, 1932) ;
first transverse muscles of the bristle trough (?) (Notonecta
and Oncopeltus; Butt, 1943).

6. M. TRANSVERSALIS LABII TERTIUS (Figs. 22, 23, 24)

A fairly well developed, unpaired muscle.

Origin: Ventral wall of the third labial segment.

Insertion: Floor of the stylet groove in the anterior part of the
third labial segment, anteriorly to the posterior sclerotized
ring.

Action: Depresses the floor of the stylet groove, possibly forcing
together the dorsal lips of the groove.

Similar muscles: Muskel der Endoskelettbildung (?) (Naucoris;
Becker, 1929); M. transversalis 2 (Cimex; Kemper, 1932) ;
second transverse muscles of the bristle trough (Notonecta;
possibly also Oncopeltus; Butt, 1943); first transverse muscles
of the labial plate (Leptocorisa; Akbar, 1957).

7. M. RETRACTOR SEGMENTI ULTIMI LABIT (Figs. 22, 23, 24)

A well developed, fan-shaped, paired muscle.

Origin: Posterolateral walls of the third labial segment, reaching
nearly to the stylet groove dorsally and nearly to the midline
ventrally.

Insertion: On a slender sclerotized strand from the posterolateral
margin of the fourth labial segment.

Action: Pulls the fourth labial segment posteriorly (simultaneous
contraction of the muscles of both sides) or laterally (contrae-
tion of only one muscle).

34 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Similar muscles: Levatormuskel des vierten Unterlippengliedes
(Naucoris; Becker, 1929); Adduktor 6 (Cimex; Kemper,
1932); retractors of the terminal segment (Notonecta and
Oncopeltus; Butt, 1943).

Muscles of the stylets

8. M. PROTRACTOR SETAE MANDIBULARIS PRIMUS (Figs. 13B,
C, 25, 26)

A paired muscle, running dorsal to the mandibular lever and to
the mandibular and maxillary stylets.

Origin: Ventromedial surface of the lorum, and the side of the
food pump.

Insertion: Lateral margin of the mandibular lever, just anterior
to M. protractor setae mandibularis secundus.

Action: Moves the mandibular lever medially, thus extending the
mandibular stylet.

Similar muscles: Protractor der Mandibularborste (Nawcoris;
Becker, 1929) ; hypopharyngeal-mandibular muscle (Notonecta
and Oncopeltus; Butt, 1943); protractor of mandible (Rham-
phocoriza; Griffith, 1945); M. protractor mandibulae (Tri-
atoma; Barth, 1952); M. protractor setae mandibularis
(Coriza; Benwitz, 1956); protractor of mandibular stylets
(Leptocorisa; Akbar, 1957).

9. M. PROTRACTOR SETAE MANDIBULARIS SECUNDUS (Figs.
13A, 25, 26)

A paired muscle, running dorsal to the mandibular lever, the man-
dibular and maxillary stylets, and the food pump.

Origin: Inner surface of the clypeus, just lateral to the midline, and
between the U-shaped ridge and the point where the elypeal
connectives of the food pump contact the cranium.

Insertion: Lateral margin of the mandibular lever, between the
insertion of M. protractor setae mandibularis primus and the
apex of the lever.

Action: Same as M. protractor setae mandibularis primus.

Similar muscle: Protractor muscles of the mandible (?) (Velia;
Ekblom, 1926).

10. M. RETRACTOR SETAE MANDIBULARIS PRIMUS (Figs. 134, 25,
26)

A paired muscle, better developed than M. retractor setae mandibu-
laris secundus.

PARSONS: HEAD OF GELASTOCORIS 35

Figure 25. Ventral view of the head, with most of the gula removed in
order to show the muscles. The compound eyes have been cut off at their
bases. Muscles nos. 12, 13, 24, and 25 are omitted on the right side, and
muscles nos. 1, 2, 8, and 9 are omitted on the left. Muscles nos. 16 and 17
are not shown.

Figure 26. Dorsal view of the head, with much of the elypeus and vertex
removed to show the more dorsal muscles. The compound eyes have been
eut off at their bases. The more ventral muscles are omitted, and the
maxillary stylets and hypopharyngeal wings are not shown.

36

Nie

12.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Origin: Roof of the cranium, on the more posterior of the two
short grooves lying lateral to the ocellus (see Fig. 1).

Insertion: Dorsal side of the base of the mandibular stylet.

Action: Retracts the mandibular stylet.

Similar muscles: Retractor Nr. 1 der Mandibularborste (Naucorts ;
Becker, 1929); retractor ‘‘rm,’’ of mandible (Aphanus and
Myrmus; Ekblom, 1926 and 1930).

. RETRACTOR SETAE MANDIBULARIS SECUNDUS (Fig. 25)

A short, slender, weakly developed, paired muscle.

Origin: Dorsal margin of the occipital foramen, just dorsal to the
origin of M. retractor setae maxillaris.

Insertion: End of a long, slender, sclerotized tendon extending
from the medial side of the base of the mandibular stylet.

Action: Retracts the mandibular stylet.

Similar muscles: Retractor muscle ‘‘rm’’ of the mandibular seta
(Salda; Nabis, Coriza, Aphanus, Myrmus, and Hydrometra;
Ekblom, 1926 and 1930); Retractor Nr. 2 der Mandibular-
borste (Naucoris; Becker, 1929); retractor of mandible
(Notonecta; Butt, 1943; and Rhamphocorixa; Griffith, 1945) ;
M. retractor mandibulae (Triatoma; Barth, 1952); M. re-
tractor setae mandibularis (Corixa; Benwitz, 1956); retractor
‘a?’ of mandibular stylets (Leptocorisa; Akbar, 1957).

bd

M. PROTRACTOR SETAE MAXILLARIS (Figs. 13A-E, 25)

An extremely well developed, paired muscle, surrounding the basal
part of the maxillary stylet, and lying ventral to the man-
dibular lever and stylet.

Origin: Along the lateral surface of the hypopharyngeal wing in
the region of the salivary pump, and along the inner surface
of the anterior part of the genal sac.

Insertion: Encircling the base of the maxillary stylet.

Action: Extends the maxillary stylet.

Similar muscles: Maxillary protractors (Salda, Nabis, Aphanus,
and Hydrometra; Ekblom, 1926); Protractor der Maxillen-
borste (Naucoris; Becker, 1929); protractor of maxilla
(Notonecta and Oncopeltus; Butt, 1943); Mm. protractores
maxillae primus et secundus (Triatoma; Barth, 1952); Mm.
protractores setae maxillaris (Coriza; Benwitz, 1956); pro-
tractors ‘‘b’’ and ‘‘c’’? of maxillary stylets (Leptocorisa ;
Akbar, 1957).

PARSONS: HEAD OF GELASTOCORIS 37

13. M. RETRACTOR SETAE MAXILLARIS (Fig. 25)

A paired muscle.

Origin: Dorsal margin of the occipital foramen, just ventral to the
origin of M. retractor setae mandibularis secundus.

Insertion: Dorsal side of the base of the maxillary stylet, posterior
to M. protractor setae maxillaris.

Action: Retracts the maxillary stylet.

Similar muscles: Maxillary retractors (Salda, Nabis, Aphanus,
Hydrometra, and Coriza; Ekblom, 1926 and 1930); Retractor
der Mavxillenborste (Naucoris; Becker, 1929); retractor of
maxilla (Notonecta and Oncopeltus ; Butt, 1943; and Rhampho-
corixa; Griffith, 1945); Mm. retractores maxillae primus et
secundus (Triatoma; Barth, 1952); M. retractor setae maxil-
laris (Coriza; Benwitz, 1956); retractor of maxillary stylets
(Leptocorisa; Akbar, 1957).

Muscles of the food pump

All but the few posteriormost strands of the muscles of the
food pump are cibarial in origin, lying anterior to the frontal
ganglion. The literature indicates that the cibarial muscles may
comprise a single group, as in Ranatra and Belostoma (Marks,
1958) or may be subdivided, as in Corixa (Benwitz, 1956) or
Leptocorisa (Akbar, 1957). It appears that the grouping of the
muscles of the food pump is a variable feature, the different
eroups shifting as new needs arise. Therefore no extensive at-
tempt will be made to list similar muscles from the literature for
each of the four groups seen in Gelastocoris. The only other
heteropteran which will be considered here is Pelocoris femora-
tus; sinee the position of the frontal ganglion is the same in this
species as in Gelastocoris oculatus (see Discussion), the muscle
groups of the two insects are directly comparable, and are almost
certainly homologous.

14. M. DILATOR CIBARIIT PRIMUS (Figs. 13D, 26, 27)

Fans out dorsal to the point of insertion, so that it appears V-
shaped in transverse sections.

Origin: Inner surface of the clypeus.

Insertion: On the epipharyngeal tendons of Group 1, anterior to
the striated plates.

38 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Action: Elevates the anterior part of the epipharynx, thus dilating
the food pump.

Similar muscle: Anterior part of Muscle Group 1 (Pelocoris;
Marks, 1959).

15. M. DILATOR CIBARII SECUNDUS (Figs. 13B, 26, 27)

Similar, in appearance, to M. dilator cibarii primus.
Origin: Inner surface of the clypeus, just posterior to M. dilator
cibaril primus.

Figure 27. Lateral view of the cibarial and pharyngeal muscles. The
head has been cut down the midline, and the left halves of the food pump,
brain, and subesophageal ganglion are shown. The lateral part of the
clypeus has been cut off, and the salivary pump is omitted.

Figure 28. Medial view of the left antenna, with the medial walls of the
scape and the pedicel removed to show the muscles. The cranial exoskeleton
has been cut away from the medial side of the base of the antenna to reveal
the articular membrane.

Insertion: On the epipharyngeal tendons of Group 2, between the
striated plates.

Action: Elevates the epipharynx in the region of the striated plates,
thus dilating the food pump; raises and lowers the striated
plates.

Similar muscle: Posterior part of Muscle Group 1 (Pelocoris ;
Marks, 1959).

PARSONS: HEAD OF GELASTOCORIS 39

16. M. DILATOR CIBARII TERTIUS (Figs. 26, 27)

This unpaired muscle is contiguous with M. dilator cibarii quartius,
differing from it only in its insertion. The two muscle groups
are difficult to separate in dissections. They form an extensive,
fan-shaped bundle, which runs posterodorsally from its points
of insertion, passing ventral to M. protractor setae mandibu-
laris secundus.

Origin: U-shaped medial ridge of the eclypeus, between the ocelli.

Insertion: On the epipharyngeal tendons of Group 3, on the united
posteromedial corners of the striated plates.

Action: Moves the striated plates back and forth in an anterior-
posterior direction.

Similar muscle: Muscle Group 2 (Pelocoris ; Marks, 1959).

17. M. DILATOR CIBARII QUARTIUS (Figs. 13A, 26, 27)

See comments under the preceding muscle. The frontal ganglion
lies within this muscle group, just anterior to the few most
posterior strands. Strictly speaking, therefore, these last few
strands (Fig. 27, 17 A) should be termed M. dilator praepha-
ryngis, and the point of their insertion represents the anterior
pharynx. The cibarial and pharyngeal elements appear, how-
ever, to be merged into one muscle group.

Origin: In common with M. dilator cibarii tertius.

Insertion: On the epipharyngeal tendons of Group 4, posterior to
the striated plates.

Action: Elevates the posterior part of the epipharynx, thus dilating
the food pump.

Similar muscle: Muscle Group 3 (Pelocoris; Marks, 1959).

Muscles of the posterior pharynx
18. M. DILATOR POSTPHARYNGIS VENTRALIS (Figs. 25, 27)

A very slender, paired muscle.

Origin: Tip of the hypopharyngeal wing.

Insertion: Ventrolateral surface of the posterior pharynx, just
posterior to where the pharynx emerges from between the
brain and the subesophageal ganglion.

Action: Acting along with M. dilator postpharyngis dorsalis,
dilates the posterior pharynx.

Similar muscles: Mm. dilatatores oesophagi ‘‘mre’’ (Naucoris ;
Becker, 1929); Mm. dilatatores postpharyngis ventrales
(Corizxa; Benwitz, 1956).

40) BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

19. M. DILATOR POSTPHARYNGIS DORSALIS (Figs. 25, 27)

A very slender, paired muscle.

Origin: Dorsal margin of the occipital foramen, just medial to
Mm. retractores setarum mandibularis secundus and maxillaris.

Insertion: Dorsolateral surface of the posterior pharynx, dorsal to
M. dilator postpharyngis ventralis.

Action: Acting along with M. dilator postpharyngis ventralis,
dilates the posterior pharynx.

Similar muscles: Mm. dilatatores oesophagi ‘‘mr,’’? (Naucoris ;
Becker, 1929); dilator of postpharyngeal region (khampho-
coriza ; Griffith, 1945) ; Mm. dilatatores postpharyngis dorsales
(Coriza ; Benwitz, 1956).

’

Muscles of the antenna

20. M. LEVATOR SCAPI (Figs. 138A, 25, 26, 28)

A slender muscle, on either side of the head, joined with M. de-
pressor scapi for most of its length. These two muscles separate
just before their insertions.

Origin: Roof of the cranium, on the more anterior of the two short
grooves lying lateral to the ocellus (see Fig. 1).

Insertion: Slender sclerotized tendon in the anterior part of the
articular membrane joining the first antennal segment (scape)
to the cranium.

Action: Raises the entire antenna (pulls it anteriorly ).

Similar muscles: Part of ‘‘Muskelziige des Kopfes inseriert an der
Gelenkhaut 1’’ (Nepa and Ranatra; Hiifner, 1939); M. levator
2 seapi (Coriaa; Benwitz, 1956).

21. M. DEPRESSOR SCAPI (Figs. 138A, 25, 26, 28)

See comments for the preceding muscle.

Origin: In common with M. levator scapi.

Insertion: Slender sclerotized tendon in the posterior part of the
articular membrane joining the scape to the cranium.

Action: Lowers the entire antenna (pulls it posteriorly).

Similar muscles: Part of ‘‘Muskelziige des Kopfes imseriert an
der Gelenkhaut I’’ (Nepa and Ranatra; Hiifner, 1939); M.
depressor scapi (Coriaa; Benwitz, 1956).

22. M. SCAPO-PEDICELLARIS MEDIALIS (Fig. 28)
A short, broad muscle.
Origin: Anterior wall of the scape, just medial to M. scapo-pedi
cellaris lateralis.

PARSONS: HEAD OF GELASTOCORIS 4]

Insertion: Ventromedial margin of the base of the second antennal
segment (pedicel).

Action: Moves the last three antennal segments ventromedially.

Similar muscles: Part of ‘‘Scapus-muskulatur’’ (Notonecta, Nau-
coris, and Aphelocheirus; Hiifner, 1939).

23. M. SCAPO-PEDICELLARIS LATERALIS (Fig. 28)

A short, broad muscle.

Origin: Anterior wall of the scape, just lateral to M. seapo-pedi-
cellaris medialis.

Insertion: Dorsolateral margin of the base of the pedicel.

Action: Moves the last three antennal segments dorsolaterally.

Similar muscles: Part of ‘‘Scapus-muskulatur’’ (Notonecta, Nau-
coris, and Aphelocheirus; Hiifner, 1939).

Muscle of the cephalic gland

24. M. DILATOR ORIS GLANDULAE CAPITIS (Figs. 13B, 25)

A long, slender muscle, lying ventral to all other structures in the
lateral part of the head.

Origin: Tip of the hypopharyngeal wing.

Insertion: Selerotized flap on the medial wall of the mouth of the
cephalie gland.

Action: Pulls the mouth of the cephalic gland medially, to lie
beneath the orifice in the exoskeleton, thus allowing secretion
to escape to the exterior (see Parsons, in press, where this
muscle is termed ‘‘muscle inserting on sclerotized flap’’).

Similar muscles: Das Reservoir Offnender Muskel (Naucoris ;
Becker, 1929).

Muscle of the salivary pump

25. M. RETRACTOR PISTILLI (Figs. 13A-C, 25)

A well-developed, paired muscle.

Origin: Posterior portion of the hypopharyngeal wing, yeutro-
medially and ventrolaterally.

Insertion: Side of the plate-like tendon of the salivary pump.

Action: Retracts the invaginated posterior wall of the pump, thus
dilating it.

Similar muscles: Muskel der Speichelpumpe (Naucoris; Becker,
1929); aspirator muscles (Nepa; Hamilton, 1931); dilator
muscles of the syringe (Notonecta; Butt, 1943); M. retractor
pistilli (Triatoma; Barth, 1952; and Coriza; Benwitz, 1956) ;
muscles of salivary syringe (part only) (Leptocorisa; Akbar,
1957).

42 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

DISCUSSION

The three littoral families Gelastocoridae, Ochteridae, and
Saldidae share several external structural similarities which sug-
gest that they are closely related to each other. China (1933) has
expressed the belief that all the littoral families had a common
origin. The evidence presented in the literature seems to indicate
that the gelastocorids arose from ochterid-lke ancestors (Reuter,
1910; China, 1933, 1955b), and that the shore-dwelling Heter-
optera are rather generalized members of that order. Several
primitive features of their anatomy support the latter view. Two
cephalic characteristics, the possession of ocelli and four-seg-
mented antennae, are regarded by Reuter (1910) and China
(1955b) as primitive. Some authors have considered the Saldidae
(Osborn, 1895; Ekblom, 1929) or the Ochteridae (Reuter, 1910)
to be the most primitive living families of the Heteroptera.

Most authorities agree that the lttoral bugs represent a stage
in the evolution of the aquatic or semi-aquatic forms, their habi-
tat being a transition between the terrestrial and aquatic environ-
ments (in the following discussion, the terms ‘‘aquatic’’ and
‘*Hydrocorisae’’ will be used for those families which live be-
neath the surface of the water, while ‘‘semi-aquatic’’ or ‘‘Am-
phibicorisae’’ will refer to those which inhabit the surface film).
However, differences of opinion have arisen as to which of the
littoral families are related to the semi-aquatie forms, and which
to the aquatie ones.

The division of the Heteroptera into two groups, based upon
the condition of the antennae, has been used by many workers
such as Handlirsch (1908). The terrestrial and semi-aquatic
families, along with the Saldidae, are included in the Gym-
nocerata because of their long, free antennae, while the totally
aquatie families plus the gelastocorids and ochterids are termed
Cryptocerata. In the latter group, the antennae are relatively
short and, exeept in Aphelocheirus and in the Ochteridae, are
concealed beneath the eyes. This system of classification was
criticized by Reuter (1910), who, feeling that the gymnocerate-
eryptocerate split was not a natural division, set up six heter-
opteran series. In the Series Hydrobiotica he grouped all the
littoral, aquatic, and semi-aquatic families, placing the rest of

PARSONS: HEAD OF GELASTOCORIS 43

the Gymnocerata in five other series. His system was based pri-
marily upon the character of the wings, the eggs, and the sternal
selerites, features which, according to Spooner (1938), have been
shown to be rather undependable for phylogenetic purposes.

The relationship between the littoral and the totally aquatic
forms has not been generally agreed upon. De la Torre-Bueno
(1923) considered the totally aquatic families to be derived from
saldid-like ancestors, the Ochteridae representing an interme-
diate condition. Ekblom (1929) also believed the saldids to
represent the ancestral aquatic bugs. China (1955b), however,
proposed that the totally aquatic families arose from ‘‘Proto-
Ochteridae,’’ while the ‘‘Proto-Saldidae’’ gave rise to the semi-
aquatic forms. His system corresponds, in this respect, to the
old gymnocerate-cryptocerate division, although he used Du-
four’s (1833) terms Hydrocorisae (for the aquatic families plus
the ochterids and gelastocorids), Geocorisae (for the terrestrial
eroups), and Amphibicorisae (for the semi-aquatic forms and
the saldids).

Spooner (1938) considered the ochterids and gelastocorids to
form a bridge between the aquatic and semi-aquatic bugs, and,
unlike China, he believed them to be-more closely related to the
latter than to the former. His classification was based primarily
upon the structure of the labrum and the mandibular lever.
Using the terminology of Dufour (1833) he placed the Ochteri-
dae, Gelastocoridae, and semi-aquatie families in the Amphibi-
eorisae, the Saldidae and terrestrial families in the Geocorisae,
and all the aquatic forms except the corixids in the Hydrocorisae.
He considered the Corixidae different enough to constitute a
fourth group, the Sandalhorrhyncha (a term proposed by Borner,
1904).

The problem of the phylogenetic position of the Saldidae and
Ochteridae deserves further attention, and more morphological
work upon these two families is needed. The present study has
revealed a few features in the structure of the gelastocorid head
which may shed some light on whether the Gelastocoridae are
more similar to the aquatic or to the semi-aquatie Heteroptera.
This purpose is somewhat hindered by the fact that only one of
the two genera of Gelastocoridae is here represented. Another
obstacle is that, in general, more morphological work has been

44 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

done upon the aquatic families than upon the semi-aquatic ones,
and hence it is easier to make comparisons with the former than
with the latter.

Several features in the head of Gelastocoris link this insect
with the aquatic Heteroptera. Some of these are seen in its ex-
ternal anatomy. First, there is the absence of a distinct ante-
clypeus; of the genera examined by Spooner (1938), only four
showed this condition; Gelastocoris, Ochterus (Ochteridae), and
the aquatic bugs Pelocoris (Naucoridae) and Notonecta (No-
tonectidae). A second similarity between Gelastocoris and the
latter two genera is the absence of external loral or paraclypeal
sclerites in all three. Spooner claimed that these sclerites are
smaller in Pelocoris and Notonecta than in any other forms which
he examined; his mistaking, in Gelastocoris, of the maxillary
plate for the external lorum has already been discussed. The
present author has examined Notonecta and Pelocoris and con-
cluded that their lora, lke those of Gelastocoris, are contained
entirely within the head. Butt (1943) reported the absence of
external lora in Notonecta, and Rawat (1939) does not mention
any loral sclerite in Naucoris (Naucoridae).

A third external resemblance to the aquatie forms lies in the
shortness and stoutness of the labium. The beaks of the semi-
aquatic bugs are, in general, longer and more slender. A fourth
similarity is the presence of the two intersegmental sclerites
between the last two segments of the labium. This character is
not diagnostic, however, since these sclerites are not restricted
to the aquatic Heteroptera; Leon (1897) has reported their pres-
ence in the semi-aquatic genera Velia and Gerris. The short,
concealed antennae form a fifth point of resemblance between
velastocorids and the Hydrocorisae, but one which may be less
significant than the first three preceding ones. Reuter (1910)
suggested that the reduced length of the antennae of Ochterus
may be an adaptation to the burrowing habit, and Todd (1955)
has also proposed this with regard to the gelastocorids. This
antennal reduction could have taken place in the ancestral Hy-
drocorisae (the littoral ‘‘Proto-Ochteridae’’ of China, 1955b)
as an adaptation to burrowing; it could also be argued, however,
that the reduction did not occur in the Hydrocorisae until after
they entered the water, and that in the gelastocorids it has been

re

PARSONS: HEAD OF GELASTOCORIS 45

an independent development. Hiifner (1939) has deseribed
much variation in the general form of the antennae of the aquatic
Heteroptera. It seems, therefore, that the length and form of the
gelastocorid antenna is a somewhat unrehable phylogenetic char-
acter.

Five internal characteristics of the gelastocorid head may in-
dicate a relationship with the aquatic Heteroptera. The food
pump of Gelastocoris is in many ways similar to that of Pelocoris,
and also shows some resemblances to that of Notonecta. Since
this is a point worthy of some discussion, it will be considered
in some detail later. Secondly, as was mentioned earlier, the
oblique plate of the labium is also found in the aquatic families
Notonectidae, Naucoridae, Belostomatidae, and Nepidae; to the
author’s knowledge, it has never been reported in any semi-
aquatie or terrestrial forms. The striking similarity of the labial
musculature of Gelastocoris to that of Notonecta forms a third
point of comparison; the two bugs appear to be identical in this
character except that Notonecta possesses muscles in the terminal
segment (Butt, 1943) which are lacking in Gelastocoris. A fourth
internal similarity is the presence, in Gelastocoris, of a maxillary
lever. Ekblom (1929) found this structure in the Geocorisae
and in some of the Hydrocorisae which he examined, but re-
ported its absence in the Amphibicorisae. The fifth resemblance
is the presence of cephalic glands in Gelastocoris; this has been
discussed in another paper (Parsons, in press). Such glands
are present in representatives of all the families of water bugs,
but have never been reported in the semi-aquatic forms.

According to Spooner (1938), two features of the heads of
velastocorids and ochterids link them more closely to the semi-
aquatic than to the aquatic forms. His first and most convincing
piece of evidence is the presence of the epipharyngeal projection
extending beyond the apex of the Jabrum. Spooner studied a
wide variety of genera, and found this projection only in
the semi-aquatic bugs and in Ochterus and Gelastocoris. The
present study confirms its presence in the last. His evidence from
the mandibular levers, however, is less convincing. Spooner cites
the work of Ekblom (1929), who distinguished four types of
mandibular levers. The first of these is typical of Geocorisae, the
second of Hydrocorisae, the third of Sandaliorrhyncha, and the

46 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

fourth of Amphibicorisae. Type II is three-branched and at-
tached directly to the stylet, while Type IV is quadrangular and
attached to the stylet by a projection from one of its faces.
Spooner considered the mandibular levers of Ochterus and
Gelastocoris to be of this last type.

Spooner’s argument on the grounds of the mandibular lever
appears weak for two reasons. First, he found much variation,
especially within Ekblom’s Types I and II; in the latter he
found three fairly distinct sub-types. China (1955a) has already
questioned the phylogenetic significance of the mandibular levers
in the Geocorisae. Secondly, Spooner admits that the levers of
Ochterus and Gelastocoris differ somewhat from the others of
Type IV. They are triangular in shape rather than quadrangu-
lar. In Ochterus, the projection joining the lever to the stylet
is very short and is attached to the edge of the lever rather than
to its face, while in Gelastocoris the projection appears to be
absent, the lever articulating directly with the stylet. In the
author’s opinion, the mandibular levers of these two insects, as
described by Spooner, bear more resemblance to the three-
branched levers of the Hydrocorisae, and should not be placed
in the same group as those of the Amphibicorisae. In Gelastocoris,
as was earlier described, the two posterior edges of the lever are
heavily sclerotized while the middle part is membranous, con-
tinuing anteriorly into the dorsal wall of the genal sac. A similar
condition is found in the lever of Pelocoris. Spooner diagrammed
only a slender, elbow-like bent rod for this genus; I have exam-
ined the Pelocoris lever, however, and found it to have a triangu-
lar, membranous, middle part which, like that of Gelastocoris,
passes into the wall of the genal sac. Spooner states that the
attachment of the levers to the stylets is direct in Ochterus and
Gelastocoris ; this is also characteristic of the levers of the Hydro-
corisae, as described by him.

The cephalic characters of the gelastocorids appear, to the
author, to be more like those of the aquatic than of the semi-
aquatic Heteroptera. The epipharyngeal projection seems to be
the best evidence in support of Spooner’s theory; however, the
condition of the lora and labium, the structure of the mandibular
levers, the absence of a distinct anteclypeus, and the presence of
maxillary levers and cephalic glands all tend to support China’s

PARSONS: HEAD OF GELASTOCORIS 47

hypothesis. The several resemblances between Gelastocoris and
the naucorid Pelocoris are interesting in view of China’s (1955b)
proposal that the Naucoridae are the most generalized of the
totally aquatic families.

The structure of the food pump of Gelastocoris is worthy of
some discussion. The very thorough studies of Marks (1958 and
1959) make it possible to compare this structure with that of
several aquatic Heteroptera. Unfortunately, no similar study
has been made of the food pumps of the semi-aquatic forms, so
that no comparisons can be made with those families.

Marks found a relatively simple pump in Ranatra and Belos-
toma and a more complicated one in Hesperocoriza, Pelocoris,
and Notonecta. The resemblances between Gelastocoris and the
last two forms are rather marked. Marks’ diagrams (1959, Figs.
2, 4) of lateral views of the pumps of Pelocoris and Notonecta
show them to be very similar to that of Gelastocoris in general
shape, and to possess certain epipharyngeal specializations in the
same region in which the gelastocorid striated plates occur. These
specializations consist of two heavily sclerotized, tooth-bearing
transverse bars, one on either side, which, like the striated plates,
are joined at the midline. At the point of junction are tendons
for muscle attachment, similar to those of Group 3 in Gelasto-
coris. Notonccta possesses a second pair of bars anterior to these,
while Pelocoris has a series of toothed folds concealed beneath
the transverse bars. Marks proposed that these epipharyngeal
elaborations are straining or filtering devices, as may be the case
in Gelastocoris.

Whether or not the striated plates ean be homologized with the
transverse epipharyngeal bars of Pelocoris or Notonecta is a
matter of conjecture. It is difficult to imagine how either could
have developed from the other, although it is possible that the
striated plates could have arisen from the bars plus the tooth-
bearing folds of Pelocoris. It seems more likely, however, that
both the bars and the striated plates, if homologous, developed
from an intermediate ancestral condition. Marks (1959) was
uncertain as to whether the transverse bars of Pelocoris are
homologous with those of Notonecta. He stated that although
the similarities between them are very great, the muscles activat-
ing the bars of Notonecta are cibarial in origin (anterior to the

48 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

frontal ganglion), while those of Pelocoris are pharyngeal (pos-
terior to the frontal ganglion).

While considering this question, the author dissected food
pumps of both Notonecta and Pelocoris and found that in the
latter the frontal ganglion is located much farther posteriorly
than Marks described it. It is in exactly the same position,
relative to the muscles of the pump, as that of Gelastocoris ; only
a few thin strands of muscle (from the posterior portion of
Marks’ Muscle Group 3) lie posterior to the ganglion, the rest
being anterior to it. Therefore, most of the Pelocoris pump is, like
that of Gelastocoris, cibarial in nature, and consequently it ap-
pears quite probable that the transverse bars of Pelocoris and
Notonecta are homologous. The great similarity between the
musculature of the food pumps of Pelocoris and Gelastocoris
makes it possible to suggest the probable muscle homologies which
were listed earlier. Mm. dilatores cibarii primus and secundus
of Gelastocoris seem to correspond to Marks’ Muscle Group 1,
and M. dilator cibarii tertius, activating the striated plates,
probably represents Marks’ Muscle Group 2, which moves the
transverse bars. M. dilator cibarii quartius resembles Marks’
Muscle Group 3; in both insects, the frontal ganglion is found
within this muscle group.

Although the resemblances between the food pumps of Gelasto-
coris and Pelocoris are strong enough to suggest these few homol-
ogies, the author will not attempt to make similar comparisons
between Gelastocoris and the other aquatic bugs, since the re-
lationships between the groups studied by Marks are not clear.
The reader is referred to that author’s two excellent papers
(1958 and 1959) for a complete discussion of this problem.

One of the chief obstacles encountered by Marks in his attempt
to homologize the parts of the food pumps in different aquatic
bugs was the position of the frontal ganglion in relation to the
various muscle groups. In the author’s opinion, too much em-
phasis has been placed on the distinction between cibarial and
pharyngeal muscles. The origins and insertions of these two
groups of muscles have been used by Snodgrass (1947) and many
other authors as criteria not only for distinguishing the cibarial
from the pharyngeal portions of the food pump but for deter-
mining the boundary between the frons and the elvpeus. Dnu-
Porte (1946), in discussing the facial sclerites, pointed out that

PARSONS: HEAD OF GELASTOCORIS 49

the origins of the pharyngeal or cibarial muscles might easily
shift from one sclerite to another, and Ferris (1944) has ex-
pressed a similar opinion. It seems possible that the insertions
of such muscles could also shift — that muscles posterior to the
frontal ganglion could come to insert on a cibarial portion of the
food pump, or that cibarial muscles might shift their insertions
to the pharyngeal portion. In the insects studied by Marks, the
frontal ganglion varies, in its position relative to the muscles,
from one species to the next. It may be that these differences are
not significant but, as Marks (1959) has suggested, can be ex-
plained by muscle shifts.

LITERATURE CITED

AKBAR, S. 8.
1957. The morphology and life-history of Leptocorisa varicornis Fabr.
(Coreidae, Hemiptera) —a pest of paddy crop in India. Part 1.
Head and thorax. Aligarh Muslim Uniy. Publ. (Zool, Ser.) Ind.
Ins. Typ., vol. 5, pp. 1-53.

BartH, R.
1952. Anatomische und histologische Studien ueber die Unterfamilie
Triatominae (Heteroptera, Reduviidae). I. Teil: Der Kopf von
Triatoma infestans. Mem. Inst. Oswaldo Cruz, vol. 50, pp. 156-
196.

BECKER, E.
1929. Zum Bau des Kopfes der Rhynchoten. I. Teil. Bau des Kopfes
von Naucoris cimicoides L. Rev. Zool. Russe, vol. 9, pp. 54-96.

BENWITZ, G.
1956. Der Kopf von Corixa punctata Ill. (geoffroyi Leach) (Hemip-
tera-Heteroptera). Zool. Jahrb. (Abt. Anat.), vol. 75, pp.
311-378.

BOoORNER, C.
1904. Zur Systematik der Hexapoden. Zool. Anz., vol. 27, pp. 511-
533.

Burt, F. H.
1943. Comparative study of mouth parts of representative Hemiptera-
Homoptera. Mem. Cornell Univ. Agric. Exp. Sta., no. 254, 20 pp.

50 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

CHINA, W. EH.

1933. A new family of Hemiptera-Heteroptera with notes on the
phylogeny of the suborder. Ann. Mag. Nat. Hist., ser. 10, vol.
12, pp. 180-196.

1955a. A reconsideration of the systematic position of the family Jop-
peicidae Reuter (Hemiptera-Heteroptera), with notes on the
phylogeny of the suborder. Ann. Mag. Nat. Hist., ser. 12, vol.
8, pp. 353-370.

1955b. The evolution of the water bugs. Bull. Nat. Inst. Sei. India,
vol. 7, pp. 91-103.

CRAMPTON, G. C.
1921. The sclerites of the head, and the mouthparts of certain imma-
ture and adult insects. Ann. Ent. Soc. Amer., vol. 14, pp. 65-

110.

Durour, M. L.
1833. Recherches anatomiques et physiologiques sur les Hémipteéres.
Mém. Savy. Etrang. Acad. Sci. Paris, vol. 4, pp. 129-462.

DuPorts, E. M.
1946. Observations on the morphology of the face in insects. J.
Morph., vol. 79, pp. 371-417.

EKBLOM, T.

1926. Morphological and biological studies of the Swedish families of
Hemiptera-Heteroptera. Part I. Zool. Bidrag fran Uppsala,
vol. 10, pp. 31-179.

1929. New contributions to the systematic classification of Hemiptera-
Heteroptera, Ent. Tidsk., vol. 50, pp. 169-180.

1930. Morphological and biological studies of the Swedish families of
Hemiptera-Heteroptera. Part II. Zool. Bidrag fran Uppsala,
vol. 12, pp. 113-150.

FERRIS, G. F.
1944. On certain evolutionary tendencies in the heads of insects.
Microentomology, vol. 9, pp. 78-84.

GRIFFITH, M. E.
1945. The environment, life history and structure of the water boat-
man, Rhamphocoriza acuminata (Uhler) (Hemiptera, Corixi-
dae). Kansas Univ. Sei. Bull., vol. 30, pp. 241-366.

HAMILTON, M. A.
1931. The morphology of the water-scorpion, Nepa cinerea Linn.
(Rhynchota, Heteroptera). Proc. Zool. Soc. London, vol. 1931,
pp. 1067-1136.

PARSONS: HEAD OF GELASTOCORIS 5]

HANDLIRSCH, A.
1908. Die fossilen Insekten und die Phylogenie der rezenten Formen.
Leipzig, pp. 1244-1249.

HuFner, B.
1939. Die Antennen und antennalen Sinnesorgane der Hydrocores mit
besonderer Beriicksichtigung der Nepiden. Zoologica, vol. 35,
no. 97, pp. 1-32.

HUNGERFORD, H. B.
1922. The life history of the toad bug (Heteroptera). Kansas Univ.
Sci. Bull., vol. 14, pp. 145-171.

Kemper, H.
1932. Beitrage zur Biologie der Bettwanze (Cimex lectularius L.).
III. Uber den Mechanismus des Stech-Saugaktes. Zeitschr.
Morph. Okol. Tiere, vol. 24, pp. 491-518.

Leon, N.
1897. LBeitrige zur Kenntnis des Labiums der Hydrocoren. Zool. Anz.,
vol. 20, pp. 73-77.

MARKS, E. P.
1958. A comparative study of the foregut of several aquatic Hem-
iptera. J. Kansas Ent. Soce., vol. 31, pp. 138-155.
1959. The food pump of Pelocoris and comparative studies on other
aquatic Hemiptera. Psyche, vol. 64, pp. 123-134 (vol. dated
1957).

Osporn, H.
1895. The phylogeny of Hemiptera. Proc. Ent. Soc. Washington, vol.
3, pp. 185-190.

PARSONS, M. C.
In press. Cephalic glands in Gelastocoris (Hemiptera-Heteroptera).
Psyche.

QapRI, M. A. H.
1951. On the anatomy of the mouth-parts and the mode of feeding in
the aquatic bugs (Cryptocerata). Proc. Zool. Soe. Bengal, vol.
4, pp. 117-128.

Rawat, B. L.
1939. Notes on the anatomy of Naucoris cimicoides L. (Hemiptera
Heteroptera). Zool. Jahrb. (Abt. Anat.), vol. 65, pp. 535-600.

52 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Reuter, O. M.
1910. Neue Beitraige zur Phylogenie und Systematik der Miriden, nebst
einleitenden Bemerkungen iiber die Phylogenie der Heteropteren-
Familien. Acta Soc. Sci. Fennicae, vol. 37, no. 3, pp. 1-172.

SNODGRASS, R. E.

1935. Principles of insect morphology. McGraw-Hill, New York and
London.

1938. The loral plates and the hypopharynx of Hemiptera. Proc. Ent.
Soc. Washington, vol. 40, pp. 228-236.

1947. The insect cranium and the ‘‘epicranial suture.’’ Smithsonian

Mise. Coll., vol. 107, no. 7, pp. 1-52 (publ. no. 3896).

SPOONER, C. S.
1938. The phylogeny of the Hemiptera based on a study of the head
capsule. Illinois Biol. Monog., vol. 16, no. 3, pp. 1-102.

SPRAGUE, I. B.
1956.

The biology and morphology of Hydrometra martini Kirkaldy.

Kansas Univ. Sci. Bull., vol. 38, pp. 579-693.

Topp, E. L.
1955.

A taxonomie revision of the family Gelastocoridae (Hemiptera).

Kansas Univ. Sci. Bull., vol. 37, pp. 277-475.

TorRE-BUENO, J. R., DE LA
1923. Family Saldidae.
Connecticut. Part IV.

In: W. E. Britton, Guide to the insects of
The Hemiptera or sucking insects of

Connecticut. Hartford, pp. 408-416.

ABBREVIATIONS USED IN FIGURES

A — antenna

AD — afferent salivary duct

AF -—anteriormost portion of hypo-
pharynx of food pump

AM — articular membrane of antenna

B-—brain (cut edge)

BL —base of first labial segment

C —clypeus

CE — compound eye

CN —clypeal connectives

CO — orifice of cephalic gland

CR —cranial exoskeleton (cut edge )

D — dorsal suture

DG — dorsal wall of genal sac

DH — dorsal surface of hypopharyn-
geal wing

FE —epipharynx of food pump

ED — efferent salivary duct

EP —epipharyngeal projection

HR —-epipharyngeal ridge

IF —food pump

I°'C — food canal

FG-—frontal ganglion (cut edge)

G — gula

G1 —tendons of Group 1

G2 — tendons of Group 2

PARSONS: HEAD OF GELASTOCORIS O38

G3 —tendons of Group 3

G4 —tendons of Group 4

GL — gular lobe

H — hypopharynx of food pump

HF — hypopharyngeal flap

HW — hypopharyngeal wing

IM — intersegmental membrane

IS — intersegmental sclerite between
third and fourth labial segments

L —labium

LA — lateral apodeme

LI — first labial segment

LIT — second labial segment

LIII — third labial segment

LIV — fourth labial segment

LB —labrum

LF —\umen of food pump

LO —\orum

MD — mandibular stylet

MDL — mandibular lever

MG — medial wall of genal sac

MX — maxillary stylet

MXL — maxillary lever

MXP — maxillary plate

O — ocellus

OC — occipital condyle

OF — occipital foramen

OP — oblique plate

OV — oval raised area on epipharyn-
geal plate

P — postocciput

PA — pedicel (second antennal seg-
ment )

PB-pigmented border of epiphar-
ynx

PL — epipharyngeal plate

PP — posterior pharynx

S — salivary pump

SA —scape (first antennal segment)

SC — salivary canal

SD-dorsal intersegmental sclerite
between first and second labial
segments

SE —subesophageal ganglion (cut
edge)

SG — stylet groove

SP — striated plate

SU — suspensory plate

SV —ventral intersegmental sclerite
between first and second labial
segments

7 — transparent area in hypopharynx
of food pump

TA —tendon for M. depressor seapi

TM-—tendon for M. retractor setae
mandibularis secundus

TS —tendon for M. retractor pistilli

U — U-shaped ridge on cranium

V — ventral suture

VG — ventral wall of genal sac

VH — ventral surface of hypophar-
yngeal wing

VX — vertex

1—M. levator labii

2— M. depressor labii primus

3 — M. transversalis labii primus

/ — M. depressor labii secundus

5 — M. transversalis labii secundus

6 — M. transversalis labii tertius

? — M. retractor segmenti ultimi labii

8 — M. protractor setae mandibularis
primus

9 — M. protractor setae mandibularis
secundus

10 —M. retractor setae mandibularis
primus

11—M. retractor setae mandibularis
secundus

12—M. protractor setae maxillaris

13 — M. retractor setae maxillaris

14 —M. dilator cibarii primus

15 — M. dilator cibarii secundus

16 — M. dilator cibarii tertius

17 — M. dilator cibarii quartius

17, A — M. dilator praepharyngis

18—M. dilator postpharyngis ven-
tralis

19 — M. dilator postpharyngis dorsalis

20 —M. levator seapi

21 —-—M. depressor scapi

22 — M. seapo-pedicellaris medialis

23 — M. scapo-pedicellaris lateralis

24 —M. dilator oris glandulae capitis

25 — M. retractor pistilli

7 ae ee
is
wali ' ee

sil TL yon

7

Tony ar 1
_ abe ote ee

arth : ' re

=e
=

an nn “oun Pay Sh
5 oi) rer

rn

| raktt OA nt SE
- -

m sate fa

VIET s, Faith
Hen isin bose

iy dt

se af a
by hss Leake

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE

Vou. 122— No:

A PRELIMINARY REVIEW OF THE FAMILY
GONOSTOMATIDAE, WITH A KEY TO THE GENERA
AND THE DESCRIPTION OF A NEW SPECIES
FROM THE TROPICAL PACIFIC

By Marion GREY

CAMBRIDGE, MASS., U.S.A.
PR Net HD. ek OFR han Ey Mat) Ske Uy MoM

Frpsruary, 1960

PUBLICATIONS ISSUED BY OR IN CONNECTION
WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

BULLETIN (octavo) 1863 — The current volume is Vol. 122.

BREVIORA (octavo) 1952 — No. 120 is current.

Memorrs (quarto) 1864-1938 — Publication was terminated with
Vol. 55.

JOHNSONIA (quarto) 1941— A publication of the Department of
Mollusks. Vol. 3, no.39 is current.

OccASIONAL PAPERS OF THE DrEPARTMENT oF MOLLUSKS (octavo)
1945 — Vol. 2, no. 23 is current.

PROCEEDINGS OF THE NEW ENGLAND ZoouocicaL CLUB (octavo)
1899-1948 — Published in connection with the Museum. Publication
terminated with Vol. 24.

The continuing publications are issued at irregular intervals in num-
bers which may be purchased separately. Prices and lists may be
obtained on appheation to the Director of the Museum of Comparative
Zoology, Cambridge 38, Massachusetts.

Of the Peters ‘‘Check List of Birds of the World,’’ volumes 1-3 are
out of print; volumes 4 and 6 may be obtained from the Harvard Uni-
versity Press; volumes 5 and 7 are sold by the Museum, and future
volumes will be published under Museum auspices.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE
Von22: No; 2

A PRELIMINARY REVIEW OF THE FAMILY
GONOSTOMATIDAE, WITH A KEY TO THE GENERA
AND THE DESCRIPTION OF A NEW SPECIES
FROM THE TROPICAL PACIFIC

By MARIoN GREY

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM

Frsruary, 1960

No. 2—A Preliminary Review of the Family Gonostomatidae,
with a Key to the Genera and the Description of a
New Species from the Tropical Pacific.

By MArIon GREY

In recent years collections made aboard the research vessel
Oregon of the United States Fish and Wildlife Service, have
yielded a number of species of the family Gonostomatidae.
Among these were the first known western Atlantic specimens of
the genera Triplophos and Argyripnus; a new maurolicid genus
and species, Sonoda megalophthalma Grey (1959) ; new and well
preserved material of Yarrella blackfordi Goode and Bean, and
Polymetme corythaeola (Aleock); and adult specimens ol
Diplophos maderensis (Johnson). A study of these specimens,
preparatory to a revision of the family Gonostomatidae, has
resulted in the following discoveries.

1. Yarrella blackfordi possesses serial photophores on the body
other than the ventral and lateral rows deseribed on the types,
and therefore all species presently placed in the genus Yarrella
Goode and Bean, except the type species, must be excluded from
the genus.

2. Lychnopoles Garman is a synonym of Yarrella.

3. The genus Polymetme McCulloch, usually considered a
synonym of Yarrella, is reinstated to include most of the species
formerly placed in the latter.

4. Yarrella mauli Poll has been assigned to a new genus,
Pollichthys (Grey, 1959).

5. Manducus Goode and Bean is reduced to the status of a
subgenus of Diplophos Gunther.

Other material studied has produced the following results.

1. Photichthys nonsuchae Beebe has been placed in a new
genus, Woodsia, and a second specimen of this species has been
described from the eastern Pacific (Grey, 1959).

2. Snellius Koumans is a synonym of Margrethia Jespersen
and Taning.

3. Gonostoma atlanticum Norman, which was first described
as a subspecies of G. denudatum Rafinesque, is given specific
rank.

58 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

4. A new species of the genus Gonostoma Ginther is described
from the Pacific, and new records of G. gracile Giimther are
listed.

5. New material of Danaphos Bruun from the eastern Pacific
has shown that D. asteroscopus Bruun is probably a synonym of
D. oculatus (Garman).

To the following people I would like to express my apprecia-
tion for the loan of specimens, or for information or advice. Dr.
Frederick H. Berry, Dr. Henry B. Bigelow, Mr. Harvey R.
Bullis Jr., Dr. Daniel M. Cohen, Mrs. Myvanwy Dick, Dr. Alfred
W. Ebeling, Dr. Carl L. Hubbs, Dr. Einar Koefoed, Mr. G. E.
Maul, Dr. Giles W. Mead, Dr. James E. Morrow Jr., Dr. George
S. Myers (in particular), Dr. Max Poll, Dr. Andreas B. Rech-
nitzer, Prof. L. R. Richardson, Dr. C. Richard Robins, Dr. L. P.
Schultz, Miss Margaret Storey, and Dr. D. W. Tucker.

I am indebted to Dr. Rainer Zanger] for a number of X-ray
photographs, which made possible vertebral counts of most of
the species studied, and to Miss Janet Wright for assistance in
the preparation of Figure 3. Figure 1 was drawn by Mrs.
Myvanwy Dick and Figure 2 by Mr. John Pfiffner, Staff Artist,
Chicago Natural History Museum.

In spite of recent discoveries the inter-relationships of gonos-
tomatid genera are still obscure, and a separation into sub-
families has not been attempted. The maurolicid genera certainly
form a natural group but it seems unnatural and misleading
to lump the remaining genera in a single subfamily, and I ean-
not see any advantage in introducing four or five new sub-
families at this time. Future discoveries may make such division
possible, as the sea undoubtedly harbours undiscovered species,
but undiscovered species are more likely to close the gaps between
the known ones.

The maurolicid genera, given family rank by some and sub-
family rank by most authors, are distinct from other gonosto-
matids in having many of the photophores grouped together
and connected to a common gland, as well as in having fewer
photophores on the branchiostegal membranes. Maurolicid fishes
usually differ further in having smaller mouths, minute teeth in
the jaws, and well developed pseudobranchiae, but there are
exceptions. The pseudobranchiae of Sonoda and Danaphos, for

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 59

example, are small, fragile, and not always discernible; and
furthermore two non-maurolicid genera, Margrethia and Diplo-
phos (subgenus Diplophos), also possess pseudobranchiae.
Maurolicid genera show more variation among themselves in some
characters than do other groups of gonostomatid genera, especial-
ly in the relative positions of the dorsal, anal and ventral fins.

The four genera Gonostoma, Cyclothone, Bonapartia and Mar-
grethia also form a natural group, distinguished principally by
the absence of photophores on the isthmus; and the genus
Ichthyococcus is so aberrant that there would be little or no
objection to placing it alone in a separate subfamily. However,
attempts to join the remaining genera into natural groups have
not been entirely successful. Diplophos, Triplophos and Yarrella
appear to be related, sharing such characters as numerous rows
of photophores on the body, a tendency toward a long premaxil-
lary, and the absence of an adipose fin; and Photichthys,
Woodsia, Vinciguerria and Pollichthys can be grouped together
in their possession of two suborbital photophores. But Polymetme
cannot be placed in either of these groups. In general appear-
ance it is much like Photichthys but the resemblance is relatively
superficial. In its dentition and rather long premaxillary Poly-
metme approaches Yarrella but in the arrangement of the
photophores, meristie characters, and the presence of an adipose
fin it is similar to Pollichthys. Polymetme also has a tendency
toward the elongation of some of the photophores, a character
usually associated with the maurolicid genera. Some genera
have one or two common characters although they do not seem
to be otherwise closely related. For example, Danaphos and
Triplophos are the only members of the family in which the
dorsal origin is more than slightly in advance of the middle of
the body length; and tubular eyes are found in Ichthyococcus,
young Vinciguerria, Valenciennellus, and Danaphos.

Generic relationships have been tentatively arranged as shown
in Figure 1. It is impossible to state which genus is the most
primitive but it is possibly either Diplophos or Yarrella. Rech-
nitzer and Bohlke (1958, p. 15) have written of the evolutionary
trend of the family Sternoptychidae (subfamilies Gonostomati-
nae, Maurolicinae, Sternoptychinae): ‘‘. . . the body becomes
deep and foreshortened, with an enlarged head region and with

OF COMPARATIVE ZOOLOGY

BULLETIN: MUSEUM

60

‘auprEeUoysoUoNy A[ruey oY} JO VOUT Jo sdrysuorjzejot o[qissod Surmoys weiserg “[ omsyy7

D//AIIDA

aW{IWA/Og soydojdig

snonpuop
DI{4DdDUu0g
D1y4a461D
SAYLYINIOYd SAYLYIS//Od me o Dwossouoyg
DISPoOoK” SOYdo/ais{
DIssaNnBIIUl) soydoay auoysosIAD
SNIDOIOANY {YI/

soydosoy]
snudisAbiy

soydoung

snoonop

SN//AUUAIIUA] DA

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 61

a reduced number of segmental structures. The interruption of
the anal photophores [i.e., their separation into groups| is in the
same line of evolution, and the out-of-line arrangement of the
anterior ventral photophores is an obvious specialization. An
enlargement of the photophores, especially in the vertical axis,
marks both the species |[/chthyococcus] sequence and the sub-
family sequence.’’? Such trends were demonstrated by these
authors in the genus Ichthyococcus, with elongatus Imai the most
primitive of three species and irregularis Rechnitzer and Bohlke
the most advaneed. A similar trend is evident within the genera
Diplophos and Yarrella but on the whole the genera of the family
Gonostomatidae, including maurolicids, cannot be fitted into such
an evolutionary picture with any feeling of confidence. It is
possible to trace a progression from Diplophos to Ichthyococcus
through the following genera: Yarrella > Polymetme — Pol-
lichthys > Photichthys > Vinciguerria > Woodsia, but in the
remaining groups of genera there may be more than one line of
evolution and these are obscure at best. Gonostoma and Bona-
partia are obviously closely related, as are Gonostoma and Cyclo-
thone, but Margrethia, which is the only short and deep-bodied
member of this group, shows relationship to all three of the other
genera without a strong resemblance to any of them. As for the
maurolicid genera, even after disposing of Argyripnus as an
aberrant offshoot, we are left with six genera which defy any
kind of orderly arrangement from primitive to specialized or
advanced except that Maurolicus does have a short, deep body
and a relatively low vertebral count and could therefore possibly
be considered the most specialized. And Neophos, having fewer
of the photophores in groups, a large mouth, and an elongate
body, is probably the most primitive (its vertebral count is not
known). In Figure 1 the maurolicid genera have been placed
near Diplophos, Margrethia and Triplophos for the following
reasons.

1. According to Brauer (1908, p. 27) the photophore structure
of Triplophos is very similar to that of maurolicids.

2. Margrethia is similar in several respects to maurolicids
(pseudobranchiae present, some of the photophores elongated
and close together) and there is further evidence also that mau-
rolicids stem from something near the Gonostoma-Margrethia

62 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

group of genera. In both groups the distance between the ventral
bases and the anal origin is short and the VAV count is corre-
spondingly low (with the exception of the newly discovered
Gonostoma ebelingt, which has 10 VAV and a rather long space
between ventrals and anal). And in both there is a tendency
toward a more anteriorly situated anus than in most other
eonostomatids.

3. Diplophos and the maurolicid group show a possible com-
mon origin in the possession of pseudobranchiae (subgenus
Diplophos) and in having the spines on the inner edge of the
first gill arch absent or reduced to rudiments. These spines are
well developed, even though usually short, in all other gonosto-
matids, and it may be significant that they are minute in
Ichthyococcus, Bonapartia and Margrethia. It is also interesting
to note that in some specimens of Diplophos (subgenus Diplo-
phos) the rows of minute photophores found below the eye and
along the lower jaw appear to be contained in a common black
membrane, although I am unable to assert that this structure
is similar to the connective membrane of the photophores of
maurolicids.

Of the twenty genera included in the family Gonostomatidae
at least nine are monotypi¢ and it is sometimes difficult to judge
whether differences between species are generic or specific. How-
ever, where several species of a genus are known the interspecific
differences are usually slight and the existence of numerous
genera seems to be natural. Diplophos, Yarrella and, especially,
Gonostoma are exceptional in having more or less sharply dif-
ferentiated species. A number of characters seem to be highly
plastie throughout the family and even sometimes within a genus.
There are also areas of stability but these are notable for their
exceptions, as can be seen by the following examples.

1. Until recently all described maurolicid genera had only 4-6
VAV photophores; and no known species of Gonostoma had more
than 5. However, the recently described maurolicid, Sonoda
megalophthalma, has a VAV count of 7-8; and the newly dis-
covered Gonostoma ebelingi has 10 VAV photophores.

2. In all genera except some maurolicids and Ichthyococcus
the ventral fins are inserted in advance of the dorsal origin. They
are behind the first dorsal ray in [chthyococcus and are variable

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 63

in position in maurolicid genera (behind the dorsal origin in
Danaphos, well ahead of it in Neophos, Thorophos, Valencien-
nellus and Sonoda, and close to a vertical from the first dorsal
ray in Argyripnus and Maurolicus).

3. The dorsal origin in non-maurolicid genera is in advance
of the anal origin in all but Gonostoma, Cyclothone and Bona-
partia. In these three genera either the dorsal and anal origins
are opposite one another or the anal is farther forward, but in
Margrethia, the fourth member of this group, the dorsal origin
is always shehtly in front of the anal origin. In maurolicid
genera, again, the relative positions of these fins is variable
(dorsal origin anterior to anal origin in Argyripnus, Danaphos
and Maurolicus, opposite or behind it in Sonoda, Valenciennellus,
Thorophos and Neophos).

4. The presence or absence of an adipose fin is a relable
generic character except in Gonostoma, in which this fin is pres-
ent in all but two species, atlanticum and gracile. It is also very
small and difficult to distinguish in several maurolicid genera.

5. In general the anus is situated close to the origin of the
anal fin in non-maurolicid genera and more remote from it in
maurolicids. Again there are exceptions, for the anus is situated
well in advance of the anal origin in the genus Cyclothone and
in two species of Gonostoma (ebclingi, gracile) ; and it is near
the anal fin in three maurolicid genera (Neophos, Valencien-
nellus, Maurolicus).

6. The mouth is large and oblique in all non-maurolicid genera
excepting Ichthyococcus, in which it is relatively small. Three
maurolicids also have large, oblique mouths (Neophos, Thorophos,
Argyripnus) but in Danaphos, Valenciennellus, Sonoda and
Maurolicus the mouth is relatively small and the gape very
oblique or almost vertical.

7. The extent to which the premaxillary enters the gape varies
in the extreme, being excluded from it in Ichthyococcus and
forming virtually all of it in Triplophes. In maurolicid genera
the premaxillary is always at least half as long as the toothed
portion of the maxillary and is usually more than half as long.
The premaxillary of Diplophos and Yarrella is equal to or longer
than the toothed portion of the maxillary; in Polymetme it is
only shehtly shorter; in Pollichthys, Photichthys and Woodsia

64 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

it is about half, or more than half as long; and in Vinciguerria,
Bonapartia, Margrethia, Gonostoma and Cyclothone it is less
than half as long.

8. The premaxillary teeth are uniserial in all gonostomatids
excepting Yarrella, Triplophos and Polymetme, in which they are
in two rows, and Neophos, in which they are arranged irregu-
larly. Maxillary teeth are always uniserial as far as known,
although in some species one or two of the most posterior maxil-
lary teeth may be situated higher than the remainder. The lower
jaw teeth also are mostly in a single row except anteriorly, where
there is usually a very short outer row of teeth. However, in
some genera the teeth on the mandible are biserial for almost
its full length (Yarrella, Triplophos, Polymetme, Argyripnus),
or for about half its length (Vinciguerria, Sonoda); and in
Ichthyococcus, probably Valenciennellus, and sometimes Mau-
rolicus, they are entirely uniserial and lack the anterior outer
row. The mandibular teeth of Danaphos are irregular anteriorly
and uniserial posteriorly.

9. Teeth on the vomer, palatines, pterygoids and tongue, when
present, are usually small or even minute, although Gonostoma
bathyphilum has a few much enlarged pterygoid teeth posterior-
ly. Cyclothone may have as many as five or six small, close-set
teeth in a lengthwise row on each side of the vomer (absent in at
least one species), and in this genus the palatine and pterygoid
teeth are present only on the anterior ends of the bones, in small
clusters. Other genera have fewer vomerine teeth in crosswise
rows (or none); the palatine teeth, if present, are linear; and
the pterygoid teeth, if present, are arranged in more or less
circular patches of minute teeth with sometimes an irregular,
sparse row posteriorly. Mawrolicus sometimes has a few ptery-
void teeth but these are apparently lacking in all other mauro-
licid genera. Only three non-maurolicid genera lack pterygoid
teeth (Ichthyococcus, Vinciguerria, Woodsia) but they are some-
times absent also in Diplophos (subgenus Diplophos), Triplo-
phos, Pollichthys and Photichthys; and are microscopic in
Yarrella, if present at all. Teeth have been found on the tongue
only in Diplophos (sometimes), Pollichthys, Photichthys (some-
times), Bonapartia, Margrethia, and in two species of Gonostoma
(denudatum, atlanticum rarely).

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 65

10. Meristic characters are shown in Table 6.

The key on page 69 has been based on all known genera except
Thorophos Bruun, of which no specimens have been seen; nor
have I seen any Indian or Pacifie Ocean material of Triplophos,
Polymetme, Vineiguerria, Ichthyococeus, Margrethia, Argyrip-
nus or Valenciennellus. It should be emphasized also that the
genus Cyclothone has not yet been studied in detail. The follow-
ing hitherto unreported material has been examined but is not
pertinent to the present publication except as it applies to the
characters contained in the key to the genera.

Photichthys argenteus Hutton. Three specimens, standard
length 100-118 mm., off Cape Palliser, Cook Strait, New Zealand,
1942, from the stomach of a groper caught in 40-50 fathoms
(73-91 meters). Received from Prof. L. R. Richardson, Victoria
University College, Wellington, New Zealand. The genus Pho-
tichthys is defined on page 100.

Ichthyococcus Bonaparte. Only a few small specimens of
I. ovatus (Coceo) are available for study (Grey, 1955, p. 273).
Included in this lot are three hitherto unrecorded metamorphos-
ing specimens. The premaxillary bones are very difficult to dis-
tinguish on specimens of 7. ovatus but a communication received
from Dr. A. B. Rechnitzer (1958, in litt.) has confirmed the
impression that the premaxillary is excluded from the gape. Dr.
Rechnitzer also found it difficult to distmguish these bones on
two eastern Atlantic specimens of J. ovatus, but an examination
of three larger specimens of J. elongatus Imai revealed that the
premaxillary terminates at the apex of the inverted V-shaped
symphysis of the upper jaw; and that the maxillary is continuous
throughout the gape and is toothed, or serrated, along its entire
edge.

Gonostoma Rafinesque 1810. A few unrecorded western At-
lantic specimens of G. elongatum Giinther 1878 and G. bathy-
philum (Vaillant) 1888, from the collection of the United States
National Museum, have been examined, as well as new material
of G. clongatum from the Gulf of Mexico, the Caribbean Sea,
and off northern South America (Oregon). The genus is de-
fined on page 102, new material of G. atlanticwm Norman from
both the Atlantic and Pacific oceans is described on page 106,

66 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

and a new species, G. ebelingi, from the tropical Pacific, is de-
scribed on page 109. I have also seen two Mediterranean speci-
mens of G. denudatum Rafinesque 1810, loaned by the Museum
of Comparative Zoology, and the following unrecorded speci-
mens of @. gracile Giinther 1878, most of them loaned by Scripps
Institution of Oceanography. The latter show that G. gracile has
a wide distribution in the northern and western Pacific.

Western Pacific: Five, standard length 60-110 mm., C.N.H.M.
No. 42780, taken during the Albatross Philippine eruise off
Hong Kong, China, 6 November 1908; eleven, standard length
40-119 mm., SIO No. H 53-367, Kai Strait, Japan, 23-24 October
1952, 10’ midwater trawl, depth not known; one, standard
length 95 mm., SIO No. H 51-871, Japan Trench, 32° 08’N.,
142° 04’ E., 26 October 1953, S. F. Baird, 10’: midwater trawl,
0-4455 fathoms (0-8147 meters) ; five, standard length 71-121
mm., SIO No. H 53-356, off Honshu, Japan, 35° 01.8’ N., 145° 12’
BE. to 34° 48.5’ N., 145° 05.4’ E., 1 October 1953, 8S. F. Baird, 10’
midwater trawl, 1000 fathoms (1829 meters).

North and middle Pacific: Three, standard length 69-87 mm.,
SIO No. H 53-335, southeast of Kamchatka, 51° 09.5’ N., 164°
32.6’ E. to 50° 59.3’ N., 164° 27.1’ E., 4 September 1953, S. F.
Baird, 10’ midwater trawl, 0-580 fathoms (0-1061 meters) ; one,
standard length ca. 83 mm., SIO No. H 53-344, off the Kurile
Islands, 45° 29.7’N., 154° 20’ EK. to 45°18.8’N., 154° 02.6’ E.,
16 September 1953, S. F. Baird, 10’ midwater trawl, 0-2600
fathoms (0-4755 meters) ; one, standard length ca. 92 mm., SIO
No. H 538-371, 48° 58.3’ —-37.4’ N., 157° 49.8’ -29’ W., 5-6 Sep-
tember 1951, Horizon, 10’ midwater trawl, 2200-2400 fathoms
(4023-4389 meters); four, standard leneth 105-116 mm., SIO
No. Hi 51-374, 37 429% —01.5'.N., 154° 03:5’ 153° 58.27 Wel 2ai3
September 1951, Horizon, 10’ midwater trawl, 700 fathoms (1280
meters) ; three, standard length 96.5-111.5 mm., SIO No. H 51-
360, 48° 08’ N., 150° W., 13 August 1951, 10’ midwater trawl,
35-175 meters; three, standard length 102-120 mm., SIO No. H
les anes 1° 54.3’ Ne 52> 267 We, tor SaaS Oso INS lb 0 oe 6. ave
15 September 1951, Horizon, 10’ midwater trawl, 1790 fathoms
(3274 meters) ; one, standard length 96 mm., SIO No. H 53-312,
44° 59’N.. 148° 46.5’ W., 4 August 1953, S. #. Baird, 1 meter
plankton net, 0-1000 fathoms (0-1829 meters) ; five, standard

food

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 67

length 107-130 mm., SIO No. H 51-358, 40° 35’ N., 147° 55’ W.,
10 August 1951, Horizon, 10’ midwater trawl, 350-600 meters ;
one, standard length 116.5 mm., SIO No. H 53-308, 39° 23’N.,
1429 51 OW 107392 1515? Ne 142° 54.92 Wel) August 1953S» F.
Baird, diving dredge No. 1 (no depth).

Bonapartia pedaliota Goode and Bean 1895. One specimen,
standard length 42 mm., U.S.N.M. No. 108303, Caroline, Virgin
Islands, 18° 44’ N., 65° 15’ 15” W., 26 February 1933, 600 fath-
oms (1097 meters).

Margrethia Jespersen and Taning 1919. Koumans (1953, p.
183) was apparently unaware of the description of the genus
Margrethia when he proposed the genus Snellius, which is an
obvious synonym of Margrethia.

Argyripnus atlanticus Maul 1952. One specimen, standard
leneth 67 mm., Oregon, September 1957, Caribbean Sea, further
data lost. The type of atlanticus was described and figured as
having four OP and was said to differ from ephippiatus Gilbert
and Cramer 1896 and iridescens McCulloch 1926 in having a
much larger posterior opercular organ. Both Pacific species were
also described as having two photophores in this area: “‘. . . one
above the other and separated by a black, metallic-hued space’’
(McCulloch, 1926, p. 170); ‘‘. . . one above the other, at the
two ends of a short vertical’steel-blue band . . .’’ (Gilbert and
Cramer, 1896, p. 415). In the specimen examined the opercular
organ has the outward appearance of a single, greatly elongated
and enlarged one encased in a deep black sheath, with a luminous
area exposed near the top. Actually there probably are two
organs involved, but they are at least encased in a common
sheath, and the structure is similar in all species, as can be seen
in the published figures.

Danaphos Bruun 1931. The genus is defined on page 112 from
new material in the collection of Scripps Institution of Ocean-
ography.

Neophos Myers 1932. The type has been examined and the
genus is defined on page 114.

Valenciennellus tripunctulatus (Esmark) 1870. One speci-
men, 30.5 mm., Stanford University No. 18713, Oregon Station
841, Gulf of Mexico, 25° 58’ N., 88° 00’ W., 6 October 1953, 830-
930 fathoms (1518-1700 meters), reported as Maurolicus miilleri

68 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

by Springer and Bullis (1956). The genus Valenciennellus
Jordan and Evermann contains two, and possibly three species.
V. tripunctulatus, of which V. stellatus Garman 1899 is possibly
a synonym, is cosmopolitan in tropical and subtropical waters.
V. carlsbergi Bruun 1931 has been found only in the tropical
Indo-Malayan area. The latter differs from tripunctulatus in
having a shorter and deeper caudal peduncle, only two OA
photophores, and only three groups of AC photophores (a group
of three above the tenth to twelfth anal rays, a group of two
above the middle of the anal fin, and a group of four on the
caudal peduncle).

V. stellatus was distinguished by Garman (1899, p. 239) from
V. tripunctulatus by two characters, a shorter dorsal fin and a
smaller number of AC photophore groups. Actually, however,
the dorsal count of stellatus is higher, numbering twelve rays
instead of the seven to ten of tripunctulatus. Garman’s con-
clusion that the dorsal fins of the two species differed was based
on an error in the figure published by Liitken (1892, pl. 1, fig. 6).
This figure shows a long dorsal fin of about seventeen rays al-
though in the text Liitken gave the dorsal count as nine or ten
rays. The error in the figure probably resulted from the artist
having drawn the dorsal and the adipose as a continuous fin.
Garman’s second distinction, the presence of only four groups
of AC photophores, may be valid. There is, however, variation
in the number of AC groups in tripunctulatus, although five is
the typical number. Only examination of a series of specimens
from the eastern Pacific will show with certainty whether or not
stellatus is a distinct species with a constantly higher number
of dorsal rays and a lower number of AC photophores.

It should be noted that Jespersen’s description (1933) of the
photophores of V. tripwnctulatus is misleading in part. His OA
count of 10 ineludes the second IV group of (4) and the lower
posterior OP. His ‘‘series of 4 organs underneath the cheek’’
refers to the BR (six in Jespersen’s figure), which are visible
beneath the transparent jaw bones in most specimens. Finally,
the ‘‘two or three on the branchiostegal membrane’’ must refer
to the first IV group of (3) on the isthmus.

Maurolicus muelleri (Gmelin) 1789. A few unreported speci-
mens from the collection of the United States National Museum

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 69

have been examined, four of them washed ashore on the Massa-
chusetts coast and two found on the beach at Juan Fernandez
in the southeastern Pacific.

In the key, and in descriptions, the following symbols are used
to represent the photophores: ORB, those situated close to the
eye; OP, opereular photophores ; SO, a pair often found near the
symphysis of the lower jaw; BR, organs on the branchiostegal
membranes; IV, pre-ventral photophores of the ventral series;
VAV, those of the ventral series found between the ventral bases
and the anal origin; AC, photophores of the ventral series pos-
terior to the anal origin; IC, total number in the ventral series,
from tip of isthmus to base of caudal; OA, photophores of the
lateral series. Photophore counts in parentheses indicate that
these organs are grouped in a common gland.

Key to the Genera and Subgenera of the Family
Gonostomatidae

la. BR 8 or more (reduced in size and number in Cyclothone obscura,
obscure and very small in Gonostoma bathyphilum). Serial photophores
separate, not grouped in common glands.

2a. Photophores present on isthmus. IV 20 or more. IC 42 or more.
[Dorsal origin always in advance of anal origin. Body always with
at least two rows of serial photophores].

3a. More than two rows of photophores on body, those above the IC
and OA frequently mostly or entirely lost with the skin. No
adipose fin. OA 40 or more. [ORB 1, close to front of eye or below
rel |

4a. Teeth of upper jaw all uniserial. Teeth of lower jaw uniserial
except for a short outer row anteriorly. On posterior half of
lower jaw a row of minute photophores in adult, preceded by a
somewhat larger photophore. Lateral line area with a row of
small photophores extending on to caudal fin (often partially
lost). Gill rakers on lower limb of first arch 7-9. Vertebrae 63

to ca. 85 (and more?).
Diplophos Giinther 1873
Atlantic, Pacific, Indian

lod

0

4h.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

oa:

Dorsal origin about in middle of body, usually slightly nearer
tip of snout than caudal base. Head and trunk approximately
same length as tail, distance between snout and anal origin
ca. 47.5-51.2%, and between anal origin and caudal base ca.
48.0-53.0% of standard length. Depth 8.2-12.3% of standard
length. Anal rays (43?) 53-68. IV 40-49. AC 43-49. IC 97-
113. OA 66-87. Lateral line ca. 80-98. Vertebrae ca. 85 (and
more?).
subgenus Diplophos Giinther 1875
Atlantic, Pacifie, Indian

Dorsal origin slightly behind middle of body (close to middle
in young). Head and trunk longer than tail (proportionately
less so in young), distance between snout and anal origin
59.0-63.0%, and between anal origin and caudal base 36.5-
41.2% of standard length. Depth ca. 15-17% of standard
length. Anal rays 36-41. IV 30-32. AC 28-30. IC 70-75.
OA 44-48. Lateral line 63-68, Vertebrae 63.
subgenus Manducus Goode and Bean 1895
Atlantic

Teeth of premaxillary biserial. Teeth of lower jaw biserial on

most of its length. No photophores on posterior half of lower
jaw. Lateral line area not marked by a row of photophores.
Gill rakers on lower limb of first arch 12-16. Vertebrae 45 to
ca. 60(?).

6a.

6b.

Trunk much shorter than tail. Dorsal origin far in advance

of middle of body length. Toothed portion of maxillary very

short, scarcely entering gape. Snout shorter than eye. Dorsal

10-12, anal 54-63. VAV 5-7. AC 35-41. Head with additional
photophores above upper jaw. Vertebrae ca. 60?

Triplophos Brauer 1902

Atlantic, Indian

Trunk slightly longer than tail. Dorsal origin about in middle
of body length or slightly behind it. Toothed portion of
maxillary entering gape. Snout longer than eye. Dorsal 14-16,
anal 26-31. VAV 9-12. AC 20-27. Photophores on upper part
of head consisting only of the ORB and OP. Vertebrae 45-54.
Yarrella Goode and Bean 1895

Atlantic, Pacific

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS all

3b. Body with only two rows of photophores. Adipose fin present.
OA 16-34.

7a. ORB 1, close to front of eye. Premaxillary teeth biserial. AC
22-25, [Vertebrae 45].

Polymetme MeCulloech 1926

Atlantic, Pacific, Indian

7b. ORB 2, one close to front of eye, one close to its hind margin
or below center (Ichthyococcus). Teeth of upper jaw all uni-
serial. AC 12-21.

8a. Eye normal (except somewhat tubular in some juvenile Vinci-
guerria). Mouth large, bordered by premaxillary anteriorly.
Jaws equal, or lower jaw projecting slightly beyond upper
jaw anteriorly. Teeth relatively well developed. Ventral fins
in advanee of dorsal origin. Gill rakers minutely denticulate
on inner edge.

9a. Anal origin beneath dorsal or close behind a vertical from
its last ray. Branchiostegal rays 9-13. BR 8-9 (12?). OA
19-25.

10a. Anal origin beneath middle or anterior portion of dorsal.
Anal 22-30. Anterior ORB larger than posterior one.

AC 19-21. Vertebrae ca. 40?
Pollichthys Grey 1959
Atlantic, Pacific

10b. Anal origin beneath middle or end of dorsal. Anal 12-16.
ORB equal in size or posterior one larger. AC 12-16.
Vertebrae (36?) 38-42.

Vineiguerria Jordan and Evermann 1895
Atlantic, Pacific, Indian

9b. Anal origin well behind end of dorsal fin. Branchiostegal
rays 17-21. BR 14-18. OA 29-34.

lla. Gill rakers normally developed, 11 + 4-5 on first arch.

Premaxillary about half as long as toothed portion of

maxillary. Body elongate, depth ca. 6-6.5 times in

standard length. Anal 23-26. VAV 15-17. AC 16-18.

OA 33-34, ending above anterior portion of anal fin.
Vertebrae 51.

Photichthys Hutton 1872

Atlantic, Pacific

baw |

2 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

8b.

1lb. Only 3-4 normally developed gill rakers at angle of first

arch. Premaxillary more than half as long as toothed

portion of maxillary. Body not elongate, depth ca. 5

times in standard length. Anal 14. VAV 11-12. AC 12.

OA 29-31, ending above end of anal fin. Vertebrae?
(myomeres ca. 45).

Woodsia Grey 1959

Atlantic, Pacific

Eye tubular. Mouth small. Premaxillary not entering into
gape. Lower jaw included anteriorly. Teeth minute. Ventral
fins behind dorsal origin. Gill rakers short and smooth, [Anal
origin well behind end of dorsal. Dorsal 10-15, anal 13-17.
Branchiostegal rays 11-12. Lateral line 34-42. ORB 2. BR
11-12. IV 25-28. VAV 9-14. AC 12-14. OA 23-31. Vertebrae
38-47].
Ichthyococcus Bonaparte 1841
Atlantic, Pacific, Indian

2b. No photophores on isthmus. IV 17 or less. IC 26-43. [Premaxillary
less than half as long as toothed portion of maxillary].

12a. Dorsal origin opposite or behind anal origin. No _ pseudo-

br

ias

anchiae. SO present or absent.

Body with at least two rows of photophores or photophores
inconspicuous or obsolete. Dorsal 10-17. Pectoral 7-13. Adi-
pose fin present or absent. Luminous glands usually present
on procumbent caudal rays.

14a. Maxillary with a series of relatively long, slender teeth, and

14b.

subequal, short teeth in the interspaces between them. Each
palatine with a single row of teeth. Pterygoid teeth in a
patch on each side anteriorly, these teeth small or minute;
and a few teeth posteriorly, these sometimes enlarged.
Vomer usually with one or two small teeth on each side,
sometimes absent. Adipose fin present or absent. Anal rays
21-31. OA 11-21. SO present (except usually absent in
bathyphilum). Vertebrae 37-40.
Gonostoma Rafinesque 1810
Atlantic, Pacific, Indian

Maxillary teeth subequal, close-set, more or less increasing
in size posteriorly, sometimes a few of them moderately
enlarged. Palatine and pterygoid teeth usually present

1b.

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 73

anteriorly only, each in a small group of a few relatively
prominent teeth; posterior pterygoid teeth, if present, mi-
eroscopic. Vomer usually with a double row of several small,
close-set teeth (absent, microscopic, or reduced to one or
two in braueri, alba and signata). Adipose fin normally
absent. Anal rays 16-21. OA 6-10. SO absent. Vertebrae
29-33.
Cyclothone Goode and Bean 1883
Atlantic, Pacific, Indian, Antarctic

13b. Body with a single row of conspicuous photophores. Dorsal
17-20. Pectoral 14-16. No adipose fin. No luminous glands

on procumbent caudal rays. [SO present. Vertebrae 37].
Bonapartia Goode and Bean 1895
Atlantic

12b. Dorsal origin slightly in advance of anal origin. Pseudobranchiae

present. SO absent. [Body with only one row of large, con-

spicuous, somewhat irregular photophores. Adipose fin present.
Vertebrae 34].

Margrethia Jespersen and Taning 1919

Atlantic, Pacific

BR (6) or less (7 on one side of one specimen of Sonoda), conspicuous.
At least some of the serial photophores grouped together in common
glands appearing as black or silvery bands. [Photophores present on
isthmus].

15a. AC composed largely of separate organs, more or less evenly spaced.

16a. Dorsal origin about in middle of body length, and a little behind
anal origin. Ventral bases well ahead of dorsal origin. Eye
normal. Dorsal 8-11. Anal 31-38. SO present.

17a. Photophores on isthmus, and VAV, in more than one group.
AC 12-13, all single. OA 1. Dorsal 8. Anal 38. No adipose

fin. Number of vertebrae unknown.
Neophos Myers 1932
Pacific

17b. Photophores on isthmus, and VAV, each in a single group.

AC 14-15, mostly single but with a group of two anteriorly

and another group of two just behind anal fin. OA 7. Dorsal

11. Anal 31. Adipose fin present. Number of vertebrae un-
known.

Thorophos Bruun 1931

Pacific

74

16b.

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Dorsal origin well ahead of middle of body length and ahead
of anal origin, which is behind end of dorsal fin. Ventral bases
behind dorsal origin. Eye tubular. Dorsal 6. Anal 24-25. SO
absent. [AC mostly single and rather widely and evenly spaced
but with the first three in a group and a group of four on the
peduncle, followed by a single organ. Vertebrae 38].
Danaphos Bruun 1931
Pacific, Indian

15b. AC composed of two to five groups of two or more organs each.

18a.

Photophores between ventral and anal fins grouped separately
from those above anterior portion of anal fin, the VAV straight
and numbering 4-8.

19a. Dorsal origin about in middle of body length. Anal origin

beneath or slightly in advance of dorsal origin. Ventral bases
well ahead of dorsal origin. Trunk shorter than tail. SO
absent. OA 7 or less.

20a. AC in three to six well separated groups of only two to four

organs each. IV 20-24. VAV 4-5. AC 9-15. IC 36-40, Adi-

pose fin present. Eye sometimes slightly tubular. Vertebrae
32-337

Valenciennellus Jordan and Evermann 1895

Atlantic, Pacifie, Indian

20b. AC in two subequal groups, each with sixteen organs or

more. IV 16. VAV 7-8. AC 36-42. IC 58-66. No adipose fin.
Eye normal. Vertebrae 40.

Sonoda Grey 1959

Atlantic

19b. Dorsal origin behind middle of body length. Anal origin be-

18b.

hind dorsal origin. Ventral bases close to a vertical from first

dorsal ray. Trunk longer than tail. SO present. OA 9. [AC

in two long groups, preceded by a single elevated organ.
Adipose fin present. Eye normal, Vertebrae 32-33].

Maurolicus Cocco 1838

Atlantic, Pacific, Indian

Photophores between ventral and anal fins continuous with those
above anterior portion of anal fin, this group rather sinuous and
numbering 19-28, followed on the tail by one group of 5 and a
second group of 12-18 organs. [Dorsal origin in advance of

on

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 7

anal origin. Ventral bases close to a vertical from first dorsal
ray. Eye normal. SO absent. IC 53-67. Vertebrae 47].

Argyripnus Gilbert and Cramer 1896

Atlantic, Pacific

DIPLOPHOS Giinther 1873

Diplophos Gimther, 1873, Jour. Mus. Godeffroy, 2: 101, type species Diplo-
phos taenia Giinther 1873, Atlantic, 22° N., 30° W. and 30°S., 24° W.;
1889, Rep. Sci. Res. Voy. Challenger, Zool., 31: 32; Goode and Bean,
1895, Ocean. Ichth., p. 104; Barnard, 1925, Ann. So. Afr. Mus., 21:
148; Norman, 1930, Discovery Rep., 2: 295; Parr, 1931, Bull. Bingham
Oceanogr. Coll., 2, (4): 11 (part, Lychnopoles in synonymy); Fowler,
1936, Bull. Amer. Mus. Nat. Hist., 70: 235; Matsubara, 1940, Suisan
Kenkiu-shi, 35: 319; Smith, 1949, Sea Fishes So. Afr., p. 105.

Manducus Goode and Bean, 1895, Ocean. Ichth., p. 514, type species
Gonostoma maderense Johnson 1890, Madeira; Norman, 1930, Discovery
Rep., 2: 293 (part, Lychnopoles in synonymy); Fowler, 1936, Bull.
Amer. Mus. Nat. Hist., 70: 221, 1202 (part, Lychnopoles in synonymy ).

Paraphotichthys Whitley, 1931, Australian Zool., 6: 334, Manducus con-
sidered preoccupied by Manduca Huebner, ca, 1806, Lepidoptera.

Generic characters. Kye normal, moderate. Snout longer than
orbit. Interorbital width at center of eye about equal to, or
shightly greater than, diameter of orbit. Mouth large, oblique;
edge of premaxillary straight; toothed edge of maxillary slight-
ly convex; maxillary nearly reaching preopercle. Premaxillary
almost, or quite as long as toothed edge of maxillary. Angle of
preoperele shghtly acute or nearly vertical. Teeth of upper jaw
uniserial, unequal. Teeth of lower jaw unequal, uniserial except
for a short outer row anteriorly. Vomer toothless or with 1-5
teeth on each side. Palatines each with a row of small teeth.
Pterygoids with or without a patch of minute teeth on each side.
Tongue with or without teeth. Gill rakers 7-9 + 3-5 = 10-14 on
first arch. Spines on inner edge of first gill arch rudimentary or
absent. Pseudobranchiae present or absent. Anus close to anal
fin. Relative proportions of head and trunk to tail variable.
Origin of dorsal fin near middle of body length, sometimes slight-
ly before or sightly behind it. Origin of anal fin beneath end of
dorsal fin or slightly behind last dorsal ray. Ventral bases in
advanee of dorsal origin. No adipose fin. ORB 1, below front

76 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

margin of eye. OP 3, upper one about level with center of eye,
lower two level with end of maxillary. Additional photophores
present on head above maxillary, in a row along posterior portion
of lower jaw (the first of these larger), and lower surface of
symphysis of lower jaw (2-4 minute, round spots, perhaps pho-
tophores). SO present, somewhat behind symphysis, hidden by
lower jaw. BR 8-16. Body with ventral and lateral rows of
photophores typical of family and also several rows above these ;
photophores present on isthmus. IV 30-49. VAV 12-17. AC
28-49, straight. IC 70-118. OA 45-87. Always a third row of
photophores along lateral line from upper edge of gill opening to
caudal fin, 65-98 organs, last one or two on caudal fin. Older
specimens with additional serial rows of photophores above and
below lateral line, and 1-3 photophores on or before pectoral base
between ventral and lateral series. A pair of narrow strips of
pale tissue, probably luminous, on ventral surface of body
below eighth to seventeenth IV photophores. Fin rays: dorsal
9-13, anal 36-68, pectoral 8-11, ventral 7-8. Branchiostegal rays
11-14 (15-16?), bases prominent but without spines. Vertebrae
63 to ca. 85, or more.

Remarks. Diplophos is probably related to both Yarrella and
fonostoma. Its affinities with the former are set forth in the
key to the genera on page 70, and in addition these two genera
have a relatively long premaxillary. The relationship of Diplo-
phos to Gonostoma is less obvious, but their common origin is
possibly indicated by the similarity of their dentition and of
their photophore structure (Brauer, 1908, p. 18), as well as in
their tendency toward the development of luminous tissue on
head and body. Diplophos may also be distantly related to the
maurolicid group of genera (see p. 61).

Manducus Goode and Bean 1895 is reduced to the rank of sub-
venus because the distinction between this genus and Diplophos
is of noticeably different value than distinctions between other
eonostomatid genera and the differences appear to be of a specific
rather than a generic nature, as shown in the key on page 70.
However, Diplophos, sensu stricto, probably contains two or more
closely related species that differ rather sharply from maderensis,
the only known species of Manducus, and the latter name is
therefore retained as a subgenus of Diplophos.

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS

Subgenus DIPLOPHOS§ Giinther 1873

In the following characters the subgenus Diplophos differs
from the subgenus Manducus (ef. p. 81).

Angle of preopercle nearly vertical or slghtly rounded.
Pseudobranchiae sometimes present. Gill rakers 7-9 + 3 on first
areh. Premaxillary about the same length as, or slightly shorter
than, toothed portion of maxillary. Head and trunk about the
same length as tail or slightly shorter. Origin of dorsal fin
about in middle of body length but as far as known always
slightly nearer tip of snout than caudal base. Origin of anal
fin beneath end of dorsal fin or just behind its last ray. Ventral
bases noticeably in front of dorsal origin. Photophores on upper
part of head, in addition to the ORB and OP, consisting of one,
level with ORB, on cheek above middle of maxillary; a row of
9-13 minute organs on cheek below eye, hidden beneath subocular
bone ; 2-4 small organs beneath posterior end of maxillary ; small
photophores also sometimes present on opercle above lower pos-
terior OP and between the two lower OP. Area just below
symphysis of lower jaw with two small round spots (photo-
phores?) on each side. A row of 12-16 minute photophores on
posterior portion of lower jaw, the first one enlarged. Patches of
luminous tissue sometimes present on opercle, maxillary and
tongue. BR 10-12 + 0-3. IV 40-49. VAV 13-17. AC 43-49, last
two or three usually, but not always, shghtly separated from the
rest. IC 97-113. OA 66-87, sometimes ending above middle of
anal fin, sometimes reaching caudal base. Lateral line row (80?)
86-98. A pair of narrow strips of pale tissue, probably luminous,
below ninth to seventeenth IV photophores. Anal rays (43?)
53-68. Vertebrae ca. 85 (one eastern Pacific specimen).

It is impossible to determine the number of species contained
in the D. taenia complex without examining a large series of
specimens from different areas. Variation in fin ray and photo-
phore counts is rather wide but specimens are too rare to de-
termine the significance of these variations. The photophores of
the ventral series are consistently fewer in Pacific specimens.
Atlantic specimens are all in pretty good agreement with one
another but even here there is enough variation, especially in
the number of anal rays, to allow the possibility at least of sub-
specifie differences. All Atlantic specimens are considered here

78 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

to belong to a single species, D. taenra Gunther. The small speci-
men reported by Brauer (1906, p. 89) from the Indian Ocean
agrees with Atlantic specimens. Norman (1939, p. 19) also
reported, as taenia, two small specimens from the Indian Ocean.
They were not described.

Brauer (1906, p. 90) described the BR as follows: ‘15-16
between the branchiostegal rays, the twelfth smaller and the last
three smaller and separated by a large interspace from the others
and from one another.’’ No trace of these smaller organs has
been found on Atlantic specimens examined but one to three are
present on several eastern Pacifie specimens. The description
quoted suggests that the first 11-12 BR are analagous with the
11-12 of most of the specimens examined.

DIPLOPHOS TAENIA Gunther 1873
Atlantic specimens

In the Atlantic, D. taenia has been reported from various
localities between about 40° N. and 380°S.; and in the Indian
Ocean off Natal, between the Seychelles and Zanzibar, and in the
Arabian Sea. Counts and proportions of some Atlantic speci-
mens are shown in Table 1. Included are two hitherto unreported
specimens, U.S.N.M. Nos. 100525 and 100616, caught in 1914,
at Grampus (Bache) Station 10182, off Bermuda, 30° 27’N.,
66° 05’ W., 19 February, surface; and Station 10196, northeast
Providence Channel, Bahamas, 25° 27’ N., 77° 16’ W., 3 March,
surface.

Both of these specimens are small, their standard lengths
being 74 and 66 mm. No pseudobranchiae are visible. There
are a few minute teeth on each side of the vomer in the larger
specimen but none can be seen in the smaller one. Both have a
single row of 6-9 small teeth on each palatine and there are no
teeth on either the pterygoids or the tongue. A pair of thread-
like strips of luminous(?) tissue is present on the ventral surface
of the body below the fourteenth to seventeenth IV photophores.
The OA number 71 and 73 and extend to a vertical from the
middle of the anal fin; the organs decrease in size posteriorly.
In addition to the IC, OA, and lateral line photophores, the
following rows are present: between the IC and OA one row
of minute dots reaching well past the anal origin in the larger

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS fis)

specimen, not quite reaching the ventral base in the smaller;
between the OA and lateral line one row of minute dots extending
about to the anal origin in the larger fish and halfway between
the ventrals and the anal origin in the smaller one; above the
lateral line three similar rows in the larger specimen, only one
in the smaller one, reaching below or beyond the dorsal fin.

Pacific specimens

The following ten specimens of the taenia complex have been
reported from the Pacific Ocean, most of them caught at the
surface.

The type of Diplophos pacificus Gunther 1889, length 37 mm.,
5° 24’ N., 147° 02’ W., found in a townet attached to the dredge
after a deep haul; inadequately described and now in poor con-
dition.

The type of Diplophos proximus Parr 1931, standard length ca.
82 mm., Gulf of California, 24° 07’ N., 108° 40’ W., 523 meters;
differs from taenia in number of photophores but not significant-
ly in proportions.

A specimen 43 mm. long from the Solomon Islands, identified
as D. pacificus but not described, by Seale (1935, p. 340).

The type of Diplophos taenia orientalis Matsubara 1940, total
length 195 mm., standard length 179.8 mm., off Huji River, near
Kambara, Japan, 306 meters; largest known specimen of the
group, differing from taenia principally in photophore counts.

A second specimen of D. orientalis, total length ca. 195 mm.,
taken off Kambara, Suruga Bay, Japan (Abe, 1958, p. 1241, pl.
238, fig. 598).

Five specimens, 34-41 mm. long, Sulu Sea, 6° 48.5’ N., 118°
51.5’ K., surface at night, identified as D. taenia and partially
described by Herre and Herald (1950, p. 314, fig. 2).

Twelve hitherto unreported Pacific specimens from the collec-
tion of Scripps Institution of Oceanography have been examined:

Western Pacific: One, standard leneth 40 mm., SIO 56-127,
Marshall Island area, 13° 03’ N., 166° 04’ E., to 13° 03’ N., 166°
32’ E., 0-400 fathoms (0-732 meters).

Kastern Pacific: Five, standard length 42 and 89-100 mm.,
SIO 54-89, off the Revillagigedo Islands, 19° 09’ N., 110° 58.5’ W.,
0-825 fathoms (0-1509 meters) ; one, standard length 100.5 mm.,

80 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

SIO 54-98, same area, 21°01’N., 115° 39’ W. to 21° 04.5’ N.,
115° 48’ W.; one, standard length 90 mm., SIO 54-88, same area,
20° 16’ N., 111° 32.2? W. to 20° 03.27 N., 111° 20’ W., 0:25 fathoms
(0-45 meters) ; two, standard length 81 and 88.5 mm. SIO 54-
92 same area, 19527 N., ass 2054 Wi tow 19> Sioa Nets
32.5’W., 0-100 fathoms (0-183 meters); one, standard length
91 mm., SIO 55-213, west of Clipperton Island, 11° 13’ N., 120°
57’ W., 0-63 fathoms (0-115 meters) ; one, standard length 97.5
mm., SIO 54-95, off Lower California, 23° 05’ N., 119° 08’ W. to
22° 23’ N., 119° 36’ W., 0-1333 fathoms (0-2438 meters).

Counts and proportions of some of these Pacific specimens are
shown in Table 2, those of Atlantic specimens in Table 1. The
consistently lower counts found in Pacific specimens suggest that
they belong to a distinct species or subspecies. It is unfortunate
that the type of D. pacificus, the first species described from the
Pacific, is small and now in poor condition, and it is also un-
fortunate that no large Atlantic specimens are available for com-
parison with eastern Pacific specimens. Small but well developed
pseudobranchiae are present on all of the larger eastern Pacific
specimens and their apparent absence in Atlantic specimens
examined is probably questionable, especially as Brauer (loc.
cit.) noted their presence in an eastern Atlantic example.

The eastern Pacific specimens have one small tooth on each
side of the vomer; and a row of four to seven small teeth and
some additional minute teeth on each palatine. There are no
teeth on the pterygoids, and usually none on the tongue, but
several specimens do have a cluster of three or four very small
teeth at the tip of the tongue. Small teeth on the tip of the
tongue were described in both specimens of D. orientalis but have
not been mentioned in any other description of a Diplophos
except maderensis. Larger specimens have three tiny photo-
phores in a vertical series above the lower posterior OP, and
several in a horizontal series between the two lower OP. All
eastern Pacific specimens except the smallest have a pair of
thread-like strips of luminous(?) tissue below the ninth to fif-
teenth IV photophores (below the fourteenth to seventeenth in
the two Atlantic specimens examined). The OA number 60-61
and reach a vertical from about the middle of the anal fin in
three specimens. On a fourth they number 84 and reach the

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 81

caudal base, but the organs on the latter half of the tail are
minute. Few of these Pacific fishes are intact but those retaining
most of their skin possess the following rows of minute photo-
phores on the body: between the IC and OA one row, reaching
sheghtly past the anal origin; between the OA and lateral line
three rows, the lower and upper reaching well past the anal
origin and the middle row to about halfway between the ventral
bases and the anal origin; above the lateral line three rows, one
of them reaching the caudal base.

The 40 mm. specimen taken off the Marshall Islands lacks both
pterygoid and tongue teeth, and has no luminous tissue on the
ventral surface of the body; the premaxillary and toothed por-
tion of the maxillary are equal in length; two small teeth are
present on each side of the vomer; and each palatine bears five
small teeth.

D. orientalis Matsubara should probably be retained as a dis-
tinct species, at least until intermediate sizes are found. In
appearance it somewhat resembles D. maderensis, although its
characters are distinetly those of the subgenus Diplophos. A
similarity to Manducus is seen in the deeper body and the heavy,
fang-like premaxillary teeth. Even in the largest eastern Pacific
specimen seen, standard length 100.5 mm., the long premaxillary
teeth are slender and needle-like. It is possible, of course, that
D. orientalis merely represents the adult Pacifie Diplophos. The
higher OA count (87), greater body depth, and even the enlarged
premaxillary teeth, might be attributed to the large size of the
only two specimens known (177 and ca. 180 mm. in standard
length). On the other hand, one of the eastern Pacific specimens
(SIO 54-88), 90 mm. in standard length, is a female with large
ovaries which contain small and probably immature eggs. It is
interesting that Abe (1958, p. 1242) described a narrow, semi-
transparent band along the mid-ventral line of this species, with
the inference that the area may be luminous.

Subgenus MANDUCUS Goode and Bean 1895

The subgenus Manducus differs from the subgenus Diplophos
in the following respects (ef. p. 77).

Angle of preoperele slightly acute. No pseudobranchiae. Gill
rakers 8-9 + 3-5 on first arch. Premaxillary at least as long as

82 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

toothed portion of maxillary, sometimes slightly longer. Head
and trunk longer than tail, proportionately more so in adult.
Origin of dorsal fin slightly behind middle of body length, close
to mid-body in young. Origin of anal fin slightly behind a ver-
tical from last dorsal ray. Ventral bases slightly in front of
dorsal origin. Photophores all relatively inconspicuous, some on
head (especially the SO and lower anterior OP) obscured by
bone. On upper part of head in addition to ORB and OP, one
small organ above anterior toothed portion of maxillary ; a second
one level with and not far behind it; and one small photophore
beneath end of maxillary. Area just below symphysis of lower
jaw blackish and occupied by two short series of 2-4 minute ring-
like spots (photophores?). A row of about 10-13 minute photo-
phores on posterior portion of lower jaw, the first one enlarged.
No patches of luminous tissue on head or body as far as known
except a pair of narrow strips of pale tissue, possibly luminous,
below eighth to eleventh or twelfth IV photophores. AC photo-
phores evenly spaced. BR 8-9. IV 30-83. VAV 12-14. AC 28-30.
IC 70-75. OA 45-48, reaching about to or beyond middle of anal
fin and followed by 12-17 much smaller organs, which reach
caudal base in complete specimens. Lateral line row 65-68. Anal
rays 36-41. Vertebrae 63.
The subgenus contains only the following species.

DIPLOPHOS MADERENSIS (Johnson) 1890

This species is known only in the North Atlantic, and the only
adults recorded in the literature have been found at Madeira.
It is therefore of interest that during a recent cruise made by
the Oregon twenty specimens, 94.5-144.5 mm. in standard length,
were caught in a trawl off Surinam at Station 2008, 7° 38’ N.,
54° 43’ W., 7 November 1957, in 250 fathoms (457 meters). I
have also examined five specimens, 98.5-133 mm. in standard
length, sent by Mr. G. E. Maul from Madeira, where they were
found on the beach at Funchal in November 1954; and one young
specimen, standard length 32 mm., from the Bahamas, reported
under the name Diplophos minutus by Parr (19387, p. 46).

The following counts have been made on adult specimens:
dorsal rays 12-13, anal rays 38-41, pectoral rays 10-11, ventral
rays 8, branchiostegal rays 11-3, gill rakers on first arch 8-9 + 3-5,

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 83

vertebrae 63 (counted on an X-ray photograph of one Madeiran
specimen). BR 8-9; IV 30-33, the last one or two smaller and
situated just in front of or on ventral bases; VAV 12-18; AC
28-30; IC 70-75; OA 45-48, followed by smaller organs to caudal
base; lateral line ca. 65-68.

Each premaxillary has a row of 3-7 enlarged teeth of varying
sizes, the second and third the largest, and a few very small
teeth between the fangs; all of these teeth are inclined slightly
backward. On the maxillary is a row of ecloser-set, unequal teeth,
all straight and all smaller than the long premaxillary teeth.
On each mandible is a row of widely spaced fangs and posterior-
ly a few smaller teeth between the fangs. Anteriorly, there are
no small teeth between the fangs but there is an outer row of
4-6 teeth, smaller than the fangs but larger than the interspace
teeth.

There are 1-6 small teeth on each side of the vomer; a row of
9-15 small teeth on each palatine, the first one or two slightly
enlarged; a patch of minute teeth on each pterygoid; and a
cluster of small teeth near the tip of the tongue. It should be
noted that Maul (1948, p. 33) found no vomerine teeth on three
specimens from Madeira although these are present in all speci-
mens examined, including six from Madeira; and that Welsh
(1923, p. 1) found no palatine teeth on juvenile specimens from
the Bahamas and Florida. Welsh also reported eleven BR photo-
phores, while specimens examined have only eight or nine, and
a higher number has not been mentioned by other authors. It is
possible that the two lower opereular organs were included in
Welsh’s count.

The scales are mostly lost but it is evident that the dorsal
and ventral surfaces are fully sealed, neither ‘‘rugosely warted”’
as described by Johnson (1890, p. 458), nor ‘‘keeled’’ as de-
scribed by Maul (loc. cit.).

In addition to the IC, OA and lateral line rows, the following
rows of photophores are present on the body: between the OA
and the lateral line, three rows of smaller organs, the first of
these reaching about to a vertical from the middle of the anal
fin and continued posteriorly as still smaller round spots, some-
times to the end of the anal fin, sometimes to the base of the
caudal; the second is traceable in one specimen nearly to the end

84 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

of the anal but is shorter in other specimens; and the third is
obscure, apparently shorter, not reaching the anal origin. Above
the lateral line are two rows of obscure spots, reaching at least
beyond the end of the dorsal fin, sometimes to a vertical from
the middle of the anal fin; and, in two specimens, a third row
commencing in one instance above the other two, in the other
ease beneath the anterior portion of the dorsal fin. There is also
one row of small round organs between the IC and OA, reaching
to beneath the last of the larger OA organs.

The following measurements, in per cent of the standard
length, are based on six western Atlantic specimens, standard
leneth 94.5-144 mm., and five specimens from Madeira, standard
leneth 98.5-133 mm. The figures in parentheses refer to the
latter. Depth 15.8-16.9 (15.5-16.8) ; head 20.1-21.8 (21.1-21..6) ;
snout 4.4-5.1 (4.5-5.1) ; orbit 3.6-4.2 (3.3-3.8) ; interorbital width
at center of eve 3.7-4.1 (4.0-5.1); upper jaw 14.6-15.5 (14.8-
15.6) ; premaxillary 7.4-7.9 (7.5-7.9) ; toothed portion of maxil-
lary 6.9-7.8 (7.2-7.8); tip of snout to dorsal origin 51.8-53.1
(49.5-52.2), to anal origin 61.5-62.5 (59.0-63.0), to ventral bases
45.5-46.8 (45.2-46.3) ; distance between first anal ray and base of
middle caudal rays 36.5-39.2 (37.6-41.2), last anal ray and base
of middle caudal rays 5.9-7.6 (6.0-7.2), last dorsal ray and base
of middle caudal rays 38.1-40.2 (38.7-40.0), ventral bases and
anal origin 14.7-16.9 (14.8-17.6) ; least depth of caudal peduncle
5.4-5.9 (4.3-5.1) ; dorsal base 7.2-8.7 (8.0-8.7); anal base 30.3-
31.8 (30.0-34.0) ; pectoral length ca. 16.2 (one specimen only) ;

ventral length 9.75-11.1 (all broken in Madeiran specimens).

YARRELLA Goode and Bean 1895

Yarrella Goode and Bean, 1895, Ocean. Ichth., p. 103, type species Yarrella
blackfordi Goode and Bean 1895, Gulf of Mexico, 324 fathoms (593
meters) ; Jordan and Evermann, 1896, Bull. U. S. Nat. Mus., 47: 583.

Lychnopoles Garman, 1899, Mem. Mus. Comp. Zool., 24: 244.

Diplophos Parr, 1931, Bull. Bingham Oceanogr. Coll., 2 (4): 11 (part,
Lychnopoles in synonymy).

Manducus Fowler, 1936, Bull. Amer. Mus. Nat. Hist., 70: 1202 (part, Lych-
nopoles in synonymy ).

Generic characters. Eye normal, moderate. Snout longer than
orbit. Interorbital width at center of eye greater than diameter
of orbit. Mouth large, oblique; edge of premaxillary straight

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 85

laterally ; toothed edge of maxillary shghtly convex, nearly reach-
ing preopercle. Premaxillary longer than toothed portion of
maxillary. Angle of preoperele quite acute. Teeth of upper jaw
biserial on premaxillary, those of inner row smaller and eurving
inward; uniserial on maxillary. Teeth of lower jaw biserial, simi-
lar to those of premaxillary but somewhat smaller. Vomer with
one to three small teeth on each side. Palatines each with a short
row of two to six small teeth. Pterygoids usually toothless but
a few specimens seen with one or several microscopic teeth
present. Tongue toothless. Gill rakers 12-16 + 6-7 = 18-22 on
first arch. Spines on inner edge of first gill arch very short, a
row of minute spines below each one. No pseudobranchiae. Scales
present, very deciduous. Anus close to anal fin. Head and trunk
longer than tail. Dorsal origin about in middle of body length
or slightly behind it. Anal origin beneath middle or front of
dorsal fin. Ventral bases well ahead of doral origin. No adipose
fin. ORB 1, in front of lower margin of eye. OP 3, obscure,
upper one level with upper edge of pupil, lower ones level with
end of maxillary. SO present, slightly behind symphysis and
somewhat laterally situated. BR 11-13. Body with ventral and
lateral rows of photophores typical of family and also several
rows above these. Photophores present on isthmus. IV 9 + 3-4
+ 11-12 = 23-25. VAV 9-12. AC 20-28, straight, 6-10 of them
behind anal fin. IC 52-64. OA probably always about 50 or
more, reaching caudal base. No additional photophores on head,
and no patches of luminous tissue on head or body as far as
known. Fin rays: dorsal 14-16 (17 in one specimen of black-
fordi), anal 28-31, pectoral 8-10, ventral 6-7. Branchiostegal
rays 13-16, no spines at bases. Vertebrae 45-54.

Remarks. Yarrella shows relationship to Triplophos in the
arrangement of the body photophores, the long premaxillary, and
the dentition; and to Polymetme in dentition, the position of the
fins, and many proportions. An examination of two specimens
from the type lot of Lychnopoles argenteolus Garman has shown
them to be congenerie with Y. blackfordi. As in the latter, the
toothed portion of the maxillary is slightly shorter than the
premaxillary, the dentition is similar (although the teeth of
Y. argenteola are somewhat smaller than those of Y. blackfordt),
and the pattern and arrangement of the body photophores are

86 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

similar in the two species. Specifie differences are set forth in
Table 3.

The discovery that Yarrella has more than two rows of photo-
phores on the body elarifies the confusion that has resulted in
the past from a misunderstanding of the genus, and necessitates
the removal of all species described under the name except the
type species, blackfordi. Although the species referred here to
Polymetme do appear to be somewhat related to Yarrella, the
rather common, small form, Pollichthys mauli (Poll), confused
with Y. blackfordi until renamed Y. mauli by Poll (1953, p. 59),
is quite different and, as noted by Jespersen and Taning (1919,
p. 223), is more closely related to Vinciguerria.

YARRELLA BLACKFORDI Goode and Bean 1895
Figure 2

Yarrella blackfordi Goode and Bean, 1895, Ocean. Ichth., p. 108, fig. 121;
Jordan and Evermann, 1896, Bull. U. S. Nat. Mus., 47: 584; 1900,
op. cit., fig. 249; Poll, 1953, Rés. Sci. Exp. Océanogr. Belge (1948-
1949), 4 (2), (3): 56, fig. 22; Springer and Bullis, 1956, Spee. Sci.
Rep. U. S. Dept. Int., Fish., 196: 50 (part).
The following unrecorded specimens have been examined.
Gulf of Mexico, Oregon: Station 126, 29° 02’ N., 88° 34.5’ W..,
23 September 1950, 195 fathoms (357 meters) ; one specimen,
standard length ca. 202 mm. Station 279, 29° 11’N., 86° 52’ W..,
24 February 1951, 305 fathoms (558 meters) ; three specimens,
180-200 mm. Station 548, 27° 38.2’N., 94° 59.4’ W., 16 April
1952, 350-400 fathoms (640-732 meters) ; two specimens, 170 and
261 mm. Station 597, 29° 13’ N., 87° 59’ W., 10 July 1952, 280
fathoms (512 meters); two specimens, 212 mm. Station 640,
20° O1’ N., 88° 24’ W., 19 September 1952, 355-475 fathoms (649-
869 meters) ; two specimens, 200 and 212 mm. Station 1019,
24° 16’ N., 83° 22’ W., 16 April 1954, 375 fathoms (686 meters) ;
one specimen, 200 mm. Station 1272, 28° 20’N., 89° 46’ W.,
8 March 1955, 250 fathoms (457 meters), 49°F. at bottom; two
specimens, U.S.N.M. No. 157908, 204 and 213.5 mm.
Off northern South America, Oregon: Station 1980, 10° 10’ N.,
59° 54’ W., 3 November 1957, 350 fathoms (640 meters), 58.6°F.
at bottom; one specimen, 163 mm. Station 2009, 07° 40’ N., 54°

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 87

Oregon Station No. 2010. Standard length 193 mm.

Figure 2a. Vertical section, enlarged,
of area indicated by arrows.

Yarrella blackfordi Goode and Bean,

Figure 2.

88 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

47’ W., 7 November 1957, 300 fathoms (549 meters) ; five speci-
mens, 190-200 mm. Station 2010, 07° 44’ N., 54° 40’ W., 7 No-
vember 1957, 350 fathoms (640 meters) ; fifty-four specimens,
133-222 mm. Station 2011, 07° 46’ N., 54° 36’ W., 7 November
1957, 400 fathoms (732 meters) ; eight specimens, 144.5-274 mm.

One specimen, 222 mm., Stanford University No. 9486, Alba-
tross Station” 23876 (ype locality), 29°03" 157 N. 887167 We
Gulf of Mexico, 324 fathoms (592 meters).

One specimen, 148.5 mm., University of Miami Marine Labora-
tory No. 1678, Antilles, Gulf of Mexico, 28° 36’ N., 89° 49” W..,
234 fathoms (428 meters).

The only certain previous records of Y. blackfordi are the type
series (three specimens) from the Gulf of Mexico and thirty-
nine specimens reported by Poll (1953, p. 56) from the eastern
Atlantic off Africa, 5°-11°S. The two larger specimens men-
tioned by Longley and Hildebrand (1941, p. 15), taken south of
Tortugas in 672-686 meters, may have been blackfordi, but the
smaller one, which I have examined, belongs to Polymetme cory-
thaeola. Other Atlantic reports were of undescribed material
and either cannot be identified or are referable to Pollichthys
mault. Two little specimens tentatively identified as Y. black-
fordi by Koefoed (1958, p. 6) have been examined and are not
referable to this species. The larger one is a young Gonostoma,
sp. indet., and the smaller, which is in very poor condition,
appears to be a juvenile maurolicid but cannot be identified
further.

The skin of this species is extremely fragile and is almost
entirely lost in most of the specimens at hand. In fact, not a
single specimen has its full complement of skin, although many
from Oregon Station 2010 retain much of it. Many of the photo-
phores, especially those of the lateral and accessory series, have
been lost with the skin and the condition of other recorded
specimens was apparently similar. Dr. Leonard P. Schultz has
been kind enough to examine the types of Y. blackfordi and has
written (in litt., 1956) that these still retain shreds of skin on
which are found portions of the rows of photophores above the
lateral series.

The photophores are moderate to small in size and are rather
inconspicuous. There are usually 12 BR (13 in one specimen)

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 89

and other counts are as follows: IV 9 + 3-4 + 11-12 = 24-25, the
three or four pre-pectoral organs smaller than those preceding
or following them; VAV 12; AC 24-27 (28 in one specimen),
eight to ten of them behind anal fin; OA probably always reach-
ing caudal base, count indeterminable on all but three specimens,
on which were counted 52, 52 and 53 photophores. Above the
OA are three rows of photophores, the first of these (just above
the OA), apparently reaching a little beyond the ventral bases and
numbering about 15-16 organs. The second and third rows reack
the caudal base but the number of photophores, which are very
small posteriorly, is indeterminable. In addition to the photo-
phores described there are clusters of minute round organs on
the edges of the scale pockets. These are associated with all serial
photophores above the ventral row (IC) and also form two rows
above the uppermost row of serial photophores, both rows of
clustered organs reaching the caudal base.

TRIPLOPHOS Brauer 1902

Triplophos Brauer, 1902, Zool. Anz., 25: 282; type species Triplophos
elongatus Brauer 1902 = Photichthys hemingi McArdle 1901; 1906,
Wiss. Ergebn. Deutschen Tiefsee Exp. Valdivia, 15 (1): 98; Norman,
1930, Discovery Rep., 2: 296; Misra, 1953, Rec. Indian Mus., 50: 398.

Generic characters. Eye normal, moderate. Snout shorter than
orbit. Interorbital width at center of eye equal to diameter of
orbit or a little shorter. Mouth large, shghtly oblique; edge of
premaxillary straight, toothed edge of maxillary slightly convex,
nearly reaching preopercle. Premaxillary much longer than
toothed portion of maxillary, which is very short and scarcely
enters the gape. Angle of preopercle very acute. Premaxillary
teeth biserial, well spaced, those of inner row curving inward

(described as uniserial by Brauer). Maxillary with only a few

teeth in a single row. Teeth of lower jaw in specimens examined

similar to those of premaxillary, outer row set slightly lower and
curving inward; described as uniserial by other authors. Vomer

toothless or with one tooth on each side. Palatines each with a

short row of small teeth, the anterior one sometimes slightly

enlarged. Pterygoids with or without teeth. Tongue toothless.

Gill rakers 14-16 + 9 = 23-25 on first arch. Spines on inner edge

of first gill arch short, a row of minute prickles below each one.

90 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

No pseudobranchiae. Scales present but very deciduous. Anus
elose to anal fin. Head and trunk much shorter than tail. Dorsal
origin far in advance of middle of body length. Anal origin
beneath last dorsal ray or just behind a vertical from it. Ventral
bases a little in front of dorsal origin. No adipose fin. ORB 1,
below center of eye. OP 3 (more easily seen from inside operele),
upper one level with upper margin of eye, lower anterior one
beneath maxillary, lower posterior one on same level. SO present,
slightly behind symphysis, hidden by lower jaw. BR 8-13. Addi-
tional photophores on head: one, large, above maxillary; and a
row of tiny organs above and close to premaxillary. Body with
ventral and lateral rows of photophores typical of family and
also several rows or partial rows above these. Photophores pres-
ent on isthmus. IV 24-30, one or two raised toward pectoral base,
those following commencing below the last raised organ. VAV
d-7. AC 35-41, straight, two or three of them behind anal fin.
IC 68-76. OA 50-56, the first 9-11 lower than those following.
Above OA three or four additional rows of photophores. No
patches of luminous tissue on head or body as far as known.
Fin rays: dorsal 10-12, anal 54-63, pectoral 9-11, ventral 6-7.
Branchiostegal rays 11-14 (17?), no spines at bases. Vertebrae
probably ca. 60 (counted from X-ray photograph of one speci-
men, indistinct on tail).

Remarks. Triplophos stands apart from all other gonostomatid
genera in the relative lengths of the premaxillary and maxillary
bones, and from all except Danaphos in the advanced position
of the dorsal fin. It is perhaps most closely related to Yarrella.
According to Brauer (1908, pp. 27, 123) Triplophos is similar
to Diplophos in some ways, but the structure of the photophores
is close to that of Maurolicus and Valenciennellus, as well as the
Sternoptychidae. The fact that the premaxillary forms almost
the entire upper jaw somewhat strengthens the relationship
between Triplophos and Yarrella, the latter having a lengthened
premaxillary as well as additional photophores on the body.
Brauer’s statement (1906, p. 99) that the premaxillary is short
in Triplophos must have been an error. In specimens at hand it
forms almost the entire upper border of the mouth, the maxillary
being reduced to a small round knob bearing five or six small
teeth.

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS of

The generic diagnosis has been based in part on previous de-
scriptions of specimens from the eastern Atlantic and the Indian
Ocean, and in part on the newly discovered western Atlantic
specimens listed below.

TripLopHos teMiInet (McArdle) 1901

Photichthys hemingi McArdle, 1901, Ann. Mag. Nat. Hist., (7) 8: 521;
Aleock and MeGilchrist, 1905, Ill. Zool. Investigator, Fishes, pl. 36,
fig. 2 (holotype).

Triplophos elongatus Brauer, 1902, Zool. Anz., 25: 282; 1906, Wiss. Ergebn.
Deutschen Tiefsee Exp. Valdivia, 15 (1): 99, pl. 7, fig. 4, text fig. 41
(elongatum) ; 1908, op. cit., 15 (2): 27, 176, pl. 22, figs. 4-7, pl. 36,
fig. 8 (elongatum) ; Misra, 1950, Rec. Indian Mus., 45: 415 (elongata) ;
1953, op. cit., 50: 399, fig. 15b.

Triplophos hemingi Lloyd, 1909, Mem. Indian Mus., 2: 150; Norman, 1930,
Discovery Rep., 2: 296; Poll, 1953, Rés. Sci. Exp. Océanogr. Belge
(1948-1949), 4, (2), (3): 61, fig. 25.

Three specimens, 106-130 mm. in standard length, Oregon Sta-
tion 1907, Caribbean Sea off Central America, 12° 25’ N., 82° 23’
W., 11 September 1957, trawl, 400-425 fathoms (732-778 meters) ;
one specimen, 122 mm., Oregon Station 1916, same area, 13° 18’
N., 82°12’ W., 12 September 1957, trawl, 350 fathoms (640
meters) ; two specimens, ca. 176 mm., Oregon Station 2007, off
Surinam, 07° 34’N., 54°49’ W., 7 November 1957, trawl, 225
fathoms (411 meters), 47.5°F. at bottom.

Although these specimens differ slightly from descriptions
based on Indian Ocean specimens, the differences do not seem
important enough to warrant their separation as a distinct
species. All six specimens examined have seven VAV _ photo-
phores, in contrast to a count of five in Indian Ocean examples.
However, Brauer’s figure (1906, p. 99, fig. 41) shows six photo-
phores between the ventral bases and the anal origin; and Poll
(1957, in litt.) has written that VAV counts in sixteen of the
eastern Atlantic specimens reported by him in 1953 are seven,
six, and (in one specimen) five. The IV count of our specimens
is lower, 24-25 (29-30 in Indian Ocean specimens). The first row
of serial photophores above the OA has been described and fig-
ured as short, and the one above it with a count of 36-43. In
specimens examined this short row appears to be continuous

92 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

with the row that extends on the tail, but the first nine photo-
phores are on a lower level than those following. Similarly, the
first nine organs of the OA series are shehtly below those follow-
ing them. The uppermost rows of photophores are smaller and
do not reach the tail.

The only significant proportional differences noted (Table 4),
a shorter pre-anal distance and a longer anal base in Indian
Ocean specimens, need to be verified by direct comparison of ma-
terial from both oceans. Further information is also needed on
the jaw teeth. The teeth on the premaxillary have been described
as biserial, as they are in specimens examined, by all authors
except Brauer. However, all authors have stated that the lower
jaw teeth are in a single row, and in the western Atlantic speci-
meus there is, along most of the mandible, an outer row of teeth,
curving inward and set somewhat below the inner row.

The vomer is toothless or has one minute tooth on each side.
Each palatine has a short row of 2-6 teeth, microscopic in size
excepting the first one or two. The pterygoids and tongue are
toothless.

The photophores are moderate in size. The row of tiny organs
above the premaxillary is usually difficult to see, or damaged.
The OA organs number 50-55, the first nine slightly lower than
those following, and those on the tail diminishing in size pos-
teriorly. These reach the end of the anal fin in one specimen (ca.
118 mm. in standard length), not quite to the end of the anal
fin in others. Above the OA are three or four rows of small
photophores, 39-47 in the lowest of these, which reaches beyond
the middle of the anal fin in all but the smallest specimen; the
first nine are on a slightly lower level. The upper rows of photo-
phores are not complete. An additional photophore is present
close to the opercle between the IC and the OA.

POLYMETME McCulloch 1926

Polymetme McCulloch, 1926, Biol. Res. Endeavour, 5: 166; type species
Polymetme illustris McCulloch 1926, southern Australia; Barnard, 1927,
Ann. So. Afr. Mus., 21: 1018; McCulloch, 1929-30, Mem. Australian
Mus., 5: 51.

Yarrella Barnard, 1925, Ann. So. Afr. Mus., 21: 148; Norman, 1930, Dis-
covery Rep., 2: 288 (part); Matsubara, 1938, Jour. Imp. Fish. Inst.
Tokyo, 33: 44 (part); Smith, 1949, Sea Fishes So. Afr., p. 104 (part) ;
Misra, 1953, Ree. Indian Mus., 50: 398.

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 93

Generic characters. Eye normal, moderate. Snout about equal
to diameter of orbit. Interorbital width at center of eye about
equal to or a little less than diameter of orbit. Mouth large,
oblique, edge of premaxillary straight ; toothed edge of maxillary
slightly convex, nearly reaching preopercle. Premaxillary more
than half as long as toothed portion of maxillary. Angle of pre-
opercle acute. Teeth of upper jaw biserial on premaxillary, uni-
serial on maxillary. Teeth of lower jaw biserial, those of outer
row larger. Vomer with one to three teeth on each side. Palatines
each with a short row of small teeth. Pterygoids with a rather
large patch of minute teeth (not described in type species, ilus-
tris). Tongue toothless. Gill rakers 9-12 + 5-8 = 15-19 on first
arch. Spines on inner edge of first gill arch very short, a cluster
of minute prickles below each one. No pseudobranchiae. Scales
present but very deciduous. Anus close to anal fin. Head and
trunk a little longer than tail. Dorsal origin about in middle of
body length. Anal origin beneath end of dorsal fin or just behind
a vertical from its last ray. Ventral bases noticeably ahead of
dorsal origin. Adipose fin above end of anal fin. ORB 1, in front
of lower margin of eye. OP 8, upper one smaller, about on a
level with upper border of pupil or higher; lower anterior one
just behind maxillary; lower posterior one shghtly higher. SO
present. BR 9-10. No other photophores on head. Body with
two rows of photophores; photophores present on isthmus. IV
19-21, ninth or tenth raised toward pectoral base. VAV 7-8.
AC 21-25, first one or two elevated and sometimes elongate, five
to eight of them behind anal fin. 1C 50-54. OA 16-18, not always
perfectly straight, ending above next to last VAV. No additional
photophores and no patches of luminous tissue on head or body
as far as known (but see p. 99). Fin rays: dorsal 11-13, anal
24-33, pectoral 9-11, ventral 7 (8?). Branchiostegal rays 12-14,
no spines at bases. Vertebrae 45, counted from an X-ray photo-
graph of one western Atlantic specimen.

Remarks. Polymetme is apparently related to Yarrella, judg-
ing by the relatively long premaxillary and the partially biserial
teeth on the jaws. It differs, however, in the number and arrange-
ment of the serial rows of photophores, in the larger and more
conspicuous individual photophores, and in having an adipose fin.

94 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Polymetme is similar to Pollichthys in the possession of an adi
pose fin, the length of the anal fin, the number of BR photo-
phores, and in having only two rows of body photophores. It
differs from that genus principally in having only one ORB, ten
pectoral rays (eight in Pollichthys) and two rows of teeth on
the premaxillary.

This genus has for many years been considered synonymous
with Yarrella and is reinstated here to include the following
species, some or most of which are probably synonymous: Diplo-
phos corythaeolum Aleock, Andaman Sea (and Yarrella cory-
thaeola Poll, West Africa) ; Yarrella africana Gilchrist and von
Bonde, Natal coast; Polymetme illustris McCulloch, Australia
(and Yarrella blackfordi illustris Matsubara, Japan) ; Yarrella
blackfordi elongata Matsubara, Japan; and Yarrella surugaensis
Matsubara, Japan. In the absence of comparative material it is
not possible to determine the number of valid species belonging
to the genus, but africana is probably a synonym of corythaeola.
surugaensis is probably the same as illustris, and actually all
four of these species may be synonymous. P. surugaensis was
distinguished from P. idlustris on the basis of the irregularity of
the OA photophores and slight differences in the number of anal
rays, gill rakers, and AC photophores. These small variations do
not seem to warrant the separation of surugaensis, particularly
in view of the fact that the OA photophores are usually irregular
in specimens examined from the western Atlantic.

In the original diagnosis of the genus, McCulloch (1926, p.
166) stated that the upper jaw is formed largely by the maxil-
lary, and the figure of the type of dlustris (McCulloch, loc. cit.,
pl. 45, fig. 1) shows this bone to be proportionately somewhat
longer than in Atlantic specimens examined. The maxillary is
also proportionately longer in Matsubara’s figures of dlustris
(1938, p. 42, fig. 4) and surugaensis (1948, p. 74, fig. 22). On
the other hand, in Alcock’s figure of the Indian Ocean corythaeola
(Ill. Zool. Investigator, 1899, pl. 25, fig. 3) the premaxillary is
nearly as long as the toothed portion of the maxillary, as it is in
specimens at hand, and the same is true of Matsubara’s figure
of the Japanese elongata (1938, p. 45, fig. 5), Poll’s figure of an
eastern Atlantic specimen (1953, p. 58, fig. 23), and Smith’s
South African specimen (1949, p. 104, fig. 152). The relative

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 95

length of these two bones was not described for africana by
Gilchrist and von Bonde (1924, p. 8, pl. 1, fig. 2) nor can they
be distinguished in the figure of the type. Norman (1930, p. 289),
who had both Australian and Indian Ocean specimens at hand,
unhesitatingly synonymized illustris with corythaeola and his
opinion should probably be upheld for the present.

Polymetme elongata (Matsubara) 1938, from Japan, is prob-
ably a distinct species. It differs from both dlustris and
corythaeola in having a smaller head, shorter snout and upper
jaw, and lesser depth, as noted by Matsubara, and it may differ
further, from illustris at least, in having a relatively longer
premaxillary.

Yarrella blackfordi microcephala Matsubara (1941, p. 1) prob-
ably does not belong to the genus Polymetme. It differs in the
following significant characters: tail very slender; anal origin
beneath fifth dorsal ray; BR 8; OA 25, ending above fourth anal
ray. It cannot, however, be assigned to any genus examined by
me. In the position of the anal fin, the slender tail, and the
number of BR and IV photophores it is like Pollichthys, but it
differs from that genus in having a single ORB, two rows of teeth
on the premaxillary, and ten pectoral rays (eight in Pollichthys ).
The species has not been figured.

POLYMETME CORYTHAEOLA (Alcock) 1898

Diplophos corythaeolum Alcock, 1898, Ann. Mag. Nat. Hist., (7) 2: 147:
1899, Ill. Zool. Investigator, Fishes, pl. 25, fig. 3; 1902, Nat. Indian
Seas, p. 239, fig. 38; Parr, 1931, Bull. Bingham Oceanogr. Coll., 2
(ye ie als}.

Photichthys corythaeolus Alcock, 1899, Cat. Indian Deep-sea Fishes, p. 142;
Brauer, 1906, Wiss. Ergebn. Deutschen Tiefsee Exp, Valdivia, 15 (1):
92, 374.

Yarrella africana Gilchrist and von Bonde, 1924, Rep. Fish. Mar. Biol. Surv.
So. Afr., 3 (7): 8, pl. 1, fig. 2; Barnard, 1925, Ann. So. Afr. Mus.,
21: 148.

? Polymetme illustris McCulloch, 1926, Biol. Res. Endeavour, 5: 167, pl. 45,
fig. 1; 1929-30, Mem. Australian Mus., 5: 51; Whitley, 1948, Fish. Bull.
West. Australia Fish. Dept., 2: 11.

Polymetme africana McCulloch, 1926, Biol. Res. Endeavour, 5: 167; Barnard,
1927, Ann. So. Afr. Mus., 21: 1018.

Polymetme corythaeolus McCulloch, 1926, Biol. Res. Endeavour, 5: 167.

96 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Yarrella corythaeola Norman, 1930, Discovery Rep., 2: 289; Barnard, 1937,
Ann. So. Afr. Mus., 32: 46; Norman, 1939, Sci. Rep. John Murray
Exp. 1933-34, 7 (1): 19; Bertin, 1939, Bull. Mus. Hist. Nat. Paris,
11: 379; Herre, 1941, Mem. Indian Mus., 13: 336; Smith, 1949, Sea
Fishes So. Afr., p. 104, fig. 152; Misra, 1950, Rec. Indian Mus., 45:
415; 1953, op. cit., 50: 398, fig. 17a; Poll, 1953, Rés. Sci. Exp. Océanogr.
Belge (1948-1949), 4 (2), (3): 58, fig. 23; Springer and Bullis, 1956,
Spec. Sci. Rep. U. S. Dept. Int., Fish., 196: 50.

? Yarrella corythaeola Kamohara, *1936, Zool. Mag., Tokyo, 48.

? Yarrella illustris Kamohara, 1938, Offshore Bottom Fishes Prov. Tosa,
p. 9; Matsubara, 1955, Fish Morph. Heir., 1: 221, pl. 16, fig. 61.

? Yarrella blackfordi illustris Matsubara, 19388, Jour. Imp. Fish. Inst. Tokyo,
33: 42, fig. 4; 1940, Suisan Kenkiu-shi, 35: 319; Kamohara, 1952, Rep.
Kochi Univ. Nat. Sei., 3: 16; Haneda, 1952, Pacific Sci., 6: 138.

Yarrella blackfordi corythaeola Matsubara, 1938, Jour. Imp. Fish. Inst.
Tokyo, 33: 44.

Yarrella blackfordi africana Matsubara, 1938, loc. cit.

Yarrella blackfordi Longley and Hildebrand, 1941, Publ. Carnegie Inst.
Washington, 535: 15 (part); Springer and Bullis, 1956, Spec. Sci. Rep.
U.S. Dept. Int., Fish., 196: 50 (part).

? Yarrella surugaensis Matsubara, 1943, Jour. Sigen. Kenk., 1: 74, fig. 22;
1955, Fish Morph. Heir., 1: 221, pl. 17, fig. 63; Kamohara, 1957, Res.
Rep. Kochi Univ., 6 (5): 1.

Photichthys argenteus Bruun, 1950, Atlantide Rep., 1: 20, fig. 18.

The following western Atlantic specimens have been examined.

Gulf of Mexico: One specimen, standard length 125 mm., Uni-
versity of Miami Marine Laboratory No. 49:769, Antilles, 29° 13’
N., 88° 2’ W., 200 fathoms (366 meters).

Gulf of Mexico, Oregon: One specimen, 113 mm., Station 382,
29° 11.5’ N., 88° 07.5’ W., 21 June 1951, 190-210 fathoms (348-
384 meters) ; five specimens, 74-112.5 mm., Station 1054, 19° 37’
N., 92° 40’ W., 15 May 1954, 200 fathoms (366 meters), 52°F. at
bottom ; two specimens, 86 and 127 mm., Station 1055, 19° 14’ N.,
93° 00’ W., 15 May 1954, 225 fathoms (411 meters), 50°F. at
bottom; two specimens, 195 and 207 mm., U.S.N.M. No. 157907,
Station 1272, 28° 20’ N., 89° 46’ W., 8 March 1955, 250 fathoms
(457 meters), 49°. at bottom; one specimen, 145 mm., U.S.N.M.
No. 157909, Station 1277, 28° 32’N., 86° 20’ W., 11 March 1955,
260 fathoms (475 meters), 48°F. at bottom; one specimen, 113.5
mm., U.S.N.M. No. 157899, Station 1407, 28° 07’ N., 89° 59’ W..,

*Starred references not seen.

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 97

20 September 1955, 258 fathoms (472 meters) ; one specimen,
129 mm., U.S.N.M. No. 157898, Station 1412, 27° 58’ N., 90° 41’
W., 21 September 1955, 150-175 fathoms (274-320 meters), 53°F.
at bottom; one specimen, 157 mm., Station 1541, 24° 28’N.,
83° 29’ W., 15 June 1956, 220 fathoms (403 meters) ; four speci-
mens, 121-134 mm., Station 1565, 29° 11’ N., 88° 02’ W., 22 June
1956, 240 fathoms (488 meters), 52.2°F. at bottom; one speci-
men, 146.5 mm., Station 1566, 29° 13’ N., 87° 54’ W., 22 June
1956, 250 fathoms (457 meters) ; one specimen, 123 mm., Station
1968, 29° 11’N., 88°03’ W., 24 September 1957, 240 fathoms
(438 meters).

Atlantic off Florida: One specimen, ca. 80 mm., U.S.N.M. No.
116937, Longley Collection (reported by Longley and Hilde-
brand, 1941, under the name Yarrella blackfordi) ; one speci-
men, 133.5 mm., Combat Station 446, 25°10’N., 79° 13’ W.,
23 July 1957, 250 fathoms (457 meters), between Florida and
the Bahama Islands.

Caribbean Sea: One specimen, 115.5 mm., collection of the
Museum of Comparative Zoology, Blake Exp. 1880, off Cayman
Brac, 247 fathoms (452 meters).

Western Caribbean Sea, Oregon: Three specimens, 113-140
mm. “Station 1871, 16°39’N., 82°26’ W.,. 22) Aucust 1957; 250
fathoms (457 meters) ; four specimens, 85.5-ca.134 mm., Station
1872, 16° 41’ N., 82° 20’ W., 22 August 1957, 300 fathoms (548
meters) ; one specimen, 158 mm., Station 1888, 16° 41’ N., 81° 02’
W., 23 August 1957, 250 fathoms (457 meters) ; two specimens,
102 and 151 mm., Station 1889, 16° 39’ N., 81° 01’ W., 24 August
1957, 250 fathoms (457 meters) ; two specimens, ca. 146 and 147
mm., Station 1902, 11° 27’ N., 83° 11’ W., 9 September 1957, 135
fathoms (247 meters) ; six specimens, 123-ca.140 mm., Station
1908, 11° 31’N., 83°09’ W., 9 September 1957, 150 fathoms
(274 meters) ; one specimen, 94 mm., Station 1919, 13° 30’N.,
82° 00’ W., 12 September 1957, 275-300 fathoms (503-549 me-
ters) ; four specimens, ca. 49-189 mm., Station 1921, 13° 33’N.,
81° 55’ W., 13 September 1957, 275 fathoms (503 meters) ; one
specimen, 134 mm., Station 1948, 16° 43’ N., 82° 44’ W., 16 Sep-
tember 1957, 275 fathoms (503 meters) ; seven specimens, 78-186
mm., Station 1945, 16°41’N., 82°40’ W., 16 September 1957,
250-300 fathoms (457-549 meters).

98 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Off northern South America: One specimen, ca. 63 mm.,
U.S.N.M. No. 44588, Albatross Station 2125, 11° 43’ N., 69° 09’
30” W., 18 February 1884, 208 fathoms (380 meters).

Off northern South America, Oregon: One specimen, ca. 108
mm. Station 199% 09° 177N,, 59° 19° W., 4° November (f9a%
250 fathoms (457 meters) ; four specimens, 83-152.5 mm., Station
1992, ca. 09° N., 59° W., 4 November 1957, 275 fathoms (503
meters); three specimens, 101.5-ca.140 mm., Station 2005,
07° 34’ N., 54°50’ W., 6 November 1957, 200 fathoms (3866
meters).

P. corythaeola was reported from the western Atlantic off
Tortugas, Florida, by Longley and Hildebrand (1941, p. 15)
under the name Yarrella blackfordi (see p. 88). The first At-
lantie specimens were described by Poll (1953, p. 58) from West
Africa, although Bertin (1939, p. 379) had earlier listed two
specimens from the Cape Verde Islands, taken by the Talisman.
Bruun (1950, p. 20) reported specimens under the name
Photichthys argenteus Hutton, from the Gulf of Guinea, in either
530-850 or 650-260 meters. The position of the anal fin and the
number and position of the lateral row of photophores (loc. cit.,
p. 21, fig. 13) show these specimens to be Polymetme corythaeola.
The species is otherwise known from South Africa, the tropical
Indian Ocean, and possibly from the Pacific off Japan and
Australia.

The following counts have been made on specimens examined :
dorsal rays 11-138, anal rays 30-33, pectoral rays 10-11, ventral
rays 7, branchiostegal rays 12-13, gill rakers on first arch 11-12
+ 5. The photophores are large and conspicuous, their counts:
BR9;1V9+1+11 = 21, the tenth elevated; VAV 8; AC 24-25,
the first one or two shehtly elevated and sometimes elongate, the
last six or seven behind the anal fin; IC 53-54; OA 17 (an ad-
ditional small, incomplete organ on one side of one specimen),
not quite reaching a vertical from the anal origin.

There is some variation in the placement of the OA photo-
phores. The first two are usually joined basally, i.e., the re-
fleectors juxtaposed; two or three above the ventral fins are
usually smaller, narrower, and irregularly disposed; and the
last eight or nine are usually smaller and narrower, and some-
times a little higher than the pre-ventral organs. However, in

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 99

some specimens the first two OA are quite separate; in some,
all of the OA are more or less equal in size and all on one level,
and in one specimen the last eight organs are on a slightly lower
level than the anterior ones. These variations occur at times on
a single specimen, the left side differing shgehtly from the right.

No luminous tissue has been noted on the head or body of
this species, but on most specimens there is a fragile tube of
transparent skin on the belly between the ventral bases and the
anus (below the VAV photophores), sometimes partially torn
and probably lost in specimens lacking it. This structure is
usually quite colorless but in large specimens it is sprinkled with
minute black spots.

Small specimens have a proportionately longer head, larger
eye, longer upper jaw and premaxillary, a shorter distance be-
tween the ventral and anal fins, and a narrower caudal peduncle.
The young are also more compressed, and the body depth de-
creases more abruptly behind the head than in the adult. In
specimens less than about 100 mm. in standard length the pre-
maxillary and the toothed portion of the maxillary are about
equal in length, while in older specimens the maxillary is always
a little longer. The following measurements, in per cent of the
standard length, are given in two groups, the first figures repre-
senting specimens with a standard length of more than about
115 mm. (fourteen, 118.5-207 mm.), the second groups of figures,
in parentheses, representing specimens about 115 mm. or less
(ten, 60.5-115.5 mm.).

Depth 14.8-17.9 (14.8-16.8); head 21.0-23.6 (23.6-ca. 26.0) ;
snout 4.55-5.86 (5.1-6.16); orbit 4.2-4.83 (5.2-ca. 6.4) ; interor-
bital width at center of eye 3.938-4.8 (4.05-4.77) ; upper jaw 15.3-
17.6 (18.2-19.1) ; premaxillary 7.0-8.4 (8.5-ca. 9.75) ; toothed por-
tion of maxillary 7.7-9.17 (8.83-ca. 9.75) ; tip of snout to dorsal
origin 46.4-49.6 (47.2-49.5), to anal origin 54.4-58.8 (54.6-57.5),
to ventral bases 38.3-42.2 (ca. 38.3-42.5) ; distance between first
anal ray and base of middle caudal rays 41.9-45.0 (42.3-46.0), last
anal ray and base of middle caudal rays 12.1-16.7 (11.8-15.8), last
dorsal ray and base of middle caudal rays 40.5-43.1 (39.0-42.5),
last dorsal ray and adipose fin 17.9-20.2 (17.7-20.6), ventral base
and anal origin 15.0-17.95 (12.7-16.0) ; least depth of caudal
peduncle 5.8-6.65 (4.25-6.16) ; dorsal base 8.8-11.2 (8.9-ca. 10.75) ;

100 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

anal base 27.9-30.9 (28.1-ca. 31.9); pectoral length 11.8-17.5
(ca. 13.0 and 17.1-20.5) ; ventral length 7.95-9.77 (9.1-ca. 11.5).

Apparently the only character in which Atlantic and Indian
Ocean specimens differ is in the number of anal rays, 30-33 in
the Atlantic, 25-30 in the Indian.

PuoticHtHys Hutton 1872

Phosichthys Hutton, 1872, Cat. Fishes New Zealand, p. 55, type species
Phosichthys argenteus Hutton, 1872, Cook Strait, New Zealand.

Photichthys Hutton, 1873, Trans. New Zealand Inst., 5: 10 (emended
spelling); Gimther, 1887, Rep. Sci. Res. Voy. Challenger, Zool., 22:
177; Goode and Bean, 1895, Ocean. Ichth., p. 104; Collett, 1896, Bull.
Soe. Zool. France, 21: 94; Barnard, 1925, Ann. So. Afr. Mus., 21: 149;
McCulloch, 1926, Biol. Res. Endeavour, 5: 166; Norman, 1930, Discovery
Rep., 2: 292; Smith, 1949, Sea Fishes So. Afr., p. 104.

Generic characters. Kye normal, moderate or large. Snout
about equal to diameter of orbit. Interorbital width at center
of eye about equal to diameter of orbit. Mouth large, slightly
oblique; toothed edges of premaxillary and maxillary straight;
maxillary nearly reaching preopercle. Premaxillary about half
as long as toothed portion of maxillary. Angle of preopercle
slightly acute. Teeth of upper jaw uniserial, premaxillary with
one or two longer teeth and a few small ones; maxillary teeth
all rather short, unequal, curving slightly inward. Lower jaw
with several widely spaced longer teeth (as long as large pre-
maxillary teeth) ; smaller teeth between the long ones except
anteriorly, where there is an outer row of small, inwardly curved
teeth. Vomer with one or two teeth on each side. Palatines each
with a lone row of unequal teeth. Pterygoids toothless or each
with a small patch of minute teeth posteriorly. Tongue with or
without a few small teeth at tip. Gill rakers 11 + 4-5 = 15-16 on
first arch. Spines on inner edge of first gill arch short (a little
longer near angle), one or two minute spines below a few
anteriorly. No pseudobranchiae. Scales present, very deciduous.
Anus near anal fin, beneath twelfth to fourteenth VAV_ photo-
phore. Head and trunk more than twice as long as tail. Origin
of dorsal fin about in middle of body length. Origin of anal fin
well behind end of dorsal fin. Ventral bases slightly in advance
of dorsal origin. Adipose fin small, above middle or anterior
portion of anal fin. ORB 2, one near lower anterior margin of

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 101

eye, the other below its posterior margin; anterior one larger.
OP 3, upper one level with upper border of pupil; lower two on
same level, behind end of maxillary, anterior one slightly larger.
SO present. BR 17-18. No additional photophores on head.
Body with two rows of serial photophores; photophores present
on isthmus. IV 24-25, straight. VAV 15-17. AC 16-18, straight,
five to seven of them behind anal fin. IC 57-58. OA 33-84,
reaching above front of anal fin (about over third AC photo-
phore). No additional photophores and no luminous tissue on
body as far as known. Fin rays: dorsal 12-18, anal 23-26, pec-
toral 9, ventral 6-7. Branchiostegal rays 20-21, no spines at
bases. Vertebrae 51, including hypural (counted from X-ray
photograph of one specimen).

Remarks. Photichthys is closely related to Woodsia but differs
in higher meristic counts, in having the gill rakers more devel-
oped, and in body proportions. The generic diagnosis is based
on three specimens from New Zealand (p. 65) and also on
published accounts. In the specimens examined the palatine
teeth are unequal in size, some of the anterior ones are enlarged,
the posterior ones are all small, and a few are curved. There are
two small patches of minute pterygoid teeth and a few tiny teeth
on the tongue.

The genus contains only one species, P. argenteus Hutton,
which is known from New Zealand and in the South Atlantic
from ca. 32° S., 8° W. to Cape Point, South Africa.

Gonostoma Rafinesque 1810

Gonostoma Rafinesque, *1810, Ind. Ittiol. Sicil., p. 64; type species Gonos-
toma denudata Rafinesque, 1810, Mediterranean; Bonaparte, *1841,
Ieon. Fauna Ital., 3; Cuvier and Valenciennes, 1849, Hist. Nat. Poiss.,
22: 278; Giinther, 1864, Cat. Fishes Brit. Mus., 5: 391; 1887, Rep. Sci.
Res. Voy. Challenger, Zool., 22: 172 (part, Cyclothone in synonymy) ;
Moreau, 1891, Hist. Nat. Poiss. France, Suppl., p. 78; Goode and
Bean, 1895, Ocean. Ichth., p. 93 (part, not G. brevidens) ; Jordan and
Evermann, 1896, Bull. U. S. Nat. Mus., 47: 578 (part; not G. brevi-
dens); Collett, 1896, Bull. Soc. Zool. France, 21: 94; Brauer, 1906,
Wiss. Ergebn. Deutschen Tiefsee Exp. Valdivia, 15 (1): 70; Weber
and de Beaufort, 1913, Fishes Indo-Austr. Arch., 2: 120; Barnard,
1925, Ann. So. Afr. Mus., 21: 148; Norman, 1930, Discovery Rep., 2:

* Starred references not seen.

102 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

281; Fowler, 1936, Bull. Amer. Mus. Nat. Hist., 70: 229; Lozano Rey,
1947, Mem. R. Acad. Cien. Madrid, Ser. Cien. Nat., 11: 158; Smith,
1949, Sea Fishes So. Afr., p. 104; Misra, 1953, Rec. Indian Mus.,
50: 396; Mead and Taylor, 1953, Jour. Fish. Res. Bd. Canada, 10: 570.

Sigmops Gill 1883, Proe. U. S. Nat. Mus., 6: 256; type species S. stigmaticus
Gill, 1883, equals Gonostoma elongatum Ginther.

Neostoma Vaillant, in Filhol, *1884, Nature, Paris, 558; 1888, Exp. Sci.
Trav. Talis., Poiss., pp. 96, 385; type species N. bathyphilum Vaillant,
1884, 1888; Collett, 1896, Bull. Soc. Zool. France, 21: 95.

Cyclothone Goode and Bean 1895, Ocean. Ichth., p. 99 (part, C. bathyphila,
C. elongata); Jordan and Evermann, 1896, Bull. U. S. Nat. Mus., 47:
581 (part, C. bathyphila, C. elongata); Collett, 1896, Bull. Soc. Zool.
France, 21: 94 (part, C. grandis Collett); Alcock 1899, Descr. Cat.
Indian Deep-sea Fisehes, p. 139 (part, C. elongata); Fowler, 1936,
Bull. Amer. Mus. Nat. Hist., 70: 222 (part, C. bathyphilum).

Generic characters. Eye normal, moderate or small. Snout
longer than or about equal to diameter of orbit. Interorbital
width at center of eye longer than or about equal to diameter
of orbit. Mouth large, oblique; toothed edge of premaxillary
straight or slightly coneave; toothed edge of maxillary convex
or nearly straight, nearly reaching preopercle. Premaxillary
less than half as long as toothed edge of maxillary. Angle of
preopercle acute. Teeth of upper jaw uniserial, longer teeth
with smaller ones in interspaces. Lower jaw with a row of teeth
similar to those of upper jaw and, anteriorly only, a very short
outer row. Vomer usually with 1-2 small teeth on each side

(absent in denudatum). Palatines each with a single row of

teeth. A patch of teeth on pterygoids on each side and usually

a few additional teeth posteriorly on roof of mouth. Tongue with

or without teeth. Gill rakers 10-17 + 5-11 = 15-27 on first arch.

Spines on inner edge of first gill arch moderate or long, a row

of prickles or minute spines below each one. No pseudobranchiae.

Seales present or absent. Anus nearer anal origin than ventral

bases. Head and trunk about the same length as, or longer than,

tail. Dorsal origin about in middle of body length or behind it.

Anal origin opposite or in advance of dorsal origin. Ventral fins

well ahead of dorsal origin. Adipose fin present or absent. ORB 1

(obsolete in bathyphilum). OP 2 or 3, lower anterior one often

absent or obscured by bone (all obsolete in bathyphilum). SO

* Starred references not seen.

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 103

present (except in bathyphilum). BR 9 (not always visible on
bathyphilum). No other photophores on head, but luminous
glands or tissue usually associated with some of the photophores.
Body with two rows of serial photophores, but in bathyphilum
and gracile the upper row irregular and placed relatively high
on sides; no photophores on isthmus. IV 11-16, prepectoral
organs irregular or arched upward. VAV 3-10. AC 15-23. IC
32-43. OA 11-21, extending beyond ventral bases, first one or
two elevated (irregular in bathyphilum and gracile). Additional
photophores sometimes present on body. Luminous tissue usually
present in the procumbent caudal rays, either dorsally, ventrally,
or both. Fin rays: dorsal 10-18, anal 21-31, pectoral 9-13, ventral
6-8. Branchiostegal rays 10-14, spines present at inner bases of
most of them. Vertebrae 37-40.

Remarks. Gonostoma is closely related to Cyclothone, from
which it differs externally principally in dentition. It is also
similar to Bonapartia in many respects (dentition, positions of
dorsal and anal fins, most meristie characters).

There is much more variation among the species of Gonostoma
than is found in any other genus of the family with the exception
of Diplophos, which has been divided into two subgenera. How-
ever, all species of Gonostoma are held together by rather strik-
ing similarities and a division, even into subgenera, is considered
inadvisable. The inter-relationships of the species are not clear,
with the exception that denudatum and atlanticum are closely
related; and one or two common characters are at times shared
by species which seem otherwise not to be closely related. For
example, the anus is typically situated close to the anal fin (below
or close behind the penultimate VAV photophore) but in both
gracile and ebelingi it is more remote from the anal origin,
although never, as far as known, nearer the ventral bases. Simi-
larly, a relatively short tail has until recently been characteristic
only of the closely related species denudatum and atlanticum,
but the newly discovered ebelingi is also a short-tailed form.

In most gonostomatid genera the presence or absence of an
adipose dorsal fin is a distinctive character, but in Gonostoma
this fin may be well developed (elongatum, bathyphilum,
ebelingt), small (denudatum), or absent (atlanticum, gracile).
In species whose life-histories are known, the adipose appears

104 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

relatively late in the developmental period, whereas in other
sonostomatid fishes with an adipose fin it is present in very early
stages.

Gonostoma is unique in showing considerable variation in
the arrangement of the photophores. In other gonostomatid
genera the position of the ight organs is relatively constant. The
reduction in size and number of these organs in bathyphilum is
correlated with the greater depth of habitat of this species, but
the irregular placement of the OA in bathyphilum and in gracile,
the additional row of photophores near the dorsal profile of
gracile, and the many minute organs found on elongatum, and
possibly also on bathyphilum, represent a departure from the
usual generic limits. The recent discovery of a new species,
ebelingt, with ten VAV and twenty-one OA, further emphasizes
the variability found in the genus, which is otherwise character-
ized by a short ventral-anal space containing no more than five
VAV, and by having only eleven to fifteen OA.

The amount of scalation present is uncertain as the scales
are usually extremely deciduous. The body of denudatum is
normally full sealed and that of atlanticum may be also. Seales
have never been reported on any specimen of either bathyphilum
or gracile, and only rarely on elongatum. The latter, however,
has seales on at least some portions of the body, although they
are very deciduous.

The following key will serve to distinguish the species of
Gonostoma but is not necessarily phylogenetic.

Key to the species of the genus GONOSTOMA

la. Anus well ahead of anal origin, sometimes alomst midway between
ventral bases and first anal ray. First 4-5 IV in an ascending line,
5th or 6th directly below 4th or 5th and level with those following it.

2a. Adipose fin present. Anal origin opposite or only slightly in advance

of dorsal origin. Head and trunk noticeably longer than tail. VAV

9-10. OA 21, all but the first on the same level. No row of photo-
phores near dorsal profile of body.

ebelingi, new species

Pacifie

1b.

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 105

2b. Adipose fin absent. Anal origin noticeably in advanee of dorsal
origin. Head and trunk about same length as tail. VAV 3-5. OA

ial

-14, somewhat irregularly arranged. A row of widely separated

photophores near dorsal profile of body.

gracile Giinther 18781
Pacifie

Anus close to anal origin. First 5-6 IV forming an upwardly arched

are, or reduced in number to 2 or 3 (bathyphilum).

3a. Photophores minute, obscure. ORB, OP and SO usually obsolete in

4a.

4b,

adult. OA irregularly arranged, some rather high on sides. A few

posterior pterygoid teeth much enlarged. Anal rays 21-24. IV
11-13. Gill rakers on first arch 15-17 + 9-11 = 24-27.

bathyphilum (Vaillant) 1888

Atlantic

Photophores obvious, if not always conspicuous. ORB, OP and
SO relatively well developed, a glandular mass present below upper
OP. OA mostly on a single level low on sides, only the first one
or two elevated. Pterygoid teeth all small, or a few of the posterior
ones slightly enlarged. Anal rays 27-31. IV 14-16. Gill rakers on
first arch 10-12 + 5-9 = 15-21.

Body and head with numerous minute photophores (not always
evident). A mass of glandular tissue associated with each SO
and ORB. Gill rakers 10-12 + 7-9=18-21, Head and trunk
almost same length as tail. Interspace teeth in jaws not espe-
cially close-set.
elongatum Giinther 1878
Atlantic, Pacific, Indian

No minute photophores on body or head. No glandular masses
associated with ORB or SO (but a small silvery reflector present
behind ORB). Gill rakers 10-11 + 5-6 =15-17. Head and trunk
noticeably longer than tail. Interspace teeth in maxillary close-
set and about equal in size.

5a. Adipose fin present. Teeth absent on vomer, present on tongue

centrally. First two AC photophores above, second two below
remaining AC. Gill rakers 10 + 5=15.

denudatum Rafinesque 1810

eastern Atlantic, Mediterranean

1See Matsubara (1938, p. 41, fig. 3) ; Mead and Taylor (1953, p. 568). Syno-
nym: Gonostoma vitiazi Rass (1950, p. 1041, fig.).

106 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

5b. No adipose fin. Teeth present on vomer, absent on tongue

centrally. First one or two AC photophores slightly elevated,
others on one level. Gill rakers 11 + 5-6 = 16-17.

atlanticum Norman 1930

Atlantic, Pacifie

GoNOSTOMA ATLANTICUM Norman 1930

Gonostoma denudatum atlanticum Norman 1930, Discovery Rep., 2: 283;
Atlantie, 8°12’N., 18° 49’ W. to 0°56’S., 14° 08’30” W., type loeality not
designated.

Gonostoma denudatum Goode and Bean 1895, Ocean. Ichth., p. 98 (part,
off northern Florida); Jordan and Evermann, 1896, Bull. U. S. Nat.
Mus., 47: 579 (part); Murray and Hjort, 1912, Depths of Ocean, pp.
604, 605, 612, 744, fig. 456 (middle and eastern Atlantic); Jespersen
and Taning, 1926, Rep. Danish Oceanogr. Exp. 1908-1910, 2, (A 12):
4 (part), fig. 2 (part) (Atlantic, ca. 12°N., 35° W.); Grey, 1956,
Fieldiana, Zool., 36: 119 (part); Koefoed, 1958, Rep. Sci. Res. M.
Sars No. Atl. Deep-sea Exp. 1910, 4, (2), (6): 13 (part?).

The following specimens have been examined.

Atlantic off Florida: One, standard length 53 mm., U.S.N.M.
No. 44582, Albatross Station 2665, 29° 47’N., 80° 05’ 45” W.,
4 May 1886, 263 fathoms (480 meters), 45.2°F. at bottom (re-
ported by Goode and Bean, 1895).

Atlantie off South Carolina, one specimen, 63.5 mm., Combat
Station 296, 32° 40’N., 77° 40’ W., 21 April 1957, 220 fathoms
(403 meters).

Western Pacific: Two, standard length ca. 55 and 46 mm.,
SIO 56-133, Marshall Island area, 12° 27’N., 164° 30’ K. to
12° 38.8’ N., 165° 09’ E., 15-16 June 1956, Horizon, 10’ midwater
trawl, 0-1150 fathoms (0-2104 meters).

Counts and proportions of these specimens are shown in
Table 5.

Head and body considerably compressed. Skin entirely or
almost entirely lost, no remains of scales or scale pockets except
in smaller Pacific specimen, where there is evidence on the tail
that scales were once present. Head and trunk longer than tail.
No adipose fin. Anus below fourth VAV photophore, near anal
origin. Spines on inner edge of first gill arch short.

10ne minute tooth found on each lateral edge, posteriorly, in one specimen of
atlanticum.

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 107

Edge of premaxillary straight, toothed edge of maxillary
convex. Maxillary with a series of 9-10 longer teeth and 4-8
small, subequal, close-set teeth in each interspace except pos-
teriorly ; last four maxillary teeth shghtly enlarged and without
interspace teeth. Teeth of lower jaw similar but interspace teeth
less close-set and slightly larger than those of upper jaw; an-
teriorly two to four small outer teeth and a few minute ones.
Vomer with two very small teeth on each side, difficult to dis-
tinguish. Atlantic specimens with three or four teeth on each
palatine, increasing in size posteriorly and followed by a few
minute teeth; Pacific specimens with three teeth on each palatine,
the second one largest. Pterygoids with a patch of very small
teeth ; none enlarged ; none visible posteriorly in smaller Atlantic
specimen, a few present in larger Atlantic specimen and in
Pacific specimens. Tongue toothless except in larger Atlantic
specimen, which has 1-2 minute teeth on each lateral edge
posteriorly.

No gland tissue associated with ORB or SO, but a small,
narrow silvery reflector present behind ORB. SO minute. OP
3, glandular material present below upper one, which is level
with center of eye; lower two about level with end of maxillary,
posterior one larger and shghtly higher than anterior one. Pho-
tophores on body without evident glandular areas. Five pre-
pectoral IV photophores forming a very low are; first one or two
AC slightly elevated, others on one level; OA all or nearly all
lost in Atlantie specimens and in larger Pacific specimen, last
three (behind ventrals) remaining on larger Atlantic specimen,
in a slightly ascending line, ending over space between third and
fourth VAV; smaller Pacific fish with 13 OA, the first elevated,
the last three (behind ventrals) in an ascending line, ending
above space between third and fourth VAV. Two infracaudal
elands present in Pacific specimens, and larger Atlantie speci-
men with a small mass of tissue below last AC photophore,
probably somewhat more extensive in life. No supracaudals
(lost ?).

Color of smaller Atlantic specimen pale yellowish brown after
long preservation, peritoneum darker, inside of mouth and
opercles pale. Skin remaining on larger Atlantic specimen black-
ish; end of peduncle blackish; fin rays colorless, base of each

108 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

dorsal and anal ray black, peritoneum black; inside of mouth
with some black puncticulations; linings of gill covers and inner
half of branchiostegal membranes black. Skin remaining on
Pacific specimens blackish brown, cheeks silvery with black
puncticulations, peritoneum and linings of opercles black, inside
of mouth pale anteriorly and dusky posteriorly.

G. atlanticum was originally distinguished by Norman as a
subspecies of G. denudatum Rafinesque on the basis of differences
in the arrangement of the AC photophores and the number of
gill rakers (usually 10+ 5 in denudatum, 11+ 6 in atlanticum).
G. atlanticum is here given specific rank. It has been found to
differ further from denudatum in lacking an adipose fin and
tongue teeth, in being more slender bodied, in having a longer
dorsal base, and in being a smaller species. Dr. D. W. Tucker
has written (in litt., 1958) that none of the specimens of the
type lot, in the British Museum, has an adipose fin; and that a
specimen 48 mm. in standard length, from Discovery Station
296, is a female with ovaries in which the eggs are readily dis-
tinguished, although not fully developed. The Albatross speci-
men has not been examined internally but the Combat specimen
and both Pacific specimens reported here are also females with
immature eggs in the ovaries.

G. atlanticum is known from the central and eastern Atlantic,
from the latitude of the Azores (ca. 89° N.) south to 0° 56’S.,
14° 08’ 30” W.., off the African coast; and in the western Atlantic
off the southern United States coast. In the Pacific it is known
so far only from the two specimens reported here from the
Marshall Island area. Norman’s statement (1930, p. 283) that
Goode and Bean reported G. denudatum off the coast of Cali-
fornia was an error.

Goode and Bean (1895, p. 98) stated that G. denudatum was
trawled off New England by the Fish Hawk in 1881, but no
further data was given. In the same work, on page 102, these
authors listed, under the name Cyclothone elongata, a specimen
from Fish Hawk Station 1048, which was made on October 10,
1881. This specimen, U.S.N.M. No. 29069, cannot now be located
but there is in existence an unpublished drawing of it, labeled
‘“Gonostoma denudata, U.S.N.M. 29069, Fish Hawk Station
1048, H. L. Todd, del., Jan. 1882.’’ The drawing undoubtedly

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 109

represents G. elongatwm Giinther and the record of denudatum
from New England was perhaps an error based on an original
mis-identification of this specimen.

GONOSTOMA EBELINGI, new species

Holotype. Standard length 97 mm. Scripps Inst. of Oceanog-
raphy No. SIO-56-133, Marshall Island area, 12°27’N., 164°
30’ E. to 12° 38.8’ N., 165° 09’ E., 10’ mid-water trawl, 1150-0
fathoms (2104-0 meters), Horizon, 15-16 June 1956, time 1725-
0830.

Paratype. Standard length ca. 87 mm. Scripps Inst. of Ocea-
nography No. SIO-56-127, same locality, 13° 03’ N., 166° 04’ E.
to 138° 03’ N., 166° 32’ E., 10’ mid-water trawl, 400-0 fathoms
(732-0 meters), 23-24 June 1956, time 2020-0700.

Description. Counts and proportions in parentheses refer to
the paratype. Dorsal 13 (12). Anal 28 (28). Pectoral 9 (9).
Ventral 8. Branchiostegal rays 13 (13), six (eight) of them
with short spines at bases. Gill rakers 12+8 (11+ 8) on first
arch. Number of vertebrae unknown.

Measurements in per cent of standard length, 97 (ca. 87) mm.:
depth ca. 11.85; head ca. 22.7; snout 3.61; orbit 2.68; inter-
orbital width at center of eye ca. 3.61 (ca. 3.45); upper jaw
18.05; premaxillary 5.15; toothed edge of maxillary 12.9; lower
jaw ca. 18.6; tip of snout to dorsal origin 58.3 (ca. 59.8), to anal
origin 58.3 (ca. 57.5), to ventral bases ca. 39.7 (ca. 40.2) ; dis-
tance between first anal ray and base of middle caudal rays
40.2 (ca. 38.8), last anal ray and base of middle caudal rays
8.76 (ca. 8.05), last dorsal ray and base of middle caudal rays
27.8 (ca. 27.0-27.6), last dorsal ray and adipose fin 7.74 (ca. 8.6),
ventral base and anal origin ca. 18.05 (ca. 16.7) ; least depth of
caudal peduncle 4.13 (ca. 4.0) ; dorsal base 12.4 (ca. 12.65) ; anal
base 31.4 (ca. 33.3) ; ventral length 8.76.

Head and body considerably compressed. Neither specimen
in perfect condition, head and body just behind head damaged
in paratype, pectoral fins broken in both specimens, skin almost
entirely lost in paratype, and with it most of the photophores
of the lateral row. No scales or scale pockets visible. Head and
trunk noticeably longer than tail. Anal origin behind middle
of body length, below or slightly in advance of dorsal origin.

MUSEUM OF COMPARATIVE ZOOLOGY

BULLETIN :

110

“UU 26 YJSUI, pavpuLyyg

“CCT-9G-OTY (ON SYdvisSOUBIG JO WOTN4TYsUT Sddt19g ‘odAjo[OF_ ‘sotoads Mou “1HUI)9qQa DWO)sSOUOL) “Eg BINS TY

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 111

Adipose fin moderate, short-based, situated above end of anal
fin. Pectorals broken. Ventrals reaching fifth VAV photophore
in holotype, broken in paratype. Anus beneath sixth VAV photo-
phore, well ahead of anal origin.

Kdge of premaxillary straight, toothed edge of maxillary con-
vex. Eye moderate. Angle of preopercle acute. Spines on inner
edge of first gill arch well developed but not quite half as long
as longest gill raker.

Premaxillary teeth of holotype all broken off short; upper
jaw lacking in paratype except the premaxillary of one side,
this bearing seven teeth, one fang-like, the others smaller and of
varying lengths. Longer teeth of maxillary mostly broken;
interspace teeth numbering six to ten, small, about equal in size,
evenly spaced and rather close together. Lower jaw with about
nine long teeth; four to six smaller teeth in each interspace,
unequal in size and not close together ; anteriorly probably three
or four teeth in outer row, mostly broken. Vomer with one small
tooth on each side. Holotype with a row of about twelve small,
well separated teeth on each palatine, smaller posteriorly ; pala-
tine bones of paratype broken, remaining teeth smaller than
those of holotype. Pterygoid teeth all small, sparse, in a small
patch anteriorly ; posteriorly a few scattered teeth. Tongue with-
out teeth.

Photophores moderate. ORB with a relatively large gland
behind it, SO with a small one. OP 2, upper one small, level
with center of eye, a narrow streak of glandular material below
it; lower posterior one larger, level with end of maxillary ; lower
anterior one absent. BR 9. No separate glandular areas asso-
ciated with ventral row of photophores on body; a small mass of
eland tissue below each photophore of the lateral row. IV 15,
first four rising toward pectoral base, fifth directly below fourth
and on the same level as the remaining ten. VAV 10, continuous
with AC. AC 19, all on one level except the third, which is
slightly elevated; possibly an additional raised organ at base
of caudal fin. IC 44. OA 21, the first elevated, others straight,
reaching a vertical from sixth anal ray; almost all missing in
paratype. Two infracaudal glands, no supracaudal (possibly
lost).

a, BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Color brownish black, abdomen and branchiostegal membrane
black. Traces of dark metallic iridescence on both head and
body. Inside of mouth and gill covers dark.

Remarks. G. ebelingi is distinet from all other species of the
venus Gonostoma in having a much greater distance between
ventral and anal fins, a correspondingly larger number of VAV
photophores, and more OA photophores. Within the genus its
relationships are obscure. Only in gracile is the anus situated
so far in advance of the anal origin and these two species are also
similar in coloration, in the arrangement of the prepectoral IV
photophores, and in having the OA extend beyond the anal
origin. On the other hand, ebelingi is like denudatum and atlan-
ficum in the relative proportions of trunk and tail, and in having
the small interspace teeth of the maxillary close-set and more or
less equal in size.

The species has been named for Dr. Alfred W. Ebeling, of
Seripps Institution of Oceanography, in appreciation of his
interest and assistance during the course of this study.

DanapuHos Bruun 1931

Danaphos Bruun 1931, Vidensk. Medd. Dansk Naturh. Foren., 92: 286.
Generic characters. Kye tubular, directed obliquely upward.
Snout shehtly shorter than orbit. Interorbital width at center
of eye much less than diameter of orbit. Mouth moderate, nearly
vertical; maxillary abruptly horizontal, its toothed edge convex,
reaching or nearly reaching a vertical from posterior margin of
eye. Premaxillary more than half as long as toothed edge of
maxillary. Angle of preopercle slightly obtuse or nearly vertical.
Teeth of upper jaw uniserial, minute on premaxillary ; slightly
longer on maxillary, very slender, much shorter teeth between
longer ones. Teeth of lower jaw minute, seen under high power
to be in two or more rows in anterior half of jaw, uniserial and
shghtly larger posteriorly. No teeth on vomer, palatines, ptery-
goids or tongue. Gill rakers 11-13 + 2 = 13-15 on first arch.
Spines on inner edge of first gill arch rudimentary. Pseudo-
branchiae present but very small and fragile. Seales present but
very deciduous. Anus a httle nearer anal origin than ventral
bases, beneath third or fourth VAV photophore. Head and trunk
slightly shorter than tail. Origin of dorsal fin well ahead of

—

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS NI

middle of body length. Origin of anal fin behind end of dorsal
fin. Ventral bases beneath dorsal fin. Adipose fin present or
absent; if present, small and poorly developed. ORB 1, in front
of center of eye. OP 2 or 3, upper one not always present;
lower two behind end of maxillary, about equal in size or the
anterior one slightly larger. SO absent. BR (6). No additional
photophores on head. Body with two rows of serial photophores ;
photophores present on isthmus. IV (3) + (4) +1+ (2) +8=
18. VAV usually (5), (4) in one specimen. AC (8) + 14-18 +
(4)1+ 1 = 22-26, straight. IC 45-49. OA (2) + 4-5 = 6-7. No
additional photophores and no luminous tissue on body as far
as known. Fin rays: dorsal 6, anal 24-25, pectoral (12?) 13-14,
ventral 6. Branchiostegal rays 9-10, bases prominent on inner
edge and occasionally with minute spines. Vertebrae 38, counted
on X-ray photograph of one specimen.

Remarks. As suggested by Bruun (1931, p. 287) Maurolicus
oculatus Garman (1899, p. 241, pl. 53, fig. 3) belongs in the
genus Danaphos. New material of this species from the eastern
and northern Pacific has shown that Danaphos asteroscopus
Bruun 1931, known from the tropical Indian Ocean and from
the central and western parts of the Pacific, is probably a syno-
nym of D. oculatus. The only apparent distinctions between the
two species are the presence, in asteroscopus, of a small and
poorly developed adipose fin, which could not be found on any
of the specimens examined; and the fact that asteroscopus was
said to have an upper OP photophore, which is not present on
any of the specimens seen. Many of the latter are mature, or
nearly so, both sexes being represented but females predominat-
ing. The generic description was based on the following speci-
mens of D. oculatus from the collection of Scripps Institution of
Oceanography.

Eastern Pacific: Two, standard length 33 and 36 mm., SIO
54-98, 26° 23.5’ N., 123° 14’ W., to 26° 53.5’ N., 128° 22’ W., 25-26
June 1954, 10’ midwater trawl, 1578 fathoms (2886 meters) ; five,
standard length 28-38 mm., SIO 57-43, 28° 52’ N., 118° 12.5’ W.
to 28° 59.5’ N., 118° 09’ W.,_ 10 February 1957, 380-0 fathoms
(695-0 meters) ; eleven, standard length 32-39 mm., SIO 57-206,
28° 34.5’ N., 126° 52’ W. to 28° 47.9’ N., 126° 24’ W., 20-21 June

1One specimen with (3) on left side, (4) on right side.

114 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

1955, 10’ midwater trawl, 0-675 meters; sixteen, all fragmentary,
SIO 57-88, 28°46’N., 126°33’ W. to 28°17’ N., 126° 45’ W;;
22 May 1955, 10’ midwater trawl, 0-348 fathoms (0-636 meters) ;
two, standard leneth ca. 29 and 26.5 mm., H-51-188, 33° 09’ N.,
118° 01’ W. to 32° 57’N., 117° 48’ 30” W., 22 May 1951, 200
fathoms over 490-500 fathoms (360 meters over 896-1006 meters).

North Pacific: Nineteen, standard length 21-39 mm., H 51-354,
40° 23’ N., 189° 23’ W., 5 August 1951, 10’ midwater trawl, 400-
650 meters; thirteen, standard length 21-43 mm., H 51-358,
40° 35’ N., 147° 55’ W., 10 August 1951, 10’ midwater trawl,
350-600 meters; fifteen, standard length 27-ca. 45 mm., H 51-
359, 41° 42’ N., 150° W., 11 August 1951, 600-800 meters.

NeopHos Myers

Neophos Myers, 1932, Copeia, p. 61.

Generic characters. Kye normal, moderate or large. Snout
shorter than orbit. Interorbital width at center of eye less than
diameter of orbit. Mouth large, oblique; premaxillary nearly
vertical, its toothed edge straight; toothed edge of maxillary
curving abruptly downward from juncture with premaxillary,
becoming almost straight posteriorly, nearly reaching preopercle.
Premaxillary about half as long as toothed portion of maxillary.
Angle of preopercle vertical. Teeth all minute, irregular on
premaxillary, uniserial and with some shghtly longer teeth on
maxillary; teeth of lower jaw uniserial but with a short outer
row of minute teeth anteriorly. Vomer with a few teeth on each
side. Palatines each with a few teeth. No teeth on pterygoids
or tongue. Gill rakers about 13-14 + 5 = 18-19 on first arch. No
spines on inner edge of first gill arch, a few minute ones on
second arch. Pseudobranchiae present. No evidence of scales.
Anus close to anal origin, beneath last VAV photophore. Head
and trunk shorter than tail. Origin of dorsal fin about in middle
of body length. Anal origin well ahead of dorsal origin. Ventral
bases well ahead of dorsal origin. No adipose fin. ORB 1, in
front of eye. OP 3, upper one elongate, about level with middle
of eye; lower two on the same level behind end of maxillary, the
posterior one larger. SO present. BR (6). No additional pho-
tophores on head. Body with one row of serial photophores
and a single organ representing the lateral row; photophores

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 115

present on isthmus. IV 1 + (2) + (3), on isthmus, + 11 = 17.
VAV 1+ (3) +1=5. AC 13, all single, straight, two of them
behind anal fin. IC 35. OA 1, above pectoral base. No addi-
tional photophores and no luminous tissue on body as far as
known. Fin rays: dorsal 8, anal 38, pectoral 13, ventral 7.
Branchiostegal rays 7 or 8, no spines at bases. Number of
vertebrae unknown.

Remarks. The type and only known specimen of Neophos
newvilis Myers (1932, p. 61) has been examined. The genus is
closely related to Thorophos Bruun and is possibly synonymous
with it, but comparative material is necessary to establish their

identity.
LITERATURE CITED

ABE, T.
1958. Diplophos orientalis Matsubara. In I. Tomiyama and T. Abe,
Figures and Descriptions of the fishes of Japan, 59: 1221-1247,
pls. 236-238. Tokyo.

BERTIN, L.
1939. Révision des Stomiatiformes (Téleostéens Isospondyles) du
Muséum. Bull. Mus. Hist. Nat. Paris, 11: 378-382.

BRAUER, A.
1906. Die Tiefseefische. 1. Systematischer Teil. Wiss. Ergebn. Deutschen
Tiefsee Exp. Valdivia, 15 (1): 1-432, 18 pls., 176 text figs. Jena.
1908. Die Tiefseefische. 2. Anatomischer Teil. Wiss. Ergebn. Deutschen
Tiefsee Exp. Valdivia, 15 (2): 1-266, 42 pls., 187 text figs. Jena.

Bruun, A. F.
1931. On some new fishes of the family Gonostomatidae. Vidensk.
Medd. Dansk Naturh. Foren., 92: 285-291, pl. 8.
1950. General Report. Atlantide Rep., 1: 7-28, 17 figs.

GARMAN, S.
1899. Reports on an exploration off the west coast of Mexico, Central
and South America, and off the Galapagos Islands... 26.
The fishes. Mem. Mus. Comp. Zool., 24: 1-431, 85 pls.

GILBERT, C. H. and F. CRAMER
1896. Report on the fishes dredged in deep water near the Hawaiian
Islands, with descriptions and figures of twenty-three new
species. Proc. U. S. Nat. Mus., 19: 403-435, pls. 36-48.

116 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

GILCHRIST, J. D. F. and C. von BONDE
1924. Deep-sea fishes procured by the S.S. ‘‘Pickle.’’ Rep. Fish. Mar.
Biol. Surv. So. Africa, 3 (7): 1-24, 6 pls.

JSOODE, G. B. and T. H. BEAN
1895. Oceanic Ichthyology. Smithsonian Contr. Knowl., 981, 982,
xxxv + 553 pp., 417 figs.

GREY, M.
1955. Notes on a collection of Bermuda deep sea fishes. Fieldiana,
Zool., 37: 265-302, figs. 45-56.
1959. Three new genera and one new species of the family Gonosto-
matidae. Bull. Mus. Comp. Zool., 121: 167-184, 3 figs.

GUNTHER, A.
1889. Report on the pelagic fishes. Rep. Sei Res. Voy. Challenger,
Zool., $1: 1-47, 6 pls.

Herre, A. W. and HE. S. HERALD
1950. Noteworthy additions to the Philippine fish fauna, with descrip-
tions of a new genus and species. Philip. Jour. Sci., 79: 309-3840,
figs. 1-12.

JESPERSEN, P.
1933. Valenciennellus tripunctulatus. Faune Ichth. Atl. Nord, 14, fig.

JESPERSEN, P. and A. V. TANING
1919. Some Mediterranean and Atlantic Sternoptychidae. Vidensk.
Medd. Dansk Naturh. Foren., 70: 215-226, pl. 17.

JOHNSON, J. Y.
1890. On some new species of fishes from Madeira. Proc. Zool. Soc.
London, 1890: 452-459.

KKOEFOED, E.

1958. Isospondyli. 2. Heterophotodermi, 1, from the ‘‘ Michael Sars’’
North Atlantie Deep-sea Expedition 1910, with addenda et
corrigenda to Isospondyli 1. Rep. Sci. Res. M. Sars No. Atl.
Deep-sea Exp. 1910, 4 (2), (6): 1-18, 1 pl., 2 text figs.

Koumans, F. P.
1953. Biological results of the Snellius Expedition. XVI. The Pisces
and Leptocardii. Temminckia, 9: 177-275, 2 figs.

LuoypD, R. E.
1909. <A description of the deep-sea fish caught by R.I.M.S. Ship
‘““Tnvestigator.’? Mem. Indian Mus., 2: 139-180, pls. 44-50.

rc

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 117

Loneaury, W. H. and S. F. HILDEBRAND
1941. Systematic catalogue of the fishes of Tortugas, Florida, Publ.
Carnegie Inst. Washington, No. 535, xiii + 331 pp., 34 pls.

LUTKEN, Cur. FR.
1892. Bidrag til Kundskab om det aabne Havs Laxesild eller Seope-
liner, med et Tillaeg om en anden pelagisk Fiskeslaegt. Vidensk.
Selsk. Skr., (6) 7 (6): 223-297, pls. 1-3, text figs.

McArpir, A. F.
1901. Natural history notes from the Royal Indian Marine Survey
Ship ‘‘Investigator’’ ... Ser. 3, No. 5. An account of the
trawling operations during the surveying season of 1900-1901.
Ann. Mag. Nat. Hist., (7) 8: 517-526.

McCuuuocnu, A. R.

1926. Report on some fishes obtained by the F.I.S. ‘‘Endeavour’’ on
the coasts of Queensland, New South Wales, Victoria, Tasmania,
south and southwestern Australia. Part 5. Biol. Res. Endeavour,
5: 157-216, pls. 43-56, text figs. 1-4.

MATSUBARA, K.
1938. On some stomiatoid fishes from Kumano-Nada. Jour. Imp. Fish.
Inst. Tokyo, 33: 37-52, 8 figs.

1940. Studies on the deep-sea fishes of Japan. 13. On Prof. Nakazawa’s
collection of fishes referable to Isospondyli, Iniomi, and Allo-
triognathi. Suisan Kenkiu-Shi, 35: 314-19.

1941. Studies on the deep-sea fishes of Japan. 13. On Prof. Nakazawa’s
collection of fishes referable to Isospondyli, Iniomi, and Allo-
triognathi. Suisan Kenkiu-Shi, 36: 1-10, 1 fig.

1943. Ichthyological annotations from the depth of the Sea of Japan.
Jour. Sigen. Kenk., 1: 37-81, pl. 1, text figs. 1-24.

MAUL, G. E.
1948. Monografia dos peixes do Museu Municipal do Funchal. Ordem
Tsospondyli. Bol. Mus. Mun. Funchal, 3 (5): 5-41, figs. 1-16.
1952. Additions to previously revised families. Bol. Mus. Mun.
Funchal, 6 (16): 51-62, figs. 13-17.

MeaD, G. W. and F. H. C. TAyLor
1953. <A collection of oceanic fishes from off northeastern Japan. Jour.
Fish. Res. Bd. Canada, 10: 560-582, 8 figs.

118 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Myers, G. S.
1932. A new gonostomatid fish, Neophos nexilis, from the Philippines.

Copeia, 1932: 61.

NorRMAN, J. R.
1930. Oceanic fishes and flatfishes collected in 1925-1927. Discovery
Rep., 2: 261-370, pl. 2, 47 text figs.
1939. Fishes. Sci. Rep. John Murray Exp. 1933-34, 7 (1): 1-116, 41
figs.

Parr, A. E.
1931. Deepsea fishes from off the western coast of North and Central
America. Bull. Bingham Oceanogr. Coll., 2 (4): 1-53, figs. 1-18.
1937. Concluding report on fishes. With species index for articles 1-7
(Fishes of the Third Oceanographic Expedition of the Pawnee).
Bull. Bingham Oceanogr. Coll., 3 (7): 1-79, 22 figs.

PouL, M.
1953. Poissons, Téleostéens Malacopterygiens. Rés. Sci. Exp. Océanogr.
Belge (1948-1949), 4 (2), (3): 1-258, 8 pls., 104 text figs.

Rass, T. 8.
1950. A new Pacific Ocean deep-water fish, Gonostoma vitiazi, nova
species (Pisces, Gonostomatidae). Doklady Akad. Nauk SSSR,
74: 1041-1043, fig.

RecunitzeEr, A. B. and J. BOHLKE
1958. Ichthyococcus irregularis, a new gonostomatine fish from the
eastern Pacific. Copeia, 1958: 10-15, pls. 1, 2, text figs. 1-3.

SHALE, A.
1935. The Templeton Crocker Expedition to western Polynesian and
Melanesian islands, 1933. No. 27. Fishes. Proe. California Acad.
Sei., 21: 337-78, 3 pls.

SmitH, J. L. B.
1949. The sea fishes of southern Africa. Capetown, 550 pp., 1100 figs.

SPRINGER, S. and H. R. BULLIS, JR.
1956. Collections by the Oregon in the Gulf of Mexico. Spee. Sei. Rep
U.S. Dept. Int., Fish, 196: 1-134.

WELSH, W. W.
1923. Seven new species of fish of the order Malacopterygii. Proc. U.S.
Nat. Mus., 62: 1-11, figs. 1-10.

gis)

PRELIMINARY REVIEW OF GONOSTOMATIDS

GREY

G
= L0¢
= OF
= 66S
G9E G'9e
G'6F 9'8F
O LF 0'9F
ans EG
as FOT
CEs "B
= 6'¢
LPI SFI
GZ'8 Gas
Cg FL
= 86
689 “Do TL
OLT BILE ILILIE
Bar Miz e-% + OF
LI OT
FF L¥
EOL PI-S1
~- e+s
8 L
OT 6-8
(9 “Do 89
6 OL
“MOS OM oz
“N ,06 “N G6 98
61z 4 9T900T
‘26ST ‘TexIn ‘ON WNSO

‘Z ayR[d JO Z vansy uo pojyuNoDe

96°T ak
00S 0'SF
£0°§ “D9 GLP
cog —
P'9E '8é
00S Ors
GCF CLF
Le vs
Tete =
¥'3 ve
Ll? US
Lor O'LT
38 SOL
YJSuolL plLpuLys JO JUo) Log
99 6S
aa 16
&L OL
eI GOT
G1 LF 6+ eh
LT 9T
LP tf
clo cOT-ST
CHL =
ra 8
x 8
89 19
IL 1H!
M99 "M ST
“N08 “D9 “NT 0G “D0
_ GZS00T _ 6gd
ON WNSN? 90G6L ‘teneig

"sy (d 9992

“M .S6
pue oot “NOL

wre:
c68E UHINT

‘Po}PBOIpPUL VLOYM OINZBIO}] OY} ULOTF Woy} “DIWaD}
soydojdig jo Suomtdeds orueyyy Fo suorytodord pue syunog

T S48

"WISUaT 1830} ATqQISSOdt

apunped yepnes
jo yAdop 4seory
aseq [vUuy
asvq [es10g
eseq [epneo
04 UISIIO [eUYy
soseq [B1IZUBA
UISLLO [BUR
UISTIO [Vs10p
204 qnous Jo dry,
£8 JO 104W9d 4B
Y}PTA [2}1q1010}0T
met todd
WQIO
ynoug
pee
yydoqg

Yjsuel prepuvys
our] *Fery

Yor’ Isai
WO SLOYeI [LD
sABI [V1}U9 A
SAI [810999
sfkel [PUY
SABI [VSIOG

Aqpeoory

MUSEUM OF COMPARATIVE ZOOLOGY

BULLETIN

120

‘SOT MOU
asoy) Surpnyouy ‘6g “v9 A] [BULSs TO S[quqoad ‘ornsy wory supspnfe —“T18} Jo FTV 197 VT WO ysol A[Quqoigdgs “80, ATQISSOd OM} 10 vUQ,
‘<yuO UamMTDeds 8VUOg ‘soroydojyoyd aynulm ¢-T OSlVe ‘SuIYOR] 1183 JO dLLy “YSU, TRIOT, “Uy [eUR Jo pus OL ‘aInsy WoT

iim SLG Pre OS == re a == apunpod pepurs
JO YAdop ysvory
ep OFF ()9F-0° SF — — —— — asuq [euy
69 cgi 0'S-9'F 7 a rae = ISB [VS1O(]
OLY PSV 0° 09-6 6F at ae G Ly vv = aseqg Pepues
O} UTSLLO [BUY
OLE = PS8E-G9E OLE ; GLEe vo naa sosVq [BIpUOA
a = 0° 1S-8°0¢ 0°0S GLP po = ULSLIO TRUL
6 OF CGP 9 LE-S'9P O'8F ; L&P vo = ULSTLO [USAOp
203 ynous Jo dry,
Vol Lit GILVTIt Cae Pa GEL ve = mel soddy
8's Goe Lig-0€ Oe GE “Do = ahd JO 1O}W9) 4V
APTA [e4rq Lo1oguy
ie 8é 9°€-0'S Ge = Gg “Dov — 1GIO
La 6¢€ TPs Sy = 0G “DO = ynoug
o LT sol VSt-0'8t GOL = GE9OT Be = peop
V6l Sol COLG6 Gol = GL8 Dp) a yydoqd.
Yysuo] PxrwvpuBys JO }U9d 9g
6 (668) FL9 98 y98 {Oks} 66-68 768 LLF OU] [BLOpRTY
g1§9 18 PS-09 Gu-LL 69-99 L9 = vo
+96 86 GOL-S6 =1-°6 GOT-L6 FOL De = OT
Le+6s £+68 6+oFr OF c+lF rae 6 + FF == OV
FI Gl CL-FI ¢l aa CT — AVA
OV OF oV- LF Lv, mer: PF DY = Al
Wt ae eL LOL Tal say él = bY ke |
e+6 e+6 e+6 i +6 a = = yore 4S.1y
ILO SIO VA [[L1)
8 8 8-L 8 = = L BABI [BI}UWO A
Ot OL 6 6 6 6 6 skBI [VIO
69 “09 69 09-SS 8S 19-6¢ éT¢ eS sfBi [VUY
ol Gl OL-6 OL IL-OL Te GL SABA [BSLO
LLT S'6LI 001-68 p+68 clP FE OF ‘De Le yASusy prvpueyg
Cd. “OUd ‘SPL
weds ¢ uvde rp WLLOSUOL ULOISVOT vag n[Nsg [TVYsreyy dj LOBq-Pl XYTpRoo'y
8S6T adh} ‘OFGL “I50URIIO JO add} ‘TSG6I OS6L ‘PICI9eTT LGT-9& ad4} ‘688T
‘Ody ‘RIRVQGUSTBIL “ysuy sddti0g ‘LIV puv oto} OIS ‘eqjunsy
*StpPDPUILLO $1)D] UdtLO “ds soydo]aq snunvord <DWUaID],, snayfiond

‘pazPBotpUL aLOYM OLNYZVLOH] OY} WOIF
oye, fsoydop dig, Jo Suouttoods oytor JO smortodoad pue syuno/

6 PIAL

GREY: PRELIMINARY REVIEW OF GONOSTOMATIDS 121

Table 3.

Counts and proportions of Yarrella blackfordi and

Standard length, mm.
Dorsal rays

Anal rays

Pectoral rays

Ventral rays
Branchiostegal rays
Gill rakers on first arch
Vertebrae (from X-ray

Yarrella argenteola.

blackfordi

170-261
14-16
29-31
8-9
ii
14-16
12-134-6-7=18-20

photographs ) 54
BR 12-13
IV 9+3-4411-12=23-25
VAV 12
AC 25-27
IC 61-63

Per cent of standard length

Depth 11.4-14.3
Head 20.5-23.8
Snout 4.4-5.4
Orbit 2.47-3.25
Interorbital width at

center of eye 9.0-5.75
Upper jaw 15.2-17.7
Premaxillary 8.4-10.0
Toothed portion of

maxillary 6.6-7.65
Tip of snout to:

dorsal origin 49.5-51.7

anal origin 56.1-58.5

ventral base 37.2-40.1
Ventral base to anal origin 16.4-18.0
Anal origin to base of

middle caudal rays 41.9-45.4

End of anal to base of
middle caudal rays

End of dorsal to base of

middle caudal rays
Dorsal base
Anal base
Least depth of caudal
peduncle

15.5-18.6 (20.2)

0-35.
11.3-12.
(21.4) 24.3-26.

Ses
Loo

6.4-7.3

argenteola

125-133

13

943411-19=293-24
9-10
20-21
52-53

—_
po oT
SSeWe

54.0-54.2
58.5-59.0
40.1-40.8
17.6-18.5

40.5-41.4
13.9-14.0
31.6
14.0-15.0
26.8-27.0

6.4

i222 BULLETIN :

MUSEUM OF COMPARATIVE ZOOLOGY

Table 4.

Counts and proportions of Triplophos hemingi, taken from

Locality

Standard length
Dorsal rays
Anal rays
Pectoral rays
Ventral rays

Branchiostegal rays

Gill rakers on
first arch

BR

IV

VAV

AC

IC

Depth

Head

Snout

Orbit

Interorbital width
at center of eye

Tip of snout to:
dorsal origin
anal origin
ventral bases

Least depth of
caudal peduncle

Dorsal base

Anal base

McArdle, 1901

Per cent of standard length
— 12.8
— 14.6
— 1.39
— 2.43

the literature where indicated.

Lloyd, 1909 Brauer, Poll,
Norman, 1930 1906 1953
Bay of Southof West Africa,
Bengal Ceylon SicilalenSe
205 and ? 144 60-161
10 10 10-11
57-61 57 54-56
10-11 10 —
6 (9?) 6 —
14 17 —
ote 13 =
29-30 30 —
5 5 5-7
35-36 41 —
69-71 76 —

Western
Atlantic
specimens

Caribbean
and Surinam

106-176
10-11
57-63

9-10
6-7
11-15

14-16+-9
8-10
24-25
7
37-39
68-70

ca. 12.0-15.0

ea. 15.0-17.5

ca. 1.7-2.8
2.4-3.4

2.3-2.9

29.6-31.5
35.2-37.9
24.7-28.4

1.3-1.9
ca. 5.7-6.8
60.0-ca. 61.5

GREY :

Table 5.

PRELIMINARY REVIEW. OF GONOSTOMATIDS

123

Counts and measurements of Gonostoma atlanticum.

Standard length

Dorsal rays

Anal rays

Pectoral rays

Ventral rays

Branchiostegal rays

Gill rakers on
first arch

Vertebrae

BR

IV

VAV

AC

IC

OA

Depth
Head
Snout
Orbit
Interorbital width
at center of eye
Upper jaw
Premaxillary
Toothed portion
of maxillary
Tip of snout to:
dorsal origin
anal origin
ventral bases
Distance between:
anal origin and
caudal base
last anal ray and
caudal base
last dorsal ray
and caudal base
inner insertion of
ventral base
and anal origin
Least depth of
caudal peduncle
Dorsal base
Anal base

Combat
specimen
South
Carolina

39

U.S.N.M.

44582

Florida

53
16
ca, 29

i

11+6
38

9
16
19
40

Per cent of standard length

15.75

ca. 23.6
3.94
ca. 3.94

3.94
18.9

3.94
15.0
58.3
58.3
ca. 48.8
42.5

9.45-10.0

23.6

ca. 9.45

6.3
18.1-18.9
33.1

ca. 16.0*

ca. 24.5
Boll U

etal

11.3

24.5

9.43

5.65
ca. 17.9
oil

SIO 56-133
Marshall Ids.
55 46
17 17
2 28
11 10?
7 wh
1] ii
11-5 11+6
9 9
16 16
5 5
17+1 19
— 40
— 113}
ca. 16.3 ca. 16.3
Gh DIS PAT)
3.64-4.54 ca. 4.34
4.54 4.34-5.43
4.54 3.26-4.34
20.0 20.6
4.54 4.34
ne: 16.3
59.0 ca. 56.5
56.5 ca. 55.4
ca. 47.2 ca. 46.7
41.0 42.4
ca. 10.0 9.77
23.6 93.9
ca. 6.36 ca. 7.6
6.36 5.43-6.5
16.3-17.3 18.45
ca. 29.8 ca. 32.6

1Lacking on end of tail.
2Unpublished drawing of this specimen, prepared in June, 1886, shows 13.
3Depth somewhat greater in drawing.

MUSEUM OF COMPARATIVE ZOOLOGY

BULLETIN

124

‘ST-9L=G-E+E1 S}UeUTIpNaA PUB S18eHBI JO JaQUINU [B}OL7

¥2-G0=L1-S+L1-ST
0$-€2=8-F +2¢-L1
1g-ST=E+8L-St
pl=c ie
CheCl—Garclarr
é
61-8T=S+FI1-EI °09

(0€) 23-FI=01-#+8T-6

L@-P6=11-6-+-L1-ST
Og Si—6-2-1-eL- 00
2G-0G=8-L+F1-E1
02-61=8+6T-IT
L1-91=9-S-F TT
GI=S+0I

¢T=S-+01
8T-91=9-G+2T-1IT
eae levatnO Gaal
pO[BUV ye F-Z
91-SI=S-F+1T
€¢-GT=1L-§+&s-1L
LI-S1=S-F+21-T1
61-SI=8-S+31-6
GZ-€6=64+ 9I-FT
o-SI=L-9+ 9T-ST
FI-3I=S-E +6-8

6L-01T=e+6-2

O[-8
01-6
8
O1-6
OL-6
é
8-L
FL-OL
EL-GL
€L-GL
IL-OL
all
II
FILS

él
Si Gate
6 Il

LT
16-06
6L-OL
6T-TL
FL-6L
FI-IL
91-§T
et-IL

TAI

OF snudiiibspy
Gest SNIYOMND JY
OF ppouoy
b€E-SE SN)]PUUWAIDWI]D A
8é soydouvg
é soydovoy
é soydoa Nn
E-66 awoy ,opaiy
LE wnpydhyyog
6€ wnypbuoja
1G 7 ajwvub
é wWuyaga
8g UWNILZUD)PD
6E-8E wnjpopruap
:DWO SOUND)
FE DIYPALD LD
LE vipwvdvuo g
L¥-88 snav000fiyzy]
b Dispoo M
TS shyqyorpoud
GF 8t DIMMING WU A
Ome? shypyoulod
CF awupawhyod
09 “De soydo day
vG-SP D)JILLD X
69 (snonpun yy snuds

-qns) soydojdiq
i+¢g°po (soydojdig snues
-qus) soydojdi]

ee

SABI

Yoie Jsig UO S19yVI [IID

SABI

‘B19UIS PI}EULOJSOUOS JO SI9}JOVIVYO OI}SLIO[L

[vZa}so;qouvig [BiIjZUVA

9 198

61-1 Gas FL-OL
03-LT 13-61 31-6
CT-1 GZ-36 6-8
eL-Zr G3-3S aie,
PI-S1 GG-FG 9

FL ‘ve 1¢ It
gL 8 8
€T-6 13-9T GT-31
IL-O1 93-36 PI-3L
S101 GE-6S FL-31
01-6 63-(b) 9% FL-OL
6 83 SL-3L
IT-O1 08-82 ST-9L
ray iy t 18-63 LI-9T
FI-S1 93-13 9T-SI
OT-FI 1€-63 0Z-LT
8-2 LI-81 CT-OL
OI-6 ial BL
6 93-&3 €L-21
01-6 (401) LI-3E = (80 T) 9T-ST
8 08-33 31-01
II-6 €E-F3 eL-UL
II-6 £9-FG SL-OL
01-8 [€-83 OL-FI
IL-01 [F-9¢ €L-61
01-8 89-E¢ 31-6
sfv1 SABI SABI
[810} 00g [vuy [esi0q

aB1qe}1IA

snuey

125

PRELIMINARY REVIEW OF GONOSTOMATIDS

GREY

‘snonuyjaoo Dy jo dnois Jsig puv AWVAgs
‘dnoais 10d $-Z ‘SuBsAO [[BUIS AIOA JO SdNoIS 9-E UIz

L 40 1=$+(F) L9-€S elS-6€=(8T-31)+(G)+(83-8T) 91T=(0T)+(9) (9-¢) snudiwhDay
6=/an a) GS-9F Gcc= (G2) (Sieh) ate iE (9) 61-8T=(€T-3T) +(9) (9) SNIVYOAND
1-9=S-$+ (2) L9-6S €h-98=(F3-61) + (12-91) (8-L) 9T=(0T) +(9) (1-9) ppowogy
e=e+(Z) 40 (3) FF-LE 291-6 (G-F) be-e0=(LT-91) + (4) + (8) (9) SN]JIUUIIDUIYD A
L9=GP+(6) 6FSH 98-63=1+ (4) F8T-FI + (8) () gI=st+(3)+I+(%) +(e) (9) soydvund
P=G-(G) LE-9S GT-PI=1+ (zs) 4OT+l+() +6 (¢) LI=1I+ (9) (9) soydo.oyy,
I CE rom Gao e 9 Mitta aaa (9) soydoa N
01-9 $866 9T-3I G-F PI-3L 11-8 auoy,ojahi)

¥1 LE-G& 13-03 C-F SI-11 6 uenqrydtiyyog

CI-SI SPv-lF €3-1G C-F cl 6 wnjobuoja

€I-I1 6E-SE 61-L1 G-§ GT 6 apo Dub

13 iad 61 OT Gl 6 Wuryaqa

rene OF-6€ 61 G 9T-ST 6 ULND17UN] 7D

GT-€T I¥-LE 03-LT G 9T-ST 6 wnyopnruap
/DUWOJSOUOL)
a 9¢-S¢ LI F cT-€T 61-6 DIY LILO LD
=< Iv-LE O036-8T G 9T-FI et-It pijivdnuog
6&6 cc-9F FL-GL F1-6 836-GG 6L-It snav000hy yo]
1¢-62 6F-8F Fail ra a G3 ial DISPOO Af
8-€E 8G-L¢ 8L-91 LT-ST GZ-F3 8T-LL shypyoujoud
CO-1G 6F-GP GL-GL Il-L $313 8 DIM LINDWUL A
12-61 0S-L¥ 13-81 6-2 E316 8 shiypyoyod
8T-9T ¥S-0S Go-1G 8-2 13-61 01-6 aujpawhyog
9G-0G 91-89 IF-Ge L-G O€-FG 61-8 soydo yay
+0¢ ‘po $9-GGS 80-02 Z1-6 GZ-&S Gilieta! D)JILMD_X

8h-SF GL-OL 0-82 FI-GT GE-0€ 6S (shonpupyy snus
-qus) soydojduy

L8-99 €11-L6 6F-&F LT-S0 6F-OF £0F6LOL (soydojdig snuos
-qns) soydo,digy

vo OI OV AVA Al ud

“RIQUIH PTZVULOJSOUOS JO SLoPBAIVYD IYSWOTAT

so10ydoj,oyg

(709) 9 AIL

snuos)

‘
‘ 1 .
:
No =
s
® 4 Apes
, x
4 via oy
eee has ’
pare seas +
eter — F

. 4 Sy]
: Ect
‘ ‘
H

Bulletin of the Museum of Comparative Zoology
AG Te eet AM VrAWsR, 21D ns @) Ola slr Ey Gedy
Vor 122, No. 3

TRICHECODON HUXLEYI (MAMMALIA: ODOBENIDAE)
IN THE PLEISTOCENE OF SOUTHEASTERN
UNITED STATES

By CuaytTon EK. Ray

Wirth Two PLatEes

CAMBRIDGE, MASS., U.S.A.
PRIN TED OHOR DHE) MUS HU M

Marcu, 1960

PUBLICATIONS ISSUED BY OR IN CONNECTION
WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

BULLETIN (octavo) 1863 — The current volume is Vol. 122.

BreEviorA (octavo) 1952 — No. 124 is current.

Memorrs (quarto) 1864-1938 — Publication was terminated with
Woolton:

JOHNSONIA (quarto) 1941 — A publication of the Department of
Mollusks. Vol. 3, no. 39 is current.

OCCASIONAL PAPERS OF THE DEPARTMENT OF MoLuusks (octavo)
1945 — Vol. 2, no. 23 is current.

PROCEEDINGS OF THE NEW ENGLAND ZoouocicaL CuuB (octavo)
1899-1948 — Published in connection with the Museum. Publication

terminated with Vol. 24.

The continuing publications are issued at irregular intervals in num-
bers which may be purchased separately. Prices and lists may be
obtained on application to the Director of the Museum of Comparative
Zoology, Cambridge 38, Massachusetts.

Of the Peters ‘‘Check List of Birds of the World,’’ volumes 1-3 are
out of print; volumes 4 and 6 may be obtained from the Harvard Uni-
versity Press; volumes 5 and 7 are sold by the Museum, and future
volumes will be published under Museum auspices.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE
Vor 122) Now 3

TRICHECODON HUXLEYI (MAMMALIA: ODOBENIDAE)
IN THE PLEISTOCENE OF SOUTHEASTERN
UNITED STATES

By CuayTton EK. Ray

WitH Two PLAtEs

CAMBRIDGE, MASS., U.S.A.
PRIN LED HOR he MUS 80 M

Marcu, 1960

*

No. 3 — TricHEcopon HUXLEYI (Mammalia: Odobenidac) in the
Pleistocene of southeastern Umted States

By CuAYTon EK. Ray

In 1877 Joseph Leidy (pp. 214-216, Pl. 30, fig. 6) deseribed
and illustrated the tusk of a fossil mammal from the Ashley River
phosphate beds, near Charleston, South Carolina. He referred
the specimen, with reservations, to the modern Atlantic walrus.
This tentative Pleistocene record for the modern walrus, together
with its climatic implications, has been accepted uneritically by
biogeographers and paleoecologists to the present day (cf.
Deevey, 1949, p. 1375; Dorf, 1959, pp. 184, 196). Recently, my
own attempts to identify a fossil tusk (University of Florida
No. 3274) found near Sarasota, Florida, have led me to re-ex-
amine the South Carolinian and other occurrences of fossil
walruses. Aside from the living walrus, here treated as a single
species (ef. Scheffer, 1958, p. 84; Davies, 1958, p. 102), and
Quaternary specimens clearly pertaining to it,! the only form
with which the Sarasota specimen can be compared usefully is
Trichecodon huxleyi Lankester 1865,2 from the Pleistocene of
Europe. The tusks of other extinct forms are either unknown
(Prorosmarus) or poorly known (Alachtherium; Hasse, 1910,
ps oUa ik

DESCRIPTION

Judging from its slight curvature in the frontal plane, which,
although variable, is characteristically mesad in walruses, the
Sarasota specimen is apparently from the right side and will be
so considered for purposes of description. The dimensions of the
specimen are given in Table 1. The natural surface of the tusk
is composed of a thin, smooth layer of cementum, still preserved

1 These inelude, with the exception of the record for the Ashley River, those
records cited by Kellogg (1922, pp. 49-51, Trichecodon was then considered to be
Pliocene) and by Hay (1923, pp. 21-29), together with more recent reports by
Allen (1930), Borissiak (1930), Dow (1954), Handley (1953), Matsumoto (1926),
Norton (1930), and Palmer (1944).

2TLankester (1880, pp. 213-216) has shown beyond doubt that the name
Trichecodon rests with his East Anglian fossils and not with the Belgian ma-
terial of van Beneden (1877). Whether Trichecodon is congeneric with Odobenus
must be determined on the basis of the European material, and is thus outside
the scope of this paper.

130 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Table 1. Measurements of tusk of Trichecodon huxleyi
from Florida, U.F. 3274.

mm.
Length along outer curve 435.0
Length in straight line 406.0
Greatest anteroposterior diameter 93.0
Greatest transverse diameter 56.9
Ratio of greatest transverse diameter to greatest anteroposterior
diameter 61
Greatest circumference 240.0
Least anteroposterior diameter 55.3
Least transverse diameter 34.5
Least circumference 90.8
Maximum thickness of cementum preserved 2.6
Long diameter of pulp cavity at proximal end 68.0
Short diameter of pulp cavity at proximal end 28.9
Long diameter of osteodentinal tube at distal end (estimated ) 25.0
Short diameter of osteodentinal tube at distal end (estimated ) 9.3
Depth of open pulp cavity (minimum ) 60.0

for the most part. The dominant surface features are two strong-
ly developed longitudinal grooves on the flattened medial surface
(Plate 1A), both of which diminish in depth distally to the point
of disappearance at the extreme tip of the tusk, perhaps due in
part to wear. A similar, but weaker, groove is present on the
posteromedial surface of the tusk. On the lateral, convex side
are several lesser longitudinal grooves, largely smoothed out on
the surface, but clearly shown where the cementum is missing
(Plate 2). The cementum is thick in the grooves and thin on the
intervening ridges, the effect of which is to mask the grooves. This
condition is probably due to differential wear rather than to an
initial variation in thickness of cementum deposited. The ridges
are exposed to, and the grooves protected from, wear. The ce-
mentum is sharply differentiated by weathering into a firm,
black, inner layer of variable thickness (owing to the fluting)
and a flaky, gray, outer layer averaging 1-1.5 mm. in thickness.
A small area of the extremely smooth outer surface of the un-
weathered layer is exposed by the exfoliation of weathered
cementum at the distal extremity of the tusk (Plate 1B). At
least some of the subsurface flutings persist to the distal ex-
tremity of the tusk as revealed in section at the broken distal

RAY : TRICHECODON HUXLEYI 131

Figure 1. Transverse sections A, B, and C of U.F. 3274 at the points
A, B, and C respectively, indicated on Plate 1B. The medial surface is
upper, and anterior is to the right in each section. x1.

132 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

end. The fluted dentinal surface is marked by transverse stria-
tions or growth lines, which appear in Plate 2 as alternating
light and dark bands (owing to the presence of light colored
clay in the valleys between ridges), averaging about 9-10 of
each per centimeter. The shape of the transverse section of the
tooth is roughly that of a flattened ovoid with the medial side
plano-convex and deeply grooved and the lateral side smoothly
convex (Fig. 1). The tusk is rather strongly curved in the
parasagittal plane. Its inner curvature forms an are of a circle
with a radius of about 45 centimeters, and its outer, of about
40 centimeters. These figures represent only the ecrudest of
approximations, since the curvature is not constant throughout
the length of the tusk. The degree of external curvature is
strongly increased in the distal 10 centimeters of the tusk, as a
result of natural wear which produced a flattened facet on the
outer curvature (Fig. 1C). The tusk is also curved very gently
mesad. The base of the open pulp cavity is oceupied by a pro-
jecting mass of globular osteodentine (Plate 2), which fills the
cavity solidly distad from the base, as revealed at the fractured
distal end of the tooth. The tooth tapers abruptly distad, a
feature enhanced by the wear on the outer curvature. Except
for the distal 10 centimeters of the tooth, the curvature and
taper shown represent very nearly the initial, unworn condi-
tion. If either feature were due largely to wear, the osteodentine
would be exposed on the outer curvature of the tooth (ef. Lankes-
ter, 1880, p. 219, figs. 1, 2, and Mansfield, 1958, p. 26, fig. 5).
The small amount of wear, deep, open pulp cavity, and uncon-
stricted proximal end indicate that the tooth belonged to a sub-
adult or adult (but not senile) individual.

The general form and structure together with the character-
istic globular osteodentine leave no doubt that the tooth belongs
to a walrus-like mammal. The presence of osteodentine, the ovoid
cross-section, and the absence of ‘‘engine-turning’’ in the dentine
rule out the possibility of its belonging to a proboseidean. How-
ever, certain discrepancies prevent the assignment of the tusk
to Odobenus rosmarus. Although it is not certain that the true
outer surface is anywhere preserved in the fossil, the sheath of
cementum is apparently thinner than that of O. rosmarus. The
maximum thickness of cementum measurable on the fossil is 2.6

RAY : TRICHECODON HUXLEYI ea

mm.; that of a sectioned tusk of Odobenus rosmarus (M.C.Z.
21638), 4.0 mm. The maximum anteroposterior diameter of the
fossil tusk is 10 millimeters greater than that of any modern
walrus in the Museum of Comparative Zoology (19 adult indi-
viduals). Its maximum circumference equals that of the world’s
record Atlantic walrus and is exceeded only by the first and
second-place Pacific walruses (Webb, et al., 1952, pp. 155, 158).
These latter two individuals may be taken as approaching the
absolute maximum in robustness for Odobenus rosmarus, since
they represent maxima obtained from the world’s major col-
lections. It is improbable that the fossil hes at the upper size
limit of its species. The deep, persistent, longitudinal surface
grooves on the medial face and the fluting of the dentinal
surtace on the lateral face cannot be matched in O. rosmarus.
The transverse striations are not interspersed with widely sep-
arated, strong, transverse (annual?) ridges, as they are in
O. rosmarus (ef. Brooks, 1954, p. 40, fig. 7, and Mansfield, 1958,
p. 31, fig. 7, p. 35, fig. 8A). The fossil is more flattened trans-
versely than is typical of O. rosmarus. The curvature of the
tusk exceeds that typically seen in O. rosmarus (ef. Rutten, 1907,
Table 1). Its curvature can be matched very nearly by the right
tusk of M.C.Z. 7301 (which owes its great curvature in part to
extreme wear on the outer curvature), but otherwise exceeds
that of any tusk in the Museum of Comparative Zoology. The
initial taper of the fossil tusk is much greater than that of any
Recent tusk examined.

It is of great interest to note that Leidy (1877, pp. 215-216)
observed almost exactly the same dissimilarities between the tusk
from the Ashley River phosphate beds and those of the modern
walrus, all of which led him to suspect that his fossil represented
a distinct species, which wisely he did not describe on the basis
of his single example of such a variable element. Comparison
between the Sarasota fossil and Leidy’s description and figure °
reveals striking similarities in gross size, curvature, taper, and
fluting. The maximum anteroposterior and transverse diameters
are almost identical. Leidy’s specimen has a reduced transverse
diameter at its proximal extremity, indicating old age and

3 The specimen was part of a temporary exhibition of fossils belonging to the

Pacific Guano Comnany. Recent correspondence has failed to reveal its presence
in the Academy of Natural Sciences of Philadelphia or in the Charleston Museum.

134 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

cessation of growth, as is seen in old specimens of O. rosmarus
and in Trichecodon huxleyi (Rutten, 1907, p. 4).

Leidy evidently did not compare his specimen with Tricheco-
don huxleyi and his paper is not cited by the subsequent writers
on Trichecodon (van Deinse, 1943-44, Hasse, 1910, Rutten, 1907,
and Lankester, 1880). Kelloge was aware of all the literature,
but his paper was a review and did not deal with any of the
material at first hand. These lapses may be explained in part by
Trichecodon’s having been assigned to the Tertiary, until recently,
and to the very general lack of transatlantic liaison. The charac-
ters of the tusks from Sarasota and Charleston are exactly those
diagnostic of Trichecodon huxleyi. These include robust form,
ereat curvature, abrupt tapering, lateral compression, prominent
surface and subsurface fluting, uniform transverse striation on
dentinal surface, and thin cementum (Lankester, 1865, 1880; Rut-
ten, 1907). In the aggregate, these features readily distinguish
T. huxleyi from O. rosmarus, but no one character can be re-
garded as absolute. For example, the degree of lateral com-
pression can be expressed in terms of the ratio of transverse to
anteroposterior diameters. A series of 30 Recent tusks of
O. rosmarus in the Museum of Comparative Zoology gives a
mean ratio of .68 and a range of .59-.82. Comparable figures for
eight tusks of 7. huxleyi are .59 and .48-.71.4 Six of the eight
tusks of 7. huxleyi have ratios of .59 or less. The Floridian and
South Carolinian tusks have ratios of .61 and .59 respectively.
Thus, on the basis of lateral compression alone, the American
fossils fit more comfortably into 7. huxleyi but cannot be ex-
cluded absolutely from O. rosmarus. That there is great varia-
tion and thus considerable overlap between species in these
non-occluding teeth is not at all surprising. On the basis of all
characters mentioned above, the specimens from southeastern
United States may be assigned confidently to the extinct form.

PROVENANCE

The Floridian tusk was found in August, 1957, ‘‘near the edge
of a marl pit in a new housine development, De Soto Lakes,
4 The standard statistical operations were not carried out, since neither species

is represented by an adequate sample. Most of the skulls of O. rosmarus are
unsexed and without locality.

RAY : TRICHECODON HUXLEYI 135

where they were digging shell for their roads. . .. De Soto
Lakes is about six miles from the present shore of Sarasota Bay,
and about the same distance northeast of the center of the city
of Sarasota, virtually on the line between Sarasota and Manatee
Counties’’ (Murray, in litt., April 27, 1959). According to the
records and well logs of the Florida Geological Survey the
Miocene Hawthorn Formation is overlain by no less than 20 feet
of Pleistocene deposits in the vicinity of De Soto Lakes (Olsen,
in litt., June 1, 1959). Exploratory cores drilled 12 miles east
northeast and 21 miles due east of De Soto Lakes reveal Pleisto-
cene terrace sands 44 and 35 feet thick respectively (Cathcart
and McGreevy, 1959, pls. 23, 24, 84). The marl is probably
Pleistocene in age and represents the reworked top of the
Hawthorn Formation. A sample of sand extracted from the pulp
cavity of the tusk and examined by Dr. Jules R. DuBar, is
considered by him to be more like the Miocene than like the
Pleistocene sands of the area.

Kew if any of the vertebrate fossils from the Ashley River
phosphate beds bear precise locality or stratigraphic data owing
to their incidental recovery in the course of commercial opera-
tions. Most of the fossil mammals are clearly of Pleistocene age,
but some of the marine mammals are of probable Miocene aspect
(Allen, 1926, p. 447).

The Enelish specimens of 7. huxleyi come primarily from a
triangular area in Suffolk delimited by lines joining the towns
of Ipswich, Oxford, and Walton-on-the-Naze. Most of the fossils
are derived from the so-called Nodule Bed or Suffolk Bone Bed,
a discontinuous, thin deposit, possibly in part reworked from
older strata, at the very base of the Red Crag Series (Chatwin,
1954, pp. 42-57; Lankester, 1880, pp. 213, 216; Newton, 1891,
p. 17). The Red Crag Series, formerly thought to include strata
as old as Miocene, is now assigned to the earliest Pleistocene
(Charlesworth, 1957, pp. 599, 1016; Chatwin, 1954, p. 42).
T. huxleyi has also been recorded from the Cromer Forest Bed
on the basis of a single tusk (Newton, 1882, p. 26; a second
fragmentary tusk showed resemblances to that of O. rosmarus),
and doubtfully from the Chillesford Beds at Aldeby on the basis
of the proximal half of a femur (Newton, 1891, p. 17). The
Cromer Forest Bed, which overlies the Red Crae Series, is

136 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

generally assigned to the first interglacial (Charlesworth, 1957,
p. 1016; Zeuner, 1959, pp. 137-138).

The fine skull referred by Rutten to 7. hualeyi was dredged
from the sea near Breskens, West Scheldt, Zeeland, Netherlands.
tutten, influenced by the supposed Pliocene age of the Red
Crag Series, attributed a Plhocene age to the Dutch fossil. The
overgrowth of barnacles, perfect preservation, and considerable
organic content suggest a Quaternary age for the specimen. Van
Deinse (1948-1944, pp. 97-101) has subsequently reported upon
two almost perfect skulls found in the same vicinity. Hooijer
(1957, p. 256) has assigned these and all the ‘‘black bones’’ of
the Scheldt estuary to the base of the Pleistocene and correlated
them with the Nodule Bed of the Red Crag Series. Thus, acecord-
ing to current stratigraphic interpretations, 7. huxleyi is known
definitely only in the Pleistocene.” I am unable to determine the
affinities of the fragmentary tusk from the vicinity of Antwerp
inadequately described by van Beneden (1877, p. 47 and Pl. VI,
fig. 8) and considered by Lankester to pertain to T. hiazley?.

PALEOECOLOGICAL SIGNIFICANCE

It is perhaps significant that Trichecodon is unknown from
the cold-water members of the Crag Series, but occurs in the
basal bed of the Red Crag Series, probably in the Chillesford
Beds, and in the Cromer Forest Bed, all of which were laid
down under temperate conditions (Zeuner, 1959, p. 148). The
occurrence of 7. huxleyi in the southeastern United States sug-
gests that it was an inhabitant of warmer waters than is the
living walrus. One hesitates to invoke boreal conditions in the
surface waters of the Gulf of Mexico, as would appear to be
necessary for Odobenus rosmarus. If such conditions prevailed,
the terrestrial Pleistocene fauna of peninsular Florida, the
climate of which is profoundly influenced by adjacent seas.
would certainly have reflected such cooling, and it does not.
It seems probable that the southerly limits of the range of
T. huxleyi extended into warm-temperate seas.

5 Although I consider it to be unlikely, it must ke admitted that the nature
of the stratigraphy at the De Soto Lakes, Charleston, and East Anglian localities
together with the nature of the fossils discovered at these localities (isolated
tusks and a fragmentary femur only) leaves open the possibility that the fossils
could be reworked from deposits as old as the Miocene. The fine Duteh material
could searcely have been redeposited.

RAY ;: TRICHECODON HUXLEYI er

The amphi-Atlantic distribution of 7. huxleyi is not unex-
pected and is analogous to the distribution of Halichoerus grypus
(ef. Davies, 1958, p. 110). Like the grey seal, it may never have
spread into the Pacific Ocean, owing to an aversion to subarctic
waters (assuming also that its range never extended far enough
south to utilize the Central American water route).

With the elimination of the South Carolinian record for
Odobenus rosmarus, the most southerly extension of the range of
this species in the Pleistocene recedes northward some 230 miles
to Kitty Hawk, North Carolina at 36°03’ north (Hay, 1923,
p. 29).6 In Europe the most southerly Pleistocene localities for
O. rosmarus are in the vicinity of Paris (Montrouge, 48° 50’
north, and Saint-Menehould, 49° 04’ north; Kellogg, 1922, p. 50).
Before commenting upen these records it is necessary to discuss
the Recent distribution of the walrus. Unfortunately, the walrus
is firmly and erroneously associated with Arctic conditions in
popular thinking and in much of the scientific literature. Allen
(1930) has ealled attention to the fact that the walrus bred on
Sable Island off Nova Scotia (44° N.) at least through 1650 and
was recorded off Cape Breton as early as 1583. The number of
specimens dredged from New England waters (Allen, 1930;
Dow, 1954; Palmer, 1944) and a live record for Plymouth in
1734 (Allen, 1930) suggest the possibility of regular seasonal
occurrence south perhaps to Cape Cod within Recent time. In
Europe the walrus apparently frequented (and bred upon?)
the Orkney Islands (59° N.) through 1550 (Ritchie, 1921, p. 8)
and is still reported occasionally in North Sea waters (Ritchie.
1921, pp. 8, 9, 77 et sqq.; Mohr, 1952, pp. 251-254). Very little
importance can be attached to the sporadic southerly occurrences
of the walrus, since occasional widely extra-limital records are
the rule among marine mammals. However, Sable Island and
northernmost Scotland clearly did he within the normal range
of the walrus in historie time. Both Sable and Orkney Islands
lie near the southern limits of the boreal sea (Cleneh and Turner,

6 There is in the University of Florida collections (U.F. 2112) a small fragment
of the worn distal end of a walrus tusk from St. Mary’s, Camden County, Georgia,
Its thin cementum. fluting. lateral compressien, and transverse striations are
suggestive of Trichecodon, but I regard the specimen as generically indeterminable.
Dorf (1959, p. 196) states that ‘walrus bones have been found .. . as far as
South Carolina and Georgia,’ whieh is correct only in the sense that Charleston
is as far south as, indeed farther south than, much of (inland) Georgia. I am
aware of no published record for the walrus in Georgia.

138 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Ms. Zoogeographic provinces of the western Atlantic; Feyling-
Hanssen, 1955, p. 25). Thus, on a uniformitarian basis, a
minimum hypothesis to account for the Pleistocene distribution
of O. rosmarus would shift the southern limit of the boreal sea
south to Kitty Hawk on the west and to Paris on the east. Kitty
Hawk hes nearly 600 miles south of Sable Island, and Paris
nearly 700 miles south of Orkney. In Recent time the range of
the walrus extended 15° farther south in the western Atlantic
than in the eastern. The fossil evidence indicates a figure of 13°
for the Pleistocene, suggesting that similar zoogeographie zones
extended farther south in the western than in the eastern At-
lantic during the Pleistocene, just as they do today.

It seems likely that 7. huxleyi ranged to the south of O. ros-
marus on either side of the Atlantic. During glacial epochs both
forms would have moved southward, with 7. huxleyi becoming
separated into American and European populations in south-
eastern United States (Florida and South Carolina records) and
(hypothetically) southwestern Europe, while O. rosmarus ex-
tended its range southward to Kitty Hawk and Paris, though
presumably maintaining its transatlantic connection in the north.
During preglacial and interglacial epochs both forms would
have moved northward, with 7. huxleyi re-establishing trans-
atlantic interchange between the previously isolated populations
in northeastern North America (hypothetical) and northern
Europe (Enelish and Dutch records), while O. rosmarus re-
treated into high latitudes, undoubtedly establishing connection
with the Pacific population at times (Davies, 1958, pp. 102-103).

ACKNOWLEDGMENTS

First and foremost I wish to express my thanks to Marian
Murray of Sarasota, Florida, who discovered the fossil tusk,
recognized its importance, and graciously deposited it in the
Florida State Museum at the University of Florida. Thanks are
also due to Mr. Richard van Frank for the photographs, Pro-
fessor Bryan Patterson and Dr. Ernest Williams for critical
reading of the manuscript, Drs. W. J. Clench and Ruth Turner

T1It might be argved that Arctic seas extended farther south at the time when

walruses occurred in these southerly localities, but they were already exterminated
or nearly so by the onset of the “Little Ice Age” about 1650 (Dorf, 1959, p. 199).

RAY : TRICHECODON HUXLEYI 139

for unpublished zoogeographic data, and Messrs. A. A. Arata,
Sy eOlsens Drs, H. Me Burton, He Daviss Jd. he DuBar, DA:
Hooijer, and K. F. Koopman for valuable information provided
via correspondence. This work was carried out while on the
tenure of a National Science Foundation predoctoral fellowship.

LITERATURE CITED

ALLEN, G. M.
1926. Fossil mammals from South Carolina, Bull. Mus. Comp. Zool.,
vol. 67, no. 14, pp. 445-468.
1930. The walrus in New England. Jour. Mammalogy, vol. 11, no. 2,
pp. 139-145.

BENEDEN, P. J. van
1877. Description des ossements fossiles des environs d’Anvers. Pt. I.
Pinnipédes ou Amphithériens. Ann. Musée Royal Hist. Nat.
Belgique, Tome 1, pp. 1-88.

BorISSIAK, A.
1930. <A fossil walrus from the Okhotsk coast. Ann. Soc. Paléont.
Russie, vol. 8, pp. 1-10 (1928-29). (Russian, with English sum-
mary. )

Brooks, J. W.
1954. <A contribution to the life history and ecology of the Pacific
walrus. Alaska Cooperative Wildlife Research Unit, Spec. Rept.
1, viii + 103 pp.

CarHcart, J. B. and L. J. McGREEVY
1959. Results of geologic exploration by core drilling, 19538, land-
pebble phosphate district, Florida. Bull. U. 8. Geol. Survey,
no. 1046-K, pp. iv, 221-298.

CHARLESWORTH, J. K.
1957. The Quaternary era. London, Edward Arnold Ltd., vol. 1,
pp. 1-592; vol. 2, pp. 5938-1700.

CHATWIN, C. P.
1954. East Angha and adjoining areas. Jn, British regional geology,
ed. 3, 99 pp., Great Britain, Geol. Surv. Mus., Dept. Sci. Indus-
trial Res., London.

Davirs, J. L.
1958. Pleistocene geography and the distribution of northern pinni
peds. Heology, vol. 39, no. 1, pp. 97-113.

140 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

DrEvry, E. S., JR.
1949. Biogeography of the Pleistocene. Part I: Europe and North
America. Bull. Geol. Soe. Amer., vol. 60, no. 9, pp. 1815-1416.

DriInsE, A. B. VAN
1943-44. Over Resten van fossiele en recente Pinnipedia, aangetroffen
in Zeeland en elders Nederland. De Levende Natuur, vol. 48,
pp. 84-87, 97-101, 119-125, 133-136.

Dorr, E.
1959. Climatic changes of the past and present. Univ. Michigan Con-
trib. Mus. Paleont., vol. 13, no. 8, pp. 181-210.

Dow, R. L.
1954. Walrus skull (Odobenus) from Bluehill Bay, Maine. Jour.
Mammalogy, vol. 35, no. 3, p. 444.

I bYLING-HANSSEN, R. W.
1955. Stratigraphy of the marine late-Pleistocene of Billefjorden,
Vestspitsbergen. Skrifter, Norsk Polar-Institutt, nr. 107, 186 pp.

HANDLEY, C. O., JR.
1953. Marine mammals in Michigan Pleistocene beaches. Jour. Mai-
malogy, vol. 34, no. 2, pp. 252-253.
THLASSE, G.
1910. Les morses du Pliocéne poederlien & Anyers. Bull. Soc. Belge
Géol., Paléont. Hydrol., Brussels, Mém., vol. 23, fasc. 3, pp.
293-322.

BUNye; @, 12%
1923. The Pleistocene of North America and its vertebrated animals
from the states east of the Mississippi River and from the
Canadian provinces east of longitude 95°. Publ. Carnegie Inst.
Washington, no. 322, 499 pp.

Hooter, D. A.
1957. Mammals and correlation of the ‘‘black bones’’ of the Schelde
Estuary with those of the Red Crag. Geol. en Mijnbouw, Jaarg.
19, nr. 7, 0.8., pp. 205-256.
IKXELLoGG, R.
1922. Pinnipeds from Miocene and Pleistocene deposits of California.
Univ. Calif. Publs. Bull. Dept. Geol. Sei., vol. 13, no. 4, pp.

23-132.

RAY : TRICHECODON HUXLEYI 141

LANKESTER, E. R.

1865. On the sourees of the mammalian fossils of the Red Crag, and
on the discovery of a new mammal in that deposit, allied to the
walrus. Quart. Jour. Geol. Soe. London, vol. 21, pt. 3, no. 83,
pp. 221-232.

1880. On the tusks of the fossil walrus, found in the Red Crag of
Suffolk. Trans. Linn. Soe. London, 2 ser., Zoology, vol. 2, art. 6,
pp. 213-221.

Leipy, J.
1877. Description of vertebrate remains, chiefly from the phosphate
beds of South Carolina. Jour. Acad, Nat. Sei. Philadelphia, 2
ser., vol. 8, pt. 3, pp. 209-261.

MANSFIELD, A. W.
1958. The biology of the Atlantic walrus Odobenus rosmarus rosmarus
(Linnaeus) in the eastern Canadian Arctic. Fish. Res. Bd.
Canada, Ms. Rept. Ser. (Biol.), no. 658, pp. 1-146.

Matsumoto, H.
1926. On two species of fossil Pinnipedia from Kazusa and Saghalin.
Sci. Repts. Téhoku Imperial Univ., Sendai, Japan, 2 ser. (geol.),
vol. 10, no. 1, pp. 13-16.

Monr, E.
1952. Die Robben der Europidischen Gewiisser. Monographien der
Wildsiugetiere, vol. 12, 285 pp.

NEwTON, E. T.
1882. The Vertebrata of the Forest Bed series of Norfolk and Suffolk.
Mem. Geol. Surv. England and Wales, 1438 pp.
1891. The Vertebrata of the Pliocene deposits of Britain. Mem. Geol.
Sury. United Kingdom, 137 pp.

Norton, A. H.
1930. The mammals of Portland, Maine, and vicinity. Proc. Portland
Soc. Nat. Hist., vol. 4, pt. 1, pp. 1-151.

PALMER, R. S.
1944. Walrus remains from northern New England. Jour. Mammalogy,

VOlZoanOn a Dales

RITCHIE, J.
1921. The walrus in British waters. Scott. Naturalist, nos. 109 and
110, pp. 5-9; nos. 113 and 114, pp. 77-86.

142 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

RUTTEN, L.
1907. On fossil trichechids from Zealand and Belgium. Koninkl.
Akad. van Wetenschappen, Amsterdam, vol. 10, pp. 2-14.

ScHEFFER, V. B.
1958. Seals, sea lions, and walruses. Stanford Univ. Press, Stanford,
California, 179 pp.

Wess, S. B., M. Baker, F. K. Barsour, A. ELy, and A. C. GILBERT
1952. Records of North American big game. Boone and Crockett Club,
Charles Scribner’s Sons, N. Y., 174 pp.
ZEUNER, FE. E.
1959. The Pleistocene period: Its climate, chronology and faunal
successions. Hutchinson and Co., London, 447 pp.

ADDENDUM

After the present paper had gone to press it was found that
J. G. D. Clark (1952, Prehistoric Europe. The Economic Basis:
London, Methuen & Co. Ltd., xix, XVI + 349 pp.) has noted
(p. 84) the association of Odobenus rosmarus with prehistoric
man at sites in northwestern Europe, including Skara Brae,
Orkney, and Jarlshof, Shetland.

‘FX FLEE AD ‘Wajwny wopoosaya.y JO Yysny Jo syoodse (q) [vio}VT puw (VY) [VIP “TL aI

Plate 2. Dorsolateral aspect of proximal end of tusk of Trichecodon
9_C

hualeyi, U.F. 3274, showing pulp cavity, globular osteodentine, and fluting
and transverse striations where cementum is broken away. x1.

Bulletin of the Museum of Comparative Zoology
PNM TEL IN UR EIN Tard) CPOE GMa (er 18

Vor l22) Nor 4

CONTRIBUTIONS TOWARD A RECLASSIFICATION
OF THE FORMICIDAE. Il. TRIBE
AMBLYOPONINI (HYMENOPTERA)

By Wiuu1am L. Brown, JR.

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM
Marcu, 1960

PUBLICATIONS ISSUED BY OR IN CONNECTION
WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

BULLETIN (octavo) 1863 — The current volume is Vol. 122.

BreEviora (octavo) 1952 — No. 124 is current.

Memorrs (quarto) 1864-1938 — Publication was terminated with
Vol. 55.

JOHNSONIA (quarto) 1941— A publication of the Department of
Mollusks. Vol. 3, no. 39 is current.

OccASIONAL PAPERS OF THE DEPARTMENT OF MOoLuusks (octavo)
1945 — Vol. 2, no. 23 is current.

PROCEEDINGS OF THE NEW ENGLAND ZOOLOGICAL CLUB (octavo)
1899-1948 — Published in connection with the Museum. Publication
terminated with Vol. 24.

The continuing publications are issued at irregular intervals in num-
bers which may be purchased separately. Prices and lists may be
obtained on application to the Director of the Museum of Comparative
Zoology, Cambridge 38, Massachusetts.

Of the Peters ‘‘Cheek List of Birds of the World,’’ volumes 1-3 are
out of print; volumes 4 and 6 may be obtained from the Harvard Uni-
versity Press; volumes 5 and 7 are sold by the Museum, and future
volumes will be published under Museum auspices.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE

Vou. 122, No. 4

CONTRIBUTIONS TOWARD A RECLASSIFICATION
OF THE FORMICIDAE. III. TRIBE
AMBLYOPONINI (HYMENOPTERA)

By Wiuu1am L. Brown, JR.

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM
Marcu, 1960

No. 4 — Contributions toward a Reclassification of the
Formicidae. III. Tribe Amblyoponini (Hymenoptera)

By Wiuui1am LL. Brown, JR.

CONTENTS

Page

IntCrOdUCtTION “Nee sia. See ce See ne Se «hc a eee ene, 145
eiripal SCharacbers)s <peee espe or ee he eae epi 3 cso isu ee 146
HMcolocyaan dabehavlOnea sia. Seee. 6 ey-bs eget segs Oaei cl yeoisicnc donee Poca eee 149
Relationships, distribution and evolution of the genera ............. ileal
Rreatmentsor Species-levell iaxonomy~s psem 4 - - cls ee a eee 153
Key to the genera of Amblyoponini (workers and females) ......... 154
NAYS) 5 PASLAVED EE rr avonecd SONG. ORS SRY GPCRS Ce OT RCESIS c, - RcMNO mR Reena site me aac eae ee a:
AMID LyOPONGs BI CHUSON Mei er eet eee Aare re eo eee 155

Moy stat nl vO COTE eto ny ec cpteer are Sn Src: | hats omar: SPs 169

Mey GDODONEP ROP CIM rie, sy cert ete AOU coe, OO RA A dat ee 170

Ee rronopelbaye Maye ees sae Ee ee re ee a ee eee 173
OnychomynmextHimerye). weet cele ce oo, ee ee anc ee ede Ae a ee 8

IN CETEAC ee NCCISs eee ee eat ee aera AS ek eA SRPREES oh eA Seer 181
Paraprionopeltamkisnezove ieee soe eee ica cle bene eee 181
Dorylozelis] Hore lyeryse rs eee ee keer ee els eet SANE: Se ee CI 181

IA Den Ciixsiee tee eta eer ear ees, ere: ore desea ce Meany rg ature else
Key to the New World species of Amblyopone ................ 191
Key to the neotropical species of Prionopelta ............... 5, wig
Key to the Indo-Australian species of Prionopelta ............. 221
Ste Otene Lerencesmeube die = «© este es.< Cie) scsi eel ecg ead = et eg oe 224

INTRODUCTION

The tribe Amblyoponini Forel as dealt with in this section has
as its core the old tribe Amblyoponini of the Emery-Wheeler
classification, with Amblyopone (including Stigmatomma and
other synonyms), Myopopone and Mystrium. From _ tribe
EKetatommini is transferred the genus Prionopelta, as already
indicated in Part II of this series (see Brown, 1958a: 177, and
below, under genus Prionopelta). Onychomyrmex, placed by
Emery in tribe Ponerini, and by Wheeler in a separate tribe
Onychomyrmicini, is here recognized as a specialized group of
amblyoponines, following Clark, 1928.

146 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

In addition to these clearcut members of the tribe, there are
two insufficiently known ants that may or may not belong to the
Amblyoponini. One of these is Dorylozelus mjoebergi Forel (see
below, p. 181), and the other is Paraprionopelta minima Kusnezov
(see p. 181). These last two genera will be treated here as
‘“incertae sedis’? members of subfamily Ponerinae, not definitely
assigned to Amblyoponini or any other tribe. However, I feel
that both genera, when better known, will probably fit into
existing tribes.

TRIBAL CHARACTERS

The Amblyoponini are rather homogeneous as compared to
many other ant tribes. Size varies from minute (ca. 2 mm. total
length) to medium (ca. 1 em.) for workers and females, and
males are within this range. The workers and females are similar
in most respects, though the latter are usually larger and more
heavily pigmented, with large eyes, three ocelli, and, when vir-
ein, well-developed wings and flight sclerites in the thoracic part
of the alitrunk. Females of Onychomyrmex form a conspicuous
exception in that they are minute-eyed, wingless and otherwise
‘*dichthadiiform.’’ In additon, ergatoid females occur in some
Amblyopone species, but no species is yet known to have ergatoids
completely replacing normal females. The workers are essentially
monomorphic, but may show wide variation in size in a single
colony.

Workers and females. Amblyoponine workers have the ocelli
absent (sometimes represented by one or more pits), and the
compound eyes are never as well developed as in most epigaeic-
ally-foraging ant groups; typically, they are reduced to very
small remnants, and they may even be absent or so shrunken as
to escape ordinary microscopic observation of the integument.
When they are present, the eyes are characteristically placed
behind the middle of the sides of the head.

The worker-female cranium is slightly depressed (elliptical
or broadly ovoid in cross section), with straight or gently con-
vex, parallel or posteriorly converging sides. The occiput has a
transverse posterior border (straight, weakly concave, or weakly
convex) and broadly rounded occipital angles merging smoothly
into the sides, not forming conspicuous lobes. The anterior genal

BROWN: ANT TRIBE AMBLYOPONINI 147

angles may or may not be toothed. Clypeus well developed, the
median portion often forming a median lobe or apron and fre-
quently with a series of teeth, tubercles or denticles along the
free margin. Mandibles varying widely with genus and species,
always strongly developed and inserted remotely, at the corners
of the head (Figs. 1, 3, 4, 10, 15, 18, 19, 29, 30, 33-40, 48).

The labrum is broad and shield-shaped, folding back up over
the smaller mouthparts when retracted, and often bearing den-
ticles near its base, presumably to help in holding prey. The
palpi are short, with segmentation ranging from 5, 3 downward,
normally the same in both worker and female (with some ex-
ceptions).

Amblyopone australis, worker from near Auckland, New Zealand. Figure

1. Head and mandibles, with left antennal scape, full-face view. Figure 2.
Petiolar node and adjacent parts, viewed from the side.

Frontal carinae short, situated more or less near the midline,
widely separated from each other or, in smaller forms, con-
tiguous. The antennae are simple, often with a rather short,
heavy seape; funiculus slender throughout to distinctly clavate
at apex; the whole antenna usually has 12 segments; more
rarely there are 11 segments, and in one species only 7.

Alitrunk, petiole and gaster conservative in form throughout
all genera. Alitrunk long and narrow, usually more or less
parallel-sided (pronotum often slightly wider) and the dorsum
more or less horizontal from front to rear; promesonotal suture
normally complete and possibly often mobile; metanotal groove

148 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

well-marked to obsolete. Propodeum with distinet dorsal and
declivitous surfaces, rounding into each other, no propodeal
teeth or angles. Metapleural glands with well-developed bulla
and orifice. Petiole with little or no extended peduncle; node
with steep anterior face and extensive dorsal face, but usually
firmly attached to postpetiole over all or nearly all of its pos-
terior extent, so that the dorsal face of the node is generally more
or less continuous in level with the postpetiolar dorsum (Figs. 2,
31, 32, 46). This type of node is very characteristic for the
tribe, and only a few scattered convergences to it exist in other
ant tribes; it may well represent a structural holdover from
the tiphioid wasps that gave rise to the ants. The postpetiole
(abdominal III) and first gastrie (IV) segments are usually
subequal and ringlike, separated by a modest constriction; the
succeeding segments are also ringlike, but shorter. The petiole,
postpetiole and gaster together form a rather regular cylinder,
usually rather long and straight (shorter in Mystrium). Sting
present and functional.

Sculpture varying from foveolate or rugulose through reticu-
late-punctulate to smooth and shining. Pilosity simple, fine,
varying from fairly sparse to dense, short, and pubescence-like ;
Mystrium is an exception, with many body hairs spatulate, clav-
ate, or squamose. Color varying from blackish to pale yellow;
depigmentation is common in this predominantly eryptobiotic
group. Tarsal claws simple.

The known males of all genera agree in general with the
description of that sex given for Amblyopone on page 161. The
wing venation for both male and female is also covered by the
variation found in Amblyopone (Figs. 5-7, 17, 47).

The internal anatomy of the Amblyopone pallipes worker and
female has been studied in some detail by Whelden (1958), and
the cytology by Whelden and Haskins (1954). Whelden thinks
there are 9-10 Malpighian tubules in A. pallipes, and he covers
the glandular system in some detail; the glands are apparently
generally similar in kind and position to those of higher ants
like Myrmuica.

Larvae of species of Amblyopone, Mystriwm, Prionopelta and
Onychomyrmex have been described and figured by G. C. and
J. Wheeler (1952: 113-117, 120, 637-639, 653-660), though these

BROWN: ANT TRIBE AMBLYOPONINI 149

authors group them according to the Emery-Wheeler classifica-
tion. The larvae belong to the more primitive group among
the Ponerinae, lacking tubercles over the body. Their hairs are
moderately to fairly abundant and simple (denticulate or
branched in most Eetatommini), though hairs are sparse or vir-
tually lacking on the cranium in Amblyopone. Thoracic seg-
ments forming a slender neck in Amblyopone; only the first see-
ment much narrowed in Prionopelta; thoracic segments broader
than the first abdominal segment in Onychomyrmex mjoebergi.
Amblyopone larvae are rather similar to those of Myrmecia, and
probably approach the primitive type of larvae among ants, while
Prionopelta is slightly and Onychomyrmex is more strongly
specialized in the direction of the Cerapachyinae or even the
Dorylinae, in agreement with their presumed nomadic existence.

Amblyoponine pupae are generally enclosed in cocoons, from
which the workers can frequently emerge unaided (unlike higher
ants). A. celata Mann apparently lacks cocoons, and the pupae
are unknown for most species.

ECOLOGY AND BEHAVIOR

The members of this tribe are all, so far as known, obligatory
predators of other arthropods. All or nearly all species are pre-
dominantly eryptie foragers in the soil, leaf litter or rotting logs.
The nests, beyond the first stages of foundation, often tend to be
diffusely spread through the substrate, and their limits may be
very ill-defined and constantly shifting. One gets the impression
that there is little permanent centralization of brood chambers,
and that the larvae are moved about a great deal. Perhaps it is
common in this tribe for the larvae to be moved to large dead
prey wherever the latter is killed, rather than the reverse. I have
several times found Amblyopone pallipes workers and larvae
clustered about large lithobiud chilopods under stones in such a
position as to suggest that the larvae had been transferred to the
prey; Wilson (1958a) reports that Myopopone castanea appar-
ently does the same thing with the large beetle larvae it preys
upon in rotten logs in New Guinea. Wilson (1958b) has already
proposed that behavior like that of some Amblyoponini could
have led to the true army-ant foraging and nomadism, and of

150 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

course we can point to Onychomyrmex as an Amblyopone-de-
rived stock that has already travelled a long way along this road.

Observations, some of which are reported in more detail else-
where in this section, indicate that amblyoponine workers and
nest-founding females generally attack living prey (in the form
of chilopods, beetle larvae, or other arthropods), seize it in the
formidable mandibles, and sting it to death with the long sting
while holding on with the jaws. The prey may be much larger
than the ant, and high potency is indicated for the ants’ poison.
The jaws, clypeus and labrum are usually toothed and well-
suited to gripping active victims. The larvae feed directly on the
prey when they are placed on it, or when it is given to them, and
they may insert their heads into the prey’s body to feed. Despite
repeated trials, no one has yet succeeded in inducing Amblyopone
workers or females to take honey, but Haskins (1928) says that
A. pallipes will sample fruit, and that the males of this species
feed avidly on honey, an interesting point in view of the im-
portance of nectar or honeydew to tiphioids and other lower
Hymenoptera, as well as to all the castes of Myrmecia.

Amblyopone and its relatives are moisture-loving species, most
abundant in forested, temperate or tropical areas. A. pallipes
is, for instance, often the commonest ant in wet Rhododendron
stands in the Appalachians, and A. australis and several other
smaller species of Amblyopone are common in wet, dark, fern
oullies in Tasmania and southeastern Australia where most other
ants are scarce. In drier areas, such as the Kansas plains or
North Africa, the ants stay deep in the soil during the dry
season, and are usually only found in flood times or by chance
in deep man-made excavations.

Nuptial flight has been observed, at least in part, by Haskins
(1928) for A. pallipes and by Haskins and Haskins (1951) for
A. australis. The sum of evidence indicates that in these species,
the female emerges from the nest, either flies or walks to an ex-
posed position on rocks or foliage, and is found and mated by the
male, which arrives in active flight. A. australis was observed
‘‘extending and arching the gaster and rubbing it with the hind
legs,’’ which suggests that it may have been releasing an at-
tractant to draw the males. After mating, the females may
return to the parent nest (A. pallipes always or usually; A.

BROWN: ANT TRIBE AMBLYOPONINI 151

australis sometimes) or may enter the soil or rotten wood after
a time and construct a cell (A. australis), from which it forages
actively for the prey it catches and stings to death. The cell may
or may not be closed by the young female when she leaves to
hunt, and prey is brought back to the cell regardless of whether
brood is yet present or not. Sometimes two or more dealate
females of A. australis apparently combine to produce a primar-
ily polygynous nest. The most complete accounts of behavior and
nest-founding are those of Haskins and Haskins (1951), and
they give all the relevant earlher references. We need more in-
formation on the behavior, food and nest-founding of these and
other species of Amblyoponini, because they are obviously so
primitive in many ways that we may hope to learn a great deal
from them about the evolution of ants in general.

RELATIONSHIPS, DISTRIBUTION AND EVOLUTION
OF THE GENERA OF AMBLYOPONINI

The present record indicates that 70-75 valid species are known
for the Amblyoponini of the world, of which about two thirds
belong to genus Amblyopone. Although it is likely that more
species remain to be discovered in this cryptobiotic tribe, the
chances are that the proportion of Amblyopone species to those
of other genera will increase rather than decrease. Amblyopone
ineludes both generalized and specialized forms within the tribe,
and it seems reasonably clear to me that all four of the remain-
ing genera can be derived separately and directly from this
large genus.

Mystriwm, with 6 species in the Malagasy area, one in West
Afriea, and one in the Indo-Australian area, looks as though it
probably arose from an African Amblyopone stock, although
Amblyopone is now represented in Africa only by a very few
rare species found along the western and northern fringe of
the continent. Prionopelta is very rare in southeastern Africa
and Madagascar, but is better represented in tropical America,
New Guinea and the southwest Pacific islands, with outhers in
New Caledonia and tropical Queensland.

Prionopelta tends to be best represented in numbers and in
species where the smaller species of Amblyopone are rare or
absent, which may indicate that Prionopelta has been competing

152 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

with small Amblyopone and winning out in some tropical areas.
In similar fashion, Myopopone castanea may be replacing Am-
blyopone australis in a specialized rotting-log niche in Melanesia
(see page 173).

Two fifths of all the Amblyopone species, and by far the most
massive populations maintained by any members of the tribe,
occur in the Australian region. Here several species, most
typically A. australis, appear to have broken out of a small-
predator adaptive zone and have secondarily enlarged in body
size and become general arthropod predators. Their reduced
wing venation and palpi indicate much smaller (and probably
more oligophagous) ancestors, presumably much like the inter-
eradient series of species, still existing in the Australian region,
that connects them with the small ‘‘ Pulakora’’ group forms. If
this reasoning is correct, A. australis represents a case, parallel
to that of Strunugenys nidifex Mann of Fiji, of ‘‘countercurrent
evolution’’ (for an explanation see Brown and Wilson, 1960).
The whole picture of the considerable Amblyopone radiation in
Australia speaks of the arrival of an early stock of the genus
in a practically unexploited series of niches in this isolated land
mass. As we see it there today, the genus still oceupies a prom-
inent place in the cryptobiotic-predator zone in the forests of
southern Australia.

One other niche left empty on the Australian continent was
the army-ant niche, but this was eventually entered by an
Amblyopone stock that developed into Onychomyrmex and
stayed within the northern forests. Onychomyrmex 1s now prob-
ably in competition with two or three species of the army-ant
venus Aenictus Shuckard that appear to have extended their
ranges into tropical Australia from the north in recent times.

Other relatively weak relictual representations of Amblyopone
occur in the Americas, especially in temperate North America
and southern South America, in New Zealand, in the Mediter-
ranean area, in southern and eastern Asia, and in the Kast
Indies. A single small species in Hawaii may have been intro-
duced by man. But the presence of undoubted endemic members
of the tribe in extralimital regions hike Chile, New Zealand, New
Caledonia, southwestern Australia, Tasmania, Cuba, Japan, Mad-
agasear and the Solomons, speaks for an old, worldwide distribu-
tion. Tribe Amblyoponini apparently is the surviving remnant

BROWN : ANT TRIBE AMBLYOPONINI 153

of a much larger and more varied group of ants, probably
dating back to the Cretaceous. Today, the tribe is reduced to
the few stocks that survived contact with more progressive ant
groups only because they became specialized for relatively con-
stant, predominantly eryptobiotie niches. Even though they thus
managed to avoid the heaviest competition, the general scarcity
of amblyoponines outside a few limited and usually more or less
peripheral areas points to their decline as a world group.

Unfortunately, we have no fossil record of the tribe, although
the morphology, habits and distribution indicate a greater age
than for the Ectatommini, which were present and widespread in
the Oligocene. The petiolar-gastric structure may indicate direct
lineage from the proto-formicid stem, but in other characters
(ineluding loss of anal lobe in hind wing, reduction of palpi and
eyes, elaboration of mandibles, simple tarsal claws, and 6-partite
proventriculus) the tribe is specialized.

TREATMENT OF SPECIES-LEVEL TAXONOMY

As already stated in Part II of this series, the primary aim
of the work is a revision of classification at the generic level.
But in the course of such work there is always uncovered a mass
of information concerning new synonymy, undescribed forms,
geographical variation, biology, ete., at the species level, which
logically should go with a survey of this kind. As in Part IH, such
information has been placed in the Appendix, where the items
are listed consecutively against boldface numbers in brackets
corresponding to the bracketed numbers in the text.

Former varieties and subspecies have been eliminated as such
in the species lists, and are either synonymized or treated as
species. Revisionary studies at species level were carried to
completion only in the cases where a clearcut decision seemed
possible on the basis of the available material. In a few other
eases, the taxonomic problems involved have been discussed, but
not settled. It is obvious that more work needs to be done at the
species level, particularly in Prionopelta, in the Madagascan
Mystrium, and among the Far Hastern Amblyopone. In the case
of the few names arbitrarily raised from infraspecific to specific
rank, this action does not necessarily carry any implication of
support for their distinctness as species.

154 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

The capital letters placed in parentheses before each specific
name indicate the kind of evidence upon which the present
generic placement is directly based.

(T) indicates that type material, nidotypes, reliably type-com-
pared material, or similarly authentic specimens have been
examined, in most cases by myself; rarely, examination has been
made by other myrmecologists.

(P) means that material identified from reasonably good de-
scriptions, or from other satisfactory evidence, has been ex-
amined and is thought to be correctly determined.

In eases where the species is not seriously questioned, but no
specimens referable to it have been seen, or if specimens seen
cannot be satisfactorily verified as to identity, no entry has been
made before the species name.

(?) signifies that, in my opinion, the species is inadequately
described for purposes of distinction and that its taxonomic
status is doubtful.

The species lists have been based on various myrmecological
compendia and basic papers, and were checked against Emery’s
Genera Insectorum list and the Zoological Record, 1908 through
1955. References through 1958 are included so far as I am
aware of them. I shall consider it a great favor if readers will
send me notice of the inevitable omissions for inclusion in a
corrective supplement.

Key to the Genera of Amblyoponim, based on the
workers and females

1. Mandibles short, narrow, closing tightly against the clypeus, their apical
borders distinct and completely occupied by 3 large teeth, of which the
middle tooth is shortest; basal border of mandible unarmed (Fig. 15;

WACLESPREAC sent O01 CS) umes) sia) See ene en ee Prionopelta Mayr?
Mandibles of another form, usually strongly projecting beyond clypeus
when closed, and with more than 3 teeth (Figs. 19, 29, 48) .......... 2.

2. Antennal funiculi markedly compressed; when head is viewed in perfect
full-face view, lobes of frontal carinae are approximately even with, or
slightly surpass, the anterior clypeal border beneath them (Fig. 10;
Indo-Melanesian area, widespread) .......... _... Myopopone Roger

1 Dorylozelus Forel, of Queensland, may be an amblyoponine, in which case it
would key out to Prionopelta.

BROWN : ANT TRIBE AMBLYOPONINI 155

Antennal funiculi not compressed, approximately round in cross section;
in full-face view of head, lobes of frontal carinae distinctly behind the
anterior median clypeal border (Figs. 19, 29, 40) .... 3.

3. Apex of mandible bluntly rounded or subtruneate as seen from above,
although the inner apical margin may bear two or more small teeth, tri-
angular in shape and often retrorse; many body hairs clavate or spatu-
late (Fig. 4; Madagascar, w. Africa, Indo-Australian area)

Mystrium Roger
Apex of mandible in the form of an acute tooth (Figs. 19, 29, 30, 33,
40, 48); body hairs simple, fine and tapered ...................... 4,

4. Tibia of posterior leg without an apical spur, or at most with a very
small, straight, non-pectinate vestigial spur; small, slender, predomi-
nantly smooth and shining species with greatly enlarged tarsal claws on
middle and posterior legs; queens wingless, small-eyed (‘‘dichthadii-
form’’); prebably with legionary habits (Figs. 46-48; Queensland)

Onychomyrmex Emery
Tibia of posterior leg with a well-developed apical spur having a curved,
broadly pectinate inner margin; species of diverse size and form, with
tarsal claws rarely enlarged; in cases where they are enlarged, the head
and alitrunk tend to be coarsely sculptured and often opaque; virgin
queens normally winged and with large compound eyes (Figs. 29, 33, 40;
widespread in temperate and tropical areas, but often local; not known
from ec. and s. Africa or Madagascar) ........ . Amblyopone Hrichson

THE GENERA
AMBLYOPONE ERICHSON

> Amblyopone Erichson, 1842, Arch. Naturg., 8(1): 260. Type: Amblyo-
pone australis Erichson, 1842, monobasic.

> Stigmatomma Roger, 1859: 250. Type: Stigmatomma denticulatum Roger,
1859, by designation of Bingham, 1903. N. syn.

> Arotropus Provancher, 1881, Naturaliste Canad., 12: 205. Type: Aro-
tropus binodosus Provancher, 1881, monobasie.

> Stigmatomma subgenus Xymmer Santschi, 1914: 311. Type: Stigma-
tomma (Xymmer) muticwm Santschi, 1914, monobasic.

> Stigmatomma subgenus Fulakora Mann, 1919: 279. Type: Stigmatomma
(Fulakora) celata Mann, 1919, by original designation.

> Amblyopone subgenus Neoamblyopone Clark, in Wheeler, 1927: 1. Type:
Amblyopone (Neoamblyopone) clarki Wheeler, 1927, by original designa-
tion, monobasic.

> Amblyopone subgenus Protamblyopone Clark, in Wheeler, 1927: 1. Type:
Amblyopone (Protamblyopone) aberrans Wheeler, 1927, by original
designation, monobasic.

156 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

> Lithomyrmex Clark, 1928: 30. Type: Lithomyrmesx glauerti Clark, 1928,
by original designation, monobasic. N. syn.

> Ericapelta Kusnezov, 1955: 273. Type: Hricapelta egregia Kusnezovy,
1955, monobasic. N. syn.
(For previous writings on synonymy and relationships of Amblyopone at.
generic and subgeneric level, see Wheeler, 1927; Clark, 1928; Brown,

1949).

Amblyopone is a curious genus containing a large and hetero-
veneous array of species, which represents the residue left after
dividing a few specialized amblyoponines among the few other
genera recognized here. Amblyopone is not only residual as a
genus; it is ‘‘central,’’ which is to say that it is probably to be
regarded as the stock from which the other genera have risen.
While its species differ widely in size, color, sculpture and in the
structure of the cranium, clypeus and mandibles, the general
form of the remainder of the body is remarkably uniform in
nearly all of them. In fact, departures such as the pedunculate
petiole of A. mutica are slight when compared to the variation
among species of other genera of ants, and stand out in Amblyo-
pone ouly because the great majority of species in the genus are
so monotonously similar in posteephalic body form.

The most characteristic trait of Amblyopone (and most other
Amblyoponini), aside from general habitus, is the structure of
the petiole and its mode of attachment to the postpetiole (Figs.
2, 31, 32). The petiole is nodiform and robust, with an abruptly
descending anterior face to the node and an approximately hori-
zontal dorsal face. No true posterior face is differentiated, the
node attaching directly, with little or no posterior constriction,
to the postpetiole. There is usually an anteroventral process
adorning the petiolar keel. The postpetiole is functionally a part
of the gaster, and is large, with only a moderate constriction
between it and the structurally similar succeeding segment (IV
abdominal). In both the postpetiole and the IV abdominal seg-
ment, the tergites are immovably fused to the sternites.

These characters of the petiole and gaster are shared by all
castes, including the males, and they are also more or less faith-
fully copied by all of the other genera of the tribe. The confor-
mation of this region appears to be of a very primitive kind
among the ants, and is similar to that seen among scolioid (ti-
phioid) wasps. This group of wasps is primitive among the

BROWN: ANT TRIBE AMBLYOPONINI Pat

aculeate Hymenoptera, and probably included, sometime during
the Cretaceous, the ancestral stem of the Formicidae. The sting
in worker and female Amblyopone is long, sharp, heavily sclero-
tized and obviously functional.

In the female and worker Amblyopone there is a fundamental
uniformity in the structure of the head and mandibles despite
the great interspecific differences one finds in detail. The man-
dibles always end in an acute (usually dentiform or spiniform)
apex, and the eyes, when present, are situated behind the middle

9

Heads of amblyoponine workers, full-face view. Figure 3. Amblyopone
sp. of reclinata group from Maeao. Figure 4. Mystriwm camillae from
near Darwin, northern Australia.

of the sides of the head. The eclypeus forms a band across the
anterior part of the head, and its median portion is usually more
or less projecting to form a low, broad lobe or apron, which in
most but not all species bears a row of denticles or small
tubercles.

The mandibles are linear (rarely triangular) and inserted
far apart at the corners of the clypeus; when at rest, their apical
portions cross over one another. In most species, the basal border

158 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

is absent or very poorly developed, and the masticatory margin
may either include it or be formed entirely of the apical border.
The dentition consists of few (more than 3) to many teeth, vary-
ing in size and form, and ranked in one or two rows along the
masticatory border (Figs. 1, 18, 19, 29, 30, 33-40). The dentition,
of course, varies widely according to species, and some species
show considerable intraspecific variation also. Fundamental gen-
eric characters are the presence of more than three teeth and
the more or less acute nature of the apex.

Most, but not all, species of Amblyopone bear on each antero-
lateral angle of the head a more or less distinct and usually acute,
dentiform extension of the gena (Figs. 3, 33, 40) which I have
previously called the ‘‘amblyoponine tooth.’’ However, the term
‘“‘oenal tooth’? used by Wilson (1958a) is more specifie and
objective, and I feel should be preferred.

The frontal carinae form narrow lobes in front; the carinae
and lobes may be widely separated (Fig. 29), close together
or even fused (Fig. 40), fusion or close proximity usually being
correlated with extreme reduction of size, depigmentation and
other traits of specialized subterranean life forms. The lobes
reach or overlap the elypeus in front, but do not surpass the
anterior border of the median lobe or apron (Figs. 29, 40).

The antennae are usually 12-segmented in the female castes,
but in one species. A. degenerata Borgmeier, fusion has reduced
the number to only 7 distinct segments. The flagellum varies
from slightly inerassate toward the tip to distinctly clavate, ac-
cording to species.

The labium is a broad, tongue-shaped flap or shield, apically
more or less deeply emarginate, hinged at the base so as to be
able to cover the smaller mouth-parts when in repose, or to swing
forward under and between the mandibles to assist in holding
prey or perhaps other objects as well. The maxillary and labial
palpi are always more or less reduced, as usual in ants with
hypogaeie tendencies; there is wide variation in the number of
segments among the different species, and some variation has also
been noticed within certain species. The highest counts noted
are those of A. impressifrons (5 maxillary, 3 labial) and the
reclinata group (5, 3 or 4, 3), and several species are known to
have counts of 4, 3, or 3, 2, or 2, 2; a female of A. sawndersi from

BROWN : ANT TRIBE AMBLYOPONINI 159

New Zealand had maxillary 1, labial 2, although a worker
assigned to the same species had 2 and 2.

The alitrunk is elongate, more or less parallel-sided, with a
distinct and complete promesonotal suture, in most species ap-
parently representing a flexible joint separating the two main
parts of the alitrunk. Metanotal groove present and distinct in
most larger forms, but lost or obsolescent in most smaller species.
Mesonotum (when fully distinguishable) short and transverse.
Propodeum unarmed, bluntly rounded into declivity. Sting long
and stiff, fully functional.

Female lke worker, and usually only slightly larger in size,
with large compound eyes, ocelli well developed, and flight scler-
ites differentiated ; however, the alitrunk is low and has a rela-
tively straight profile. Wings present; venation varying widely
with the species, and often within species, but relatively complete
even in the smaller species. Some of the larger or medium-sized
species have all of the ‘‘primitive’’ ant venation in both wings,
except for the first radial crossvein, which is present only as an
occasional atavism (Brown and Nutting, 1950:125). In many
of the smaller species, as well as in the large A. australis and some
of its large and medium-sized relatives, the forewing has lost
Rsf2°3 partly or entirely (compare Figures 5 and 6 with Figure
7). Venation is particularly variable in A. australis (of both
sexes), and this appears to be the normal situation, not just due
to ‘‘abnormality’’ as suggested by Kusnezoy (1955 :268). In this
species, Rsf2-3 may be completely absent, present as a spur of
varying length in its apical portion, or, more rarely, present only
as a variable free section attached to an atavistice first radial
crossvein; Mr. John Clark long ago sent me many sketches of
forewings of this last type. The anal vein varies in development
with the species, and may extend to or very nearly to CuA (thus
closing a ‘‘second discoidal’’ cell), or may fall short by a
greater or lesser distance. The ‘‘open’’ or ‘‘closed’’ condition
of ‘‘cubital’’ or ‘‘discoidal’’ cells, upon which Kusnezov so
largely based his 1955 classification, is in my opinion a virtually
useless character-system because of the wide variation and
subtlety of the occurrence of the veins themselves, among species
as well as within certain single species. I should like to re-
emphasize the position taken in 1950 by Brown and Nutting: the

160 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Wings of males of the reclinata group, species indeterminate. Figure
5. Specimen from Los Bafios, Luzon, (terminalia shown in Figs. 22, 24, 27).
Figure 6. Specimen from Chipon, Formosa (terminalia shown in Figs. 23,
25s 28).

BROWN : ANT TRIBE AMBLYOPONINI 161

most attention should be paid to the veins themselves; descrip-
tions of ‘‘eells’’ in ant wings convey much less information, and
convey it less precisely. In the Amblyopone species so far
ehecked, the hind wing always lacks an anal lobe.

The seulpture of worker and female is much alike in general,
but varies according to species and species-groups. The head,
and to a lesser extent the alitrunk, is usually much more dis-
tinctly sculptured than the gaster; frequently the gaster is pre-
dominantly smooth and shining, with scattered fine punctulae.
The sculpture of the head varies from smooth and shining, with
or without some foveolae or costulae, through coarse or fine stria-
tion (in most cases predominantly longitudinal), subvermiculate
rugulation, to fine, dense, reticulopunctation or reticulo-striola-
tion. The mandibles may be striate, punctate or smooth. The
posterior sides of the alitrunk, especially near the metapleural
elands, are frequently superficially striate, a feature that may be
connected with distribution of material from the glands. Pilosity
is simple and fine, and ranges from short, dense, pubescence-
like pile to longer and sparser hairs of irregular length.

The male of Amblyopone is normally smaller and more slender
than the corresponding female, and has an entirely different
head, usually or always broader than long, including the large,
convex compound eyes. The mandibles are slender, curved, strap-
like, the apex simple and acute or with two coarse teeth; when
closed, their tips meet or overlap, and the entire blades are
usually tucked away (rarely not) under the clypeus in such a
way that only their external margins show externally along the
anterior elypeal border. Clypeus usually convex and most often
with a rounded anterior border or apron that frequently carries
serially-arranged denticles or tubercles homologous to those of
worker and female, where they are present in these castes. An-
terior border more rarely transverse and approximately straight.

The male antennae are slender, but often slightly to moder-
ately incrassate toward their apices; the scape is short, but
usually more or less slender-cylindrical ; funiculus 12-seemented
in all of the species for which the male is known, but could con-
ceivably be of a smaller count in unknown males of species like
A. degenerata (see under Paraprionopelta, below). The funiecu-
lar segments increase in length toward the tip, the apical segment
being the longest one,

Amblyopone australis, male taken with workers at Dorrigo, New South
Wales. Figure 7. Wings. Figure 8. Mandible dissected out of head. Fig-
ure 9. Left half of genitalia as viewed from midline; valve of aedeagus
drawn as a very light line.

The under-mouthparts are in many ways similar to those of
the corresponding workers and females, and frequently the
palpi have the same segmental counts as the workers (2, 2 to
5, 3, of course varying with the species). In Amblyopone austra-
lis, however, the male has a formula of 4, 3, while the worker

BROWN: ANT TRIBE AMBLYOPONINI 163

has only 2 and 2. A similar situation probably holds in some of
the other species, particularly those related most closely to A.
australis. The formula of the male may therefore be said to be
either like that of the corresponding worker, or else more
conservative.

Alitrunk with well (but not excessively) developed flight
sclerites; notauli present and distinct, usually forming a com-
plete Y. Propodeum unarmed. Petiolar node in general shape
as in worker-female, but may be more slender or differ otherwise
in proportions; a ventral lobe or process is usually present an-
teriorly. Petiolar attachment and general form of anterior gaster
much as in worker; posterior portion modified, of course, to
receive genitalia. The VIII, and often the VII, sternites are
broadly emarginate posteriorly. The subgenital plate (or hypo-
pygium, sternite IX) varies greatly with the species; sometimes
it is rather simply tongue-shaped, but it is highly modified in
some species, and may even (some species of reclinata group)
possess a more or less well developed slender median caudal
process (Fig. 23). The pygostyles are usually present, but occa-
sionally are absent, or at least so reduced that they were not
noticed in the dissections. The phallus is constructed on the
usual formicid plan, but there is a great deal of modification of
the particular components in many of the species, so that the
genitalia in this group furnish much better species characters
than is usual in ants. A few of the variations are illustrated in
Figures 9, 24, 25, 27, 28 and discussed in [8]. The genitalia are,
however, not very useful in distinguishing Amblyopone from
other genera of the tribe, because the intrageneric variation 1s
much greater than the intergenerie.

Males tend to be blackish in general color, or at least darker
than the corresponding workers, but some of the forms from
Old World desert areas are lighter, more brownish or even yel-
lowish in color. Probably many of the males are nocturnal, so
that more species will be found to be light in color. Male
sculpture often follows in a rough way that of the workers of
the same species, but tends to be much better developed and
coarser over the head and alitrunk, though often less regular and
more rugose or vermiculate.

164 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

In general size, the species of Amblyopone vary considerably,
the smallest workers (A. degenerata) being in the neighborhood
of 2 mm. outstretched length, while large workers and females
of A. australis are robust insects, reaching a centimeter and
more in length. Some of the species are very variable in size in
the worker caste, but polymorphism is not well developed even in
these. Allometric differences affect mostly things like propor-
tions (width) of head and petiole, and size of eyes.

The synonymy of Amblyopone and Stigmatomma was dis-
cussed in detail, and previous references to the subject men-
tioned, in Brown, 1949: 86; see also Brown and Nutting, 1950:
124. In the 1949 paper, Stagmatomma was reduced once again
to subgeneric status under Amblyopone. Xymmer and Fulakora
were synonymized with Stigmatomma. Stigmatomma had previ-
ously been separated on the basis of its ‘‘double-ranked’’ (vs.
‘“sinele-ranked’’ in Amblyopone) mandibular dentition ; its finer,
more opaque cephalic sculpture; and the supposed absence of
teeth on the anterior clypeal border of Amblyopone. But, as is
now fairly widely appreciated by myrmecologists, the double-
ranked condition grades through, especially in some Australian
species of the ferruginea group. Most of the ‘‘true’’ Amblyopone
as formerly separated have denticulate anterior clypeal borders,
although this denticulation is often very fine, and may oceasion-
ally be absent, as explained below. The sculptural distinction is
worthless, since some species with Stigmatomma characters other-
wise, e.g. A. normandi, A. elongata, have the dorsal surface of the
head predominantly smooth and shining. On the other hand,
some Amblyopone s. str. have densely sculptured heads (e.g.,
ferruginea group). There remains the question of the presence
or absence of Rsf2°3 (which determines whether there are one
or two cubital cells). In 1949, I maintained this as a provisional
separatory character at the subgeneric level, stating at the same
time my doubts as to its usefulness when more material became
available. Now that we have more material of the winged castes
of various species, it is clear that Rsf2:3 has been lost independ-
ently at least twice, and possibly more than three times, that is,
in the Australian ‘‘true’’? Amblyopone (where variable remnants
of Rsf2°3 are common), in a South American species (the
‘“Bricapelta egregia’’ of Kusnezoy), and in certain of the smaller

BROWN: ANT TRIBE AMBLYOPONINI 165

species (e.g., A. saundersi of New Zealand). A generic split
along these lines would certainly be discordant with divisions
that might possibly be drawn on the basis of other characters,
and as a matter of fact, it turns out that some species that would
be placed in Amblyopone on the basis of venational characters
have workers that would go into Stigmatomma. The Amblyo-
pone-Stigmatomma dichotomy certainly loses its former supposed
sharpness, and I see no reason to maintain a separate formal
subgenus for Stigmatomma any longer.

The subgenus Xymmer, based on the sole West African
species Stigmatomma (Xymmer) muticum Santschi, has a median
clypeal apron absolutely devoid of teeth along its anterior mar-
gin, but in 1949 I synonymized this subgenus with Stigmatomma
because of the similar condition reported by Forel for S. belli of
southeastern Asia. However, the types of S. belli have been ex-
amined again, and they have been found to possess denticulation
of the clypeal apron, although the denticles are small and rather
indistinct. Since muticum also has the petiole conspicuously nar-
rowed into a slender, short peduncle at its base, a case might
be made for the resurrection of Xymmer as a generice or sub-
eeneric name for this aberrant species. However, we now know
at least one more species, A. gingivalis sp. nov., that definitely
lacks teeth or denticles on the clypeal margin; gingivalis is
aberrant in the form of its mandibles, but the petiole is normal
for the genus. Several species, especially A. bruni, have elypeal
denticulation reduced very nearly to the vanishing point. I
have examined types of S. (X¥.) muticum, and in the light of
what we now know about variation among Amblyopone species,
I believe that this species should be included in the genus with-
out subgeneric distinction. (In practice, I have found formal
subgenera to be awkward and confusing, and I much prefer to
use informal species-groups that do not clutter the nomencla-
ture.) Actually, A. mercovichi and A. gingivalis are more aber-
rant in some respects than is A. mutica, to say nothing of A.
degenerata, which Borgmeier wisely placed in Amblyopone when
he first described it.

Clark has suggested (in litt., 1948) that A. reclinata, A.
rothneyt and allies with more or less large compound eyes should
be segregated in a separate genus, but here again there seems

166 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

to be much variation leading toward more ‘‘normal’’ kinds of
Amblyopone, of which perhaps A. pallipes or A. silvestrii would
represent an average species. Probably further investigation
will only strengthen the links between this group and the rest
of the Amblyopone species.

Lithomyrmex is another generic name proposed for a single
species, glawerti Clark from Western Australia. This species, as
Clark suggests, may indeed represent an intermediate form lead-
ing from Amblyopone to Onychomyrmex, but the females are
normal (winged as virgins), and the three castes cannot be
separated from the corresponding ones of Amblyopone by any
really important character. As an Amblyopone, glauerti cannot
even be considered as more than a moderately aberrant species.

The species of Amblyopone are found in many countries
widely scattered over the earth, ranging from tropical to cool
temperate in climate, and from wet to rather arid. The
Australian Region, with about two-fifths of the known species, is
the headquarters of the genus. Another fifth of the species is
found in tropical Asia, and still another fifth is distributed
through the Americas, from Canada to Chile. The remainder
of the species are scattered; several are in the Mediterranean
area, two are on the western fringe of Africa in Nigeria and
Senegal, one is in Japan, and one, possibly introduced from
Melanesia, is in Hawaii. The concentration of species in extra-
limital regions such as Chile, Argentina, Cuba, Japan, Aus-
trala, the Solomons and New Zealand speaks for the age of the
genus. Its absence or great rarity in the central Ethiopian
Region is parallel to that of other ancient groups, such as tribe
Eetatommini and the ‘‘Notomyrmex’’ group of Monomorium,
that are well developed in the peripheral areas listed above with
Amblyopone. Unlike the more open-foraging ectatommines, how-
ever, the amblyoponines are not represented in Tertiary fossil
faunas, indicating that even during those periods, Amblyopone
did not include prominently arboreal foragers, at least in the
Northern Hemisphere.

This leads us again to the obvious observation that Amblyo-
pone species are mostly small and hypogaeie specialists, although
it should be noted that some of the members of the tropical
Oriental reclinata group are fairly large in size, are heavily

BROWN: ANT TRIBE AMBLYOPONINI 167

pigmented, and have fairly large eyes in the worker caste, like
workers of some known epigaeie foragers in other tribes of ants.
However, we have no direct information on the habits of any
member of the reclinata group, and it seems unlikely that species
of this group are strongly epigaeic in foraging habits.

Most species of Amblyopone so far known have been found in
moist, forested areas, where they may nest in rotten wood, in
the leaf litter, or in the soil under stones or logs. Nevertheless,
several species are known to be tolerant of rather arid conditions
in treeless regions in Australia, North America, Argentina,
North Africa and the Middle East, where the subterranean
habits are strongly developed, and some species are known only
from males taken at light [1, 12].

So far as my reading and experience go, the Amblyopone
species feed nearly or quite exclusively on arthropods, dead or
alive at time of interception. A. pallipes, and probably other
species, specialize on chilopods [2, 16]. A. ausiralis, judging from
remains in the nests, collects various arthropods, including
beetles.

AMBLYOPONE Species

(T) aberrans Wheeler, 1927:26. W. Australia [16, Fig. 35]
amblyops Karawajew, 1935:57. N. comb. Indo-China [11]
(P) armigera Mayr, 1887:547. se. Brazil, n. Arg. [1]
(P) australis Erichson, 1842, Arch. Naturg., 8(1):261 [13, Figs. 1, 2, 7-9]
(P) = obscura Fr. Smith, 1858, Cat. Hym. Brit. Mus., 6:109.
(P) = cephalotes Fr. Smith, 1876, Trans. Ent. Soc. Lond., p. 490.
(T) = laevidens Emery, 1887, Ann. Mus. Civ. Stor. Nat. Genova, 25 447.
(T) = fortis Forel, 1910:1.
(P) =maculata Stitz, 1911, Sitzb. Ges. Naturf. Freunde, Berlin, p. 351.
(P) = nana Emery, 1914, Nova Caledonia, Zool., 1:394.
(T) = minor Forel, 1915:1.
(T) = foveolata Wheeler, 1927 :9.
(T) = pallens Wheeler, 1927:11.
(T) = queenslandica Wheeler, 1927:12.
(T) = norfolkensis Wheeler, 1927:15.
(T) = howensis Wheeler, 1927:15.
(T) bellii Forel, 1900:55. India [8]
bierigi (Santschi), 1930, Bull. Soc. R. Ent. Egypte, (n. s.) 14:17.
Cuba [7]
(P) bruni (Forel), 1912, Ent. Mitt., 1:45. Formosa [9. Fig. 21]

168 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

(T) = juergi (Forel), 1922, Rev. Suisse Zool., 30:87. N. syn.
(T) celata (Mann), 1919:279. Solomon Is. [18]
(P) chilensis Mayr, 1887:547. Chile [1]
(T) clarki Wheeler, 1927:24. sw. Australia [16, Figs. 33, 37, 38]
(T) degenerata Borgmeier, 1957:111. se. Brazil [7]
(P) denticulata (Roger), 1859:251. N. comb. s. Europe [12]
= gheorghiefi Forel, 1892, Verh. zool.-bot. Ges. Wien, 42:309. Syn. by
Emery, 1916:100.
(P) =gracilicornis (Menozzi), 1936, Boll. Lab. Zool. Portici, 29:268. N.
syn.
(P) elongata (Santschi), 1912:519. se. Brazil to n. Arg. [1]
= barretoi (Bruch), 1921, Rev. Mus. La Plata, 26:184. Syn. Borg-
meier, 1957.
= minor (Santschi), 1922, An. Soc. Cient. Arg., 94:241, nec Forel,
1915. Syn. Borgmeier, 1957.
(P) = paranensis (Santschi), 1924, Ann. Soc. Ent. Belg., 64:6. Syn. Borg-
meier, 1957.
(LT) egregia (Kusnezov), 1955:274. N. comb. n. Argentina [1]
(T) emeryi (Saunders), 1890, Ent. Mon. Mag., 26:203. N. comb. n. Africa
[12]
(P) exigua Clark, 1928:35. Australia: Victoria [17, Fig. 44]
feai (Emery), 1894, Ann. Mus. Civ. Stor. Nat. Genova, 34:454. Burma
[8]
(T) ferruginea Fr. Smith, 1858, Cat. Hym. Brit. Mus., 6:110. Australia:
vie. Melbourne [16, Fig. 36]
(T) = mandibularis Clark, 1928:33. Syn. Brown, 1952.
(T) gingivalis Brown, sp. nov. e. N.S. Wales [15, Figs. 30, 31]
(T) glauerti (Clark), 1928:31. N. comb. W. Australia
(T) gracilis Clark, 1934b:52. Australia: Victoria [16, Fig. 41]
(T) hackeri Wheeler, 1927:22. se. Queensland [16, Fig. 39]
(P) impressifrons (Emery), 1869, Ann. Acead. Aspir. Natural., Napoli,
Cy2e13e Ltaliyas [2]
(T) leai Wheeler, 1927:16. Lord Howe I. [13]
(T) longidens Forel, 1910:1. se. Australia [16, Fig. 34]
(T) lucida Clark, 1984a:27. Australia: Capital Terr. [17, Fig. 45]
(T) luzonica (Wheeler & Chapman), 1925:56. N. comb. Philippines [11]
(T) = williamsi (Wheeler & Chapman), 1925:56. N. syn.
(T) mercovichi Brown, sp. nov. se. Australia [14, Figs. 29, 32]
(P) michaelseni Forel, 1907, Fauna SW Australia, 1:264. sw., se. Australia
[13]
(T) minuta (Forel), 1913, Zool. Jahrb. Syst., 36:4. E. Indies [10]
(T) monrosi Brown, sp. nov. Chile [5]
(T) mutica (Santschi), 1914:311. N. comb. Nigeria

BROWN: ANT TRIBE AMBLYOPONINI 169

(T) mystriops Brown, sp. nov. Guatemala [4, Fig. 19]
(T) normandi (Santschi), 1915:54. N. comb. Tunisia [12]
(T) oregonensis (Wheeler), 1915:389. n. California to Brit. Columbia. N.
status [2]
(T) orizabana Brown, sp. nov. Mexico: Mt. Orizaba [6]
(P) pallipes (Haldeman), 1844, Proc. Acad. Nat. Sei. Philadelphia, 2:54.
s. Quebee and Iowa to Florida and Arizona [2]
[= serrata (Roger), Arotropus binodosus Provancher, arizonensis
(Wheeler), wheeleri (Santschi), montigena (Creighton). Synonymy
in Creighton, 1940; Brown, 1949:84.]
(T) = subterranea (Creighton), 1940:8. N. syn.
(P) punctulata Clark, 1934a:28. Tasmania [17]
(2?) quadrata (Karawajew), 1935:57. N. comb. Gulf of Siam [8]
(P) reclinata Mayr, 1878, Verh. zool.-bot. Ges. Wien, 28:667. Java [8]
(T) rothneyi Forel, 1900:56. India: Bengal [8]
santschii (Menozzi), 1922, Ann. Mus. Civ. Stor. Nat. Genova, 49:347.
N. comb. Senegal [12]
(T) saundersi Forel, 1892, Mitt. schweiz. ent. Ges., 8:336. N. Zealand [17]
(T) silvestrit (Wheeler), 1928, Boll. Lab. Zool. Portici, 21:97. Japan [11]
(T) smithi Brown, sp. nov. S. Australia: Lofty Range [17, Fig. 40]
(T) trigonignatha Brown, 1949:81. N. Carolina [3% Fig. 18]
wilsoni Clark, 1928:34. se. Australia [17]
zwaluwenburgi (Williams), 1946, Proce. Hawaii. Ent. Soc., 12:639. N.
comb. Hawaii: Oahu [19]

Mystrium Roger
= Mystrium Roger, 1862:245. Type: Mystrium mysticum Roger, 1862,
monobasic (+7 spp.).
= Mystrium, Wheeler, 1922:758, 1006, synonymic catalog of African and
Malagasy spp.
= Mystrium, Menozzi, 1929:518-536, revision of the genus.

This most aberrant genus is close to Amblyopone. The work-
ers and females differ from those of Amblyopone in their shorter,
thicker bodies and broader heads, as well as their odd, blunt-
tipped mandibles (Fig. 4). The sculpture is rather coarse and
consists of bold reticulation or costulation over head and much
of alitrunk. Within the reticular basins and elsewhere are situ-
ated the peculiar body hairs, which are generally squamose-
pointed, clavate, or otherwise broadened or bizarre. Menozzi
notes that the worker-female palpal formula is 4, 3, a count
that I have confirmed for an M. rogeri worker in the Museum of
Comparative Zoology.

170 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Male like that of Amblyopone, coarsely sculptured and with
notauli distinct. Wing venation of both sexes of the ‘‘complete’’
type (Fig. 5).

Genitalia not studied. Pygostyles apparently lacking.

So far as known, Mystrium is limited in distribution to Mada-
gvasear and adjacent islands (six nominal species), Cameroons
(one species), and the Indo-Australian area, where one species
M. camillae, has been found in Burma, the East Indies, Luzon,
New Guinea and the Darwin district of northern Australia. It
is assumed that Mystriwm is predaceous, but there is no direct
evidence known to me on its feeding habits. Almost equally
scanty 1s information on colony size, nest site and structure, and
most aspects of ecology and behavior.

Menozzi’s revision (1929) includes a key to the species, plus
illustrations and descriptions. It is difficult to judge the status
of the species from Madagascar, since most of these are known
from inadequate material. Since Menozzi gives full references
to all species, only the dates of description are given in the
list below.

Mystrium Species

(P) camillae Emery, 1889. Burma to n. Australia. [Fig. 4]
(P) = javana Karawajew, 1925. N. syn.
fallax Forel, 1897. Madagascar: Nossi Be.
(P) mysticum Roger, 1862. Madagascar, Comoro Is.
(T) oberthueri Forel, 1897. Madagascar.
(T) rogeri Forel, 1899. Madagasear.
silvestrii Santschi, 1914. Cameroons.
stadelmanni Forel, 1895. Madagascar.
(T) voeltzkowi Forel, 1897. Madagascar: Nossi Be.

Myopopone Roger

= Myopopone Roger, 1861:49. Type: Myopopone castanea var. maculata =

Myopopone maculata Roger = Amblyopone castaneus Fr. Smith, 1860, by

designation of Bingham, 1903.

In general habitus, to the naked eye, Myopopone workers look
very much like those of Amblyopone australis, and they show
variation of similar scope. In fact, Myopopone is probably
monotypic [20], and may be regarded as a more than usually

BROWN : ANT TRIBE AMBLYOPONINI yak

aberrant species of Amblyopone. The characters are, however,
sufficiently marked and numerous in the worker-female castes
to justify the retention of Myopopone as a genus apart from
Amblyopone. The head shape (Fig. 10) is like that of Amblyo-
pone, without the ‘‘amblyoponine teeth’’ at the corners anter-
iorly, but the lobes of the frontal carinae are large and placed
well forward, so as to overreach slightly the concave median
lobe or apron of the elypeus (in Amblyopone the lobes never
reach the anterior border of the median clypeal lobe). The
antennal funiculi are strikingly broadened and flattened, differ-
ing in this from Amblyopone. The legs are short and with spini-
form processes or spine-like setae developed at several points; in
particular, the extensor surface of the middle tibia is provided
with a number of sharp peg-like spines, also the metatarsus of
the posterior leg. The mandibles, as can be seen from Figure 10,
are different from those of any given species of Amblyopone,
but are not strikingly outside the range of variation seen among
Amblyopone species.

The female is winged and is markedly larger and darker
than the worker; there are also differences in sculptural detail.
The differences between these two castes have been responsible
for much of the synonymy at species level. The male is decidedly
smaller and more slender than the female, and is typically
amblyoponine in its habitus and general characters, with rugu-
lose head and alitrunk and piceous to blackish in color. Notauli
present.

Pygidium and subgenital plate both subtriangular, with
broadly rounded apices. Parameres rather long, tapered and in-
eurved so that their apices meet or even slightly overlap at the
half-retracted position. Volsellae (Fig. 12) much like those of
Amblyopone australis (Fig. 9); digitus with a flat, plate-like
apical portion, which is convex and tuberculate over its lateral
surface; cusp reduced to a vestigial swelling at the base of the
digitus, but the heel well developed and bearing a sharp erect
tooth. Aedeagus of a peculiar shape, as shown in Figure 11.
In all castes, the middle and posterior tibiae have two pairs of
spurs, one of the posterior pair being larger than its mate,
slender, curved, narrowly barbulate and with a peculiar ob-
liquely truncate tip. Antennae stout, but not flattened as in the
worker.

172 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Myopopone castanea. Figure 10. Head of worker, full-face view (from
type series of synonymous MM. rossi Donisthorpe, Finschafen, New Guinea).
Figure 11, aedeagal valve of male from Pematang Siantar, Sumatra. Fig-
ure 12. Volsella of same specimen, ventral view. Figure 13. Volsella in
place in left paramere, viewed from midline, aedeagal valve omitted; same
specimen. Figure 14, sub-genital plate of same specimen.

BROWN : ANT TRIBE AMBLYOPONINI NF

Workers and females with 4 maxillary and 3 labial palpal
segments. Wings in both sexes similar to those of ‘‘complete-
veined’’ Amblyopone species, narrow, glassy, with dark veins;
Mf2 usually completely or nearly completely contracted in fore-
wing.

I consider that the present evidence [20] indicates the proba-
bility that all of the Myopopone specimens so far collected belong
to one species, M. castanea (Fr. Smith), which ranges from
Sikkim, Ceylon, and the Nicobars in the west to the Philippines,
New Guinea through the Solomons, and central Cape York
Peninsula of northern Australia in the east. In Melanesia, M.
castanea is predominantly a lowland species, occurring mostly
at or below 500 m.; apparently it replaces Amblyopone australis
in this zone. A. australis in Melanesia is predominantly a mid-
mountain rain forest species, living chiefly at altitudes of 1000 m.
or more. In southeastern Asia and the Philippines, where no
similar Amblyopone species is known to exist, M. castanea exists
at altitudes up to more than 2000 m.

M. castanea, like rain forest populations of A. australis found
in the tropics, inhabit rotting logs, where they feed on large
beetle larvae and probably on other comparatively helpless
arthropods. Wilson’s observations indicate that, as in the case
of at least some Amblyopone, the colonies are very loosely organ-
ized, and that the workers may bring their brood to the site
where the prey has been found, stung and killed, rather than
attempt removal of the largest kinds of prey to a central nest or
brood area. As has been suggested, such behavior may grade
into primitive nomadism (Wilson, 1958 a, b).

PRIONOPELTA Mayr

= Prionopelta Mayr, 1866:503. Type: Prionopelta punctulata Mayr, 1866,
monobasic (+9 spp.).

> Renea Donisthorpe [23], see under P. majuscula,

> Examblyopone Donisthorpe [23], see under P. majuscula.

When Mayr originally described Prionopelta, he demonstrated
clearly the amblyoponine affinities of the sole species then in-
cluded (P. punctulata), rendering inexplicable Emery’s later
placement of the genus in tribe Ectatommini. Wheeler (1922)
followed Emery in this placement, making it “‘generally ac-
cepted.’’ But when the relationships of the known Prionopelta

174 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

species are seriously studied now, there is simply no question
about Mayr’s opinion that Prionopelta is close to Amblyopone.
In fact, the characteristic narrow, short, 3-toothed mandible is
the only thing really consistently marking Prionopelta (Fig. 15)

Figure 15. Prionopelta brocha, full-face view of holotype worker.

off from Amblyopone of the Fulakora group of species, as repre-
sented by A. smithi (Fig. 40). Exeept in P. brocha, which is very
Amblyopone-like in everything but mandibular structure (Fig.
15), the Prionopelta species run to more nearly parallel-sided
heads, usually broadest near the middle; that is, not broadest
anteriorly near the anterior corners. The genal tooth at the
corner on each side is, again with the exception of P. brocha,
absent or present only as a minute vestige.

BROWN: ANT TRIBE AMBLYOPONINI iS

The anterior clypeal border is normally convex or moderately
projecting and armed with a row of inconspicuous denticles cor-
responding to the clypeal armament of Amblyopone. The eyes
are small to minute in the worker, moderate-sized in the female,
and are normally placed behind the middle of the sides of the
head in both worker and female. The antennae are 12-segmented
in worker and female, or more rarely 11-segmented; apical 3-4
segments incrassate, forming a more or less distinct club. Male
antennae 13-segmented, filiform, with feebly thickened apex;
scape very short. Maxillary and labial palpi both 2-segmented in

Prionopelta punctulata male from Tucuman, Argentina. Figure 16. Left
aedeagal valve, volsella and paramere, partly detached as flattened on slide.
Figure 17. Wings.

all castes. Male mandibles like those of Amblyopone, 2-toothed.
Notauli present in male. Genitalia of male not unusual in shape
(Fig. 16); volsella without second cusp. Wings as shown in
Figure 17, though the veins are more distinct in some species.
The female usually has slightly heavier and more complete vena-
tion than in the male of a given species. Variation extends to
Mf4, which may be completely lacking (e.g., P. kraepelini
female), and to CuA in the hindwing, present in the females of
some species as a distinct spur (e.g. P. majuscula female).

The Prionopelta species are all small to minute in size, and
mostly have depigmented (testaceous to reddish-brown) work-
ers, though in some species (e.g., P. majuscula), the female is
piceous. Males are black or piceous; the three castes are usually

176 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

similar in size (length), but in majuscula, the female is distinetly
larger. The sculpture consists of close punctulation, varying in
coarseness with the species, but most distinct on the dorsum of
the head, and weakening posteriorly toward the gaster, the last
usually being wholly or in part smooth and shining. Pilosity
simple, short, reclinate to erect, usually rather abundant.

The characteristics of the worker Prionopelta are mostly
adaptations indicating a eryptobiotic mode of life, and the direct
information available on the ecology of several species all goes
to show that these are inhabitants of the soil and soil cover. Most
of the species are at home in moist tropical forest, but some of
them can survive in semiarid areas such as northwestern Argen-
tina (Kusnezov, 1955: 276) by keeping to a strictly hypogaeic
mode of existence in the moister microhabitats. Kusnezov (loc.
cit.) has questioned the inelusion of the Old World Prionopelta
species In one genus with those from the New World, but I see no
reason to question the judgment of older authorities in this mat-
ter. Leaving the somewhat primitive P. brocha to one side, the
present species of Prionopelta form an exceptionally homo-
geneous group so far as we can tell from the known examples of
all castes and from the known biology. P. brocha, while sharing
a number of characters with Amblyopone, is nevertheless clearly
a Prionopelta in the form of the mandibles. Since the mandibu-
lar differences are complex ones that have probably accompanied
an adaptive shift (in feeding habits?) of some significance, I see
no need to exclude P. brocha from Prionopelta.

The characters of Prionopelta include many that are reduc-
tions from average or primitive characters of Amblyopone, so
that it seems lkely that the latter is the parent genus. In the
tropics of both hemispheres, Prionopelta appears to be dis-
tributed geographically and ecologically roughly so as to replace
the smaller Amblyopone of the ‘‘ Fulakora’”’ group. As examples
of this may be cited New Guinea (with two widespread Priono-
pelta species) and the Solomons (with Amblyopone celata) plus
southeastern Australia (with several of the small ‘‘Fulakora’’
species). A somewhat similar situation apparently holds in the
American tropics, where in some areas Prionopelta is very com-
mon [21], but the smaller, depigmented Amblyopone species are
either unknown or tend to be restricted to geographically or
ecologically peripheral ranges.

BROWN : ANT TRIBE AMBLYOPONINI nar

Prionopelta, on the other hand, does not extend significantly
into the temperate or cool montane regions (except perhaps in
Australia), and is represented on the Asian-African mainlands
only by a single collection from Zululand. So far as is known,
of course, amblyoponines are in general very poorly represented
in Africa below the Sahara, and no species have yet been found
in the Congo rain forests. Prionopelta ranges in the New World
from southern Mexico south into northwestern Argentina, in the
West Indies extending only into the Lesser Antilles, where P.
antillana may be introduced. In Cuba, a small species of
Amblyopone that has been collected only once may take the place
of the absent Prionopelta [7]. Two species of Prionopelta are
widespread and fairly abundant in New Guinea, but these tend
to have different ranges within the area, or at least different
modes of abundance. One of the species, P. opaca, has a close
counterpart in the East Indies (Java) and the Philippines (P.
kraepelini), and these two species may intergrade in the Micro-
nesian islands, where they have probably been introduced by
man [23]. The curious species P. brocha is known only from
the type collection on New Caledonia. Since New Caledonia
appears to have received by far the greatest bulk of its ant fauna
by way of eastern Australia, the recent discovery of a species
in Queensland [23] is not surprising. In this connection, the
little-known Dorylozelus must be re-examined (see below).

PRIONOPELTA Species

aethiopica Arnold, 1949, Occ. Pap. Nat. Mus. S. Rhodesia, 2(15) :263,
fig. 4, 4a, worker. N. status. Zululand. [23]

(T) amabilis Borgmeier, 1949, Rev. Brasil. Biol., 9:203, figs. 3-5, worker.
Costa Rica. [21]

(T) antillana Forel, 1909:239. N. status. Lesser Antilles, n. S. America,
ete. [21]

(T) brocha Wilson, 1958a:147. New Caledonia. [23, Fig. 15]

descarpentriesi Santschi, 1924, Rev. Zool. Afr., 12:195, worker. Mada-
gasear. [23]

(P) kraepelini Forel, 1905, Mitt. Naturh. Mus., Hamburg, 22:3, female,
worker. E. Indies, Philippines, Micronesia. [23]

(P) majuscula Emery, 1897:595. N. Guinea & nearby islands. (= Ponera
simillima Fr. Smith, Prionopelta poultoni Donisthorpe, Reneca testacea
Donisthorpe, Examblyopone churchilli Donisthorpe, synonymized by
Brown, 1953b:12; see also Wilson, 1958a+148.) [23]

178 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

(P) modesta Forel, 1909:241. C. America, s. Mexico. [21, 22]

(T) opaca Emery, 1897:596. N. Guinea, Micronesia, Australia. [21, 23]

(T) = mocsaryi Forel, 1907, Ann. Mus. Nat. Hungar., 5:1, worker. Synony-
mized by Brown in Wilson, 1958a:149.

(P) punctulata Mayr, 1866:505. s. Brasil, n. Argentina. [21]

(P) = mayri Forel, 1909:239. N. syn.

(P) = bruchi Santschi, 1923, Rev. Suisse Zool., 30:245, ‘‘female’’ (recte
worker?). N. syn.

ONYCHOMYRMEX Emery

Onychomyrmex Emery, 1895:349. Type: Onychomyrmex hedleyi Emery,

1895, monobasic.

Onychomyrmesx, Forel, 1915:2, characters.
Onychomyrmexz, Wheeler, 1916; revision of genus, ecology, ethology, larva.

This genus evidently represents a development of the army-
ant or legionary life-form that has arisen independently from an
Australian Amblyopone stock. The worker of Onychomyrmez
(Figs. 46, 48) is accordingly more slender and has longer ap-
pendages than is usual for Amblyopone, and the integument is
smooth and shining, the mandibles more down-curved and hook-
hke, and the tarsal claws much enlarged, particularly those of
the last two pairs of legs. The spurs of the posterior pair of
tibiae are reduced to minute, straight, almost setiform vestiges,
or else are lacking altogether. In O. hedleyi, the spur vestiges
may be present or absent in different workers from the same
nest. To complete the resemblance to certain true army ant
genera (subfamily Dorylinae) the Onychomyrmez female (known
for two of the three species) is dichthadiiform, that is, it has the
head peculiarly broadened, very small eyes, workerlike alitrunk
without wings or corresponding sclerites, and elongate, bulky
gaster.

Despite these characters, which misled Emery and most other
authors to consider the genus as an aberrant and independent
tribal group within subfamily Ponerinae, the amblyoponine
affinities are so clear that one wonders why it was so long before
they were properly interpreted. The mandibles, clypeus, posi-
tion of the eyes on the posterior half of the sides of the head,
the long sting, and the basie form and structural relationships of
alitrunk, petiole, postpetiole and remainder of gaster in the
worker are all unmistakably amblyoponine. The larva also shows

BROWN : ANT TRIBE AMBLYOPONINI 179

no features that seem to contradict an amblyoponine affinity
(Wheeler, 1916; G. C. and J. Wheeler, 1952: 637), although
the Wheelers consider it to show specialization in the direction
of the Cerapachyinae larva.

Males from southeastern Queensland are believed to belong
to O. mjoebergi; these are similar in size to the worker, and
have the general characters of some of the slender small Amblyo-
pone species. The wing venation is shown in Figure 47; the
genitalia are peculiar in having short, broad parameres, which
are bent sharply inward, then caudad, the apical portion form-
ing a curved vertical plate on each side with a cultrate trans-
parent margin. The subgenital plate is narrowly tongue-shaped
and projecting at its apex. The sculpture is predominantly
smooth and shining. Mandibles with 2 apical teeth; palpi seg-
mented 2,2.

John Clark (1928) was the first to relate Onychomyrmez to
Amblyopone unequivocally when his relict Lithomyrmex (now
Amblyopone) glauertr, from Western Australia, turned up and
was recognized as intermediate between Amblyopone and
Onychomyrmex in many worker characters. A. glauerti, however,
has a normal winged female, and in other ways follows Amblyo-
pone so closely that I was unable to separate it from that genus.
On the other hand, the intermediacy of A. glauerti confirms be-
yond reasonable doubt the amblyoponine affinities of Onycho-
myrmex. Emery’s position in isolating Onychomyrmesx can be
explained, I think, by his belief in the taxonomie significance of
the tibial spurs. At present, we know that these spurs are not
always reliable classificatory guides, particularly among the
Ponerinae. Clark’s paper appears to have been seen by few
myrmecologists, and A. glawerti by still fewer, so Emery’s and
Wheeler’s classifications have in the main been followed until
very recent years.

As now known, Onychomyrmex has three species, all occurring
in eastern Queensland in rain forest, chiefly in rotten logs. Prob-
ably the genus extends into northeastern New South Wales. It is
much rarer in the southern than in the northern part of its range.

Wheeler (1916) relates his finding workers of O. mjoebergi
‘biting and stinging to death a huge lamellicorn beetle larva
more than two inches in length, which they had just found in a

180 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY
cavity’’ in a rotten log. It is not clear from Wheeler’s account
whether the beetle larva was attacked after he broke the log
open, or before. Furthermore, it would seem that a two-inch
lamellicorn (probably of family Passalidae, common in rotten
logs in North Queensland rain forests) would be an impossible
burden for even several hundred of the very small Onychomyr-
mex to move through the narrow passages of the usual rotten log.
Until the habits of Onychomyrmesx can be checked in detail, it
is interesting to speculate that the colonies of this genus may
migrate from one large victim to the next, after subduing the
prey by mass stinging attacks.

Wheeler found small companies of workers moving through
logs in the manner of army ants mass-foraging, and I myself
have found aggregations answering this description well in the
species O. hedleyi and O. doddi. One party of O. hedleyi found
in a log at Malanda on the Atherton Tableland numbered several
hundred workers, without queen or brood.

The behavior of these workers reminded me very much of
the actions of foraging groups of Leptogenys and Aenictus that
I have seen in Australia, Western China and Assam. A colony
of O. mjoebergi found by Wheeler at Kuranda ‘‘comprised at
least 400 workers, a single queen, with the abdomen greatly dis-
tended with eggs, and a large number of nearly mature larvae
but no pupae.’’ These circumstances suggest that Onychomyr-
mex may resemble the New World army ants studied by
Schneirla in having brood-rearing and nomadism synchronized.

Key to the Species of Onychomyrmex — workers

1. Mandibles predominantly smooth and shining, with a few coarse punec-
tures; robust species, with broad head (CI usually > 78) ; color yellowish-

Peds bOered dash= [1 O\walweew ee sn eee Ba OTOEVERG
Mandibles densely striate above; more slender species, with narrower
head (CI usually < 78) ..... Ne ae Ee ee ae eee, oO

blackvoridarks pie Ouse se ee eee ee Re hg a hedleyn
Smaller species; length of alitrunk (WL) < 1.2 mm.; full adult color
déepareddish-browmnl Veo 2.5. 1. LSet eee ters, © re doddi

BROWN: ANT TRIBE AMBLYOPONINI 181

ONYCHOMYRMEX Species

(T) doddi Wheeler, 1916:53. n. Queensland [24]
(T) hedleyi Emery, 1895:350. e. Queensland [24, Figs. 46, 48]
(T) mjoebergi Forel, 1915:2. e. Queensland [24, Fig. 47]

INCERTAE SEDIS
PARAPRIONOPELTA Kusnezov
= Paraprionopelta Kusnezoy, 1955:270. Type: Paraprionopelta minima
KKusnezov, 1955:271, figs. 1, 2, 5a, male; monobasic.

This monotypic genus was based on males taken separately,
presumably at light, at Tucuman, Argentina. These males are
minute (under 2 mm. TL), dark in color, and have a somewhat
Amblyopone-like petiole. They differ from the known males
of Amblyoponini in the oblong shape of the head, in the shape
of the mandibles, and especially in having 10-segmented anten-
nae. There are no distinct teeth on the anterior eclypeal margin,
but the hind tibiae do bear broadly pectinate spurs. Possibly this
genus really is an amblyoponine, but since a similar type of
petiole occasionally appears convergently in other groups of
ants, I do not consider this a certainty. If it is in the Amblyopo-
nini, then one is tempted to match it with the worker of
Amblyopone degenerata from southern Brazil; the two are
similar in size, and both have reduced antennal segmentation.

DoryLozeuus Forel
= Dorylozelus Forel, 1915:24. Type: Dorylozelus mjoebergi Forel, 1915:25,
fig. 4, worker; monobasic.

This genus is known only from the type (or types?) of D.
mjoebergi, taken by Mjoberg in the Blackall Range in southern
Queensland. Despite all efforts to locate the type material in the
Forel Collection at Geneva, in the Naturhistoriska Riksmuseet
in Stockholm, and in various Australian collections, it could
not be found.

Failing in the effort to find the type, I spent a week searching
the rain forests in and around the Blackall Range for more speci-
ments, but found none. T. Greaves and the Darlingtons have
also looked for Dorylozelus in the Blackalls without success. It
is important that this ant be rediscovered, and the sexes and
larvae examined, because its systematic position at the moment

182 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

is anything but clear. Wheeler placed it in a separate tribe
Dorylozelini, but probably he never saw a specimen of D.
myjoeberg.

As characterized by Forel, this species is a mixture of charac-
ters like those of Dorylus and Ponera. The petiole and gaster are
supposed to be like those of tribe Ponerini. The frontal lobes
are contiguous, as in Ponerini, but also as in some amblyopo-
nines. The mandibles, as drawn by Inez Forel, are like those of
Prionopelta in general pattern, narrow and straplike, with three
apical teeth, of which the middle tooth is smallest. The antennal
funiculi are incrassate, with only 10 segments. Inez Forel’s
drawing is highly diagrammatic and difficult to interpret. One
even wonders whether there is any possibility that the specimen
is a compound one, so disharmonious is the image created by the
original characterization. Dorylozelus remains one of the most
puzzling anomalies among the ants. Only with more material
will we be able to place it more satisfactorily.

APPENDIX

The pages of this appendix are reserved for notes and desecrip-
tions dealing chiefly with species-level taxonomy and _ biology.
In the descriptions, the abbreviations for measurements and
indices are as follows: TL, total outstretched length of body,
including mandibles (sum of all tagmata) ; HL, maximum meas-
urable length of head, including eclypeus, but not mandibles;
HW, maximum measurable width of head, ignoring eyes if
present (head measurements are made from dorsal full-face
view) ; Cl, or cephalic index, is HW/HL X= 100. Wii wor
Weber’s length, is the diagonal length of the alitrunk as meas-
ured from side view. L, of course, generally stands for length,
and W for width (in mm.).

Places where types or other specimens are deposited are indi-
cated within brackets where convenient, especially the abbrevia-
tion MCZ, which stands for the Museum of Comparative Zoology
at Harvard College, Cambridge, Massachusetts.

[1] Borgmeier (1957: 108-112) has discussed A. armigera and
A. elongata and brought their deseriptions up to date. His
synonymy for elongata is followed here. He found the worker
of elongata to have the palpi segmented 3, 2.

BROWN: ANT TRIBE AMBLYOPONINI 183

Kusnezov’s Ericapelta egregia, described from isolated males,
is about the right size to be the male of A. armigera, and I expect
that association of the sexes will eventually prove this synonymy.

A. chilensis has been collected by Ross and Michelbacher in
southern Chile on the northern shore of Lake Lilanquihue and
in valley forest 18 km. west of Purranque, both localities to the
south of the type locality, Valdivia.

[2] Amblyopone pallipes is a rather common ant in forested
areas of the eastern United States and in the St. Lawrence
Valley near Montreal. Brown (1949:84) has shown that the
subspecies montigena Creighton is based on individual nest vari-
ants occurring sporadically in the eastern United States. In the
1949 paper, subterranea Creighton, based on samples from deep
subterranean collections made in the plains states, was raised
to species rank on the basis of what were thought to be diagnostic
sculptural characters. Now, however, after the examination of
many more collections from all over the eastern United States
and the Mississippi Valley, and considering samples from Ari-
zona (near base of Huachuca Mts., R. G. Wesson leg.) and
Iowa (W. F. Buren leg.), it appears that the characters thought
in 1949 to be diagnostic of a distinet species are in fact only
average differences, connected by clinal variation to the eastern
characters. Actually, the differences are slight; the underlying
pigmentation affects the appearance of the fine sculpture. Sam-
ples from the Plains are frequently lighter in color than those
from the eastern forests, but even some samples from the eastern
coastal areas are medium brownish-ferruginous in color.

Although it may be that the distribution of Amblyopone is
continuous between Arizona and the Plains and the Pacifie Coast,
we have no records between Arizona and northern California,
where the population of oregonensis commences and runs north-
ward through the moist coniferous forest belt. Although ore-
gonensis has been considered to be a subspecies of pallipes, the
differences between the two remain constant so far as known,
and I shall call oregonensis a species until we have more infor-
mation on possible range contacts between it and pallipes.

On several occasions in eastern Massachusetts, A. pallipes has
been observed under stones in hardwood forest together with

184 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

dead centipedes, usually lithobiids or geophilomorphs. The
association in each case left no doubt that the ants were using
chilopods for food; usually larvae were attached to the centi-
pedes and appeared to be feeding actively. In another ease,
W. L. Nutting found A. pallopes workers attacking an asilid
larva in leaf litter. Most observers agree, however, that chilopods
are the main source of food for this species. In my experience,
the centipedes used for food showed considerable variability in
size and form, and lithobiids many times the size of the ant are
found with the ants as often as smaller prey. It is assumed that

Figure 18. Amblyopone trigonignatha, holotype worker, dorsal view of
left mandible and anterior Glypeal apron.

the long, sharp sting of A. pallipes, which can penetrate even
the human skin in some places, is used to subdue living prey, but
we do not know all of the details of hunting behavior. A. pallipes
workers in the artificial nest are very timid, and can scarcely be
stimulated to attack centipedes of any kind enclosed with them.

When the ants are found with dead centipedes in the field, it
usually is in a situation under a stone such that transport of
the prey by the ants seems unlikely. I believe that the larvae
are often brought to the spot where the larger sort of prey dies,

BROWN : ANT TRIBE AMBLYOPONINI 185

and that they feed there for considerable lengths of time, rather
than the reverse situation usual for ants, where the food is
brought by the ants into more or less central brood chambers.

[3] A. trigonignatha was described in 1949 from a single worker
taken in a leaf mold berlesate from Concord, North Carolina.
It is very distinctive in the form of the clypeus and mandibles
(Fig. 18). More than 200 separate collections of Amblyopone
from all over the eastern half of the United States have been
examined since 1949, but all these proved to be the common A.
pallipes. A. trigonignatha still remains known only from the
holotype [MCZ].

[4] AMBLYOPONE MYSTRIOPS Sp. nov.
(Fig. 19)

Holotype female (alate) : TL 7.2 (gaster expanded), HL 1.47,
HW 1.88 (CI 94), WL 2.21, petiole L 0.66, petiole W 0.76, scape
L (without basal neck) 0.91, straightline outside L left mandible
1.50, forewing L 4.3 mm.

Habitus and details of head, mandibles and scape as in Figure
19. Frontal lobes distinctly separated by an extension of the
elypeus and the pit-like ‘‘frontal area.’’ Genal teeth reduced
to low rounded eminences. Clypeal apron short, convex, with 8
slender oblique teeth. Mandibles as shown in the figure; broken
line on each shaft represents a rather indefinite dorsal carina or
costa; between this and the tooth rows, the surface is concave.
Viewed from the side edge-on, the mandibles are seen to be
markedly flattened in an oblique ventrolateral direction, and the
apical quarter is incrassate. Note the low, inconspicuous basal
lamella, the small size of the apical tooth, and the convex ridge
or swelling formed ventral to the subapical teeth, near the apex;
the principal (median) teeth on the blade are in two separate
ranks.

Antennal scapes curved and arched, slightly narrowed toward
their midlength. Funiculus slightly but gradually thickened
toward apex; segments I through VI longer than broad, I
(pedicel) about 11% times as long as II through X, which are
approximately equal among themselves in length; VIT and Vail
about as long as broad; IX and X slightly broader than long;

186 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Figure 19. Amblyopone mystriops, holotype female, full-face view of head,
including mandibles and left antennal seape.

XI (apical) twice as long as X. Under-mouthparts retracted,
but 3 segments of the maxillary palpi and a single segment of
each labial palpus are exposed; probably there are more basal
segments hidden beneath the folded labrum.

BROWN : ANT TRIBE AMBLYOPONINI 187

Alitrunk very similar to that of large females of A. pallipes.
Petiole also similar, but a little broader, with a nearly vertical
anterior face, slightly concave in profile, rounding sharply into
the gently convex dorsal face. Seen from above, the front and
side outlines of the node are convex, although the front is
shghtly indented in the middle. Ventral process as in A. pallipes.
The (normally exposed portion of the) postpetiole is a little
longer than the petiolar node, and distinctly wider; abdominal
segment IV is slightly longer and wider than III (postpetiole).
Apex of gaster slightly laterally compressed; sting slender,
curved.

Seulpture of head much like that of A. pallipes, reticulate-
punctate and predominantly opaque; on anterior half of head,
some of the interpunctural ridges form fine, parallel costulae on
the lower, inner genae and on the frontal carinae, but these less
distinet and extensive than those of pallipes. Cervical face of
occiput shining, punctate. Clypeus longitudinally — striate;
‘‘frontal area’’ a round-bottomed pit, smooth and shining, as is
also a small impression in front of the anterior ocellus. Mandi-
bles coarsely oblique-longitudinally striate, moderately shining.
Antennae, tibiae and tarsi densely punctulate, subopaque. Ali-
trunk densely punctulate, shining (punctures not contiguous),
the punctures rare on the pleura; lower half of lateral propodeal
surfaces coarsely longitudinally striate; declivity smooth. Peti-
ole, gaster and femora smooth, shining, with spaced punctures,
those on the postpetiolar dise trailing shallow sulci.

Pilosity consisting mainly of soft, reclinate to oblique, pu-
bescence-like hairs, densest on head, legs and gaster, but not
hiding sculpture. Longer fine hairs on mandibles (especially
inner surfaces), front of head, antennae (including scapes), legs
and apex of gaster. Head and body dark reddish brown. to
piceous; mandibles, antennae and legs sordid yellow.

Wings hyaline, with yellowish veins and dark brown ptero-
stigma in forewing. Venation of forewing of the ‘‘complete’’
type, without first radial crossvein. Rs joining wing margin
(‘‘radial cell closed’’), Mf4 short, Mf2 present but short.

Holotype [MCZ] a unique female found among unidentified
miscellany; labelled: ‘‘Guatemala:/Los Amates/Kellerman.”’
Presumably this refers to Los Amates in the valley of the Rio
Motagua, near the Honduras border.

188 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

In the form of its mandibles and their dentition, this species
differs widely from all other Amblyopone species, and makes
an approach to the condition in Mystrium (compare Figs. 4 and
19). The pilosity and other characters are, however, typical for
Amblyopone rather than Mystriwm. Whether this aberrant
species represents a phyletically transitional form, or merely
a New World convergence toward the Mystrium type, I cannot
say. It would be interesting to know more about the habits
and ecology of this odd species.

[5] AMBLYOPONE MONROSI sp. Nov.
(Fig. 20)

Holotype worker: TL 4.3, HL 0.92, HW 0.77 (CI 84), WL
1.19, petiolar node lL as seen from above 0.47, petiole W 0.43,
scape L without basal neck 0.51, exposed straight-line outside
L of left mandible 0.54 mm.

20 21

Lvmrmnannnn |

Amblyopone species, dorsal views of anterior border of clypeal apron.
Figure 20. A. monrosi, paratype worker. Figure 21. A. bruni, holotype of
synonymous A. bruni juergi.

A rather typical-appearing smallish Amblyopone of the
‘‘Fulakora group,’’ with sides of head gently convex, posterior
border weakly concave, greatest width of head at about the
anterior third. Genal teeth acute, but small and almost hidden
in pilosity. Compound eyes represented by a small patch of
unpigmented, indistinct facets in the usual position. Frontal
lobes small, contiguous, the line of separation deeply impressed.
Clypeal apron with corner teeth unusually large and broad,
blunt at apices and projecting over the masticatory borders of
the closed mandibles (Fig. 20). Between the corner teeth are
6 smaller, obliquely truncate teeth socketed on low tubercles,
the two median teeth about even with the apices of the corner
teeth. (One or both of the outer small teeth may be fused to the
corner tooth adjacent in different specimens. )

BROWN: ANT TRIBE AMBLYOPONINI 189

Mandibles basically of the typical Amblyopone pattern, but
shorter and broader than usual, especially the thick, blunt-
tipped apical tooth. Inner margin convex. Basal quarter or so
of masticatory border occupied by a round-edged translucent
lamella, about as high as the acute double teeth. The right
mandible (crossed above the left in both holotype and paratype)
has a small single tooth following the lamella, and after this, at
the midlength of the mandible, a large, slightly retrorse double
tooth; after this three large, apparently single teeth before the
apical tooth. The left mandible, after the basal lamella, has
three moderate-sized double teeth, followed by two subapical
single teeth. (The paratype has similar dentition. )

Antennal scapes short, thick, only very slightly curved and
feebly inerassate toward apex. Funiculus 11-segmented,
strongly incrassate toward apex, but thickened gradually from
segment II on to apex. All segments except the pedicel (1)
and the apical (XI) appear to be broader than long, though VI,
VII, VIII and IX are only slightly so. Apical segment almost
as long as the preceding 3 segments taken together.

Alitrunk slender, straight in dorsal profile, rounded downward
gently at the front of the pronotum and, posteriorly, into the
plane, sloping declivity. Mesonotum almost twice as broad as
long; promesonotal suture impressed; metanotal groove present
but not conspicuous. Seen from above, the mesonotal area is
gently constricted. Petiole seen from above subcireular, truncate
behind. Seen from side, anterior face steep, curving broadly
into gently convex dorsum. Ventral lobe rounded, with a long
posterior slope. Postpetiole only a little longer than petiole, and
only a little wider in its posterior part. Next segment (abdom-
inal IV) slightly longer than postpetiole, and slightly wider,
making the widest part of the gaster. Remainder of segments
tapering caudad, subconical; sting stout, curved.

Body with head predominantly smooth and shining ; head with
abundant, small, close but not contiguous punctures, becoming
much feebler and fewer on alitrunk and especially on gaster.
Lower 2/3 of sides of propodeum and most of upper mandibular
surfaces longitudinally striolate. Antennae very finely and
densely punctulate, opaque.

190 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Erect pilosity generally distributed, rather sparse and short,
becoming a little longer and more abundant near gastric apex.
Accompanying the longer pilosity, and replacing it on the legs,
is a short fine reclinate pubescence. Color yellowish ferruginous.

The holotype [California Academy of Sciences] and the very
similar paratype worker [MCZ] were taken together about 10 miles
northeast of Pucon, Chile, by E. S. Ross and A. E. Michelbacher.
This species is readily separated from A. chilensis by the shining
sculpture of the head and the different clypeal and mandibular
armament. Differences from other New World species are given
in the key. This is the second species to be found in Chile.

The name is given in memory of my late good friend, Dr.
Francisco de Asis Monrés, of Tucuman, Argentina, whose tragic
death has deprived the world of a gifted and devoted scientist.

[6] AMBLYOPONE ORIZABANA Sp. NOV.

Holotype worker: TL 2.7, HL 0.58, HW 0.48 (CI 83), WL
0.74, petiole L 0.26, petiole W 0.29, seape L (without basal
neck) 0.31, straightline outside L of left mandible 0.34 mm. A
very small, yellow member of the ‘‘Fulakora group’’ with dor-
sum of head densely and evenly reticulate-punctate and opaque
or nearly so. The habitus and shape of head, body and append-
ages are all substantially as in A. snuthi [17, Fig. 40], although
the posterior occipital border is more strongly concave in the
middle in orizabana. Also, the mandibles are more slender in
orizabana, and can close tightly against the clypeal apron (as
in the holotype). Clypeal apron convex, with 4 truncate teeth
socketed on low tubercles in the middle (median pair smaller,
bases fused), flanked by broader corner teeth on each side; the
latter cannot be seen clearly, and may possibly be subdivided.
The mandibular dentition is difficult to see, but it appears to
consist of a triangular basal tooth or lamella, 4 sharp double
teeth, and a reclinate, acute subapical tooth in addition to the
slender apical tooth. The inner borders are convex. Antennae
much as in A. smithi.

Alitrunk constricted at the narrow, transverse mesonotum,
which is continuous with the propodeal dorsum; metanotal
eroove almost obsolete, visible only in certain lights. Petiolar
node with a vertical anterior face, convex in both directions;

BROWN : ANT TRIBE AMBLYOPONINI 191

dorsum only weakly convex. Petiole and postpetiole about
equal in length; abdominal IV shghtly longer. Ventral process
of petiole with a small angular anterior lobe and a larger, bluntly
subtriangular posterior lobe that slopes upward gradually pos-
teriad. Gastric apex not compressed; sting stout.

Mandibles and lower sides of posterior half of alitrunk longi-
tudinally striolate. Underside of head densely reticulate-punc-
tate, but weakly shining. Rest of body, including legs, scapes
and frontal groove, smooth and shining, with numerous, spaced
small piligerous punctures, best seen on pronotum and petiole.
Pilosity fairly abundant, but very short and mostly oblique;
longer hairs on mandibles and gastric apex. Color light ferru-
ginous yellow.

The holotype, a unique [MCZ], was taken by E. O. Wilson on
Pico Orizaba, Veracruz, Mexico, on August 24, 1953. The worker
was found under a large mossy rock in an open grassy strip along
the trail between La Perla and Rancho Somecla, on the southern
slope of the mountain, at about 2700 to 2800 m. altitude. At this
altitude, the original forest cover is mainly broadleaf temperate
trees, with Carpinus abundant and some pines. This rather
ordinary small Amblyopone is easily distinguished from all the
American species by its small size and yellow color; only A.
degenerata is smaller, and this is very different in antennae and
sculpture.

[7] Key to the New World Species of Amblyopone —
Workers and Females

1. Antennae 7-segmented; worker eyeless, minute, yellow in color (holo-
type only 1.7 mm. total outstretched length) (se. Brazil)
degenerata Borgmeier

Antennae 12-segmented; worker larger, usually over 2 mm. in out-

stretchedmleng theese ioe. 5c: c uae. ote ete eee etek Peacoat hes Grew ayaa sain ee Ota 2.
2. Lobes of frontal carinae separated by a distinct gap (Fig. 19) ..... Be
Lobes of frontal carinae contiguous or fused, as in Figure 40 ...... 6.

Co

Mandibles on inner surfaces each with two sparse separate rows of
small, sharp teeth (Fig. 19, Guatemala) ............ mystriops Brown
Mandibles on inner surfaces with much larger teeth, those near the
midlength fused at their bases so as to form heavy double teeth (Figs.
SAO Die renga eb. cen Ceeen es ee RC RU AMIN Sea RMN Me Senay Saal se ote 4.

192

nN

10.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Inner borders of mandibles angulately produced, so that the blades are
triangular in shape without the apices; large double teeth with rounded
apices; genal teeth reduced to inconspicuous obtuse angles (Fig. 18;
Riedmont ot NortheCarolina)) eee eee eee eee trigonignatha Brown
Inner borders of mandibles straight to convex, not angulately produced,
the blades linear; large double teeth predominantly acute; genal teeth
acute and ;pEOjeChing”. pre mete: ig ee aeons et Soe eee nS

Inner borders of mandibles and anterior clypeal apron straight, or at
most only feebly convex (n. California to Brit. Columbia)

oregonensis (Wheeler )
Inner borders of mandibles, and usually also the anterior clypeal apron,
decidedly convex in outline (temperate N. America w. at least to Iowa
ANGPATIZ ONDA! Mea ee eee Pee ee ee ae pallipes (Haldeman)

At least the anterior 3/5 of the head (as seen in full-face view) pre-

dominantly, densely, sculptured andiopaque) 4.4...) ).s oe eee Ue
Entire or nearly entire dorsal surface of head smooth and shining, with
Spaced SpUNCtURES Fes scat che. ee eeke MLSE oe ee ire Pee Par 10.

Anterior 3/5 to 3/4 of dorsum of head coarsely longitudinally striate
with intermixed punctures, occiput smooth and shining, with spaced
punctures; full adult color piceous or black (se. Brazil to n. Argentina)
armigera Mayr
Dorsum of head densely and uniformly punctate or striolate-punctate
and opaque throughout, except that the cervical border or median
frontal groove may be shining in some cases; full color of worker yellow
LOsreTEU GAN OUS Hos 105 ea, See ein cae CO ie eae ee 8.

Size very small (TL of holotype worker 2.7 mm.) ; alitrunk very smooth,
punctation sparse; propodeum with very few punctures on the dorsum,
and its lateral striation restricted to the lower third of the sides; color
of worker yellow (Mexico: Mt. Orizaba) ........... orizabana Brown
Size larger, TL of worker > 3.0 mm.; punctures more abundant and
distinct on alitrunk; sculpture of lateral faces of propodeum covering
INANGE ie WeAONHs) (xe TNE) SUUATNOO 2c oc ckacncoussoancsecnsaseesnacen: 4).

Size larger; unique holotype worker 4.5 mm. long (according to the

Orginaledeseripion) (Cuba) ae eee eee eee bierigi (Santschi)

Size smaller; worker TL 3.0-4.1 mm., female TL 4.1-4.5 mm. (Chile)
chilensis Mayr

Larger and more robust, worker head width > 0.70 mm.; clypeal apron
having the middle teeth small and not advanced beyond the much larger
teeth that form the lateral corners (Fig. 20, Chile) .... monrosi Brown
Smaller and more slender, worker head width < 0.60 mm.; clypeal apron
convex in outline, the middle teeth distinctly advanced beyond the
corner teeth (se. Brazil to n. Argentina) ........ elongata (Santschi)

BROWN : ANT TRIBE AMBLYOPONINI 193

[8] The large, relatively large-eyed, dark-colored workers with
retrorse mandibular teeth in double rows (Fig. 3) form the group
of A. reclinata. Belonging to this group are the nominal forms
reclinata, feai, rothneyi, bellu and quadrata. I have been able
to examine and compare directly the types of rothneyi and
bellvi, and I find them to differ scarcely at all. Forel’s deserip-
tion of bellii is in error on several crucial points. First of all,
the bellii types have denticulae on the anterior clypeal margin
that are approximately as distinct as in the rothneyi type, and
similar in form. The eyes of the rothneyi type are slightly
larger (0.22 mm. greatest diam.) than in bellu (0.18 mm.
greatest diam.), but not nearly so much so as Forel claims. I
count about 70-90 facets in the rothneyi type before me (here
designated and labeled as lectotype) ; though the count is very
difficult, I find the two bella syntypes to have almost as many
facets (50-70) as rothneyr. In fact, under the best circumstances
of counting, I seriously doubt whether the difference is sig-
nificant.

The bellii types are slightly more coarsely and quite opaquely
sculptured over head, alitrunk, petiole and postpetiole; in the
rothneyi type, these areas are very densely punctate and nearly
opaque, but many of the punctures do have narrow shining
spaces between them, and the postpetiole is more definitely
shining. Both of these species have the anterior genal angles
bluntly subrectangular, and not projecting. The palpal count
for both of the bellii syntypes is maxillary 4; labial 3 (not 2
and 3, as Forel states) ; the palpi cannot be seen in the rothneyi
type. These two species may well be geographical variants;
bellii is from Kanara, on the western side of the Peninsula
south of Bombay, while rothneyi is from Barrackpore, near Cal-
eutta. From Poona, central Madras Presidency, and Orissa, all
on the Indian Peninsula, I have numbers of large males that
probably belong to bellii and/or rothneyt. These males have dis-
tinctive genitalia and a long, narrow caudal process on the
subgenital plate. The whole terminalia most resemble those of
a series of large males of another species from Formosa (Figs. 23,
25), probably the same as the male doubtfully attributed to
bruni by Forel (1913, Arch. Naturg., 79 (A6) : 183), but which
is more likely a member of the reclinata group.

194 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Males of two indeterminate species of the reclinata group. Figure 22.
Subgenital plate, specimen from Los Bafios, Luzon (wings shown in Figure
5). Figure 23. Same, specimen from Chipon, Formosa (wings shown in
Figure 6). Figure 24. Right paramere, mesial surface, and A, detached
volsella in oblique view, Los Bafios specimen. Figure 25. Left paramere
with volsella attached, oblique dorsal view, Chipon specimen. Figure 26.
Left mandible, Chipon specimen. Figure 27. Aedeagal valve, Los Bafos
specimen. Figure 28. Same, Chipon specimen.

BROWN: ANT TRIBE AMBLYOPONINI 195

There are several other series in the reclinata group in the
MCZ. A series of workers from Mt. Makiling, Luzon (L.
Uichaneo leg.), and another accompanied by males from vir-
tually the same locality (Los Banos, Luzon, F’. X. Williams leg.)
have small but protruding genal teeth, the apices of which are
blunt or truneate. The sculpture of these two series is coarse
and opaque, somewhat lke that of the belli: types, and the eyes
are about the size of bellu. However, the accompanying males
have different terminalia; the subgenital plate (Fig. 22) has a
broad, narrowly rounded lobe instead of the long, slender
process of the males discussed above from India, which I take
to be A. bellii on circumstantial evidence. A single male from
San Carlos, Philippines, resembles the other Philippine males.
The Philippine samples, and a series of workers of this same kind
from Macao, were referred to A. rothneyi by Wheeler, but I
think it is more likely that they are another species. Workers
from both the Philippine and Macao series prove to have 5 seg-
ments in the maxillary palpi and 3 in the labial palpi, in com-
parison with the 4,3 count in the bella types. A worker from
the Cuernos Mts., near Dumaguete, Negros, Philippines, has
small, protruding genal teeth, but the sculpture is somewhat
lighter than in the Luzon and Macao series, more as in the
rothneyi type. This specimen is smaller and has smaller eyes
(greatest diameter about 0.14 mm.) with perhaps 35-40 facets,
and agrees in this, in its smaller size, its sculpture, ete. with the
description of fear by Emery.

Another small worker with rather small eyes, but this time
with blunt genal angles like those of belli, comes from Mao
Marroe, 450 m., Soemba Island. This specimen is weakly shining
over the alitrunk and quite shining over the petiolar node and
postpetiole, these parts being cribrately punctate.

It is obvious that differences in eye size, petiolar width and
so on are allometrically variable, and hence untrustworthy as
species characters in the absence of more detailed information
based on large series. Judging from the material before me, and
taking into account the male characters, I would guess that the
reclinata group consists of not less than two, and probably
not more than three or four species. How the names are applied
will of course have to be left to future revisers.

196 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

[9] Of the two forms described as bruni and bruni subsp.
juergi, | have seen only the latter (holotype worker). When he
described jwergt in his ‘‘Glanures’’ paper, Forel apparently
did not have before him the bruni type, and he probably used
only the bruni description for the comparison. At any rate, he
was wrong in describing the anterior clypeal margin of jwergi
as without denticulation; actually, very fine denticles are
present. Of the other differences cited, at least some are allo-
metric characters. Thus, relatively greater width of head and
petiole and larger eye size are only to be expected of a larger
worker as compared to a smaller one, as in this case. Since both
of the types came from Pilam, Formosa (H. Sauter leg.), I think
it likely that they are conspecific. This species has blunt double
teeth on the mandibles, and the clypeal apron is characteristic
in shape (Fig. 21).

[10] Two syntypes of A. minuta (type locality: Soengei Bam-
ban, Sumatra, in termite nest) were examined through the
courtesy of Dr. Besuchet. This is a very small species (TL
2.7 mm.) with slender mandibles; inner border feebly convex,
teeth short, 3 double teeth with sharp retrorse points; apex
slender. Clypeal apron nearly straight, with 6 separated teeth,
inner two largest, outer two smallest. Head finely and densely
reticulate-punctulate, opaque; alitrunk, petiole and gaster densely
punctulate (apex of gaster less so), only moderately shining ;
the alitrunk subopaque. Color brown; appendages ete. yellowish.

A very small, dark brown male [MCZ] from Sandakan, Borneo
(Baker leg.) may belong to this or a related species.

[11] Although they differ in size, the types of luzonica and
williamsi in the Museum of Comparative Zoology are otherwise
much more similar than the original characterizations and faulty
figures indicate. In addition to the types, we now have further
series from the Cuernos Mts. and vicinity, near Dumaguete,
Negros Island, collected by J. W. Chapman and D. Empeso.
This additional material, while closest to the luzonica type in
size, somewhat bridges the gap between the two species. Con-
sidering the similarities, I cannot see that williamsi is more than
a large ‘‘nest variety’’ of luzonica. Size differences of greater

BROWN: ANT TRIBE AMBLYOPONINI 197

magnitude occur between series in other species of the genus
(e.g., australis, denticulata, pallipes). A. luzonica is related to
A. silvestrii and A. amblyops, but has the clypeal teeth in a
distinctive pattern. In all three of these species, four subequal
truneate teeth occur between a pair of larger ‘‘corner teeth,’’
which may themselves be more or less subdivided. In silvestru
and amblyops, the four middle teeth are more or less separate
and autonomous, but in luzonica they are grouped into two
partially fused pairs, one pair on each side of the midline.
The frontal lobes are very close together, though not completely
contiguous, so that this species is intermediate between the
‘<Pulakora’”’ and ‘‘Stigmatomma’’ groups in this respect. The
worker of lwzonica has the palpi segmented 4,3. Apparently this
is the common small Amblyopone of the Philippines.

A. silvestrii has the four middle teeth very regular, equal in
size, and close-set, separated by a wide gap on each side from the
blunt, bipartite corner tooth. This gap is absent or weakly de-
veloped in amblyops, according to Karawajew’s figure. A.
silvestrii also has a densely punctate and more or less opaque ali-
trunk (petiole and postpetiole subopaque) ; the sculpture may
not be quite so dense in amblyops; in silvestrii, the propodeal
declivity is finely transversely striolate or shagreened, whereas

9

in amblyops it is said to be ‘‘very smooth and shining.’

[12] A group of species, related to A. denticulatum, from Ku-
rope, the Near East and North Africa, has workers medium to
small in size, with very small eyes and reduced pigmentation in
the worker. All of these species have the frontal lobes separated,
if only narrowly so, and the clypeal apron normally has 5
truneate teeth, all socketed on tubercles, between the large
‘“corner teeth;’’ the latter often bipartite. Santschi (1915) gives
a reasonable key to the known species of the Mediterranean area.
In A. denticulata the palpi are segmented 4,3.

The variety of denticulata called gracilicornis is only a small
variant common in the Balkans and the Aegean region; I have
seen the series in the Finzi Collection [MCZ] mentioned in the
original description by Menozzi. I agree with Emery (1916)
that gheorghieffi is only the male of denticulatum.

198 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

It seems likely that A. santschu, from Senegal, belongs with
the denticulatum group, although the description is too sketchy
to make this certain.

Dr. Kenneth Christiansen has given me some interesting small
yellowish males with light brown heads, which he took in July
and August, 1952 and 1953, in NW Tripoli Province, Lebanon,
at 1100 m. altitude, and below the Turkish border, Latakia,
Syria, at 800 m.; presumably both collections were made at
light. These may be the males of an undescribed Amblyopone
or of one of the species known from workers in North Africa.

[13] The variation, distribution and synonymy of A. australis
and its close relatives has been discussed by Wheeler (1927:
1-20), Brown (1958b: 18-15) and Wilson (1958a: 142-143).
A. australis shows strong geographical variation in several char-
acters, some of which can be summarized in brief.

Seulpture of dorsum of head. In samples from southwestern
and southeastern Australia, the longitudinal costulation of the
front of the head usually extends only to near the level of the
eyes or a little beyond. In New South Wales and Queensland
the costulation varies widely by locality. A large proportion of
the samples from the subtropical forests of northeastern New
South Wales and southeastern Queensland have the costulation
very extensive, reaching well beyond the eyes and often extend-
ing almost to the posterior cephalic border. This is also the con-
dition in the Lord Howe and Norfolk Island populations, and
in the New Zealand lots; presumably these populations represent
historical introductions from eastern Australia. In samples from
northern Queensland and New Guinea, the costulation is well
developed, but does not usually extend very far beyond the eye;
at least, it is distinctly shorter than in the southeastern Queens-
land samples. In southwestern Australia the costulae and punc-
tures tend to be much coarser (‘‘race foveolata’’) than else-
where, but samples I collected in South Australia (Kangaroo
Island — Kinescote, Rocky River and Ravine des Casoars; Lofty
Range—Aldgate; Northern Flinders Ranges—Wilpena Pound)
approach the Western Australian material in this respect, and
make transition to the Victorian samples. Series from Tasmania,
south central Victoria and the Australian Alps tend to have
both the costulae and the punctures reduced.

BROWN : ANT TRIBE AMBLYOPONINI 199

Seulpture of mandibles. In southwestern Australia, South
Australia, Tasmania, Victoria and the Alps generally, the dorsal
and lateral surfaces of the mandibles are normally completely
and regularly coarsely longitudinally striate. New Hebridean
specimens are mostly of this type. In eastern New South Wales,
one notes that some samples show a weakening of the striation
along the external margin of the mandible, and in series from
the New South Wales-Queensland border (National Park), many
specimens have almost entirely smooth and shining mandibles,
with scattered punctures. These may occur together in the
same nests with specimens having nearly or quite completely
striate dorsal and lateral mandibular surfaces. Such specimens
look much like the sibling species A. michaelseni, which has pre-
dominantly smooth mandibles, and some of them further con-
verge toward muichaelseni in having the head slightly longer than
broad and the genal tooth small in size. This extreme intranidal
variation more or less matches the extensive variation in cephalic
sculpture of the same series.

Farther north in Queensland, one finds a more consistent pro-
eression from largely striate to almost wholly smooth mandibles,
the latter type prevailing in northern Queensland and New
Guinea. The Lord Howe-Norfolk Island-New Zealand popula-
tions are of the intermediate type, with mandibles predominantly
striate, but tending to be smooth along the outer borders.

Color. Pigmentation is difficult to evaluate in insects having
as long a callow period as this species has, but the abundant
material available does indicate some broad trends. In general,
populations from most forested areas in southwestern Australia,
Tasmania, Victoria, and the mountains of New South Wales have
more or less reddish workers, although associated females may
be much darker. However, in intermediate areas that are drier
or without many trees, such as the western districts of Victoria,
Kangaroo Island, and the Lofty and Flinders Ranges of South
Australia, the workers are usually dark brown. In New South
Wales and southern Queensland the color darkens, and most
workers from Queensland, as well as all the mature individuals
of New Guinea, Lord Howe, Norfolk Island and New Zealand,
are dark brown or piceous, with lighter appendages. The New
Hebrides series are intermediate on the reddish side. Some series

200 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

taken from rotten logs in wet forests near Melbourne (Dande-
none Range) are uniformly very light, and more yellowish than
red. Most red series come from beneath stones. In Queensland,
where nearly all australis come from rotten logs, the prevailing
color is dark brown.

Size. The workers show extreme size variation, even at single
localities or within single nest series. The only trend that I can
see in the available material is one toward a wider spread of
variation on the mainland of Australia, particularly in the south-
eastern areas. Perhaps much of this variation is phenotypic in
origin.

Other characters: There is considerable variation in the
mandibular dentition, but it is in large part ‘‘random’’ and diffi-
cult to analyze. New Guinea specimens often tend to form a
blunt tooth between the two largest teeth, and in Australia there
may be one, two or even three teeth basad of the largest tooth ;
however, these always decrease in size toward the mandibular
insertions. Variation in compound eye size is partly regional,
partly allometriec. Even in the smallest workers, however, there
are 12-15 or more facets backed by pigment, and usually more
than 20. In the New Zealand (‘‘cephalotes’’) populations, the
second and third funicular segments are unusually slender, but
can be matched more or less closely by samples from eastern
New South Wales.

Without detailing other variation in minor characters, I think
it can be seen that variation in this species, whether individual
or geographical, shows a high degree of discordance among the
different character-systems. Certainly a character-by-character
analysis in this case shows up the essential lack of a realistic
basis for the old system of ‘‘races’’ and ‘‘varieties,’’? which is
here discarded. Aside from the possibility of cryptic allopatric
species, always a consideration to be respected in wide-ranging
forms, A. australis looks in its totality like one species.

The two obscure forms A. michaelsent and A. leat appear to be
peripheral siblings of A. australis now undergoing displacement
pressure from A. australis after an expansion of this species into
their respective ranges.

As might be expected from elementary considerations of
character displacement (see Brown and Wilson, 1956), A.
australis shows the maximum amount of difference from each

BROWN: ANT TRIBE AMBLYOPONINI 201

of the siblings where it overlaps their respective ranges. Outside
the zone of sympatry, australis sometimes displays some charac-
teristics of the siblings, e.g., the very smooth mandibles of
many Queensland australis, away from the range of A. michael-
seni, Which has predominantly smooth mandibles.

A. michaelsent was described from southwestern Australia,
and I have seen a topotypic specimen collected there (Jarrah-
dale, J. Clark leg.). Clark also recorded samples from Victoria
and New South Wales, but I have never seen any michaelseni
from these eastern states, despite intensive collecting in Victoria
and the close examination of several hundred nests of what all
proved to be australis. Wheeler saw a specimen from Lucindale,
South Australia. Apparently michaelseni is very rare and local,
especially in the southeast, and its similarity in the field to
australis makes it hard to spot. It has a slightly longer head
than australis, the eyes are smaller (15 facets or less, not or
scarcely pigmented), the genal teeth are completely lacking, and
the costulae and punctures are much reduced.

The Lord Howe sibling, A. leai, is known from two series from
the island, of which I have seen three syntype workers from
the summit of Mt. Gervis. As compared to the australis (‘‘how-
ensis’’) populations from the same island, the leai sample is
reddish-yellow in color (vs. dark reddish-brown), has smaller
eyes with fewer and coarser facets (11-12 pigmented facets vs.
18-25), has short cephalic costulation, not or barely reaching
beyond the eyes (vs. reaching nearly to posterior border), has
slender second and third funicular segments (vs. segments II
and III nearly or quite as broad as lone), and has two well-
developed subequal teeth basad of the large median mandibular
tooth, while the latter tooth and its large distal neighbor are
rather close together (vs. one distinct tooth basad of large median
tooth, with sometimes an indistinct denticle basad of this; two
large median teeth of mandible widely separated). It would be
most interesting to investigate the relationship between these
two species on Lord Howe, if they both still exist there.

[14] AMBLYOPONE MERCOVICHI sp. Nov.
(Figs. 29, 32)

Holotype worker: TL 11.0, HL 2.04, HW 1.92 (CI 94), WL
2.98, petiole L 0.95, petiole W 1.11, scape L 1.12, exposed

202 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

straightline L of mandible 1.19 mm. In general size, color and
shape of body resembling at first glance A. australis samples
from the Victorian Alps, but with entirely different mandibles
and other divergent characters.

Head as shown in Figure 29. Cheeks shallowly impressed in
region outside frontal carinae and anterior to compound eyes.
Eyes much reduced, with 6-7 facets in greatest diameter. Frontal

Two new species of Amblyopone from Australia, heads in full-face view.

Figure 29. A. mercovichi, paratype worker. Figure 30. <A. gingivalis,
holotype worker. Drawn to same seale.

lobes broad, nearly horizontal, separated by a narrow median
groove. Clypeus with a short but distinct median lobe, its apex
rounded and denticulate, slightly overhanging the ventral or
free clypeal border, which is also denticulate and is weakly
concave in the middle. Genal angles rounded, unarmed. Labrum
with the usual bilobed free margin, but also having a prominent
transverse, bilobate ridge or welt near its middle, half of which
is shown in Figure 29 as the stippled area below the eclypeal
margin. Mandibles subtriangular, much broader than those of

BROWN : ANT TRIBE AMBLYOPONINI 203

other Amblyopone species, the apical (masticatory) margin curv-
ing evenly into the basal margin. Apical margin with 5 distinct
teeth, the apical largest, size of teeth decreasing basad; in ap-
proximate region of basal angle, the teeth are continued as a
series of 5-6 low rounded teeth or crenulations passing onto the
basal margin. The teeth and crenulations are single-ranked
along the thin free inner mandibular margins.

Antennal scapes slightly incrassate apically, almost straight
in dorsal view, but strongly curved to fit the genal curvature;
when extended back, just barely reaching the compound eyes.
Funiculi gradually incrassate toward apex, but without a
distinct club ; funicular segment II (counting pedicel as I) small,
distinctly broader than long, the segments gradually increasing
in length apicad to X, which is slightly longer than broad.
Apical segment (XI) approximately twice as long as X. None
of the specimens has the under-mouthparts opened out, but the
apical segments of the maxillary palpi are visible in two speci-
mens; this segment is flattened, curved and apically tapered;
its length and position suggest that it is one of two segments in
the maxillary palpus. The labial palpus is not visible, indicating
that it is short and probably not more than 2-segmented.

Alitrunk rather narrow, with horizontal dorsal surface; an-
terior face of pronotum and sides of alitrunk steep; inferior
pronotal angles acutely pointed. Alitruneal profile horizontal ;
pronotum and propodeum weakly convex (nearly plane), meso-
notum weakly convex; metanotal groove distinct and impressed,
with the slightly elevated mesonotum breaking an otherwise
nearly straight alitruneal profile. Mesonotum from above sub-
circular, slightly broader than long, its boundaries distinct ; econ-
striction of alitrunk distinct, centered at metanotum; sides of
propodeum diverging slightly caudad. Declivity of propodeum
steep, flat, on each side with a low, rounded margin. Petiolar
node with vertical, concave anterior face, meeting the almost flat
(gently convex from side to side) dorsal face at an acutely
rounded angle (Fig. 32).

Postpetiole a little shorter than the (normally exposed and
pilosity-bearing part of the) succeeding segment, abdominal IV.
Terminal segment of gaster conical, not compressed.

Body and appendages chiefly smooth and shining, with numer-
ous piligerous punctures over upper part of head, dorsum of

204 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

alitrunk, sides of pronotum, petiole, postpetiole (where strongest
and most numerous) and abdominal segment IV. The cheeks
anterior to the eyes, the gula, and the mandibles are rather
coarsely longitudinally striate, with coarse interspersed pune-
tures. Median lobe of clypeus more finely striate. Sides of pos-
terior alitrunk finely striolate-punctate, chiefly longitudinally.
Anterior coxae shagreened subopaque.

Most surfaces of body, especially gaster and dorsal surfaces
elsewhere, with a thin pubescence of rather coarse, short, reclin-
ate hairs. Longer, more erect tapered hairs, mostly on anterior
dorsum of head and clypeus, masticatory and ventrolateral bor-
ders of mandibles, gula, antennae, and apical and ventral parts
of gaster; also a few of these longer hairs on pronotum and fore
coxae.

Color rich ferruginous red; slightly infuseate spot on vertex.

The holotype, with three paratype workers, was taken by
Father C. Merecovich at Kinglake West (—‘‘Tommy’s Hut’’),
Victoria, on November 15, 1951. The four workers were taken
from the heart rot of a large, rather dry eucalypt log lying on
the forest floor. One of the workers was found in what appeared
to be the ants’ midden, containing large numbers of heads of
termites, which may possibly therefore have been a main item of
food. In this connection, it is interesting to note that A. merco-
vicht bears certain resemblances to presumed cryptic termite-
predators in other ant groups (e.g., Cylindromyrmex Mayr,
Metapone Forel, Gnamptogenys of the mordax group).

This large and curious ant, although it occurs in country
near Melbourne, is known only from the type collection. Father
Mercovich’s finds, in this and the case of A. gingivalis sp. nov.
(below) prove that even the relatively well-collected parts of
Australia hold many a surprise for the enterprising and per-
sistent collector.

A. mercovichi is so different from all other Amblyopone known
that it almost warrants being placed in a new genus. Still, its
affinities with the australis and saundersi groups seem clear
enough, despite the aberrant mandibles. We need to know more
about this species.

The holotype and 2 paratypes (one with no head) have been
returned to Father Mercovich for deposit in an Australian col-
lection. (Smallest paratype: TL 9.38, HL 1.80, HW 1.72 (CI 96),

BROWN : ANT TRIBE AMBLYOPONINI 205

WL 2.64, petiole L 0.83, petiole W 0.95, seape L 1.02, exposed
straightline L mandible 1.05 mm.). One paratype in MCZ.

[15] AMBLYOPONE GINGIVALIS sp. Nov.
(Figs. 30, 31)

This new species is known from a single worker collected by
Father C. Mercovich, 8.J., at Calga, New South Wales, during
April, 1956. Father Mercovich sent me the specimen together
with the sample of A. mercovichi, described as new above, and
Mrs. Buffler prepared figures of the head (Fig. 30) and petiole
(Fig. 31) under my supervision. I[ then returned these speci-
mens to Father Mercovich in Australia, after which, learning of
my progressing revision of the tribe, he very generously returned
them to me so that I might present the description with the
rest of the material on the amblyoponines. Unfortunately, the
shipment suffered damage in transit, and the sole worker of
gingivalis lost its entire head and the anterior part of the pro-
thorax. It might seem wiser to forego description of a unique
holotype missing its most important parts, but in this case we
do have a good drawing of the head available, and the species
is of such particular interest in connection with the hmits of
variation within the genus, that I have decided to give it a name.

Holotype worker: TL 8.6 (estimated), HL 1.5 (estimated),
HW 14 (estimated) (CI 91 from drawing); WL 2.1 (esti-
mated), petiole L 0.82, petiole W 0.68 mm. Estimated straight-
line L of mandible 0.9 mm. Form of head shown in drawing.
Note extremely reduced eyes and clypeal apron, the latter
completely unarmed, and the subtriangular mandibles, with
fairly distinct basal border and mostly blunt teeth, partly
double-ranked. Seapes short, subclavate; genal angles unarmed.

Remainder of body typically amblyoponine in form, slender ;
alitrunk constricted at metanotum (seen from above). Mesono-
tum nearly three times as broad as long, bounded behind by a
distinct, shallowly impressed metanotal groove. Alitruneal dor-
sum in profile horizontal; promesonotum and propodeal dorsum
forming separate feeble convexities. Propodeal declivity sloping
caudad, forming on obtuse angle with the dorsum, into which it is
gently rounded. Petiolar node longer than high and longer

206 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

than broad (Fig. 31). Postpetiole nearly as long as the suc-
ceeding segment; last visible segment (abdom. VII) laterally
compressed ; extruded sting very long, slender, curved.

Seulpture of head (as recalled) densely and finely punctate
or reticulate-punctate and predominantly opaque. Alitrunk,
petiole and gaster mainly smooth and shining, with sparse pune-
tures. Sides of alitrunk and anterior sides of petiole finely
striate with scattered punctures, the striae principally longi-
tudinal. Body overall with numerous fine, short, obliquely erect
hairs and smaller, reclinate, pubescence-like hairs, not obscuring
the sculpture. Color medium reddish-brown, head more brown-
ish, appendages lighter.

31 32

M7

Two new species of Amblyopone from Australia, petioles and adjacent
parts in side view. Figure 31. A. gingivalis, holotype worker. Figure 32.
A. mercovichi paratype worker. Not drawn to same scale.

This species is so distinct in the shape of its clypeus and
mandibles that there should be no chance of confusing it with
any other Amblyopone. Undoubtedly it is a specialized relict,
probably local in distribution. Father Mercovich took the sole
worker under a stone ina gully with Angophora and scribbly gum
(Eucalyptus haemastoma) trees as cover. The holotype will be
returned to Father Mercovich for deposit in one of the Austral-
ian collections.

[16] The ferruginea group as presently known includes five
Australian species: aberrans, clarki, ferruginea, hackeri and
longidens, the first two of which are southwestern and the last
three southeastern. The workers of all are stoutly built, small-
eyed, yellow forms, smaller than A. australis and relatives, but
not much different in general habitus. In fact, the ferruginea

BROWN : ANT TRIBE AMBLYOPONINI 207

group occupies a position intermediate between the sawndersi
and australis groups in size, structure and presumably in
biology, and may well represent the descendants of the line
that led from sawndersi-like ancestors to A. australis. Like both
contiguous groups the members of the ferruginea group have a
palpal formula of 2,2 in the workers, while the clypeal apron, its
dentition, and the male-female wing venation are reduced as in
australis. The females, where known, are distinctly larger than
the workers and have the head, alitrunk and petiole more or less
dark brown or blackish. The males are smaller and blackish. The
species are all strongly cryptic in habits, and normally nest and
forage below the soil surface, so that they are rarely seen except
when flooding forces them to the surface, or in the cold of
winter, when they may rise beneath stones to seek the heat stored
there from solar radiation. There are no good observations
known to me on nest size, behavior or food preferences, but I
think that some colonies must attain a large size. Like many
subterranean ants, members of this group are found in a great
variety of habitats, including wet forest, alpine scrub, and
arid woodland. The systematics of the group were partly clari-
fied by Brown (1952), who showed that Clark’s mandibularis
was the same as ferruginea Smith (Fig. 36), and that ferruginea
is a local species, so far found only in Melbourne and vicinity.
Forel’s variety longidens is really a good species (Fig. 34)
(confused by Wheeler in 1927 with ferruginea), widespread but
local in southeastern Australia.

A. hackert (Fig. 39), the most ‘‘Fulakora-like’’ of the ferru-
ginea group, occurs in the country on the border between
southeastern Queensland and northeastern New South Wales.
Portions of a dismembered chilopod were found in a nest of this
species taken by P. F. Darlington in southeastern Queensland.

Of the southwestern forms, aberrans (Fig. 35) and clarki
seem distinct enough in mandible type, despite the considerable
variation in dentition shown by clarki (Figs. 33, 37, 38). In
addition, samples having the general clarki habitus (Fig. 33)
show unusually strong geographical variation in sculpture,
which may in fact represent differences between two or even
three separate species. Specimens from the type series (Ludlow,
W. A.) (Fig. 37) and a series from Busselton, W. A. (J. Clark

208 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Ne

Australian Amblyopone & the igi group, workers. Figure 33. A.
clarki or near, Norseman, W. A., full-face view of head and mandibles.
Figure 34. Right mandible and anterior eclypeal margin, A. longidens,
Grampians Range, Vict. Figure 35. Same, 4. aberrans syntype, Mundaring,
W. A. Figure 36. Same, A. ferruginea, specimen compared with type, Re-
search, Vict. Figure 37. Same, A. clarki syntype, Ludlow, W. A. Figure 38.
Same, A. clarki or near, Perth, W. A. Figure 39. Same, 4. hackeri syntype,
National Park, Queensland.

BROWN : ANT TRIBE AMBLYOPONINI 209

leg.) have the coarse, longitudinally striate or costulate sculpture
of head and pronotum usual in species of the group. A worker
taken by C. P. Haskins under a log in the National Park, W. A.,
of which the mandible and elypeal margin are shown in Figure
38, has the upper half of the head and the pronotum with numer-
ous coarse, separated punctures in place of the costulae. Three
workers (Fig. 33) taken by E. O. Wilson in arid woodland at
Norseman, W. A. (which is far inland from the other clarki
localities) have even more reduced sculpture; the punctures are
much finer than in the National Park specimen, and the inter-
spaces smooth and shining. More material is needed from
southwestern Australia in order to clarify this situation.

[17] A number of small forms occurring in southeastern Aus-
tralia and New Zealand are placed in the saunders: group. Al-
ready described as species are exigua, wilsom, lucida, punctulata
and gracilis from Australia, and sawndersi from New Zealand.
These forms all have the ‘‘Fulakora habitus;’’ i.e., they are
small and slender, depigmented and with eyes much reduced
or absent in the worker; the mandibles, clypeus and genal teeth
are of the ‘‘typical Stigmatomma’’ pattern, and the lobes of
the frontal carinae are fused or contiguous at the midline of
the head. All have dense, fine punctate sculpture of the
head. The palpi are apparently all segmented 2,2 (rarely 1,2).

It seems impossible for the time being to say how many good
species there really are in this group. I am adding a new
species that is quite distinct from any of the others (A. smithi
sp. nov., described below). Only for sawndersi do I have material
from more than two or three localities, and this species is so
variable that doubts must remain as to whether it may not repre-
sent two or more sibling species existing in New Zealand (Brown,
1958b).

There are differences among the species in head shape (es-
pecially C1), mandibular dentition, overall size and sculpture,
but the best character is often the form and dentition of the
anterior clypeal apron. In two series I have from Wallaby
Creek in the Hume Range, central Victoria (D. Ashton leg.) , and
from Sherbrooke Forest in the Dandenong Range, Victoria
(Brown leg.), the center of the clypeal apron is produced into

210 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

42

Australian Amblyopone of the saundersi group, workers. Figure 40. A.
smithi sp. nov., paratype, full-face view of head. Figure 41. Anterior
margin of elypeal apron, 4. gracilis syntype. Figure 42. Same, dA. sp. from
Binna Burra, southeastern Queensland (left anterior corner of head with
genal tooth also shown). Figure 43. Same, A. sp. from the Hume Range,
central Victoria. Figure 44. Same, A. exigua from Kallista, Victoria,
determined by Clark. Figure 45. Same, A. lucida syntype.

a definite median lobe (Fig. 43), while in paratypes of A. lucida
from Coree (or Corrie) Creek in the Alps of the Capital Terri-
tory (G. Hill leg.) and a series from Bonang, Victoria, also in
the eastern Alps (P. F. Darlington leg.), the apron is very nearly
straight and has a differently arranged dentition (Fig. 45).

BROWN: ANT TRIBE AMBLYOPONINI PALI

Most samples from Australia have clypeal aprons intermediate
between these last two types (Figs. 41, 44). I have specimens
of exigua Clark (Fig. 44), collected and determined by Clark
himself, from Kallista, Victoria, which is adjacent to the type
locality (Belgrave) and close to Ferntree Gully, where I col-
lected another series of this same species. It should be noted
here that Sherbrooke Forest, where the specimens (noted above)
with the mesally lobate clypeal apron were found, is immedi-
ately adjacent to Belgrave and Kallista, and it seems likely
that all of the specimens came from the same wet forest that
clothes this limited area. That we are dealing here with a pair
of sympatric species seems assured. Possibly the lobate form
from Sherbrooke Forest and Wallaby Creek is an undescribed
species, but since Clark’s descriptions and figures are so vague,
I eannot be sure that this form does not match his A. wilsoni,
of which I have seen no authentic specimens.

Paratypes of gracilis (from Beech Forest on the Otway Penin-
sula of Victoria) are larger in size, but have a clypeal apron
(Fig. 41) similar to that of exigua. A series matching the gracilis
paratypes very well was taken by P. F. Darlington at Corinna,
Tasmania. Another series taken by the same collector at Corinna
is Similar, but the insects are smaller in size; about half of them
are ergatoid females with larger eyes.

This last collection agrees well with Clark’s characterization
of punctulata. In spite of their similarity, the two Corinna
samples may belong to different species (gracilis and punctu-
lata).

One more series, this from Binna Burra, near Beechmont in
southeastern Queensland (P. F. Darlington leg.), is similar in
size to gracilis, but has much larger genal teeth (Fig. 42), and
the sides of the propodeum are largely smooth and shining
(striolate, opaque in gracilis, saundersi and other similar
species). Possibly this species is undescribed. It is clear that
we need much more material of this group in order to be sure
of the species.

AMBLYOPONE SMITHI sp. nov.
(Fig. 40)

Holotype worker: TL 2.6, HL 0.56, WW 0.44 (CI 79), Wh
0.69, petiole L 0.29, petiole W 0.29, scape L (without basal

2A, BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

neck) 0.26, exposed straightline L left mandible measured from
lateral insertion 0.36 mm. Habitus of saundersi group, but
much smaller than any of the other species known, and with ca.
posterior 2/5 of the head distinctly shining, the punctures here
coarse and numerous, but separated by predominantly smooth
and shining interspaces; a median smooth strip extends forward
to about the center of the head. The anterior half or slightly
more of the head is densely punctulate, with intermixed indis-
tinct longitudinal-oblique striolation, nearly or quite opaque.

The general shape of head, mandibles and antennae is shown
in Figure 40. The eyes are reduced to minute, indistinct pig-
mented spots, each backing 2 or 3 indistinct facets. Mesonotum
transverse, its posterior limits marked only by a feeble constric-
tion; metanotal groove obsolete, a faint trace visible in certain
lights. Propodeal dorsum meeting declivity through an obtuse,
rounded angle; declivity plane, oblique, broader than high.
Petiolar node sessile, a little longer than high; ventral process
subquadrate, with rounded anterior corner and sharply rec-
tangular posterior corner. Anterior face of node vertical, round-
ing into horizontal, weakly convex dorsum; seen from above,
the node is approximately square, with rounded anterior cor-
ners. Postpetiole about as long as petiole, and only shghtly
wider; succeeding segment (abdominal IV) about as long as
postpetiole, but shghtly wider. Apex of gaster conical; sting
stout, curved.

Sculpture, apart from head, predominantly smooth and shin-
ing, with seattered piligerous punctures. Lower mesepisterna
and almost all of sides of propodeum, as well as mandibles,
longitudinally striolate, but still moderately shining. Clypeus
longitudinally striate. Pilosity as described for A. monrosi.
Color ferruginous yellow.

Holotype [MCZ] and one very similar paratype worker |to be
deposited in a major Australian collection] collected together
from under a stone in clayey soil in dry eucalypt forest (pre-
dominantly Lucalyptus Barter, E. obliqua and E. cosmophylia)
at Aldgate, near Mt. Lofty in the Lofty Ranges of South
Austraha, altitude about 550 m., December 8, 1950 (leg. W. L.
Brown, Jr.). The weather at the time of collecting was very
hot and dry; much digging turned up no further specimens.

BROWN: ANT TRIBE AMBLYOPONINI 213

This species is named for my long-time friend and adviser
in myrmecology, Dr. Marion R. Smith.

[18] A. celata is a small yellow species, widespread in the Solo-
mons. I was able to count only the maxillary palpal segments:
there are four. Mann found that the pupae are naked, as in
myrmicines. Apparently this species is replaced on New Guinea
by Prionopelta.

[19] A. ewaluwenburgi is a minute species (under 2 mm. total
length) that was found in the soil of a sugar cane field at the
Hawaiian Sugar Planters Association Experiment Station, Hono-
lulu. I think it is likely that the species has been introduced into
Hawaii from Melanesia or the East Indies.

[20] Wilson (1958a) has discussed the variation of Myopopone
castanea in some detail, emphasizing the Melanesian popula-
tions. He has synonymized five of the species, all described by
Donisthorpe from 1938 to 1949, and there is every reason to
accept his synonymy. Wilson’s study extended only inci-
dentally to the more western populations of M. castanea, which
fall outside Melanesia proper, and it is now possible to study
these in the light of additional material from the Philippines
and elsewhere. Most of this material is included in the collec-
tion of Dr. J. W. Chapman, and comes principally from the
vicinity of Dumaguete on Negros, Leyte, Palawan and southern
Luzon (collected by Chapman, Baker, Williams, Brues and Me-
Gregor). Other material comes from northern Borneo (Mjo-
berg), Sumatra (Brues), ete.

The varieties bakeri Viehmeyer and proxima Stitz are repre-
sented by type material in the Museum of Comparative Zoology ;
comparison of these with the rest of the Philippine series shows
that they are both part of the intraspecific variation of the
castanea of that area. The characters of sculpture, and especially
the shape of the subpetiolar process, show wide variation over-
lapping that of the New Guinea-Solomons samples.

The females, which are even more variable than the workers
in sculptural characters, and which differ considerably from the
workers of the same series, have been a source of synonymy

214 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

even dating back beyond Donisthorpe’s time. Bingham described
moelleri from Sikkimese examples taken at light, and Stitz added
a variety striatifrons based on females from Lombok and western
Sumatra; the latter were compared only with Bingham’s very
sketchy description. Neither description mentions any features
outside the usual variation of the sample of castanea in the
Museum of Comparative Zoology.

Roger described M. rufula from Batjan, thinking even then
that it was ‘‘perhaps only a local variety’’ of M. maculata, de-
scribed from Ceylon; later (1863, Verzeichniss der Formiciden
Arten und Gattungen, Berlin, p. 20) he synonymized rufula with
Smith’s castanea. Probably the rufula and castanea types came
from the same series collected on Batjan. Emery (1911) con-
sidered maculata to be a subspecies of castanea, which is more
in line with the present conception of geographical and indi-
vidual nest variation within a single species. Since we know
that the ‘‘diagnostic’’ characters of these forms vary discord-
antly among themselves, there seems little reason to recognize a
Ceylonese or a more widespread variant under a formal name.
The variety bugniom Forel, described from Ceylon, was never
satisfactorily separated from maculata, and is undoubtedly the
same geographical variant.

There remains only Myopopone beccaru Emery, described from
worker material from Ternate. According to the original de-
seription, M. beccaru is supposed to differ from M. castanea
chiefly in having the mesonotum and petiole longitudinally
rugulose and subopaque. In view of his qualifying remarks fol-
lowing the diagnosis, it seems that Emery may have meant
‘‘metanotum’’ instead of ‘‘mesonotum.’’ A sample of workers
from central Cape York Peninsula, Queensland, has faint stria-
tion or longitudinal rugulation between the coarse punctures of
the propodeum (—‘‘metanotum’’) and petiolar dorsum, and the
striation of the sides extends farther up onto the dorsum than
in many other castanea samples. It seems that this may repre-
sent an approach to the condition Emery found in beccarv. In
any case, beccaru comes from the middle of the range of castanea,
and it seems unlikely that it is more than an extreme sculptural
variant of the latter. Emery himself apparently lost some con-
fidence in the distinctness of beccarvi, because his Genera In-
sectorum entry (1911) rates beccarti as a mere subspecies of

BROWN: ANT TRIBE AMBLYOPONINI PAS)

castanea, along with moelleri and maculata. Wilson (1958a)
made no attempt to evaluate beccarii, and recognized it pro-
visionally as a species only on the basis of the original description
and grounds of virtual sympatry with castanea. Although I have
not myself examined the types, I feel that the study of the
variation in both eastern and western populations of castanea
reveals sculptural extremes that are probably about as aberrant
as described by Emery for beccarvi, and I therefore consider that
beccaru should be placed provisionally in the synonymy of
castanea.

I have listed below the synonyms of M. castanea in chronologi-
eal order. In those cases where the date is enclosed in parenthesis,
Wilson (1958a: 144) has already given the full reference in
indicating the synonymy, and the details will accordingly be
omitted here. (T) before the entry indicates that either Wilson
or myself has reviewed type material.

(T) castanea (Fr. Smith) (1860).
maculata Roger, 1861:50, worker, female. Type loc.: Ceylon. N. syn.
rufula Roger, 1861:52, worker. Type loc.: Batjan. Syn. Roger, 1863.
beccarti Emery, 1887, Ann. Mus. Civ. Stor. Nat. Genova, (2)4:447,
worker. Type loc.: Ternate. N. syn.
moellert Bingham, 1903:34, female. Type loc.: Sikkim, 7000 ft. N. syn.
bugnioni Forel, 1913, Zool. Jahrb. Syst., 36:5, nota, worker, female,
male. Type loc.: Peradeniya, Ceylon. N. syn.
(T) bakeri Viehmeyer, 1916, Ent. Mitt., 5:283, worker. Type loc.: Tacloban,
Leyte, P. I. N. syn.
(T) proxima Stitz, 1925, Sitzb. Ges. naturf. Freunde, Berlin, 1923, p. 110,
nomen nudum; descr. Wilson, 1958a:145. N. syn.
striatifrons Stitz, 1925, Sitzb. Ges. naturf. Freunde, Berlin, 1923, p. 110,
female. Type locs.: Lombok and w. Sumatra. N. syn.
(T) picea Donisthorpe (1938)
(T) wollastoni Donisthorpe (1942)
smithi Donisthorpe (1946)
(T) rossi Donisthorpe (1947)
(T) similis Donisthorpe (1949)

[21] The species-level taxonomy of the Neotropical Prionopelta
has been much confused. Mayr described the holotype female of
P. punctulata as having 11 antennal segments. In his generic
diagnosis of 1911, Emery noted that Prionopelta (so far as he
personally was acquainted with the species) had 12 antennal

216 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

segments in worker and female, and he felt that Mayr had been
misled by an ‘‘anomaly.’’ But material now available proves
that southeastern Brazil and northern Argentina are inhabited
by a species having 11 segments and corresponding reasonably
well to Mayr’s description of the type female and his later
(1887) description to the worker. This species, which I consider
to be P. punctulata, varies somewhat both inter- and intranidally
in size, strength of sculpture, and intensity of the yellow pig-
ment. Forel, utilizing differences he supposed to exist on the
basis of a comparison of Mayr’s two (female and worker) deserip-
tions, gave the name mayri to the worker from Santa Catarina.
The differences cited are in minor sculptural and pilosity fea-
tures, and it seems best to accept Mayr’s original opinion that
his worker and female belong to the same species, rather than
Forel’s judgment based on the descriptions. Santschi described
P. bruchi in 1923 from Alta Gracia, Cérdoba, Argentina, ap-
parently from the worker caste, although the description is
headed by the symbol ‘‘ ° ’’ — probably a typographical error.
Santschi’s description is in the form of a comparison with
‘*mayri,’’ although he does not state how he verified the deter-
mination of the mayri specimens used for the comparison. The
differences mentioned are minor ones, subject for the most part
to differences of viewing angle and interpretation, as well as to
allometric variation. Santschi does not give a count of the anten-
nal segmentation, the most important diagnostic character in this
case. Samples taken since in Argentina by Kusnezov and others
are all 1l-segmented and appear to represent a single species,
which I refer to P. punctulata. I myself have seen workers and
males from Tucuman, Argentina (N. Kusnezoyv, P. Wygodzin-
sky leg.), and from Loreto, Misiones, Argentina (A. Ogloblin
leg.). Kusnezov has determined the males from Tucuman as
bruchi. It seems to me that there is no reason any longer to
doubt the synonymy of bruchi with punctulata. Two workers
from Agudos, S. Paulo, Brazil (W. W. Kempf leg.) belong to
this species also.

Two other Neotropical Prionopelta species have already been
eliminated from consideration: P. mocsaryi Forel, supposedly
from Paraguay, appears to be a mislabeled sample of the
Melanesian P. opaca Emery (Wilson, 1958a:149); P. marthae

-

BROWN : ANT TRIBE AMBLYOPONINI 217

Forel is a synonym of Typhlomyrmex rogenhofert Mayr, accord-
ing to Brown, 19538a.

All of the other forms of Prionopelta from the Americas have
12-sezmented antennae in worker and female. They may be
divided on the basis of cephalic sculpture into two groups. One
kind corresponds to the first lug of couplet 2 in the short key
below; this is the clearcut species P. modesta, with coarsely
punctulate and opaque head. KE. O. Wilson collected it at Las
Hamacas, 17 km. north of Santiago Tuxtla, Veracruz, and at
Pueblo Nuevo, near Tetzonapa, in the same state. At both
localities, it was a common ant in the leaf litter of tropical
evergreen forest; notes on the ecology and behavior of this
species are given below. Other samples of modesta in the Museum
of Comparative Zoology come from Finca El Real, Ocosingo
Valley, Chiapas (C. and M. Goodnight and L. J. Stannard
lee.), and from the vicinity of Guatemala City in orchid plants
intercepted by U.S. Plant Quarantine at San Francisco. A series
from Barro Colorado Island, Panama Canal Zone (J. Zetek
leg.) is the only sample of modesta known from outside the
Guatemala-southern Mexico area.

The remainder of the twelve-segmented forms, including
types of the two nominal species antillana and amabilis, form
a rather heterogeneous lot. The antillana type (St. Vincent,
B. W. I.) in the Museum of Comparative Zoology has the an-
terior clypeal margin projecting somewhat forward and bluntly
subangulate in the middle, and the cephalic sculpture is weak
and subopaque, with sparsely punctulate, smooth and shining
areas just mesad of the eyes. Specimens from Tumupasa, in
the Amazon watershed of Bolivia (W. M. Mann leg.) agree
well with the antillana type. Other series from Colombia, Costa
Rica, Honduras, Guatemala, and even (one worker) Pueblo
Nuevo, Veracruz, are very variable in cephalic sculpture, rang-
ing from sparsely punctulate and shining to densely but very
indistinctly punctulate and subopaque. In some of these speci-
mens, the anterior clypeal margin can be seen, and is gently
and evenly rounded; in other cases the anterior clypeal margin
cannot be seen satisfactorily because the mandibles have not
been opened. The relationship of antillana to amabilis remains
obscure as far as I am concerned; much more material will be
needed to decide the variation and limits of these two species.

218 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

The present distinctions among the American Prionopelta
species are summarized in the following key to the workers and
females :

Key to Neotropical species of Prionopelta
based primarily on workers

1. Antennae 1l-segmented (n. Argentina, se. Brazil) .... punctulata Mayr

o

Antennae hi2sserinentedy aya: si 9) 1a. ee a ee ee ae ee 2.

2. Punctulation of head coarse, uniform over entire dorsum, deep and
densely arranged, rendering the whole surface nearly or quite opaque
(3; Wieaeo, Gommmomne Cl, AWIEMICR)) oc eeeccccso ocean edaec modesta Forel
Punctulation of dorsum of head fine and shallow, often varying in dis-
tinctness and density from area to area; either spaced, with smooth,
shining intervals, or combined in an indefinite, subopaque roughening of
the surface (s. Mexico, C. America) .............. amabilis Borgmeier
(Lesser Antilles, Bolivia, ?C. America) .............. antillana Forel

[22] A colony of P. modesta, taken at Pueblo Nuevo, Veracruz,
by E. O. Wilson, was studied for several weeks in a glass-topped
plaster nest in the laboratory. The colony contained eggs, which
hatched in the laboratory, but the dealate female apparently was
not fertilized, and was not the colony queen. At the time of
capture, other females were found in the colony with wings, as
also in other colonies found at the same locality (August, 1953).
The colonies often appear to be split into sections separated
more or less from one another, and occupying small bits of rot-
tine wood or bark in the leaf litter. Some of the fragments con-
tain only workers, or workers and larvae, indicating that the
colonies may be very loosely organized (like those of Amblyo-
pone) and more or less nomadie.

Newly hatched larvae in the laboratory nest spent much time
raising and moving about the anterior part of the body, possibly
an activity functioning to attract workers with food. Several
attempts were made to get the Prionopelta workers to accept vari-
ous kinds of live and dead arthropods, but the workers are ex-
ceedingly timid foragers, and almost always recoiled violently
when their antennae came into contact with the potential prey.
Even such delicate animals as live Tomocerus (Collembola) and
fresh-killed Drosophila adults were treated in this way by the
ants. After repeated contacts within a very cramped foraging

BROWN : ANT TRIBE AMBLYOPONINI 219

chamber, the ants finally killed a partly-grown geophilomorph
centipede about 15 mm. in length — that is, much longer than
the ants, which are about 2 mm. long. The ants grasped the
centipede by a leg with a very rapid lunge of body and mandibu-
lar closure too quick to be observed in detail, and immediately
doubled up and attempted to sting in the leg joints or body of
the centipede. The centipede showed obvious agitation each
time it was stung, and raced about, doubling and redoubling its
body until the ant was finally dislodged. After being stung, the
centipede showed impairment of locomotor activity of the seg-
ments behind the point where the ant had been attached, and
ran rather crookedly, but after a few minutes showed partial
recovery. Discontinuous observations covered several more indi-
vidual attacks by different ants, and the following day, the ants
were found to have removed the centipede, now motionless and
apparently dead, to the brood chamber, where some larvae were
now attached and feeding. Other later trials with small centi-
pedes were made at a time when the colony was in its final stages
of disintegration, and the ants simply fled upon the slightest
movement of the centipedes.

The general impression gained from these observations was
that P. modesta is capable of biting, hanging onto, and stinging
into immobility even relatively large arthropods (such as a
geophilomorph of the size indicated), but that such large prey
animals would not normally be available in circumstances of
confinement allowing the Prionopelta to attack them. Especially
is this indicated when one considers that the usual result of physi-
cal contact between centipede and ant was the hasty retreat of
both. Perhaps the usual prey will be found to include even
smaller, juvenal centipedes or symphylans; search of the nests
should be made to determine what Prionopelta feeds upon.

[23] The Prionopelta of the Old World occur in Madagascar,
southeastern Africa and the Indo-Australian region. The Mada-
gascan species (descarpentriesi) and the one from the mainland
of Africa (aethiopica) are both known from single collections.
They have not been compared directly with each other or with
most other species of the genus, and the characters separating
them as species are not clear from the descriptions. Since |

220 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

have seen no material from these regions, I am merely placing the
two names in the list as species without judging them further.

In the Indo-Australan region there are four nominal species,
of which brocha, isolated on New Caledonia, is a very distinet
relict (Fig. 15). The other three species are majuscula (New
Guinea), opaca (New Guinea) and kraepelini (Java). P. maju-
scula and its synonymy have been discussed by Brown (1953b)
and Wilson (1958a). So far as known, it is confined to northern
New Guinea. It has moderately punctate, shining workers, yel-
lowish in color and somewhat larger than kraepelini and opaca
workers. Its females are distinctly larger than its own workers,
and are a contrasting dark brown in color.

In kraepelint and opaca, the queens are searcely larger than
the workers and tend to have similar sculpture and color if from
the same nest or locality. I have now seen a great deal of new
material of these two forms, and the distinction between them is
becoming blurred. Together, they turn out to have a vast range,
from northern Queensland (at least) in the southeast to Java
and the Philippines to the west and north. The typical opaca
come chiefly from New Guinea and parts of Micronesia. They
are usually brownish in color and have coarse, dense punctation
over the body from head onto the first gastric segment at least.
Workers and females of kraepelini usually have much less dense
punetation, and the alitrunk and gaster are predominantly
smooth and shining. I have seen samples more or less closely
agreeing with kraepelini from several Micronesian localities;
from Buitenzorg, Java (the kraepelini type locality); from
Dumaguete, Negros, and Mt. Makiling, Luzon in the Philippines
(J. W. Chapman Coll.) ; and from Mt. Bartle Frere, 1000-1600
m. (P. J. and P. F.. Darlington leg.) and the Elliot Range, ca.
1000 m., near Townsville (Darlingtons leg.) in Queensland.
Father J. J. McAreavey also sent me a couple of workers of this
species which, he wrote me, came from the Grampians Range in
western Victoria. I questioned this range at the time (which
was before the Queensland records were available), but Father
McAreavey has assured me (in. litt.) that the Grampians record
is not based on a labelling error. Still, the Grampians record
remains so astonishingly far beyond the other localities that only
further collections there will remove all doubts about a possible

BROWN : ANT TRIBE AMBLYOPONINI 221

mixup. The specimens are very similar to some in the Museum
of Comparative Zoology from Dumaguete in the Philippines.
Perhaps in favor of the Grampians record is the fact that the
two Queensland collections both came from rather high levels
on mountains.

In Micronesia, there are found forms intergradient between
opaca and kraepelim. While I have not studied the Micronesian
material at any length, I believe that probably most of the
kraepelini specimens really are only smoother, light-colored vari-
ants of opaca, situated peripherally to New Guinea. It is inter-
esting to note that the workers of majuscula have the smooth
integument and light color approached by kraepelini outside
New Guinea; many extra-New Guinea workers also appear to be
more robust, in this tending toward majuscula. Apparently we
have in opaca another case of centrifugal variation based on
New Guinea, and involving character displacement against the
parapatrie species majuscula.

The relationship of kraepelini to opaca needs to be reviewed
with more and better material than I now have before me. Also
to be considered is the relationship of these species to the neo-
tropical species; the kraepelini topotypes, although not well-
preserved, are rather smooth and unusually pale, and do resemble
P. antillana. It is not beyond possibility that this population may
have reached Buitenzorg from the New World in plantings for
the Botanical Gardens there. The Dumaguete samples were taken
in an ‘‘Old Cemetery,’’ and this and other circumstances suggest
strongly that Prionopelta species like this one are often trans-
ferred accidentally by human commerce.

Key to Indo-Australian species of Prionopelta,
based primarily on workers*

1. Genal teeth well developed and acute (Fig. 15), more than 0.02 mm. long;
size larger, head width of holotype worker 0.64 mm. (New Caledonia)

brocha Wilson

Genal teeth minute to obsolete, never more than 0.01 mm. long; smaller

spp., head width not exceeding 0.55 mm. in worker .........--------- 2.

*Dorylozelus mjoebergi Forel, from Queensland, may possibly be a Prionopelta.
It is supposed to have only 11 antennal segments, which should separate it
from the species in the key, all 12-segmented.

222; BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

bo

Size larger; head width > 0.47 mm. in worker; worker feebly sculptured,
shining, yellowish in color; female distinctly larger than worker, dark
browne NewsGtined)) sees ee ice ee eae eae ar ete majuscula Emery
Size smaller; head width < 0.47 mm. in worker; color pale yellow to
brown; female nearly same color and size as workers from same nest;
sculpture variable, but always coarse on New Guinea (e. Australia, New
Guinea, Micronesia, Java, Philippines)

opaca Emery and kraepelini Forel

Onychomyrmex, Figure 46. O. ‘hedleyi, worker, outline of body from side
view. Figure 47. Male wings, probably O. mjoebergi, specimen from the
National Park, southeastern Queensland. Figure 48. O. hedleyi worker,
full-face view of head.

BROWN : ANT TRIBE AMBLYOPONINI 223

[24] The known ranges of the three Onychomyrmex species have
been increased in eastern Queensland in recent years, mainly
through the activities of P. J. Darlington, Jr. (PJD) in 1932 and
1957-1958, and his son, P. F. Darlington (PFD) in 1957-1958.
I (WLB) made a few collections in northern Queensland in
1950.

O. doddi: western slope of the Macalister Range, east of the
Black Mountain Road running north from Kuranda, about 100
workers in a small rotten log in rain forest (WLB). This is
only a few miles from Kuranda, the type locality.

O. mjoebergi: Mt. Spee Plateau, 2000-3000 feet altitude, about
40 miles north of Townsville (PFD). National Park, in Me-
Pherson Range, on the southern border of Queensland, 3000-
4000 feet (PJD); workers and males were taken separately.
The petiole of the workers from southern Queensland differs
slightly from that of northern samples, but not enough to con-
vince me that they are separate species.

O. hedleyi: Mt. Bellenden Ker, east side, 3000-4500 feet
(PJD). Mt. Spee Plateau, 2000-38000 feet (PFD). Malanda,
2200 feet (WLB). Kuranda, 1100 feet (WLB). Mt. Spurgeon,
3500-4000 feet (PJD). Thornton Peak, near Daintree, 1000-
4000 feet (PFD). Eungella Range, west of MacKay (PFD).

ACKNOWLEDGEMENTS

To those who aided in the research and in the writing of this
section, I owe many thanks. Valuable samples from Australia
were turned over to me by several collectors, especially Dr. D.
Ashton, Mr. P. F. Darlington, Dr. P. J. Darlington, Jr., Dr. C.
P. Haskins, Father J. J. McAreavey, S.J., and Father C. Mer-
covich, S.J. For neotropical material, I must thank Rev. Fr.
T. Borgmeier, O.F.M., and Dr. N. Kusnezov, and for series from
both the Old and New World, Dr. E. S. Ross and Dr. E. O.
Wilson. Dr. J. W. Chapman contributed interesting Philippine
material. Dr. Claude Besuchet loaned important types from
Forel’s Collection in the Museum d’Histoire Naturelle, Geneva,
and Dr. M. R. Smith sent a box of specimens from the United
States National Museum at Washington. The drawings were
mostly done by Mrs. Nancy Buffler under my supervision. Dr.
C. P. Haskins and Dr. E. O. Wilson read and criticized parts of

224

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

the manuscript and added valuable biological observations, but
this is not meant to imply their responsibility for any of the
views expressed in this paper.

LIST OF REFERENCES CITED

BINGHAM, C. T.

1903.

BORGMEIER,
1957.

3ROWN, W.
1949.

Brown, W.
1950.

Brown, W.
1956.
1960.

CLARK, J.
1928.

1934a.

1934b.

CREIGHTON,
1940.

Fauna of British India, Hymenoptera, 2. London.

ANs
Myrmecologische Studien, I. An. Acad. Brasil. Ciene., 29:103-
128.

Ibs die

A new American Amblyopone, with notes on the genus. Psyche,
56:81-88.

The status of some Australian Amblyopone species. Ent. News,
63:265-267.

Composition of the ant tribe Typhlomyrmicini. Psyche, 59:104.
Characters and synonymies among the genera of ants. Breviora,
Mus. Comp. Zool., no. 11:1-13 (ef. 11-13).

Contributions toward a reclassification of the Formicidae. II.
Tribe Ectatommini. Bull. Mus. Comp. Zool., 118:175-362.

. A review of the ants of New Zealand. Acta Hymenopterologica,

Fukuoka, 1:1-50.

L., JR. and W. L. NuTTING
Wing venation and the phylogeny of the Formicidae. Trans.
Amer. Ent. Soc., 75:113-132, 2 pls.

L., JR. and E. O. WILSON
Character displacement. Syst. Zool., 5:49-64.
The evolution of the dacetine ants. (In Press.)

Australian Formicidae. Jour. R. Soc. W. Australia, 14:29-41,
ipl

New Australian ants. Mem. Nat. Mus. Vict., Melbourne, 8:21-47,
2 pls.

Ants from the Otway Ranges. Mem. Nat. Mus. Vict., Melbourne,
8:48-73, 1 pl.

W.S.
Revision of the forms of Stigmatomma pallipes. Amer. Mus.
Novit., no. 1079:1-8.

BROWN : ANT TRIBE AMBLYOPONINI Dee,

EMERY, C.

1895. Descriptions de quelques fourmis nouvelles d’Australie. Ann.
Soe. ent. Belg., 39: 345-358 (ef. 349-351).

1897. Formicidarum. ...in Nova Guinea. ... collegit L. Biré. Ter-
mészetr. Fiiz., 20:571-599, pls. 14, 15.

1911. Ponerinae. Gen. Insect., 118.

1916. Hymenoptera — Formicidae. Jn, Fauna Ent. Ital. Bull. Soe.
Ent. Ital., 47:79-201.

FOREL, A.

1900. Les formicides de 1’Empire des Indes et de Ceylan. Part VI.
Jour. Bombay Nat. Hist. Soc., 13:52-65.

1909. Ameisen aus Guatemala usw., Paraguay und Argentinien.
Deutsch. ent. Zeitschr., pp. 239-269.

1910. Formicides australiens recus de MM. Froggatt et Rowland
Turner. Rev. Suisse Zool., 18:1-94 (ef. 1-2).

1915. Ameisen. Jn, Results of Dr. E. Mjobergs Swedish Scientific
Expeditions to Australia 1910-1913. Ark. f. Zool., 9(16) :1-119,
3 pls.

HASKINS, C. P.
1928. Notes on the behavior and habits of Stigmatomma pallipes
Haldeman. Jour. N. Y. Ent. Soc., 36:179-184.

Haskins, C. P. and E. F. HASKINS
1951. Note on the method of colony foundation of the ponerine ant
Amblyopone australis Erichson. Amer. Mid]. Nat., 45:432-445.

KaARAWAJEW, W.
1935. Neue Ameisen aus dem Indo-Australischen Gebiet, nebst Revi-
sion einiger Formen. Treubia, 15:57-117, 1 pl. (ef. pp. 57-59).

KusNEZOV, N.

1955. Zwei neue Ameisengattungen aus Tucumaén (Argentinien). Zool.
Anz., 154:268-277.

Mann, W. M.
1919. The ants of the British Solomon Islands. Bull. Mus. Comp. Zool.,
63:273-391, 2 pls. (cf. pp. 279-281).

Mayr, G.
1866. Myrmecologische Beitriige. Sitzb. Akad. Wiss. Wien, 53:484-517,
1 pl. (ef. pp. 503-505).
1887. Siidamerikanische Formiciden. Verh. zool.-bot. Ges. Wien,
37:511-632 (cf. 546-549).

226

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

MENOZzZI, C.

1929.

RoGeEr, J.

1859.

1861.

1862.

Revisione delle formiche del genere Mystrium Roger. Zool. Anz.,
82:518-536.

Beitrige zur Kenntniss der Ameisenfauna der Mittelmeerliinder.
Berlin. ent. Zeitschr., 3:225-259 (ef. 250, 251).

Die Ponera-artigen Ameisen (Schluss).
5:1-54 (ef. 49-53).

Einige neue exotische Ameisen-Gattungen und Arten.
ent. Zeitschr., 6:233-254, pl. 1 (ef. 245-248).

Berlin ent. Zeitschr.,

Berlin.

SANTSCHI, F.

1912.

1914

1915.

Quelques fourmis de 1’Amerique australe.
20:519-534 (cf. 519-521).

Formicides de ]’Afrique occidentale et australe. Boll. Lab. Zool.
Portici, 8:309-385 (ef. 310-312).

Nouvelles fourmis d’Algérie, Tunisie et Syrie. Bull. Soc. Hist.
Nat. Afr. N., 7:54-63 (cf. 54-55).

Rev. Suisse Zool.,

WHEELER, G. C. and J. WHEELER

1952.

The ant larvae of the subfamily Ponerinae iPenas IE
Midl. Nat., 48:111-144. Part II. Ibid., 48:604-672.

Amer.

WHEELER, W. M.

1915.

1916.

Some additions to the North American ant-fauna. Bull. Amer.
Mus. Nat. Hist., 34:389-421 (cf. 389).

The Australian ants of the genus Onychomyrmec.
Comp. Zool., 60:45-54, 2 pls.

Ants of the American Museum Congo Expedition. Parts VU,
VIII, IX. Bull. Amer. Mus. Nat. Hist., 45:631-1055.

Ants of the genus Amblyopone Erichson.
Arts Sei., 62:1-29.

Bull. Mus.

Proce. Amer. Acad.

WHEELER, W. M. and J. W. CHAPMAN

1925.

The ants of the Philippine Islands. Part I, Dorylinae and
Ponerinae. Philippine Jour. Sci., 28:47-73, 2 pls.

WHELDEN, R. M.

1958.

Notes on the anatomy of the Formicidae. I.
pallipes (Haldeman).

Stigmatomma
Jour. N. Y. Ent. Soc., 65:1-21, 2 pls.

WHELDEN, R. M. and C. P. HASKINS

1954.

Cytological and histological studies on the Formicidae. I.
Chromosome morphology and the problem of sex determination.

Ann. Ent. Soe. Amer., 46:579-598 (cf. pl. 5).

BROWN: ANT TRIBE AMBLYOPONINI 227

WILSON, E. O.
1958a. Studies on the ant fauna of Melanesia. I. The tribe Lepto-
genyini. II. The tribes Amblyoponini and Platythyreini. Bull.
Mus. Comp. Zool., 118:101-153.
1958b. The beginnings of nomadie and group-predatory behavior in the
ponerine ants. Evolution, 12:24-31.

228 BULLETIN :

MUSEUM OF COMPARATIVE ZOOLOGY

INDEX

Included here are names of ant species, genera and higher categories men-
tioned in the body of this paper. Names cited in the Appendix and captions
to figures are excluded where reference is made to them through bracketed
numbers at the primary (species-list) reference. The pages of the primary
references are given in boldface below. Abbreviations for generic names are

as follows:

aberrans, A., 155, 167

Aenictus, 152, 180

aethiopicea, P., 177

amabilis, P., 177

Amblyopone, 145, 146, 148, 149, 150-
153, 155-169, 170-179 181, 188, 198,
218

Amblyoponini, 145-153, 156, 181

amblyops, A., 167

antillana, P., 177

arizonensis, A., 169

armigera, A., 167, 192

Arotropus, 155

australis, A., 150-152,
164, 167, 170-173

155, 159, 162-

bakeri, Myop., 213, 215
barretoi, A., 168

beecarii, Myop., 214, 215
bellii, A., 165, 167
bierigi, A., 167, 192
binodosus, A., 155, 169
brocha, P., 174, 176, 177
bruchi, P., 178

bruni, A., 165, 167
bugnioni, Myop., 214, 215

camillae, Myst., 170

castanea, Myop., 149, 152,
213-215

celata, A., 149, 155, 168, 176

cephalotes, A., 167

Cerapachyinae, 149

chilensis, A., 168, 192

170, 173,

A. = Amblyopone, Myop. = Myopopone, Myst. = Mystrium,
. = Onychomyrmex, P. = Prionopelta.

cehurehilli, P., 177
clarki, A., 155, 168

degenerata, A., 158, 161, 164, 165, 168,
181, 191

denticulata, A., 155, 168

descarpentriesi, P., 177

doddi, O., 180, 181

Dorylinae, 149, 178

Dorylozelus, 146, 154, 177, 181-182

Dorylus, 182

Eectatommini, 145, 149, 153, 173
egregia, A., 156, 164, 168
elongata, A., 164, 168, 192
emeryi, A., 168

Ericapelta, 156

Examblyopone, 173

exigua, A., 168

fallax, Myst., 170

feai, A., 168

ferruginea, A., 164, 168

fortis, A., 167

foveolata, A., 167

Fulakora, 152, 155, 164, 174, 176, 188,
190, 197, 207, 209

gheorghieffi, A., 168
gingivalis, A., 165, 168
glauerti, A., 156, 166, 168, 179
gracilicornis, A., 168

gracilis, A., 168

BROWN: ANT TRIBE AMBLYOPONINI 299

hackeri, A., 168
hedleyi, O., 178, 180, 181
howensis, A., 167

impressifrons, A., 158, 168

javana, Myst., 170
juergi, A., 168

kraepelini, P., 175, 177

laevidens, A., 167

leai, A., 168

Leptogenys, 180
Lithomyrmex, 156, 166, 179
longidens, A., 168

lucida, A., 168

luzonica, A., 168

maculata, A., 167

maculata, Myop., 170, 214, 215
majuscula, P., 173, 175, 176, 177
mandibularis, A., 168

mayri, P., 178

mereovichi, A., 165, 168
michaelseni, A., 168

minima, Paraprionopelta, 146, 181
minor, A., 167, 168

minuta, A., 168

mjoebergi, Dorylozelus, 146, 181-182
mjoebergi, O., 149, 179, 180, 181
moesaryi, P., 178

modesta, P., 178

moelleri, Myop., 214, 215
Monomorium (Notomyrmex), 166
monrosi, A., 168, 192

montigena, A., 169

mutica, A., 155, 156, 165, 163
Myopopone, 145, 154, 176-173, 213-215
Myrmecia, 149, 150

Myrmica, 148

mysticum, Myst., 169, 170
mystriops, A., 169, 191

Mystrium, 145, 148, 151, 153, 155, 169-

170, 188

nana, A., 167
Neoamblyopone, 155
nidifex, Strumigenys, 152
norfolkensis, A., 167
normandi, A., 164, 169

oberthueri, Myst., 170

obseura, A., 167

Onychomyrmex, 145, 146, 148, 149,
150, 152, 155, 166, 178-181

Onychomyrmicini, 145

opaca, P., 177, 178

oregonensis, A., 169, 192

orizabana, A., 169, 192

pallens, A., 167

pallipes, A., 148, 149, 150,
169, 192

paranensis, A., 168

Paraprionopelta, 146, 161, 181

picea, Myop., 215

Ponera, 182

Ponerinae, 146, 149, 178

Ponerini, 145, 182

poultoni, P., 177

Prionopelta, 145, 148, 149, 151, 153,
154, 173-178, 182

Protamblyopone, 155

proxima, Myop., 213, 215

punctulata, A., 189

punctulata, P., 173, 178

166, 167,

quadrata, A., 169
queenslandica, A., 167

reclinata, A., 158, 165, 166, 167, 169
Renea, 173

rogeri, Myst., 169, 170

rossi, Myop., 215

rothneyi, A., 165, 169

rufula, Myop., 214, 215

santschii, A., 169
saundersi, A., 158, 165, 169

230 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

serrata, A., 169 testacea, P., 177

silvestrii, A., 166, 169 trigonignatha, A., 169, 192
silvestrii, Myst., 170
similis, Myop., 215
simillima, P., 177
smithi, A., 169, 174
smithi, Myop., 215

voeltzkowi, Myst., 170

wheeleri, A., 169
williamsi, A., 168
wilsoni, A., 169

stadelmanni, Myst., 170 Wollaston, Myope2is

Stigmatomma, 145, 155, 164, 165, 197,

209 Xymmer, 155, 164, 165
striatifrons, Myop., 214, 215
subterranea, A., 169 zwaluwenburgi, A., 169

PUBLIGATIONS ISSUED BY OR IN CONNECTION
WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

BULLETIN (octavo) 1863 — The current volume is Vol. 122.

BreEviorA (octavo) 1952 — No. 124 is current.

Memorrs (quarto) 1864-1938 — Publication was terminated with
Vol. 55.

JOHNSONIA (quarto) 1941 — A publeation of the Department of
Mollusks. Vol. 3, no. 39 is current.

OccASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo)
1945 — Vol. 2, no. 23 is current.

PROCEEDINGS OF THE NEW ENGLAND ZooLoGicaAL CLUB (octavo)
1899-1948 — Published in connection with the Museum. Publication
terminated with Vol. 24.

The continuing publications are issued at irregular intervals in num-
bers which may be purchased separately. Prices and lists may be
obtained on application to the Director of the Museum of Comparative
Zoology, Cambridge 38, Massachusetts.

Of the Peters ‘‘Cheek List of Birds of the World,’’ volumes 1-3 are
out of print; volumes 4 and 6 may be obtained from the Harvard Uni-
versity Press; volumes 5 and 7 are sold by the Museum, and future
volumes will be published under Museum auspices.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE

Vor 122, No. 5

LAND SHELLS OF NAVASSA ISLAND, WEST INDIES

By Ruts D. TURNER

Wit SEVEN PLATES

CAMBRIDGE, MASS., U.S.A.
PRN DED) FOR. TH EMS EM

Marcu, 1960

No. 5 — Land Shells of Navassa Island, West Indies
By Rutu D. TuRNER

No study of the mollusks of Navassa Island has been made
since 1866 when only a few dead specimens were available. This
paper is based on a large series of preserved material which has
made it possible to determine the relationships of these mollusks
to those of other islands in the West Indies.

Navassa is a small, isolated island located about 40 miles west
of the Tiburon Peninsula of Haiti and about 85 miles northeast
of Morant Point, Jamaica, at north latitude 18° 25,’ west lonei-
tude 75° 05’. It is an elevated coral reef of about 1.5 square
miles in extent, one and one-fourth miles at its longest point, and
with an elevation of 250 feet at its highest point. It is pear-
shaped in outline and in profile is said to look like a huge battle-
ship with the lighthouse for a mast, or a gigantic straw hat with
a low, flat crown. The island rises abruptly from fairly deep
water, the depth at the shoreline averaging about 12 fathoms.
On all except the north coast there is a deep undereut at the
water level. The sea cliffs average about 40 feet in height above
which there is a bench of some 100 yards in width and then a
steep slope which leads to the rather flat crown. The width of
the bench varies somewhat, being very narrow at the northwest
point but widening gradually toward the south and east. The
island is virtually inaccessible from the sea and today is
‘boarded’ by climbing a chain ladder which hanes from the
cliffs at Lulu Bay on the lee side. The entire surface of the
limestone island is pitted with holes varying from five to more
than forty feet in depth and from a foot to several yards in
diameter. As a result, though the island receives a fair rain-
fall, it is physiologically dry. There are no ponds and frequent
drilling of wells has never produced any fresh water, for all rain
water is rapidly drained off by the numerous fissures and caves.
This type of terrain makes walking on the island most difficult
and W. J. Clench (1930) wrote, ‘‘Most of our excursions con-
sisted of a long series of jumps from the rim of one hole to that
of another,’’ and he says that it could take as long as four
hours to walk a mile.

234. BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

The island of Navassa was practically unknown until 1857
when Captain E. K. Cooper of Baltimore, Maryland, and Peter
Dunean discovered the phosphatic guano deposits while making
soundings around the island. They took possession of the un-
inhabited island and worked the deposits for a time before the
formation in September 1864 of the Navassa Phosphate Company
of New York, of which E. K. Cooper was a member of the
executive council. The reports to the company in 1864 of Dr.
G. A. Liebig, chemist, and Augustus H. Fick, mining engineer,
eave an excellent picture of the rich phosphate deposits and of
the condition of the the island at that time. In November 1865
Kugene Gaussoin, mining engineer and metallurgist, sailed from
Baltimore on the company brig with the vice-president of the
company and Peter Dunean, captain, to visit the island. The
purpose of his visit was to advise the company on means of
finding water, improving mining techniques, shipping facilities
and other matters. In his report, he wrote that at that time there
were ‘‘30 white men, officers and mechanics, and 180 black
laborers’’ on the island and all food, water and other supplies
had to be shipped in from great distances — a formidable task!
However, the company was reasonably successful for a time,
though working under. great difficulties, but in 1898, during
the Spanish-American war, the company failed and the island
was abandoned. During World War I a detachment of marines
was stationed there for a time. With the building of the
Panama Canal, the amount of shipping through the Windward
Passage increased greatly and it became most important to
have a lighthouse on the island to protect shipping from this
treacherous rock rising out of the sea. The hghthouse was com-
pleted in October 1917 and from that time until 1929, when the
light was made automatic, the island was inhabited by the three
families who attended the light. Again during World War II
the island was garrisoned by American troops. Today, the island
is uninhabited and is seldom visited except by members of the
U.S. Coast Guard who service the light, and occasional hunters
who go there to shoot wild goats — descendants of those left by
the lighthouse keepers.

Though few biologists have ever visited the island, the flora
and fauna are now quite well known and are remarkable for the

TURNER: LAND SHELLS OF NAVASSA ISLAND 235

number of endemic species in several phyla which occur on such
a small area. In December 1929, W. J. Clench visited the island
and later wrote, ‘‘It is a bird island and thousands of boobies
were nesting all over the place, in the small trees, on the low
bushes and even among the low vegetation. A few frigate birds
were also nesting and to see these beautiful birds close at hand
was a royal experience.’? Wetmore and Swales (1931) listed
20 species of birds from Navassa of which one, the Navassa
eround dove, is peculiar to the island.

In July 1917, R. H. Beck collected there for the American
Museum of Natural History, and from this collection of reptiles
K. P. Schmidt (1919, 1921) described a new genus and five
new species of lizards, all endemic. He listed 13 species from
the island and related them to species found in Jamaica, His-
paniola and Cuba. Cope (1886) described an iguana, probably
collected by Gaussoin, but this large lizard has apparently been
extinet for a long time; it was undoubtedly exterminated by the
laborers working the phosphate deposits who would have used
it for food. Proctor (1959) states that ‘‘no less than 10 endemic
species of lizards have been reported.”’

In an excellent paper on the flora of Navassa Island, E. L.
Ekman (1929) wrote, ‘‘needless to state, this vegetation, for all
its freshness in the rainy season, is well suited to survive even
the severest of droughts. The roots of the trees like the Pieus
penetrate into the rock fissures to astonishing depths. The
Metopium sheds, if necessary, all its leaves in winter. The
savanna plants belong to different types of xerophytes . The cacti
are succulents, the grasses and the sedges survive by means of
their drought- and fire-resisting rhizomes, and the weeds defy
the dry season here in the same way as they do the winter in
the north. The unproportionately great number of annuals on
Navassa, ¢. 30, or about 33 per cent of the total number of plants,
bear mute witness to the efficacy of their means of protection.”’
An interesting and rather surprising thing about the flora
of Navassa is the almost complete lack of halophytes, for only
two species can be classed as such. Ekman listed 102 species otf
vascular plants for the island, 44 of which are probably imdi-
genous and 8 species and two varieties are endemic.

236 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Only three species of land shells are known from this small
island and these were all described by George Tryon in 1866.
He had received a small series collected by Eugene Gaussoin, the
nmuning engineer mentioned above. His collecting, however, was
casual and all of the material was dead and worn. As a result,
the systematie position and island relationship of these species
have remained uncertain.

In December 1929 a Harvard University Expedition, led by
William J. Clench accompanied by William E. Schevill and
Harald A. Rehder, visited the island. They landed December
29, and spent two weeks of intensive collecting of all groups of
plants and animals. This was during the dry season and the land
shells were probably not as active as at other times so that it was
necessary ‘‘to move and turn over several tons of rock before
sufficient numbers were secured for study.’? A single pupillid
was found the day they arrived while they were laboriously
earrying their supplies up the steep slope to the hghthouse
quarters where they were going to live. Not having a proper
container available the minute specimen was placed in a match
box and unfortunately it fell out and was lost. Continued in-
tensive search failed to turn up another specimen. It is possible
that collecting on the island during the rainy season might add
to the known mollusean fauna. As a result of this collecting we
are now able to show that two of the three species of land shells,
all of which are endemic to the island, are related to species in
Jamaica, one to Haiti.

Considering the unreceptive shores of Navassa, its undercut and
precipitous rocky cliffs, it would be virtually impossible for any
plant or animal, with the possible exception of lizards, to reach
the island by rafting. Consequently, whatever the population (ex-
cluding those species introduced by man), it seems safe to con-
elude that the original species were carried there by hurricanes.
The large number of endemie species on the island would suggest
that it is old and has been isolated for a very long time. Versey
(in Proctor, 1959), reporting on the Foraminifera, stated that
‘‘the micro-facies appear to be very similar to that encountered
in the Pliocene Coastal Limestones of the North Coast of Jamaica.
The only foraminifera present are Operculinoides and Hetero-
stegina, both of which range from Eocene to Recent. The age
of the limestone probably lies within the range Miocene-Recent.”’

TURNER: LAND SHELLS OF NAVASSA ISLAND 231

It is also possible, of course, that these species have differenti-
ated rapidly as may happen in small isolated populations. What-
ever the factor involved — long isolation or rapid evolution —
the molluscan species are all well differentiated.

Since Tryon had only a few dead and worn specimens at the
time he described these species he could not give the range of
variation nor even describe them completely. Consequently they
are redescribed here.

HEUTROCHATELLA CIRCUMLINEATA Tryon
Plate 1, figs: 1-3; Plate 7, figs. 3-5

Helicina circumlineata Tryon 1866, American Journal of Conchology, vol.
2, p. 305, plate 20, fig. 13 (Navassa Island).

Description. Shell trochoid, reaching 11.5 mm. in height,
solid, heavy, imperforate, pale ivory to salmon in color and
sculptured with spiral cords. Whorls 6 and slightly convex.
Suture shehtly impressed. Spire conic and produced at an angle
of about 77°. Columella short, nearly straight and curving into
the basal lip. Aperture oval and cast at an angle of about 55°
from the base. Outer lip simple, not reflected. Inner lip consist-
ing of a thinly glazed area on the body whorl. Seulpture con-
sisting of numerous, fine, evenly spaced spiral threads. There
are about 32 on the body whorl, those above the periphery being
slightly coarser and more widely spaced than those on the base.
Color ranging from a pale ivory to salmon, the body whorl
usually lighter in color than the earlier whorls. Parietal area
and lip white. Interior of the aperture a medium to rather
deep orange. Embryonie whorls 114, very small, smooth and
white. Operculum subquadrate with a thin chitinous base and a
well developed caleareous outer surface which has a thickened
ridge on the parietal margin. Nucleus near the parietal margin,
erowth lines concentric. Color of the operculum ranging from
orange to salmon, becoming lighter as the calcareous deposit
thickens near the parietal margin.

height width
i) 9.5 mm. all adult specimens
10 SILI)
10.5 aL)

ial fal

238 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Types. The holotype is in the Academy of Natural Sciences,
Philadelphia, from Navassa Island, collected by Mr. Eugene
Gaussoin.

Remarks. Tryon described this species from two dead, bleached
specimens and he did not see the operculum. The present series
shows a variation in color from light ivory to salmon. The
operculum is similar to that of other species placed in this genus
and the radula fits well within the range of variation for the
genus Hutrochatella as given by H. B. Baker (1922). However,
it seems to combine the characters of his subgenus Troschelviana
(type species Helicina erythraea Sowerby) and the subgenus
Kutrochatella (type species, Helicina pulchella Gray). The lat-
eral teeth of H. circumlineata have large, well developed cusps
as in erythraea but the marginals are very numerous as in pul-
chella. In addition, the large fourth lateral (called the T-lateral
by Baker) of circumlineata bears a number of well developed
cusps as in erythraca. This tooth in the subgenus Lutrochatella
as given by Baker is smooth or nearly so.

This species seems to be most closely related to Hutrochatella
costata Sowerby from the vicinity of St. Ann’s Bay, Jamaica, but
differs in being considerably larger though less coarsely sculp-
tured. The radulae of circumlineata and costata are also nearly
identical, both species having numerous marginals and denticu-
late laterals.

In his field notes, Dr. Clench stated that circumlineata was
found ‘‘mostly under stones and flat Hmestone slabs. They were
found in quite barren areas under the top layer of loose flattish
stones and were moderately abundant.”’

CHONDROPOMA (CHONDROPOMA) NAVASSENSE Tryon
Plate 1, figs. 4-7 ; Plate 2; Plate 7, figs. 7-9

Chondropoma navassense Tryon 1866, American Journal of Conchology,
vol. 2, p. 305, pl. 20, fig. 12 (Navassa Island).
Chondropoma (Chondropoma) navassense Tryon. Henderson and Bartsch
1920, Proceedings United States National Museum, vol. 58, p. 62.
Description. Shell reaching 20 mm. in length (truncated
specimen) rather thin in structure but strong and finely sculp-
tured. Spire extended, truncate except in very young specimens,

TURNER: LAND SHELLS OF NAVASSA ISLAND 239

and produced at an angle of about 32°. Umbilicus small, ex-
tending to the embryonic whorls, and nearly covered by the
reflected lip in adult specimens. Color a uniform dull yellowish
brown. Interior of aperture a shiny yellow-brown, lip white.
Whorls remaining 4 to 5, and moderately convex. Suture mod-
erately impressed with, in some specimens, a shallow channel
on the body whorl which increases slightly in width and depth
toward the aperture. Aperture subcireular. Outer lip simple,
very slightly reflected and with a small angular projection in the
region of the anal canal. Inner lip narrow, simple, not appressed
against the body whorl and partially covering the umbilicus.
Axial sculpture consisting of numerous, fine, more or less evenly
spaced ridges. Spiral sculpture consisting of evenly spaced
threads of about the same strength giving a rather uniform
reticulated pattern. Small nodules are produced where the ridges
and threads cross. Umbilicus bordered by 3 or 4 prominent
spiral cords. Opereulum subcireular, paucispiral and with a
thin granular caleareous deposit on the outer surface.

height width
20 20 mm. adult, truncated specimen
15.5 10 ce oe oe

Type. The holotype is in the Academy of Natural Sciences,
Philadelphia, collected by Eugene Gaussoin. The type locality is
Navassa Island. Paratype in the Museum of Comparative Zool-
ogy no. 78164, received from Mr. Tryon.

Remarks. The original description of this species was based
upon five dead specimens; in fact, Tryon was not sure of the
color of the shell and he did not see the operculum. To my
knowledge no additional collections of this species were made
until Navassa was visited by W. J. Clench, W. E. Schevill and
Hl. A. Rehder in 1929-30. In his field notes Dr. Clench wrote
that these snails were ‘‘most abundant under rocks, especially
200 to 300 feet N.W. of the lighthouse. They were quite abund-
ant in the tufts or clumps of grass about the roots and adjacent
edges of stones. On the S.E. side of the island many were found
climbing trees to a height of 5 to 6 feet and aestivating in pro-
tected nooks about the roots or on the branches.’’

240 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Asa result of this collecting we are now able to figure the oper-
culum and the radula in addition to gaining an understanding
of the range of variation within the species. On the basis of the
rather large series collected, this species appears to be remark-
ably uniform in color for this group, there being no evidence
of any eolor pattern, and all specimens being a uniform yellow-
ish brown. The variation in size and proportions is also within
rather narrow limits, the greatest variation coming in the amount
of truncation.

This species is probably most closely related to Chondropoma
(Chondropoma) brownianum Weinland from Gonave Island,
Haiti. However, it is readily differentiated by its uniform color,
the striking but variable color pattern of brownianum consisting
of axial bands of red brown on a pale buff ground color. These
bands may be entire, interrupted or in the form of triangular
spots, and the intensity of the color may also vary considerably.
Chondropoma (C.) molense Bartsch from the Le Mole River
on the northern peninsula of Haiti and Chondropoma (C.) mon-
talbense Bartsch from Coteaux on the southern peninsula are
also closely related. In fact, these last two species seem to be
at best only subspecies of brownianum. The size, sculpture and
shape of the aperture of all are very similar.

The radula of navassense shown in Plate 7, figure 9 is similar
to that shown by Hidalgo (1947) for species in the related genus
Chondrothyra, except that the denticles on the second laterals are
well developed in navassense whereas on the second laterals of
Chondrothyra the denticles are lacking entirely or are very small.
The radula of navassense is almost identical to that shown by
Hf. B. Baker (1924) for Tudora.

Chondropoma navassense feeds on the thin coating of grayish
lichens on the bark of the bushes, trees and rocks. Their feeding
tracks are shown, greatly enlarged, on Plate 2.

ZAPHYSEMA (ZAPHYSEMA) GAUSSOINI Tryon
Plate 3, figs. 1-3; Plate 4; Plate 7, figs. 2, 6

Helix gaussoini Tryon 1866, American Journal of Conchology, vol. 2, p. 304,
pl. 20, fig. 11 (Navassa Island).

Helix (Coryda) gaussoini Tryon. Pilsbry 1889, Manual of Conchology,
ser. 2, vol. 5, p. 47.

TURNER: LAND SHELLS OF NAVASSA ISLAND 24]

Hemitrochus gaussoini Tryon. Pilsbry 1889, ibid., ser. 2, vol. 5, p. 197.
Cepolis (Dialeuca) gaussoini Tryon. Pilsbry 1895, ibid., ser. 2, vol. 9, p. 183.
Sagda gaussoini Tryon. Clench 1945. Mollusea, vol. 1, no. 5, p. 65,

Description. Shell depressed-globose, reaching 11 mm. in
greatest diameter, imperforate, smooth and a lght straw-yellow
in color. Whorls 514, moderately convex and increasing rapidly,
the body whorl being large and somewhat inflated. Suture
moderately impressed. Spire depressed and produced at an angle
of about 113°. Columella short, slightly thickened, curved and
merging into the basal portion of the outer lip. Aperture oval
and east at an angle of 58° from the base. Outer lip thin, simple
and not reflected. Inner lip consisting of a very thin glaze on
the body whorl. Sculpture consisting only of indistinct growth
ridges. Color of shell beneath the periostracum hght ivory.
Periostracum thin, a light straw-yellow in color and persistent.
Columella area white. Embryonic whorls 114, white, smooth
and shining.

height greatest diameter
9 11 mm.
8.5 10

Type. The holotype is in the Academy of Natural Sciences,
Philadelphia, from Navassa Island, Mr. Eugene Gaussoin, col-
lector.

Remarks. Tryon described this species from a single dead
specimen, and in his original description related it to Helix
melanocephala Gundlach from Cuba. Over the years its syste-
matic position has been quite uncertain. As noted in the syn-
onymy above, Pilsbry placed this species in three different groups
all belonging in the family Fruticicolidae [Cepolidae]. Fortu-
nately the Harvard Expedition found this species alive and col-
lected a large series of them. They were carefully relaxed and
preserved so that it has been possible to make a study of the
anatomy and so determine its proper systematic position. There
is now no question that gaussoini belongs in the family Sagdidae ;
W. J. Clench placed it in the genus Sagda on the basis of the shell
texture and structure. A study of the reproductive anatomy
has shown that it belongs to the genus Zaphysema which is also

242 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

in the Sagdidae but which until now was not known to oceur
outside of Jamaica. It was also fortunate that preserved speci-
mens of Z. tencrrima C. B. Adams, the type species of Zaphy-
sema Pilsbry, were available for dissection so that comparisons
could be made and the figures of both could be published to-
evether. The lectotype of Helix tenerrima C. B. Adams is figured
on Plate 3, figure 6.

The illustrations of the anatomy of the reproductive system
of Z. tenerrima C. B. Adams from Jamaica given here differ
somewhat from that given by Pilsbry (1892, pl. 13, fig. F).
However, Pilsbry had only a single specimen and was never
able to check his work. He later states that this was one of his
first land shell dissections and he obviously was not satisfied
with it. So far as I know no other figure of this species has been
published. Fortunately, I was able to make dissections of five
specimens in all stages of reproduction of which three stages are
figured.

The anatomy of the reproductive system of Z. gaussoini is very
close to that of Z. tenerrima. It differs mainly in having a pro-
portionately very much larger penial appendix, in having only
three lobes in the ovotestis, in having the penial retractor muscle
inserted near the end of the epiphallus fairly close to the opening
of the vas deferens rather than near the base as in tenerrima.
In both species the middle portion of the spermathecal duct is
swollen to form a secondary spermatheca. From its side near
the apical end arises the slender duct leading to the spermatheea.
The spermathecal retractor is inserted very close to the end of
the secondary spermatheca. In its normal position the entire
spermatheeal complex is interwoven with the prostate, uterus,
and free oviduct; the spermathecal sae lies at the base of the
albumen gland and the retractor muscle is attached to the sheath
of the gland. In young or non-breeding specimens the penial
appendix lies above the entire visceral mass and can be clearly
seen through the body wall when the mantle is removed. In
specimens full of eggs or young snails the appendix is pushed
inward and the greatly enlarged uterus comes to le on top of it.
In one of the specimens of tenerrima dissected, the uterus was
distended with 53 fully developed eggs with white, calcareous
shells. The specimen figured in Plate 6, figure 1 contained over

TURNER: LAND SHELLS OF NAVASSA ISLAND 243

40 young. Most of the specimens of gauwssoini dissected were
apparently young and none had more than one ege@ in the uterus.
All of them were aestivating at the time they were collected. It
is probable that fertilization takes place before aestivation so
that the young are developed in the uterus during the dry
season, hatching immediately when the adults emerge and thus
having the benefit of the entire rainy season for growth.

Three aestivating colonies of gaussoint were found, each con-
taining from 60 to 120 specimens. They were 114 to 2 feet
below the surface under loosely piled rocks, and the specimens
were clustered together forming irregular balls. To my knowl-
edge such a habit has not been recorded for any other member
of the Sagdidae.

On the basis of shell characters as well as anatomy, Z. gaussoini
seems to be most closely related to Z. tenerrima, and the radulae
of the two species as shown on Plate 7, figures 1 and 2 are very
similar.

REFERENCES

Baker, H. B.

1922. Notes on the Radulae of the Helicinidae. Proce. Acad. Nat. Sei.
Philadelphia, vol. 74, pp. 29-67, pls. 3-7.

1924. Land and Freshwater Molluses of the Dutch Leeward Islands.
Oce. Pap. Mus. Zool. Uniy. Michigan, no. 152, pp. 1-158,
pls. 1-21.

1934- Jamaican Land Shells. Part I, Nautilus, vol. 48, no. 1, pp.

1936. 6-14; Part II, ibid., no. 2, pp. 60-67; Part III, ibid., no. 3, pp.
83-88; Part IV, ibid., no. 4, pp. 135-140; Part V, ibid.. vol. 49,
no. 1, pp. 21-27; Part Vi, tbid., pp. 52-58.

CLENCH, W. J.
1930. The Harvard Expedition to Navassa Island. Harvard Alumni
Bull., vol. 32, no. 24, pp. 684-687.
1945. Harvard Navassa Expedition. Mollusea, vol. 1, no. 5, pp. 64-66.

Corr, E. D.
1885. On the Large Iguanas of the Greater Antilles. American Nat-
uralist, 19, pp. 1005-1006.
1886. On the Species of Iguaninae. Proc. Amer. Philos. Soc., vol. 23,
pp. 261-271.

244 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

EKMAN, E. L.
1929. Plants of Navassa Island, West Indies. Arkiv fér Botanik.
vol, 22A, no. 16, pp. 1-12, pls. 1-2.

Fick, Aueustus H.
1864. Report of the Phosphatic Mineral of Navassa Island, W. I.
Navassa Phosphate Company of New York Report to Stock-
holders, J. B. Rose & Co., Baltimore, pp. 11-16.

GAUSSOIN, EUGENE
1866. Memoir on the Island of Navassa (a report to the Navassa
Phosphate Company). J. B. Rose & Co., Baltimore, pp. 1-32.

Hipaueo, A. A.
1947. Estudio radular del genero Chondrothyra. Rey. Soe. Malaco-
logica ‘Carlos de la Torre,’ vol. 5, no. 1, pp. 6-8, figs. 1-2.

LigBic, G. A.
1864. Report to the Navassa Phosphate Company. Navassa Phosphate
Company of New York Report to Stockholders, J. B. Rose
& Co., Baltimore, pp. 1-10, 2 pls.

Piussry, H. A.
1892. On the Anatomy of Sagda, Cystiocopsis, Aegista and Dentel-
laria. Proe. Acad. Nat. Sei. Philadelphia, 1892, pp. 213-215,
pl. 13.

Proctor, G. R.
1959. Observations of Navassa Island (with Appendices by L. J. Chubb
and D. J. Burns, H. R. Versey and J. B. Williams). Geonotes,
Quarterly Journal of the Jamaica Group of the Geologists’
Association, London, vol. 2, no. 2, pp. 49-54.

PUTNAM, GEORGE R.
1918. An Important New Guide for Shipping — Navassa Light.
Nat. Geogr. Mag., vol. 34, pp. 401-406, 4 pls.

ScHMIptT, K. P.
1919. Deseriptions of New Amphibians and Reptiles from Santo
Domingo and Navassa. Bull. Amer. Mus. Nat. Hist., vol. 41,
pp. 519-525,
1921. The Herpetology of Navassa Island. Bull. Amer. Mus. Nat.

Hist., vol. 44, pp. 555-559, pls. 25-26, text figs. 1-5.
TRYON, G. W.
1866. On the Terrestrial Mollusca of the Guano Island of Navassa.
Amer. Jour. Conchology, vol. 2, pp. 304-305.

WetTMoRE, A. AND B. H. SWALES
1931. The Birds of Haiti and the Dominican Republic. Bull. U. 8.
Nat. Mus., vol. 153, 483 pp., 26 pls.

PLATES

Plate 1

Figs. 1-3. EHutrochatella circumlineata Tryon, Navassa Island, West Indies
(45 D-Oe

Figs. 4-7. Chondropoma navassense Tryon, Navassa Island, West Indies.
Fig. 5, Paratype, MCZ no. 90,010 (all 3 X).

Plate 1

Plate 2

Feeding tracks of Chondropoma navassense Tryon (20 X).

In the area covered by this enlargement the snails were progressing from
right to left in upward swinging tracks. When feeding the snail moves its
head forward and upward, seraping off the whitish lichen, then moving its
head slightly to one side, repeats the process so that it produces a small arc.
Having reached as far as possible to the left and right it moves forward
and clears another are. A series of such ares indicates the direction in which
the snail was progressing. Each rectangular mark indicates a single sweep
of the radula.

Plate 3

Figs. 1-38. Zaphysema (Zaphysema) gaussoini Tryon, Navassa Island,
West Indies (3.5 X).

»

Figs. 4-6. Zaphysema (Zaphysema) tenerrima C. B. Adams. Jamaica.
Fig. 6, Lectotype of Helix tenerrima C. B. Adams, MCZ no. 222184 (all 2 X).

WwW

Plate ¢

Plate 4
Zaphysema gaussoini Tryon

Fig. 1. Anatomy of the reproductive system of an adult specimen (with
a fully developed lip on the shell) at the beginning of egg production. The
appendix has been stretched out slightly to show the basal portion. A single
large egg can be seen in the uterus.

Fig. 2. Anatomy of an immature specimen in non-breeding condition to
show differences in the proportions and shape of the various organs as
compared with the adult. The appendix is in its normal position.

spermathecal retractor

| ovid
ok Ky prostate gland
ES Pr seminal vesicle if Z)
ee a vas deferens —
4 @ YH
a Vas ovotestis duct Ca flagellum
4 y °
{ 5 m
albumen gland & SS b accessory
1 in = i; appendix flagellum
ovotestis E> ee | | |
a Pe eis penial retractor | |
1 ; : & iets ae j ee a Se et
atrium €piphallus -

OY

y
TZ

Plate 4

Plate 5
Zaphysema tenerrima C, B. Adams

Fig. 1. Anatomy of the reproductive system of an adult specimen.
The uterus was filled with eggs most of which fell out before the drawing
could be made. There was a total of 53 eggs in the uterus of this
specimen. The vas deferens was broken during the dissection. The appendix
was flattened beneath the enlarged uterus. The prostate gland appears
small in proportion to the greatly enlarged uterus. In all specimens of this
species dissected the coiled portion of the appendix did not cover the basal
portion. The spermatheca is collapsed.

Fig. 2. Anatomy of the reproductive system of an immature, non-breeding
specimen. The ovotestis unfortunately could not be dissected. The prostate
at this stage is nearly as large as the uterus and the spermatheca nearly
cireular in cross-section.

spermatheca—
DD

thecal d
Se ee TAP sets WIP
WSZ, R)

\ WZ,
GA

SE
SS flagellum
NZ accessory flagellum

appendix
epiphallus A

TF

yi yas deferens

ovotestis yon?
vagina }

Plate 5

Plate 6
Zaphysema tenerrima C, B, Adams

Fig. 1. A specimen with fully developed young ready to emerge. The
drawing was made after the removal of the shell and the anterior portion
of the mantle which forms the pulmonary cavity.

Figs. 2-3. Shells of Zaphysema tenerrima at the time of hatching.

intestine

ate 6

|

]

Plate 7
Fig. 1. Radula of Zaphysema tenerrima C, B. Adams.
Fig. 2. Radula of Zaphysema gaussoini Tryon.
Fig. 3. Radula of Eutrochatella circumlineata Tryon.

Figs. 4-5. Operculum of Eutrochatella circumlineata Tryon.
outer surface. Fig. 5, inner surface.

Fig. 6. Jaw of Zaphysema gaussoini Tryon.

Figs. 7-8. Opereulum of Chondropoma navassense Tryon. Fig.

surface. Fig. 8, inner surface.

Fig. 9. Radula of Chondropoma navassense Tryon.

Fig. 4,

7,

outer

TAR

ea ER VA
@ NG ¢

OER gGOR

ee (om mini ae

|
a

oN

=, \\\ |

}

Wj y
Yi, 5

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE

Vor 1227 No.6

A TRANSCRIPTION OF DARWIN’S FIRST NOTEBOOK
ON ““‘TRANSMUTATION OF SPECIES”’

Edited by

PAUL H. BARRETT
Michigan State University

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM

APRIL, 1960

PUBLICATIONS ISSUED BY OR IN CONNECTION
WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

BULLETIN (octavo) 1863 — The current volume is Vol. 122.

BREVIORA (octavo) 1952 — No. 124 is current.

Memoirs (quarto) 1864-1938 — Publication was terminated with
Vol. 55.

JOHNSONIA (quarto) 1941—A publication of the Department of
Mollusks. Vol. 3, no. 39 is current.

OCCASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo)
1945 — Vol. 2, no. 25 is current.

PROCEEDINGS OF THE NEW ENGLAND ZooLoGicAL CLUB (octavo)
1899-1948 — Published in connection with the Museum. Publication
terminated with Vol. 24.

The continuing publications are issued at irregular interva!s in num-
bers which may be purchased separately. Prices and lists may be
obtained on application to the Director of the Museum of Comparative
Zoology, Cambridge 38, Massachusetts.

Of the Peters ‘‘Check List of Birds of the World,’’ volumes 1-3 are
out of print; volumes 4 and 6 may be obtained from the Harvard Uni-
versity Press; volumes 5 and 7 are sold by the Museum, and future
volumes will be published under Museum auspices.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE
Vou. 122, No. 6

A TRANSCRIPTION OF DARWIN’S FIRST NOTEBOOK
ON “‘TRANSMUTATION OF SPECIES’’

Edited by

PAUL H. BARRETT
Michigan State University

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM

‘APRIL, 1960

VW Sei, =e a

7 lise: 7 e | iif ES eh OER

=

No. 6—A Transcription of Darwin’s First Notebook on
“Transmutation of Species’’

Edited by

PAUL H. BARRETT
Michigan State University*

FOREWORD
From July 1837 until July 1839, Darwin filled a number of
notebooks on the subject of evolution. Six of these notebooks are
now kept among the Darwin Papers at the University Library,
Cambridge, England. The library has catalogued these books as
follows:

INDEX TO MSS OF CHARLES DARWIN

Volume Contents
121 ‘“B’’ Notebook dealing with evolution theory.
(GUO emi')
22 ‘“C”’ Notebook dealing with evolution theory.
February-July 1838. (10x17em.)
123 ‘*T)’’ Notebook dealing with evolution theory.

‘July 15th 1838, finished October 2nd.”’
(10x17em. )

124 ‘i’? Notebook dealing with evolution theory.
‘“Minished July 10th 1839.’’ (10x17em.)
125 ““M’’ Notebook dealing with evolution theory.

‘July 15th 1838.’’ ‘‘This book full of metaphys-
ics on morals and speculations on expression.”’
(10x17em. )

126 ‘*N’’ Notebook dealing with evolution theory.
Begun October 2nd 1838. ‘‘Metaphysies and ex-
pression.’’ (10x17em.)

Through the courtesy and assistance of Sir Charles G. Darwin,
the University Library of Cambridge, and the All University Re-
search Committee of Michigan State University, microfilmed

* Department of Natural Science, Contribution 143.

248 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

copies of these notebooks were made available to the present
editor. A transeription of the first notebook ‘‘B’’ has been made
and is presented herein. In making the transcription, Darwin’s
text has been reproduced as faithfully as possible, although pune-
tuation has been added where necessary. In his rough notes,
Darwin often did not use any punctuation at all and used capital
letters indiscriminately. Since he wrote in his most hasty and
elliptical style, some words at first were illegible. After reading
the collateral works of Darwin and other authors of the period,
and through the very generous and invaluable assistance of Sir
Charles Darwin, Dr. Mary Alice Burmester of Michigan State
University, and Dr. Sydney Smith of Cambridge University, it
became possible to transcribe the obscure and illegible words.
In some instances I have inserted articles, ete., to complete sen-
tences or to correct grammar; such words are enclosed in square
brackets, e.g., [the]. All words or phrases that Darwin enclosed
in parentheses or scrawled between lines as afterthoughts, I have
put in parentheses.

Most bibliographic references cited by Darwin in the notebooks
were incomplete. These have been traced, and are included in the
appended notes in their completed form. I have purposely kept
editorial annotations and amendments to a minimum, seeking
to leave unprejudiced the integrity, arrangement and complete-
ness of the original notes insofar as possible.

There were originally 280 numbered pages in this notebook,
of which 63 were later cut out by Darwin when he began writing
The Origin of Species. Wherever pages are missing from the
notebook, a notation has been inserted in the appropriate place
in the text.

Excerpts from this notebook have been published previously
in: The Life and Letters of Charles Darwin, 2 vols., New York,
1897, edited by Francis Darwin; Darwin and the Darwinian
Revolution, Garden City, New York, 1959, by Gertrude Himmel-
farb, and From Darwin’s Unpublished Notebooks, Centennial
Review of Arts and Science, Vol. 3, No. 4, Fall, 1959, East Lan-
sing, Michigan, by Paul H. Barrett.

Grateful acknowledgments are due to Sir Charles Darwin and
Lady Nora Barlow who kindly granted permission to edit and
publish transcriptions of these notebooks.

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 249

|Transmutation of Species |
B [Notebook]
C. Darwin

(All useful pages cut out December 7, 1856, and again looked through
April 21, 1873. This book was commenced about July, 1837. Page 235 was
written in January 1838. Probably ended in beginning of February.)

ZOONOMIA

Two kinds of generation [1.e., reproduction]: the coeval kind,
all individuals absolutely similar, for instance fruit trees, prob-
ably polypi, gemmiparous propagation, bisection of Planaria,
ete., ete.; the ordinary kind, which is a longer process, the new
individual passing through several stages (typical or shortened
repetition of what the original molecule has done?). This ap-
pears highest office in organization (especially in lower animals,
where mind, and therefore relation to other life, has not come
into play) ; see Zoonomia! arguments; [reproduction] fails in
hybrids where everything else is perfect; mothers apparently
only born to breed, annuals rendered perennial, ete., ete. (Yet
Eunuch, nor cut stallion, nor nuns are longer lived.)

Why is life short? Why such high object generation? We
know world subject to cycle of change, temperature and all cir-
cumstances which influence living beings. We see the young of
living beings become permanently changed or subject to variety,
according to circumstances, [e.g.,] seeds of plants sown in rich
soil, many kinds are produced, though new individuals produced
by buds are constant ; hence we see generation here seems a means
to vary, or adaptation. Again we believe (know) in course of
generation even mind and instinct become influenced. Child
of savage not civilized man. Birds rendered wild through gen-
eration acquire ideas ditto. V. [i.e., Vide] Zoonomia.

There may be unknown difficulty with full grown individual
with fixed organization thus being modified ; therefore generation
to adapt and alter the race to changing world. On other hand,
generation destroys the effect of accidental injuries, which if
animals lived forever, would be endless (that is, with our present
systems of body and universe) ; therefore final cause of life.

250 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

With this tendency to vary by generation, why are species
are [sic] constant over whole country? Beautiful law of inter-
marriages (separating) partaking of characters of both parents,
and then infinite in number. In man it has been said, there is
instinct for opposites to like each other. Agyptian cats and
dogs, ibis, same as formerly, but separate a pair and place them
on a fresh [1.e., geologically new] island. It is very doubtful
whether they would remain constant. Is it not said that marrying
in deteriorates a race; that is, alters it from some end which is
gvood for man? Let a pair be introduced and increase slowly,
[safe] from many enemies, so as often to intermarry; who will
dare say what result? According to this view animals on sep-
arate islands ought to become different if kept long enough apart
with slightly differing circumstances. Now [e.g.,] Galapagos
Tortoises, Mocking birds, Falkland Fox, Chiloe fox, English and
Irish Hare.

As we thus believe species vary, in changing climate, we ought
to find representative species; this we do in South America
(closely approaching), but as they inosculate, we must suppose
the change is effected at once, something like a variety produced
(every grade in that ease surely is not produced?). (Granting)
species according to Lamarck? disappear as collections made
perfect ; truer even than in Lamarck’s time. Gray’s*? remark, best
known species (as some common land shells) most difficult to
separate. Every character continues to vanish: bones, instinct,
eles, enc. ebe:

Nonfertility of hybridity, ete., ete.

If species (1) may be derived from form (2), ete., then (re-
membering Lyell’s * arguments of transportal) island near con-
tinent might have some species same as nearest land, which were
late arrivals; others old ones (of which none of same kind had in
interval arrived) might have grown altered. Hence the type
would be of the continent, though species all different. In cases as
Galapagos and Juan Fernandez, when continent of Pacific existed.
might have been monsoons, when they ceased, importation ceased,
and changes commenced; or intermediate land existed, or they
may represent some large country long separated.

On this idea of propagation [i.e., evolution] of species we can
see why a form pecular to continents, all bred in from one

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 251

parent; why Megatheria several species in 8. America; why 2
of ostriches in S. America. This is answer to Decandoelle [sic]?
(his argument applies only to hybridity), genera being usually
peculiar to same country; different genera, different countries.

Propagation explains why modern animals same type as ex-
tinet, which is law almost proved. We can see why structure is
common is common [sic] in certain countries when we can hardly
believe necessary, but if it was necessary to one forefather, the
result would be as it is. Hence Antelopes at C. of Good Hope,
Marsupials at Australia. (Will this apply to whole organic
kingdom when our planet first cooled?) Countries longest sep-
arated greatest differences; if separated from immense ages®
possibly two distinct types, but each having its representatives,
as in Australia. This presupposes time when no Mammalha
existed; Australian Mamm. were produced from propagation
from different set, as the rest of the world.

This view supposes that in course of ages, and therefore
changes, every animal has tendency to change. This difficult
to prove, [e.g.] cats, ete., from Egypt. No answer because time
short and no great change has happened. I look at two ostriches
as strong argument of possibility of such change; as we see them
in space, so might they in time. As I have before said, zsolate
species, especially with some change, probably vary quicker.

Unknown causes of change. Volcanic Island? Electricity.

Each species changes. Does it progress?

Man gains ideas.

The simplest cannot help becoming more complicated ; and if
we look to first origin, then must be progress. If we suppose
monads are constantly formed, would they not be pretty similar
over whole world under similar climates, and as far as world has
been uniform at former epochs? How is this Ehrenberg?‘ Every
successive animal is branching upwards, different type of or-
ganization improving, as Owen® says, simplest coming in and
most perfect and others occasionally dying out; for instance,
secondary terebratula may have propagated recent terebrat-
ula, but Megatherium nothing. We may look at Megatherium,
Armadillos and sloths as all offsprings of some still older type,
some of the branches dying out.

252 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

With this tendency to change (and to multiplication when iso-
lated) requires deaths of species to keep numbers of forms
equable; (but is there any reason for supposing numbers of
forms equable? — this being due to subdivisions and amount of
differences, so forms would be about equally numerous). Changes
not result of will of animal, but law of adaptation as much as
acid and alkali.

Organized beings represent a tree irregularly branched, some
branches far more branched; hence Genera. As many terminal
buds dying as new ones generated. There is nothing stranger in
death of species than individuals.

If we suppose monad definite existence, as we may suppose in
this case, their creation being dependent on definite laws, then

a ed

[Fig. 1]

those which have changed most (owing to the accident of po-
sitions) must in each state of existence have shortest life. Hence
shortness of life of Mammalia.

Would there not be a triple branching in the tree of life owing
to three elements — air, land and water, and the endeavour of
each typical class to extend his domain into the other domains,
and subdivision three more — double arrangement? If each
main stem of the tree is adapted for these three elements, there
will be certainly points of affinity in each branch. A species as
soon as once formed by separation or change in part of country
[has] repugnance to intermarriage; [that] settles it. We need
not think that fish and penguin really pass into each other (?).

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 23

The tree of life should perhaps be called the coral of life, base
of branches dead, so that passages cannot be seen. This again
offers contradiction to constant succession of germs in progress
(no, only makes it excessively complicated). [Fig. 1.]

Is it thus fish can be traced right down to simple organization,
birds not? [Fig. 2.]

We may fancy according to shortness of life of species that in
perfection, the bottom of branches deader, so that in Mammalia,
birds, it would only appear like circles, and insects amongst
articulata; but in lower classes perhaps a more linear arrange-
ment. How is it that then come aberant [sic] species in each
genus (with well characterized parts belonging to each) ap-
proaching another? Petrels have divided themselves into many

wh}
wb My
—>47 qlee
2 ioe

[Fig. 2]

species, so have the Awks [sic], then is particular circumstances
to which .. .° Is it an index of the point whence two favourable
points of organization commenced branching ?

As all the species of some genera have died, have they all one
determinate life dependent on germs, that germ upon another?
Whole class would die out, therefore . . .1°

In island neighbouring [a] continent where some species have
passed over, and where other species have ‘‘air’’ of that place,
will it be said, those have been then created there? Are not all
our British Shrews diff[erent] species from the continent? Look
over Bell! and L. Jenyns.!”

Falkland rabbit may perhaps be instance of domesticated ani-
mals having effected a change which the Fr[ench] naturalists
thought was species.

Study Lesson!? Voyage of Coquille.

Dr. Smith ™ says he is certain that when White men and Hot-
tentots or Negros [sic] cross at C. of Good Hope the children

254 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

cannot be made intermediate. The first children partake more
of the mother, the later ones of the father. (Is not this owing
to each copulation producing its effect, as when bitches puppies
are less purely bred owing to having once born mongrels?) He
has thus seen the black blood come out from the grandfather
(when the mother was nearly quite white) in the two first chil-
dren. How is this in West Indies? Humboldt,’ New Spain.

Dr. Smith always urges the distinct locality or metropolis of
every species; believes in repugnance in crossing of species in
wild state.

No doubt (C.D.) wild men do not cross readily, distinctness
of tribes in T. del Fuego; the existence of whiter tribes in centre
of S. America shows this. Is there a tendency in plant’s hybrids
to go back? If so, man and plants together would establish Law,
as above stated. No one can doubt that less trifling differences
are blended by intermarriages; then the black and white is so
far gone, that the species (for species they certainly are ac-
cording to all common language) will keep to their type; in
animals so far removed, with instinct in leu of reason, there
would probably be repugnance, and art required to make mar-
riage. As Dr. Smith remarked, man and wild animals in this
respect are differently circumstanced.

Is the shortness of life of species in certain orders connected
with gaps in the series of connections? (If starting from same
epoch certainly.) The absolute end of certain forms, from con-
sidering 8S. America (independent of external causes) does ap-
pear very probable : —

Mem. |[i.e., Memento or Memorandum]: Horse, Llama, ete.,
ete.

If we suppose (grant) similarity of animals in one country
owing to springing from one branch, and the monucle!® has
definite life, then all die at one period which is not the case,
therefore monucule not definite life.

I think: [Fig. 3.]

Thus between A and B immense gap of relation; C and B the
finest gradation, B and D rather greater distinction. Thus genera
would be formed, bearing relation to ancient types, with several
extinct forms. For if each species (as ancient @)) is capable
of making 13 recent forms,’ twelve of the contemporarys [sic]

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK Zao

must have left no offspring at all,'® so as to keep number of
species constant.

With respect to extinction, we can easy see that variety of os-
trich, Petise, may not be well adapted, and thus perish out, or
on other hand like Orpheus being favourable, many might be
produced. This requires principle that the permanent varieties
produced by confined breeding and changing circumstances are
continued and produced according to the adaptation of such

B
D C

AL

[Fig. 3]

Case must be that [in] one generation there should be as many living as
now. To do this and to have [as] many species in same genus (as is) requires
extinction.

circumstances, and therefore that death of species is a conse-
quence (contrary to what would appear from America) of non-
adaptation of circumstances. Vide two pages back — Diagram
[Fig. 3].

The largeness of present genera renders it probable that many
contemporary [species] would have left scarcely any types of
their existence in the present world. Or we may suppose only
each [i.e., certain] species in each generation only breeds; like
individuals in a country not rapidly increasing.

If we thus go very far back to look to the source of the Mam-
malian type of organization, it is extremely improbable that any
of the successors of his relations shall now exist. In same manner,

256 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

if we take (a man from) any large family of 12 brothers and
sisters (in a state which does not increase) it will be chances
against any one (of them) having progeny living ten thousand
years hence; because at present day many are relatives, so that
by tracing back the fathers would be reduced to small percent-
ages. Therefore the chances are excessively great against any
two of the 12 having progeny after that distant period.!®

Henee if this is true, that the greater the groups, the greater
the gaps (or solutions of continuous structure) between them.
For instance, there would be great gap between birds and mam-
malia, still greater between Vertebrate and Articulata, still
greater between animals and Plants. But yet, besides affinities
from three elements, from the infinite variations, and all coming
from one stock and obeying one law, they may approach; some
birds may approach animals, and some of the vertebrates in-
vertebrates. Just a few on each side will yet present some anom-
aly, and bearing stamp of some great main type, and the grada-
tion will be sudden.

Heaven knows whether this agrees with Nature: Cuidado.?°

The above speculations are applicable to non-progressive de-
velopment, which certainly is the case at least during subsequent
ages.

The Creator has made tribes of animals adopted [adapted ?]
preeminently for each element, but it seems law that such tribes,
as far as compatible with such structures, are in minor degree
adapted for other elements. Every part would probably be not
complete, if birds were fitted solely for air and fishes for water.
If my idea of origin of Quinarian System?! is true, it will not
occur in plants which are in far larger proportion terrestrial ;
if in any, in the Cryptogamie Flora (but not atmospheric types.
Hence probably only four. Is not this Fries’ ?? rule? What sub-
ject has Mr. Newman? the (7) man studied?)

The condition of every animal is partly due to direct adap-
tation and partly to hereditary taint; hence the resemblances
and differences for instance of finches of Europe and America,
tes ele: ele:

The new system of Natural History will be to describe limits
of form (and where possible the number of steps known). For

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK PAT

instance among the Carabidae. Instance in birds. Examine good
collection of insects with this in view.

Geogr. Journal, Vol. VI, P. II, p. 89, Lieut. Wellsted?* ob-
tained many sheep from Arabian coast. ‘‘These were of two
kinds, one white with a black face, and similar to those brought
from Abyssinia, and the others dark brown, with long clotted
hair resembling that of goats.’’

Progressive developement [sic] gives final cause for enormous
periods anterior to man. Difficult for man to be unprejudiced
about self, but considering power, extending range, reason and
futurity it does as yet appear. . .*°

In Mr. Gould’s?® Australian work some most curious cases of
close but certainly distinct species between Australia and Van
Diemen’s Land,?? and Australia and New Zealand. Mr. Gould
says in subgenera they undoubtedly come from same countries.
In mundine [sic] genera the nearest species often . . .78

... great S. American quadrupeds part of some great system
acting over whole world; the period of the great quadrupeds de-
elining as great reptiles must have once declined.

Cuvier 7° objects to propagation of species (read his Theory
of the Earth attentively) by saying, why not have some inter-
mediate forms been discovered between paleotherium, Meg-
alonyx, Mastodon, and the species now living? Now according
to my view, in S. America parent of all armadilloes might be
brother to Megatherium; uncle now dead.

Bulletin Geologique, April 1837, p. 216, Deshayes*° on change
in shells from salt and F. water; on what is species. Very Good.
(Has not Maceculloch *! written on same changes in fish?)
Mem.: Rabbit of Falklands described by I. and L.*? as new

species. Cuvier examined it .. .*°
Same thing oeeurs with regards to other tribes in that same
family.

(NB. I see Waterhouse thinks Quinary only three elements. )

How far does Waterhouse’s** representatives agree with
breeding in irregular trees and extinction of forms?? It is in
simplest case saying every species in genus resembles each other
(at least in one point, in truth in all excepting specific charac-
ter) ; and in passing from species to genera, each retains some
one character of all its family; but why so? I can see no reason

258 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

for these analogies; from the principle of atavism where real
structures obliged to be altered, I can conceive colouring re-
tained; therefore probably in some (heteromera) colouring of
erysomela [sic]*° may be going back to common ancestor of
Crysom. and Heterom., but I cannot understand the universality
of such law.

It would be curious to know in plants (or animals) whether,
in races have tendency to keep to either parent (this is what
French call atavism). Probably this is first step in dislike to
union; offspring not well intermediate.

duyelle® Vols Sp. 3719:

Mammalian type of organization same from one period to an-
other, preeminently Pachydermata, less so in Miocene and so on.

As I have traced the great Quadrupeds to Siberia, we must
look to type of organization; extinct species of that country
parents of American. Now Genera of these two countries ought
to be similar.

Law?: existence definite without change, superinduced, or new
species; therefore animals would perish if there were nothing
in country to superinduce a change (?).

Seeing animal die out in S. America, with no change, agrees
with belief that Siberian animals lived in cold countries and
therefore not killed by cold countries. Seeing how horse and
elephant reached S. America, explains how Zebras reached South
Africa.

It is a wonderful fact, Horse, Elephant and Mastodon dying
out about same time in such different quarters. Will Mr. Lyell
say that some [same?] circumstance killed it over a tract from
Spain to South America? Never! They die, without [1.e., un-
less] they change; like Golden Pippens [sic] ** it is a generation
of species like generation of individuals. Why does individual
die? To perpetuate certain peculiarities (therefore adaptation),
to obliterate accidental varieties, and to accomodate [sic] itself
to change (for of course change even in varieties is accomoda-
tion [sic]). Now this argument applies to species. If individual
cannot procreate, he has no issue; so with species.

I should expect that Bears and Foxes, ete., same in N. America
and Asia, but many species closely allied but different, because

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 259

country separated since time of extinct Quadrupeds. Same
argument applies to England. Mem.: Shrew, Mice.

Animals common to South and North America. Are there any?

Rhinoceros peculiar to Java, and another to Sumatra. Mem.:
Parrots peculiar, according to Swainson,?* to certain islets in
East Indian Archipelago. Dr. Smith considers probable true
northern species replace the southern kinds.

(1) [sic] Gnu reaches Orange River and says so far will only
#0 and no further.

Prof. Henslow says that when race once established so diffi-
cult to root out. For instance ever so many seeds of white flax, all
would come up white, though planted in the same soil with blue.
Now this is same bearing with Dr. Smith’s fact of races of men.*?

Strong odour of negroes, a point of real repugnance. Water-
house says there is no TRUE connection between great groups.

Speculate on land being grouped towards centre near Equator
in former periods, and then splitting off.

If species generate other species their race is not utterly cut
off; like golden pippen [sic], if produced by seed go on, other-
wise all die. The fossil horse generated in 8. Africa Zebra, and
continued, perished in America.

All animals of same species are bound together just like buds
of plants, which die at one time, though produced either sooner
or later.

Prove animals like plants, trace gradation between associated
and nonassociated animals, and the story will be complete.

It is absurd to talk of one animal being higher than another.
We consider those, when the intellectual faculties (cerebral
structure) most developed, as highest. A bee doubtless would
when the instincts were . . .4°

_.. there appears in Australia great abundance of species of
few genera or families (long separated).

Proteaceae and other forms (?), being common to Southern
hemisphere, does not look as if S. Africa peopled from N. Africa.

An originality is given (and power of adaptation is given by
true generation) through means of every step of progressive
increase of organization being imitated in the womb (which has
been passed through to form that species). (Man is derived
from Monad each fresh — ) *!

260 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Mr. Don* remarked to me, that he thought species became
obscurer as knowledge increased, but genera stronger. Mr.
Waterhouse says no real passage between good genera. How
remarkable spines, like on a porcupine, on Hehidna.

Good to study Regne Animal for Geography.*

The motion of the earth must be excessive up and down: Ele-
phants in Ceylon, East India Archipelago; West Indies. Opos-
sum and Agouti same as on continent (3 Paradiseini are common
to Van Diemen’s Land and Australia). From the consideration
of these archipelago’s ups and downs [they are] in full con-
formity with European formations (England and Europe, Ire-
land, common animals); for instance tertiary deposits between
East India islets.

Ireland longer separated; hare of two countries different.
Ireland and Isle of Man possessed elk, not England. Did Ireland
possess Mastodons?? (Negative facts tell for little.)

Geographie distribution of Mammalia more valuable than any
other, because less easily transported. Mem.: plants on Coral
islets. Next to animals, land birds, and life shorter or change
greater.

In the East Indian Archipelago it would be interesting to
trace limits of large animals.

Owls transport mice alive?

Species formed by subsidence: Java and Sumatra Rhinoceros ;
elevate and join, keep distinct, two species made. Elevation and
subsidence continually forming species. (Man and wife being
constant together for life is in accordance with . . .)

(The male animal affecting all the progeny of female insures
of the mixing of individuals. )

South Africa, proof of subsidence and recent elevation: pray
ask Dr. Smith to state that most clearly.

Fox tells me that beyond all doubt seeds of Ribston Pippin
produce Ribston Pippin, and Golden Pippin, goldens, hence sub-
varieties, and hence possibility of reproducing any variety, al-
though many of the seeds will go back. Get instances of a variety
of fruit tree or plant run wild in foreign country. (Here we
have avitism [sic, 1.e., atavism] the ordinary event and success-
ion the extraordinary. )

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 261

When one sees nipple on man’s breast, one does not say some
use, but sex not having been determined, so with useless wings
under elytra of beetles, born from beetles with wings, and modi-
fied. If simple creation surely would have been born without
them.

In some of the lower orders a perfect gradation can be found
from forms marking good genera by steps so insensible that each
is not more change than we know varieties can produce. There-
fore all genera may have had intermediate steps. Quote in detail
some good instance.

But it is other question, whether there have existed all those
intermediate steps especially in those classes where species not
numerous. (N.B., in those classes with few species greatest Jumps
strongest marked genera? Reptiles?) For instance there never
may have been grade between pig and tapir, yet from some com-
mon progenitor. Now if the intermediate ranks had produced
infinite species, probably the series would have been more perfect,
because in each there is possibility of such organization (spines
in Echidna and Hedgehog). As we have one marsupial animal
in Stonefield [sic]** slate, the father of all Mammalia in ages
long past, and still more so known with fishes and reptiles. In
mere eocine [sic] rocks we can only expect some steps. I may
ask whether the series is not more perfect by the discovery of
fossil Mammalia than before, and that is all that can be expected.
This answers Cuvier.

Perhaps the father of Mammalia as Heterodox as ornithorhyn-
cus [sic]. If this last animal bred, might not new classes be
brought into play ?

The father being climatized, climatizes the child. Whether
every animal produces in course of ages ten thousand varieties
(influenced itself perhaps by circumstances) and those alone
preseved which are well adapted? This would account for each
tribe acting as in vacuum to each other.

p. 306. Chamisso on Kamtschatka quadrupeds; Kotzebue first
voyage.” Entomological Magazine, paper on Geographical
range. Richardson,*® Fauna Borealis.

It is important the possibility of some island not having large
quadrupeds. Humboldt * has written on the geography of plants.
Kssai sur la Geographie des Plants, I Vol. in 4° —

262 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

I have abstracted Mr. Swainson’s*® trash at beginning of
volume on Geographical distribution of animals.

Geograph. Journal,!? Vol. I, p. 17-21, says from Swan river
along South coast all the remarkable Australian genera collected
together.

Man has no hereditary prejudices or instinct to conquer, or
breed together. Man has no limits to desires; in proportion in-
stinet more, reason less, so will aversion be.

L’Institut,°° (1837, no. 246) a section of fossil ‘‘singe,’’*! it
cannot be made to approach the Colobes [sic] ** which in South
Africa appears to represent the semnopitheque ** of India. Tooth
of Sapajou.®* Sapajou is S. American form. Therefore it is like
ease of great edentate (has been doubted) and opossum found
in Europe now confined to southern hemisphere. (If these facts
were established it would go to show a centrum for Mammalia. )
I really think a very strong case might be made out of world
before Zoological divisions. Mem.: Species doubtful when known
only by bones. Mem.: Silurian fossils. How are South American
shells?

Do not plants which have male and female organs together,
yet receive influence from other plants? Does not Lyell® give
some argument about varieties being difficult to keep on account
of pollen from other plants?; because this may be applied to
[all], show all plants do receive intermixtures. But how with
hermaphrodite shells!!! ?

We have not the slightest right to say there never was com-
mon progenitor to mammalia and fish, when there now exists
such strange forms as ornithorhyncus [sic].

The type of organization constant in the shells.

The question if creative power acted at Galapagos it so acted
that birds with plumage and tone of voice purely American,
North and South; so permanent a breath cannot reside in space
before island existed. Such an influence must exist in such spots.
We know birds do arrive, and seeds. (And geographical divisions
are arbitrary and not permanent. This might be made very
strong, if we believe the Creator created by any laws, which I
think is shown by the very facts of the Zoological character of
these islands. )

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 263

The same remarks applicable to fossil animals same type;
armadillo-like cray [i.e., crayfish?] created; passage for verte-
brae in neck same cause. Such beautiful adaptations, yet other
animals live so well. This view of propagation gives a hiding
place for many unintelligible structures; it might have been of
use in progenitor, or it may be of use, —like mammae on man’s
breast.

Tow does it come wandering birds such [as] sandpipers not
new at Galapagos? Did the creative force know that these spe-
cies could arrive? Did it only create those kinds not so likely to
wander? Did it create two species closely allied to Mus. coro-
nata,°® but not coronata? We know that domestic animals vary
in countries without any assignable reason.

Astronomers might formerly have said that God ordered each
planet to move in its particular destiny. In same manner God
orders each animal created with certain form in certain country,
but how much more simple and sublime power: let attraction act
according to certain law; such are inevitable consequences. Let
animals be created, then by the fixed laws of generation, such
will be their successors.

Let the powers of transportal be such, and so will be the forms
of one country to another. Let geological changes go at such a
rate, so will be the number and distribution of the species!! It
may be argued representative species chiefly found where [there
are| barriers (and what are barriers by?), interruption of com-
munication, or when country changes. Will it [be] said that vol-
eanic soil of Galapagos under equator, that external conditions,
would produce species so close as Patagonian Cha — *7 and Gala-
pagos Orpheus? Put this strong, so many thousand miles distant.

Absolute knowledge that species die and others replace them.
Two hypotheses: [1] fresh creations is mere assumption, it ex-
plains nothing further; [2] points gained if any facts are con-
nected.°* No doubt in birds, mundine [sic] genera (Bats, Foxes,
Mus) are birds that are apt to wander and of easy transportal.
Waders and waterfowl. Serutinize genera, and draw up tables.
Instinct may confine certain birds which have wide powers of
flight; but are there any genera mundine which cannot trans-
port easily (it would have been wonderful if the two Rheas had
existed in different continents)? In plants I believe not.

264. BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

It is a very great puzzle why Marsupials and Edentata should
only have left off springs [sic] in or near South Hemisphere.
Were they produced in several places and died off in some? Why
did not fossil horse breed in 8. America? It will not do to say
period unfavourable to large quadrupeds; horse not large.®*®

... but not vice versa. (Could plants live without carbonic
acid gas?) Yet unquestionably animals most dependent on vege-
tables, of the two great Kingdoms.

Principes de Zool. Philosop. :*° I deduce from extreme diffi-
culty of hypothesis of connecting mollusca and vertebrata, that
there must be very great gaps, yet some analogy. The existence
of plants, and their passage to animals appears greatest argu-
ment against theory of analogies.

States there is but one animal: one set of organs; the other
animals created with endless differences; does not say propa-
gated, but must have concluded so. Evidently considers (or
hints) generation as a short process, by which one animal passes
from worm to man (highest) as typical of changes which can be
traced in same organ in different animals in scale. In monsters
also organs of lower animals appear, yet nothing about propa-
gation. I see nothing like grandfather of Mammalia and birds,
ete. P. 32,°' reference to M. Edwards’ law of crustacea, with re-
spect to mouth, those beautiful passages from one to other organ ;
Cuvier on opposite side: V. [i.e., Vide] Vol. of Fish, p. 59,8
Cuvier has said each animal made for itself does not agree with
old and modern types being constant. Cuvier’s theory of Con-
ditions of existence is thought to account [for] resemblanees, ete.,
therefore Quinary system, or three elements (p. 68).

With unknown limits, every tribe appears fitted for as many
situations as possible (conditions will not explain states) for in-
stance take birds, animals, reptiles, fish. (Perhaps consideration
of range of capabilities past and present might tell something.)

Pts (Gs St: Bolaire:*

Insects and Molluses allowed to be wide hiatus; states in one
the sanguineous system, in other nervous developed; (Owen’s
idea) ; states these class[es] approach on the confines? Balani-
dae? I cannot understand whether G. H. [Geoffroy St. Hilaire]
thinks development in quite straight lines, or branching.

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 265

G. H.: What does the expression mean, used by Cuvier, that
all animals (though some maybe) have not been created on the
same plan? (Second resumé well worth studying.) H. says
grand idea God giving laws, and then leaving all to follow con-
sequences. I cannot make out his ideas about propagation. His
work: Philosophie Anatomique® (2nd Vol. about monsters
worth reading).

N.B. Well to insist upon (different animals)® large Mam-
malia not being found on all islands (if act of fresh creation why
not produced on New Zealand?) ; if generated an answer can be
given.

It is a point of great interest to prove animals not adapted to
each country. Provision for transportal (otherwise) not so
numerous; quote from Lyell. Assuming truth of quadrupeds
being created on small spots of land, of the same type with the
great continents, we get a means of knowing of movements.

How can we understand excepting by propagation that out
of the thousands of new insects all belong to same types already
established? Why out of the thousand forms should they all be
classified [i.e., classifiable]? Propagation explains this.

Ancient Flora thought to [be] more uniform than existing.
Bd) N- Philosp. Jr: p. 191, no. 5, Apr: 1827.

F. Cuvier ® says, ‘‘But we could only produce domestic indi-
viduals and not races, without the occurrence [1.e., concur-
rence] ®8 of one of the most general laws of life, the transmission
of [the organic or intellectual modifications by generation. Here
one of the most astonishing phenomena of nature manifests itself
to us, the transformation of] ® a fortuitous modification into a
durable form, of a fugitive want into a fundamental propensity,
of an accidental habit into an instinet.’’ Ed. N. Phi. J. p. 297,
No. 8, Jan.-Apr. 1828. I take higher grounds and say life is
short for this object and others, viz., not too much change.

In Number 6 of Ed. N. Philo. Journ., Paper by Crawford on
Mission to Ava, account of hairy man (because ancestors hairy),
with one hairy child, and of albino (disease) being banished and
given to Portuguese priest. In first settling a country, people
very apt to be split up into many isolated races! Are there any
instances of peculiar people banished by rest? (Therefore most
monstrous forms have tendency to propagate, as well as diseases. )

266 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

In intermarriages, smallest differences blended; rather
stronger tendency to imitate one of the parents; repugnance
generally to marriage before domestication; afterwards none or
little, with fertile offspring; marriage never probably excepting
from strict domestication, offspring not fertile, or at least most
rarely and perhaps never female. No offspring: physical im-
possibility of marriage.

Whether those genera which unite very different structure, as
petrel and alk [sic], do [they] not show the possibility of com-
mon branching off?

Accra.” Coast of Africa. Clay slate. Strike SSW and NNBE,
dip'30°-80° (2). Hd. Phil, N. J. p. 410; 1828;

It is daily happening that Naturalist[s] describe animals as
species, for instance Australian dog, or Falkland rabbit. There
is [1.e., are] only two ways of proving to them it is not; one,
when they can [be] proved descendant, which of course most
rare, or when placed together they will breed. But what a charac-
ter is this? ”

The relation of Analogy of Macleay,” ete., appears to me the
same, as the irregularities in the degradation of structure of
Lamarck, which he says depends on external influences. For
instance he says wings of bat are from external influence. Hence
name of analogy, the structures in the two animals bearing re-
lation to a third body,’* or common end of structure.

A Race of domestic animals made from influences in one
country is permanent in another. (Good argument for species
not being so closely adapted.)

Near the Caspian (Province of Ghilan) wooded district cattle
with humps as in India. Geograph. J., Vol. III, P. I, p. 17.
(Lat. 37° about.)

Vol. IV, P. I, Geograp. Journal, Voyage up the Massaroony,
by W. Hillhouse [sic].7* Demerara. In note, Demerara, 10 (12)
feet beneath surface, forest trees fallen (kind well known, car-
bonized), clay fifty feet, then forest 120 feet, micaceous rocks;
subsidence appears indicated, p. 36.

Geograp. Journ. Vol. IV, P. II, p. 160, Melville Island. ‘‘The
buffaloes, introduced from Timor, herded separate from the Eng-
lish eattle, nor could we get them to associate together.’’

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 267

There is long rigmarole article by G. Hilaire*® on wonder of
finding monkey in France, of genus peculiar to East Indian
islands. Compares it to fossil Didelphus (S. American genus)
in plaster of Paris. Now this is exception to law of type, like
horse in §. America, or like living Edentata in Africa, etc., ete.
Now if suppose world more perfectly continental, we might have
wanderers (as Peceari in N. America) ; then if it is doomed that
only one species of family has offspring, the chance is that these
wanderers would not, but where original forms most numerous,
then would be wanderers. Some however might have offspring,
and then we should have anomalies, as Cape Anteater (V. [i.e.,
Vide] L’Institut, p. 245, 1837). This supposes world divided
into Zoological provinces — united —and now divided again.
Weakest part of theory death of species without apparent
physical cause. Mem.: Mastodon all over 8. America. Hilaire does
not seem (?) to consider the monkey as a wanderer, but as pro-
duced by climate?

M. Baer ®® (thinks) the Auroch was found in Germany and
thinks even now in central and Eastern Asia beyond the Ganges
and perhaps even in India, p. 261, L’Institut, 1837.

Mem.: Sir F. Darwin,*! cross-breed boars were wilder than
parents, which is same as Indian Cattle, therefore tameness not
hereditary? Having been gained in short time.

Milvulus forficatus*? has a wide range, is a Tyrant flycatcher
doing the service of a swallow.

I think we may conclude from Australia and 8. America that
only some mundine cause has destroyed animals over the whole
world. For instance gradual reduction of temperature from geo-
eraphical or central heat. But then shells —.

Mr. Yarrell®* says that old races when mingled with newer,
hybrid variety partakes chiefly of the former.

Kyton’s** paper on Hybrids, Loudon’s Magazine.

Gould ®° on Motacilla, Loudon Mag., September or October,
1837. Species peculiar to Continent and England.

Westwood *® has written paper on affinity and analogy in Lin-
naean Transactions.

Mr. Wynne * distinctly says that the mixture between Chinese
and English Breed decidedly exceedingly prolific, and hybrid

268 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

about half-way. Eyton says hybrid about half aways [sic] and
result the same.

Indian cattle and common produced very fine hybrid offspring,
much larger than the dam, from those imported by Dr. (Ld.)
Powis.*®

Hybrid dog’s offspring seldom intermediate between parents.
How easily does wolf and dog cross? Mr. Yarrel [sic] thinks
oldest variety impresses the offspring most foreibly, Esquimaux
dog and Pointer. (Game fowls have courage independently of
individual foree. )

Mr. Wynne has crossed Ducks and Widgeon and offspring,
either amongst themselves, or with parent birds. (W. Fox‘?
knew of case of male widgeon, winged, and turned on pool, first
season bred readily with common ducks. )

Kirby ®° all through Bridgewater errs greatly in thinking
every aninal born to consume this or that thing. There is some
much higher generalization in view.

In marsupial division do we not see a splitting in orders, car-
nivora, rodents, etc., JUST COMMENCING!! ?

Kirby says (not definite information) west of Rocky Moun-
tains Asiatic types discoverable.

Bridgewater Treatise, p. 85 [Vol. I]. Parasite of Negroes dif-
ferent from European. Horse and ox have different parasite in
different climates.

Humb. (Humboldt?)®! Vol. V, P. Il, p. 565. Consult. Says
types most subject to vary where intermixtures precluded.

There are some good accounts of passages of legs into mouth-
pieces of Crustacea. (Kirby, Bridgewater Treatise.) Vol. II,
p. 75. A Fish which emigrates over land is a silurus, p. 123. A
echmbing fish, p. 122.

A Terrestrial annelidous animal, p. 347. Vol. I? compare
with my planaria; leaches out of water.

Does the odd Petrel of T. del F. [Tierra del Fuego] take form
of auk because there is no auk in Southern Hemisphere? Does
this rule apply?

A Treatise on Form of animals by Mr. Cline.®? ‘‘The character
of both parents are observed in their offspring, but that of the
male more frequently predominates.’’ p. 20, ditto. ‘‘If hornless
ram be put to horned ewe almost all the lambs will be hornless.’’

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 269

(Does this apply to where same animal breeds often with same
female?) p. 28, ‘‘It is wrong to enlarge a native breed of animals,
for in proportion to their inerease of size they become worse in
form, less hardy, and more liable to disease.’’

If population of place be constant (say 2000) and at present
day every ten living souls on average are related to the 200th
year degree, then 200 years ago there were 200 people living
who now have successors. Then the chance of 200 people being
related within 200 years backward might be calculated and this
number eliminated. Say 150 people four hundred years since
were progenitors of present people, and so on backwards to one
progenitor, who might have continued breeding from eternity
backwards.

If population was increasing between each lustrum, the num-
ber related at the first start must be greater, and this number
would vary at each lustrum, and the calculation of chance of the
relationship of the progenitors would have different formula for
each lustrum. [A lustrum is a five year census. ]

We may conclude that there will be a period, though long
distant, when of the present men (of all races) not more than a
few will have successors. At present day in looking at two fine
families one will [have] successors for centuries, the other will
become extinct. Who can analyze causes: dislike to marriage
(some delay), disease, effects of contagions and accidents. Yet
some causes are evident, as for instance one man killing another.
So is it with varying races of man; then races may be overlooked ;
many variations consequent on climate, ete.; the whole races act
towards each other, and are acted on, just like the two fine fam-
ilies (no doubt a different set of causes must act in the two cases).

May this not be extended to all animals? First consider species
of cats, other tribes, ete., ete. (Exclude mothers and then try
this as simile.)

In a decreasing population at any one moment fewer closely
related; therefore (few species of genera) ultimately few genera
(for otherwise the relationship would converge sooner), and
lastly, perhaps some one single one. Will not this account for
the odd genera with few species which stand between great
groups, which we are bound to consider the increasing ones ?

270 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

N.B. as illustrations, are there many anomalous lizards living,
or of the tribe fish extinct, or of Pachydermata, or of coniferous
trees, or in certain shell cephalopoda? Read Buckland.*!

L’Institut, 1837, p. 319, Bronegniart,®® no dicotyledonous
plants and few monocot. in coal formation (?). P. 320, states
Cryptogam. Flora formerly common to New Holland?! P. 320,
says Coniferous structures intermediate between vascular or
Cryptogam. (original Flora) and Dicotyldones, which nearly
first appear (p. 321) at Tertiary epochs. P. 330, Fossil Infu-
soria found of unknown forms, a circumstance undiscovered by
Ehrenbergh [sic ].°°

Indian cow with hump and common; between Esquimaux and
European dog? Yet man has had no interest in perpetuating
these particular varieties.

If species made by isolation, then their distribution (after
physical changes) would be in rays from certain spots. (Agrees
with old Linnaean doctrine and Lyell’s to certain extent.)

Von Buch, Canary Islands, French Edit.:°* Flora of Islands
very poor (p. 145) 25 plants; 36 St. Helena, without ferns, anal-
ogous to nearest continent; poorness in exact proportion to dis-
tance (?) and similarity of type (?) (Mem.: Juan Fernandez).
From study of Flora of islands, ‘‘ou bien encore on pourrait au
plus en coneclure quels sont les genres qui, sous ce climat, se
divisent le plus aisément en espéces distinetes et permanentes.’’
p. 145. In Humboldt % great work de distribut., Plantarum, re-
lation of genera to species in France is 1:5, 7; in Laponia 1:2, 3.

(Mem.: Lyell? on shells. ) genera
Ini: North Africa 1 :4,2

Iles Canaries 1:1,46

St. Helena NS

Calculate my Keeling case; Juan Fernandez, Galapagos,
Radack Isle? Therefore islands and arctic are in same relation.
We find species few in proportion to difficulty of transport. For
instance the temperate parts of Teneriffe, the proportion of genera
1:1. I can understand in one small island species would not be
manufactured. Does it not present analogy to what takes place
from time? Von Buch distinctly states that permanent varieties
become species, p. 147, p. 150, not being crossed with others.

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 271

Compares it to languages. But how do plants cross? — admir-
able discussion.1°°

Mr. Owen !°! suggested to me, that the production of monsters
(which Hunter 1°” says owe their origin to very early stage) and
which follow certain laws according to species, present an anal-
ogy to production of species.

Animals have no notions of beauty, therefore instinctive feel-
ings against other species (for sexual ends), whereas man has
such instinets very little.

In Zoology Proceedings, Jan. 1837 by Eyton, 1! account of
three kinds of pigs; difference in skeletons; very good.

Apteryx, a good instance probably of rudimentary bones. As
Waterhouse 1°? remarked, mere length of bill does not indicate
affinity because similar habits produce similar structures. Mem. :
Ornitho Rhyneus [sic]. Would not relationship express a real
affinity, and affinity, [e.g.?] whales and fish?

Progeny of Manks cats without tails ; some long and some short ;
therefore like dogs. Ogleby [sic] 1°° says Wolves at Hudson Bay
breed with dogs, the bitches never being killed by them, whilst
they eat up the dogs.

L’ Institut. Curious paper by M. Serres °° on Molluseous ani-
mals representing foetuses of vertebrata, etc., 1837, p. 370. Owen
says nonsense.

The distribut[ion] of big animals in East Indian Archipelago,
very good in connection with Von Buch Voleanic chart and my
idea of double line of intersection. At India House collection
of Birds from Java, at Leyden series from several islands. Bear
peculiar to Sumatra and not found on Java. Monkey peculiar to
latter, not to former. (Dr. Horsfield.) 1%

Consult Dr. Smith History of 8. African Cattle ;!°8 Phillips!
Geology, p. 81 in Lardner’s Eneyelop., proportion between fossils
and recent shells, between herbivorous and zoophagous mollusca
according to periods. N.B., was Europe desert (like S. Africa)
after Coal Period?

In those divisions of molluses where species now least in num-
ber (as Cephalopods) in last tertiary epochs, [were] most genera
dead? Examine into this (in Phillips). According to this,
formerly there would have been many genera of monobranchous
animals, p. 82. (There are many tables in Phillip’s of numerous

he BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

genera in fossil and recent state well worth consideration. )
Tabulate Mammalia on this principle.

Man (Varieties-Races) in savage state may be called species
(races) ; in domesticated [state] species (races). If all men were
dead, then monkeys make men. Men make angels.

Those species which have long remained are those ( ?Lyell?)
which have wide range and therefore cross and keep similar. But
this is difficulty: this immutability of some species. [Fig. 4. ]

[Fig. 4]

In Phillips p. 90, it seems the most organized fishes lived far
back. Fish approaching to reptiles at Silurian age.

How long back have insects been known? As Gould remarked
to me, the ‘‘beauty of species is their exactness,’’ but do not
known varieties do the same? May you not breed ten thousand
ereyhounds and will they not be greyhounds?

Yarrell’s remark about old varieties affecting the cross most
well worthy of observation.

I think it is certain, strata could not now accumulate without
[including] sealbones and cetaceans, both found in every sea,
from Equatorial to extreme poles. Oh, Wealden, — Wealden.'!”

Do the N. American Tertiary deposits present analogies to
shells of living seas? 1!!!

A breed of Blood Hounds from Aston Hall close to Birmine-
ham, and supposed to be descended from a breed known to be
there since the time of Charles, and now in the possession of Mr.
Howard Galton,” have one of the vertebra, about 24 from base
of tail, enlarged two [sic] very considerably, so that any person
would say the tail was broken. This came so often that it was
difficult to obtain a litter without this defect. Very curious case
W. D. Fox.

When dogs are bred into each other the females loose [sic] de-
sire, and it is required to give the cantharinides.1?

Bull. Soe. Geolog., 1834, p. 217 Java. Fossils 10 out of twenty
have analogues (uses this word for similar) in the Indian sea.
Deshayes.!"4

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK Ad ( 33

Mr. MeClay ! is inclined to think that offspring of Negro and
white will return to native stoek (the cross often whiter than
white parent). The mulattos [sic] themselves explain it by inter-
marriages with people either a little nearer black or white as it
may happen. Dr. Smith !'® says he is sure of the case at Cape.
MecClay argues from it Black and White species. For says he
seeds of hybrid lillies, ete., ete., ete., (V. Herbert 17 on hybrids)
thus act. Now the point will be to find whether known varieties
in plants do so, as in cacti, ete., ete.; as in dogs; investigate case
of pidgeons [sic], fowls, rabbits, cats, ete., ete. When black and
white men cross some offspring black, others white, which is more
closely allied to case of cross of dogs. See paper in Philosoph.
Transactions on a quagga and mare crossing, by Ld. Moreton
[sic] 48 where mare was influenced in this cross to after births,
like aphides. Case of boy with foetus developed in breast looks
as if many ova impregnated at once. Dr. Smith considers the
Caffers [i.e., Kaffirs] (like Englishmen), men of many counten-
ances, as hybrid race. Is not this contradiction to his view of
races not mingling ?

In Foxes [sic] case of Blood Hounds, a little mingling would
probably have been good, namely such as blood hounds from
other parts of England. Mr. Bell of Oxford St. had a very fine
bloodhound bitch which would never take the dog. But at last
a rough-haired shepherd dog lined [i.e., bred] her and produced
a very large litter; never afterwards went in heat. This is good
instance of same fact in Mr. Galton’s case. It explains the loss
(and expense) (must probably have occurred to everyone) of
rare breeds of dogs from owners great care of them. Fox says
when two dogs of opposite breeds are crossed, sometimes off-
spring quite intermediate, sometimes take strongly after either
parent, about as often one way as other. He has known ease
of good pointer and rough water spaniel produce litter like both
parents and Mr. Bell has half bloodhound and grey hound.

Where two dogs have lined bitch directly one after the other,
puppies differ and like both parents. Fox told me of case of
mare covered by blood horse and ecarthorse, two folds [sic].1°

Mr. Herbert’s !”° papers are in the Horticultural Transactions
and a distinct work on hybridity under title of Amaryllidae and
Narcissus. (Mr. Donn ™1 considers Mr. H. rather wild.)

214 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Mr. Donn remarked to me that give him a species from Ire-
land, England, Scotland, and other localities and each one will
have a peculiar constant aspect. That is, varieties though of
trifling order are formed by nature.

Carmichael,! Tristan D’Acunha, a list of its Flora is given.
Mr. Don [sic] remarked to me that some good African and some
eood S. American forms (and daresays some of these forms would
have some [same?] peculiarity). Now when we hear that the
whole island is voleanie surmounted by crater and studded with
others, we see a beginning to island. Graham Island. We know
many seeds might be transported, some blown, — floating trees.
Thrushes (Turdus Guyanensis?), and bunting (Emberiza Bra-
siliensis?), and coots (Fulica chloropus) might bring in stomach,
ete., ete. (Mem.: discover what kinds of seeds, then plants.)
Mem.: Fact stated by Mr. Don in island Teneriffe, St. Helena, J.
Fernandez, Galapagos. Many trees compositae (Ferns ditto),
because seeds first arrived and hence formed trees, and would
creator (on voleanic island) make plants (grow closely) when
this voleanie point appeared in the great ocean, have made plants
of American and African form merely because intermediate po-
sition? We cannot consider it as adaptation because volcanic
island, whilst Africa [is] sandstone and granite (that is genera
near Cape); see if there are any species same as T. del Fuego
and C. of Good Hope; show possibility of transport. If some
cannot be explained, more philosophical to state we do not know
how transported. Glaciers might have acted at Tristan D’Acunha.
Carmichael, Linn. Transact., Vol. XII.

The Alpine plants of the Alps must be new formations because
snow formerly descended lower, therefore species of lower genera
altered, or northern plants. No. Mem.: The Antarctic flora
must formerly have been separated by short space from moun-
tains low down, therefore plants common. Take an example from
T. del. Fuego.

Ellis (?)* says Tahitian kings would hardly produce from
Incestuous intercourse, a parallel fact to Blood Hounds.

Before attract. of gravity discovered, it might have been said
it was as great a difficulty to account for movement of all [plan-
ets],1°* by one law, as to account for each separate one, so to say
that all Mammalia were born from one stock, and since distributed

I

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK ai

by such means as we can recognize, may be thought to explain
nothing, it being as easy to produce (for the creator) two quad-
rupeds (Jaguar and Tiger) at S. America.'”°

When species cross and hybrids breed, their offspring show
tendency to return to one parent, this is only character, and yet
we find this same tendency (only less strongly marked) between
what are called varieties. N.B., one mother bringing forth young
having very different characters is attempt at returning to par-
ent stock. I think we may look at it so, — ?? It holds good even
with trifling differences of expression, one child like father, an-
other like mother.

Has Lowe ® written any other paper besides one in Latin,
one on Madeira? Any general observations? Difference of
species between land shells of Porto Santo and Madeira. I be-
heve very curious.

My idea of propagation almost infers what we call improve-
ment. All Mammalia from one stock, and now that one stock
cannot be supposed to be most perfect (according to our ideas
of perfection), but intermediate in character. The same reason-
ing will allow of decrease in character (which perhaps is case
with fish, as some of the most perfect kinds, the shark, lived in
remotest epochs). Lizards of secondary period in same predica-
ment? It is another question, whether whole scale of Zoology
may not be perfecting by change of Mammalia for Reptiles, which
ean only be adaptation to changing world. I cannot for a mo-
ment doubt but what cetaceae and Phocae now replace Saurians
of Secondary epoch; it is impossible to suppose such an accumu-
lation at present day and not include Mammalian remains. The
Father of all insects gives same argument as father of Mammals,
but have improvement in system of articulation. Whether type
of each order may not be supposed that form which has wan-
dered least from ancestral form? If so are present typical spe-
cies most near in form to ancient? In shells alone can this com-
parison be instituted.

People often talk of the wonderful event of intellectual man
appearing. The appearance of insects with other senses is more
wonderful; its mind more different probably, and introduction
of man nothing compared to the first thinking being, although
hard to draw line; not so great as between perfect insect, ete.

276 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Forms hard to tell, whether articulata, or intestinal, or even mite.
A bee (compared with cheese mite) with its wonderful instincts
(might well say how —). The difference is that there is a wide
gap between man and next animals in mind, more than in strue-
tures.

If the skeleton of a negro had been found, what would Anato-
mists have said? Where is Pentland’s?!*" account of ... ?1°8

A, B, C, D. (A) crossing with (B), and (B) being crossed
with (C) prevents offspring of A becoming a good species well
adapted to locality, but it is instead a stunted and diseased form
of plant, adapted to A, B, C, D. Destroy plants B, C, D, and A
will soon form good species!

The increased fertility of shghtly different species and inter-
mediate character of offsprings accounts for uniformity of spe-
cies. We must confess that we cannot tell what is the amount
of difference which improves and checks it. It does not bear any
precise relation to structure. (Mem.: Eyton’s!?® Hogs and
Dogs. )

The passage in last page explains that between species from
moderately distant countries there is no test but generation (but
experience according to each group) whether good species, and
hence the importance naturalists attach to Geographical range
of species.

Definition of species: one that remains at large with constant
characters, together with other beings of very near structure.
Hence species may be good ones and differ scarcely in any ex-
ternal character; for instance two wrens forced to haunt two
islands, one with one kind of herbage, and one with other, might
change organization of stomach and hence remain distinet. Where
country changes rapidly we should expect most species.

The difference [in] intellect of man and animals not so great
‘as between living things without thought (plants), and living
things with thoughts (animals). (Therefore my theory very
distinct from Lamarck’s.) Without two species will generate
common kind, which is not probable ;!°° then monkeys will never
produce man, but both monkeys and man may produce other spe-
cies. Man already has produced marked varieties and may some-
day produce something else, but not probable owing to mixture of
races. When all mixed and physical changes (intellectual[ity |

.

bo
-~l
=~]

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK

being acquired alters case?), other species or angels produced.

Has the Creator since the Cambrian formation gone on creating
animals with same general structure? Miserable, limited view.
With respect to how species are, Lamareck’s ‘‘willing’’ doctrine
absurd. As equally are arguments against it, namely how did
otter live before being modern otter? Why to be sure there were
a thousand intermediate forms. Opponent will say, show them
me. I will answer yes, if you will show me every step between
bull Dog and Greyhound. I should say the changes were effects
of external causes, of which we are as ignorant as why millet
seed turns a Bullfinch black (or iodine on glands of throat), or
colour of plumage altered during passage of birds (where is this
statement, I remember L. Jenyns talking of it?), or how to make
Indian cow with hump and pig’s foot with cloven hoof.

Ask Entomologists whether they know of any ease of intro-
duced plant, which an insect has become attached to, that insect
not being called omniphitophagous [sic]. But it will be said
there are latent insects, as crows against man with gun, and
Bustards, ete., ete.!!!

An American and African form of plant being found in Tristan
D’Acunha may be said to deceive man as likely as fossils in old
rocks for same purpose! !

Can the wishing of the Parent produce any character on off-
spring? Does the mind produce any change in offspring? If so
[is] adaptation of species by generation explained? N.B. Look
over Bell! on Quadrupeds for some facts about dogs, ete., ete.
N.B. Animals very remote, ass and horse produce offspring ex-
actly intermediate. Reference to pig and dogs.

My theory will make me deny the creation of any new quad-
ruped since days of Didelphus in Stonefield [sic] !°? therefore all
lands united (Falkland Fox, ice). Mauritius, what a difficulty,
where elevation, subsidence near is only hope. New Zealand
(compare to Van Dieman’s land), glorious fact of absence of
quadrupeds. Kast India Archipelago very good on opposite ten-
deney.

Study Ellis and Williams,'* Zoology of South Sea islands. Any
animals? I believe none. Canary islands? Madeira? Tristan
D’Acunha? Iceland? The connection between Mauritius and
Madagascar very good; Fernando Po and Coast of Africa equally

278 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

good. Small island off New Guinea same fact. See Coquille’s Voy-
age.18+ Galapagos mouse (brought by canoes) (?). Ceylon and
India; Van Diemen’s land; Australia; England and Europe. It
will be well worth while to study profoundly the origin and his-
tory of every terrestrial Mammalia, especially moderately large
ones.

Is the Flora of Tierra del Fuego like that of North Europe,
many genera and few species?

The number of genera on islands and on Arctic shores evidently
due to the chance of some ones of the different orders being able
to survive or chance having transported them to new station.
When the new island splits and grows larger, species are former
[i.e., formed] of those genera (and hence by same chance few
representative species). (This must happen, and then enquiry
will explain representative Systems.) Of this we see example in
English and Irish Hare, Galapagos’ shrews, and when big con-
tinent, many species belonging to its own genera. Therefore if
in small tract we have many species, We may insure mass con-
tinental or many large islands. Hence this must have been con-
dition of Paris basin land. (How is this with Fernando Po, with
plants of St. Helena and Tristan D’Acunha?) If on one island
several species of same genus, subsided land, Mauritius? (Re-
solves itself into question of proportion of species to genus.)
Although the Horse has perished from 8. America, the jaguar
has been left, and fox, and bear. If I had not discovered channel
of communication by which great Edentata might have roamed
to Hurope and Pachydermata from Europe to America, how
strang [sic] would presence of Jaguar [have] been in 8S. America?

W. Coast of Africa and KE. of America ought to present great
contrast in forms. India intermediate, see how that 1s.

Are shell-boring [shell-bearing] molluses lke Carnivorous
Mammalia in their wide range and in their duration of species?
(Are carnivorous Mamm. in Paris basin allied to present, [1.e., ]
more like present Carnivora than Pachydermata? )

If my theory true, we get: (1Ist.) a horizontal history of earth
within recent times and many curious points of speculation ; for
having ascertained means of transport, we should then know
whether former lands intervened; (2nd.) by character of any
two ancient fauna, we may form some idea of connection of those

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 279

two countries (hence India, Mexico, and Europe, one great sea;
coral reefs ; therefore shallow water at mtns. [form] the islands) ;
(3rd.) we know that structure of every organ in A.B.C., three
species of one genus, can pass into each other (by steps we see) ;
but this cannot be predicated if structures in two genera. (We
then cease to know the steps.) #° Although D.E.F. follow close
to A.B.C., we cannot be sure that structure (C) could pass into
(D). We may foretell species, limits of good species being
known. It explains the blending of two genera. It explains typi-
cal structure. Every species is due to adaptation and hereditary
structure (latter far chief element, therefore little service habits
in classification, or rather the fact that they are not [by] far the
most serviceable). We may speculate of durability of succession
from what we have seen in old world and on amount [of] changes
which may happen. It leads you to believe the world older than
GEOLOGISTS think. It agrees with excessive inequality of numbers
of species in divisions. Look at articulata!!! It leads to nature
of physical change between one group of animals and a successive
one. it leads to knowledge what kinds of structure may pass into
each other; now on this view no one need look for intermediate
structure, say in brain between lowest Mammal and Reptile (or
between extremities of any great divisions). Thus a knowledge
of possible changes is discovered for speculating on future.
Therefore fish never become a man. Does not require fresh crea-
tion. If continent had sprung up round Galapagos on Pacific
side, the Oolite 1°° order of things might have easily been formed.
With belief of transmutation and geographical grouping we are
led to endeavour to discover causes of change, the manner of
adaptation (wish of parents??); instinct and structure become
full of speculation and line of observation. View of generation
being condensation, test of highest organization intelligible; may
look to first germ, led to comprehend two affinities. My theory
would give zest to recent and fossil Comparative Anatomy, it
would lead to study of instincts, heredity and mind heredity,
whole metaphysics. It would lead to closest examination of hy-
bridity and generation, causes of change in order to know what
we have come from and to what we tend, to what circumstances
favour crossing and what prevents it; this, and direct exam-
ination of direct passages of structures in species, might lead

280 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

to laws of change, which would then be main object of study, to
euide our speculations with respect to past and future. (The
grand question which every naturalist ought to have before him
when dissecting a whale or elassifying a mite, a fungus, or an
infusorian is, ‘‘What are the Laws of Life?’’)

Where we have near genera far back, as well as at present
time, we might expect confusion of species. Important. For in-
stance take Valvata and Conus (??) 1" which now run together ;
were not both genera formerly abundant?

Seed of Ribston Pippin tree producing erab.!** (Is [it like]
the offspring of a male and female animal of one variety going
back?) Whether this going back may not be owing to cross from
other trees???? Do the seeds of Ribston Pippin and Golden Pip-
pin produce real erabs, and in each ease similar or mere mongrels?

It really would be worth trying to isolate some plants under
glass bells and see what offspring would come from them. Ask
Henslow for some plants whose seeds go back again, not a mon-
strous plant, but any marked variety. Strawberry produced by
seeds?? Universality of generation strongly shown by hybridity
of ferns, hybridity showing connexion of two plants.

Animals whom we have made our slaves we do not like to con-
sider our equals. Do not slave holders wish to make the black
man other kind? Animals with affections, imitation, fear of
death, pain, sorrow for the dead, respect.

We have no more reason to expect [to find] the father of man-
kind, than Macrauchenia, ™° yet it may be found. We must not
compare chances of embedment in man in present state with what
he is as former species. His arts would not then have taken him
over whole world. The soul by consent of all is superadded; ani-
mals not got it, not look forward. If we choose to let conjecture
run wild, then animals our fellow brethern in pain, disease, death
and suffering, and famine, our slaves in the most laborious works,
our companions in our amusements; they may partake from our
origi in one common ancestor; we may be all netted 14° together.

Hermaphrodite animals couple; argument for true molluses
coupling.'!

Geograph. Journal, Vol. V, P. I, p. 67: Dr. Coulter 44? on de-
crease of population in California, cessation of female offspring ;
applicable to any animal.

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 281

Athenaeum.,!*? p. 154, 1838. Hybrid Ferns.

It may be argued against theory of changes, that if so, in ap-
proaching desert country or ascending mountains you ought to
have a gradation of species; now this notoriously is not the case.
You have stunted species, but not such as would make species
(except perhaps in some plants and then a chain of steps is found
in same mountain). How is this explained by law of small dif-
ferences producing more fertile offspring? Ist. All variation of
animal is either effect or adaptation, therefore animal best fitted
to that country when change has taken place. Nature .. .1*

Any change suddenly acquired is with difficulty permanently
transmitted. A plant will admit of a certain quantity of change
at once, but afterwards will not alter. This need not apply to
very slow changes, without crossing. Now a gradual change can
only be traced geologically (and then monuments imperfect) or
horizontally, and then cross-breeding prevents perfect change.
It is scarcely possible to get evidence of two races of plants run
wild (for we know that such can take place without impregnating
each other), for if they are different then they will be called
species, and mere producing fertile hybrids will not destroy that
evidence, as so many plants produce hybrids, or else whole fabric
will be overturned. Hence extreme difficulty, argument in circle.
Falkland Island case good one of animals not soon being sub-
jected to change in Americas; perhaps merely gone back previous
to fresh change. Get a good many examples of animals and plants
very close (take European birds, instance Gould’s case of Willow
wren) and others varying in wild state to show that we do not
know what amount of difference prevents breeding; or as others
would express it, amount of varying in wild state.

When breaking up the primeval continent, Indian Rhinoceros,
Java, and Sumatra ones, all different. Join Sumatra and Java
together by elevation now in progress and you will have two
Tapirs existing in East Indian seas. Marsupial animals all show
greater connexion in Quadrupeds, but plants do not follow by
any means. Ostriches, Hippotomus [sic] only African; American
and African forms maybe in India and Kast Indian Islands?
Monkeys different, not travellers? ?

Royle’s*° case of Himalayan plants; migratory birds? He
told me some story of Crane from Holland!!!, in stomach, or in
feathers — seeds.

282 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Two inhabitants of the Tropies (whether one fossil or not) are
related by real relationships, as well as effect of similar tempera-
ture; now those of temperate regions and tropics are only re-
lated by one connection, viz., descent. Hence far greater dis-
cordance in latter. Hence change in form. This probably ex-
plains erag'® and miocene. The descendants left in cooling
climate might change twice over, whereas those which migrated
a little to the southward would merely be specifically different,
if so. Now this is difficult to explain by creation, so we must sup-
pose a multitude of small creations.

Will Dromedaries and Camels breed?

As man has not had time to form good species, so cannot the
domesticated animals with him! Modern origins shown by only one
species, far more than by non-embedment of remains. Agrees
with non-blendine of languages? Till man acquired reason he
would be limited animal in range, hence probability of starting
from one point. In the crag we see the process of change of those
forms which have succeeded in becoming habituated to colder cli-
mate whilst others died out, or moved toward equator (or some
species might then have been wanderers.) There ought to be
fewer species in proportion to genera than in present seas. All
the species which survive any change may undergo indefinite
change (marking in their history an Eocene, Miocene, and Plio-
cene epoch), whilst others may die out or move southward. There-
fore species must be compared to neighbouring sea. For change
of species does not measure time, but physical changes (we as-
sume like weather on long average tolerably uniform). Com-
paring fossils with whole world would be like in a Meteorologie
table (in comparison of temperature of two countries), finding
a very hot day in one; oh, we will take a day from the equator
to add to the mean of the other. If the world had cooled by secu-
lar refrigeration in chief part instead of change from insular
to extreme climate, Iceland would have possessed a most peculiar
Flora (and north of Europe). As European forms have travelled
towards Equator, so would the plants from extreme north, which
according to all analogy would have been very unlike southern
BKuropean ones, — ‘‘a variation played on secular refrigeration.”’

Experimentise on land shells in salt water, and lizards ditto.
Ask Eyton to procure me some.

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 283

Get Hope! to give me an account of parasitic animals of
beasts varying in different climates.

Those will not object to my theory, those the philosophy who
soar above the pride of the savage, they perceive the superiority
of man over animals, without such resorts. 148

M. Jarred and Dumeril '** great work on Reptiles. M. J. says
some reptiles same from Maurice [Mauritius] and Madagascar,
and of Good Hope. His book probably worth studying.

Wingless birds [of] S. Continents, Ostriches, Dodo, Apteryx,
Penquin, Loggerheaded Duck; larger proportion of water and
small[er] of land, and a few quadrupeds.

Study productions of great Fresh Water Lakes of North
America.

If parasites different whilst man and his domesticated quad-
rupeds are not so, [then] greater faculties of change in the ar-
ticulata than Vertebrata. But how does this agree with the lon-
vevity of species in Molluscs!!!

When we talk of higher orders, we should always say intel-
lectually higher. But who with the face of the earth covered
with the most beautiful savannahs and forests dare to say that
intellectuality is only aim in this world... ? 4°

.. of all genera (in all classes) are not a few only cosmopoli-
tan, and in genera peculiar to any one country do not species
generally affect different stations? This would be strong argu-
ment for propagation of species.

Again, is there not similarity even in quite distinct countries
in same hemisphere more than in others? Are there any cases
when domesticated animals separated and long interbred having
great tendency to vary? Is not man thus cireumstanced? Va-
rieties of dogs in different countries a case in point. All cases
like Irish and English Hare bear upon this.

Why do Van Diemen’s land people require so many imported
animals?

At what part of tree of life can orders like birds and animals
separate, etc., etc.?

Work out Quinary system according to three elements.

How is Fauna of Van Diemen’s Land and Australia... ? ©!

Falconer’s '°* remarks on influence of climates, situations, etc.,
on — (242)

284 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Smellie 1? Philos. of Zoology (842).

White !* regular gradat. in man (1024). (Poor, trash; Lyell.)

Fleming’s © Philosophy of Zoology. Royle ® on Himalayan
Plants.

Would it not be possible to work through all genera and see
how many confined to certain countries?; so on with families.

Ask Royle about Indian cattle with humps.

To be solved: if horses sent to India and long bred in and no
new ones introduced, would not change be superinduced? Why
is everyone so anxious to cross animals from different quarters
to prevent them taking peculiar characters? Indian Bull?

Do species of any genus as American or Indian genus inhabit
different kind of localities? If so change.

The grand question: Are there races of plants run wild or
nearly so, which do not intermix, any cultivated plants produced
by seed? Lychnis, Flax.

In production of varieties, is it not per saltum ?

Islands bordering continents same type; collect cases. African
Islands? How is Juan Fernandez— Humming Birds? Types
of former dogs; character of Miocene Mammalia of Europe.

Mem.: Mr. Bell’s!*® case of Sub Himalayan land emys, de-
cidedly an Indian form of Tortoise. On other hand fresh water
tortoises from Germany (where Mr. Murchison’s® fox was
found) decidedly next species to some South American kinds.

Are the closest allied species always from distant countries, as
Decandelle [sic] 1® says? (No, he only says sometimes.) We
might expect disseminated species to vary a little, but such should
not be general circumstance. In insects in England surely it is
not; intermediate genera we might expect.

Lindley 16 Introduct.

Dict. Scien. Naturelle.1&

Geographie Botanique, De Candoelle [sic ].1%*
Geol. Soe.

Horae Entomologicae.

Linn. Soe.

Geoff. St. Hilaire, Philosophy of Zoology.
Waterhouse.1®®

bo

10.

iil.

13.

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 285

NOTES AND REFERENCES

Darwin undoubtedly has reference to the book of his grandfather:
Darwin, Erasmus. 1794-1796. Zoonomia; or, the laws of organic life,
2 vols. London, J. Johnson.

Probably: Lamarck (J. B. P. A. de Monet de). 18380. Philosophie
zoologique. Paris.

Gray, John Edward. 1835. Remarks on the difficulty of distinguishing
certain genera of testaceous Mollusca by their shells alone, and on the
anomalies in regard to habitation observed in certain species. Philo-
sophical Transactions of the Royal Society of London: 301-310.

Lyell referred to dispersal of plants by icebergs, or ‘‘ice-islands.’’
Lyell, Charles. 1837. Principles of geology, 5th ed., vol. 3, p. 16.
London, Murray.

Probably Candolle, Augustin Pyramus de, and K. Sprengel. 1821.
Elements of the philosophy of plants. (Translated from the German. )
Edinburgh, Blackwood.

Francis Darwin transcribed ‘‘immense ages’’ as ‘‘immersage,’’ see:

Darwin, Francis. 1898. Life and letters of Charles Darwin, vol. 1, p.
367. New York, Appleton.

Darwin undoubtedly has reference to Ehrenberg, C. G. 1837. On the
origin of organic matter from simple perceptible matter and on organic
molecules and atoms; together with some remarks on the power of
vision of the human eye. Scientific Memoirs, Selected from the Trans-
actions of Foreign Academies of Science and Learned Societies, and
from Foreign Journals; edited by R. Taylor, 1:555-583,

This no doubt is Owen, Richard, author of: Zoology of the voyage of
H.M.S. Beagle, Part 1, Fossil Mammalia. London, 1840.

This entire sentence is unfinished and crossed out.

Remainder of page 29 and first half of page 30 are missing in the
notebook.

Bell, Thomas. 1837. A history of British quadrupeds, including the
Cetacea. London.

Jenyns, Leonard. 1835. A systematic catalogue of British vertebrate
animals, Cambridge.

Duperrey, Louis Isidore. 1826-30. Voyage autour du monde, exécut?
sur la Corvette du Roi La Coquille .. . pendant les années 1822-25 .. .
publie... par L. I. D. Zoologie par Lesson et Garnot. Paris.

86

14.

15.

16.

19.

20.

21.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Smith, Andrew, author of: Jllustrations of the zoology of South
Africa, consisting chiefly of figures and descriptions of the objects of
natural history collected during an expedition into the interior of South
Africa, in the years 1834-1886; ete. London, 1838. Darwin visited
with Smith while in Capetown, South Africa in 1836.

Humboldt, Alexander von. 1811. Political essay on the Kingdom of
New Spain. Transl. from the original French, by John Black, 2 vols.
New York.

I am not sure to what Darwin has reference here. He spells the word
‘‘monucle’’ in this line and ‘‘monucule’’ in the next. I have not
found either word in such context in any other writings by Darwin or
his contemporaries. Perhaps he means ‘‘monocule’’ and is suggesting
that the progenitor of all animals was a cyclops-like creature or per-
haps he means to say ‘‘molecule,’’ as he did previously on page 249.
Possibly he meant to use ‘‘minuscule,’’ See
also Ehrenberg, op. cit., note 7.

or even ‘‘monticule.’’

In the diagram there are 13 lines that have a perpendicular line at
the end.

In the diagram there are 12 lines that are without a perpendicular
line at the end.

In this and in the preceding paragraphs Darwin to all intents and
purposes formulates the postulates of his theory of natural selection,
viz., the survival of the fittest and the constancy of populations genera-
tion after generation.

Cuidado — Be careful!

Darwin no doubt has reference to the Quinary System of classification
described by Macleay, William Sharp, in Horae entomologicae; or,
essays on the Annulose animals, etc., vol. 1, pts. 1, 2, London, 1819,
1821. William Swainson also described the system in 4 treatise on
the geography and classification of animals, London, Longman, etc.,
in Lardner, D., The Cabinet Cyclopaedia. 1835. The system was based
on the postulate that the animal kingdom may be classified into five
major divisions and that these divisions have such a relation to each
other that they form a circle when grouped together according to
morphological similarities. Thus each division includes some taxonomic
forms which bear a close resemblance through affinities or analogies
to two of the other major divisions. Since each of the five major divi-
sions is related by some of its members to two of the remaining divi-
sions, the entire group of five must have a ecireular arrangement

29

24,

31.

32.
33.
34.

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 287

amongst themselves. Also each of the five major divisions itself was
thought to be composed of five smaller taxonomic units, each of which
similarly was circularly related to two of the five units within its
division.

Fries, Elias Magnus, author of: Systema Mycologicum, etc., 3 vols.,
1821-32. Fries, according to Swainson (op. cit., note 21, p. 216), also
suggested that taxonomic groups are arranged in circles, but in Fries’
system there were four major groups instead of five.

Swainson, op. cit. note 21, p. 220, mentions Newman (Edward), author
of Sphina Vespiformis: An Essay. London, 1832. Newman, according
to Swainson also supported the circular theory of classification, but
believed the ‘‘magic’’ number of circles to be seven.

Wellsted, Lieutenant R. 1836. Observations on the coast of Arabia
between Ras Mohammed and Jiddah. Jowrnal of the Royal Geographi-
cal Society of London, 6:51-96.

Remainder of page missing.

Gould, John. 1837-38. A synopsis of the birds of Australia and the
adjacent islands. Pts. 1-4. London.

Van Diemen’s Land is Tasmania.
Remainder of page missing.

Cuvier, G. 1827. Essay on the theory of the earth, 5th ed. (Trans-
lated from the French by R. Kerr.) Edinburgh, Blackwood.

Deshayes, Gérard Paul. 1836-37. Séance du 17 avril 1837. Bulletin de
la Société Géologique de France, 8:212-224.

This reference probably is Macculloch, John. 1831. A system of
geology, with a theory of the earth, and an explanation of its con-
nexion with the Sacred Records, 2 vols. London.

Probably Isidore Geoffroy Saint-Hilaire or René P. Lesson, or both.
Pages 55 and 56 missing.

Waterhouse, George R., author of: Zoology of the voyage of H.M.S.
Beagle, Part 2, Mammalia. London, 1839. I could not locate a refer-
ence to a specific paper by Waterhouse to which Darwin probably refers
here.

Heteromera is a division of beetles including darkling and blister
beetles; the Chrysomelidae are leaf beetles.

43.

44,

45.

46.

47.

48.

49.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Lyell, Charles. 1833. Principles of geology, vol. 3. London, Murray.
Pippin, a variety of apple.

Swainson, op. cit., note 21.

Pages 69 and 70 missing.

Pages 75 and 76 missing.

Crossed out.

Probably Don, George, author of A general system of gardening and
botany. London, 1832-1838, or possibly Don, David, both botanists, and
brothers.

Cuvier, G. 1829-30. Le régne animal distribué d’aprés son organisa-
tion, etc., Tomes 1-5. Paris.

Stonesfield slate, a geological formation of the Jurassic. See Lyell,
Charles. 1850. Principles of geology, 8th ed. p. 148. London, Murray.
See also Lylell, op. cit., note 4 (in vol. 1, p. 237).

Kotzebue, Otto von. 1821. A voyage of discovery, into the South Sea
and Beering’s Straits, for the purpose of exploring a North-East passage

5 at . 1815-1818. (Remarks and opinions of the Naturalist of the
Expedition: A. von Chamisso, vol. 2) 3 vols. London.

Richardson, John. 1829-37. Fauna Boreali-Americana; or the zoology
of the northern parts of British America. 3 vols. London.

Humboldt, F. H. Alexander von, and Aimé Bonpland. 1807. Voyage de
Humboldt et Bonpland. Part 5. Essai sur la géographie des plantes,
accompagné d’un tableau physique des régions équinowiales
Rédigé par A. de Humboldt. Paris.

Swainson, op. cit., note 21.

Brown, Robert. 1831. General view of the botany of the vicinity of
Swan River. Journal of the Royal Geographical Society of London,
1:17-21.

Geoffroy Saint-Hilaire. 1837. Singe fossile de Sansan. L’Institut,
Journal des Académies et Sociétés Scientifiques de la France et de
l’Etranger, Paris, 5:242-244,

””)

““singe,’’ ape or monkey.

Colobus, a genus of African monkey.

oO
ba |

68.
69.

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 289
Semnopitheque, a name now technically invalid for the Old World
langur monkeys.

Sapajou, a monkey of the genus Cebus of S. America.
See Lyell, op. cit., note 4 (in vol. 2, p. 402).

Muscicapa coronata? Lath., a tyrant-flycatcher of the Galapagos
Islands. See Darwin, Charles, 1839. Journal of researches into the
geology and natural history of the various countries visited by H.M.S.
Beagle, p. 461. London, Colburn,

This word is crossed out, but looks as if it might have been ‘‘Chat’’
or ‘‘Chais.’?’? Perhaps Darwin meant ‘‘Calandria’’ a mocking-bird of
Patagonia. See Darwin, op. cit., note 56, pp. 62-63, where Orpheus
modulator and O. Patagonica D’Orbigny are mentioned; see also p.
475, where various species of the Galapagos’ Orpheus are mentioned.
Possibly Darwin intended to write ‘‘Chatham’’ for Chatham Island.

Perhaps this should read, ‘‘when points are finally gained, if any, then
the facts become connected.’’

Pages 107 and 108 missing.

Geoffroy Saint-Hilaire, Etienne, 1830. Principes de philosophie zoologi-
que, discutés en Mars 1830 aw sein de l’Académie Royale des Sciences.
Paris.

Op. cit., note 60.

Op. cit., note 60.

Op. cit., note 60.

Geoffroy Saint-Hilaire, Etienne, 1818. Philosophie anatomique. Paris.
The words ‘‘ different animals’’ are crossed out.

Sternberg, Count, 1827. On the distribution of living and fossil plants.
Edinburgh New Philosophical Journal, April to October: 190-192.

Cuvier, Frédéric. 1827-28. Essay on the domestication of mammi-
ferous animals, with some introductory considerations on the various
states in which we may study their actions. Edinburgh New Philosoph-
ical Journal, October to April: 45-60; 292-297.

The correct quotation is ‘‘coneurrence.”’’

Darwin, apparently inadvertently, omitted this portion of the quotation.

290

70.

ele

81.

84.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Crawford. 1827. Account of Mr. Crawford’s mission to Ava. Edin-
burgh New Philosophical Journal, April to October: 359-370.

Park, Thomas. 1827-28. Mr. Thomas Park’s journey into the interior
of Africa. Edinburgh New Philosophical Journal, October to April:
410.

Pages 123 through 128 missing.

Probably has reference to Macleay, op. cit., note 21, who discussed
analogy and affinity.

Thus the wings of birds and bats are related to the air.

Monteith, Colonel (William), E. I. C. 1833. Journal of a tour through
Azerdbijan and the shores of the Caspian. Journal of the Royal Geo-
graphical Society of London, 3:1-58.

Hilhouse, William. 1834. Journal of a voyage up the Massaroony in
1831. Journal of the Royal Geographical Society of London, 4:25-40.

Campbell, Major (James). 1834. Geographical memoir of Melville
Island and Port Essington, on the Cobourg Peninsula, Northern
Australia; with some observations on the settlements which have been
established on the north coast of New Holland. Journal of the Royal
Geographical Society of London, 4:129-181.

Op. cit., note 50.
Undoubtedly Geoffroy Saint-Hilaire, op. cit., note 50.

Baer. 1837. Aurochs du Caucase. L’Institut, Journal des Académies
et Sociétés Scientifiques de la France et de l’Etranger, Paris, 5:260-261.

Sir Francis Sacheverel Darwin (1786-1859). For biographical notes,
see Pearson, Karl. 1914. The life, letters and labours of Francis Gal-
ton, vol. 1:22-25. Cambridge Univ. Press.

See Darwin, op. cit., note 56 (p. 163 in Darwin).

Yarrell, William, author of numerous papers including: On the laws
which regulate the change of plumage in birds. Transactions of the
Zoological Society of London, 1:13-20, 1835.

Eyton, Thomas C. 1837. Some remarks upon the theory of hybridity.
Magazine of Natural History, and Journal of Zoology, Botany, Mineral-
ogy, Geology, and Meteorology, London, 1:357-359. Conducted by J. C.
Loudon.

86.

88.

89.

90.

ile

93.
94.

o
ii

96.

BARRETT : DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 291

Gould, John. 1837. Observations on some species of the genus Motacilla
of Linnaeus. Magazine of Natural History, and Journal of Zoology,
Botany, Mineralogy, Geology, and Meteorology, London, 1:459-461.
Conducted by J. C. Loudon.

Westwood, J. O. 1837. On Diopsis, a genus of dipterous insects, with
descriptions of twenty-one species. Transactions of the Linnean Society
of London, 17:283-314.

I could find no reference to Wynne in the published literature.

Undoubtedly Lord Powis. See Darwin, Charles. 1868. Variation of
animals and plants under domestication, vol. 1, p. 83; vol. 2, p. 45.
London, Murray.

Darwin’s cousin, William Darwin Fox.

Kirby, William. 1835. On the power, wisdom and goodness of God as
manifested in the creation of animals and in their history, habits and
instincts, 2 vols. London, William Pickering. (Bridgewater Treatises.)

Humboldt, Alexander de, and Aimé Bonpland. 1821. Personal narvra-
tive of travels to the equinoctial regions of the New Continent during
the years 1799-1804. Vol. 5, pt. 2. (Transl. by Williams) London,
Longman, ete.

Kirby, op. cit., note 90.
Cline, Henry. 1805. On the form of animals. London.

Buckland, William. 1836. Geology and mineralogy considered with
reference to natural theology. Wondon, William Pickering. (Bridge-
water Treatises. )

Brongniart, Ad. 1837. Végétaux fossiles. L’Institut, Journal des
Académies et Sociétés Scientifiques de la France et de l’Etranger, Paris,
5:318-321.

Pages 151 through 154 missing in the notebook. The reference is:
Ehrenberg. 1837. Infusoires fossiles du tripoli d’Oran. L’Institut,
Journal des Académies et Sociétés Scientifiques de la France et de

l’Etranger, Paris, 5:330-331.

Buch, Leopold von. 1836. Description physique des Iles Canaries,
suivie d’une indication des principaux volcans du globe. Traduite de
1’Allemand par C. Boulanger. Revue et augmentée par |’auteur. Paris,
Levrault.

102.

103.

106.

MOM

108.

109.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Humboldt, F. H. Alexander von. 1817. De distributione geographica
plantarum secundum coeli temperiem et altitudinem montium, prole-
gomena. Lutetiae Parisiorum.

Lyell, op. cit., note 4 (p. 367 in Lyell).
Pages 159 and 160 missing.
Probably Richard Owen.

Undoubtedly John Hunter. For various references, see Abernethy, J.
1817. Physiological lectures exhibiting a general view of Mr. Hunter’s
physiology and of his researches in comparative anatomy, etc.; Hunter,
John. 1835-37. The works of John Hunter. With notes. Edited by
J. F. Palmer. 4 vols. London.

Eyton, Thomas C. 1837. Notice of some osteological peculiarities in
different skeletons of the genus Sus. Proceedings of the Zoological
Society of London, Part 5:238-24.

See note 34.

This reference possibly is one of the following: Ogilby, W. 1833.
Characters of a new genus of carnivorous Mammalia (Cynictis), from
the collection of Mr. Steedman. Proceedings of the Zoological Society
of London, Part 1:48-49; Ogilby, W. 1836. Remarks upon the probable
identity of Cynictis melanurus Mart., with a species noted by Boshman
under the name of Kokebog. Proceedings of the Zoological Society of
London, Part 4:56.

Serres. 1837. Anatomie des mollusques. L’Institut, Journal des
Académies et Sociétés Scientifiques de la France et de l’Etranger,
Paris, 5:370-371.

Pages 165 and 166 missing from the notebook. The reference un-
doubtedly is Horsfield, Thomas, author of several articles including:
Notice of a species of Ursus (U. isabellinus) from Nepal. Transactions
of the Linnean Society of London, 15:332-334, 1827.

I could find no such specific reference, see however Smith, op. cit.,
note 14.

Phillips, John. 1837. A treatise on geology. London, Longman, etc.,
in Lardner, D. The Cabinet Cyclopaedia, etc., vol. 1.

110.

1k

ile

116.
ily

118.

Is)
120.

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 293

Wealden, a freshwater geological formation underlying the Lower
Cretaceous in England. See Lyell, op. cit., note 4 (vol. 4:302 in Lyell).

Pages 173 and 174 missing.

For further citations on Aston in the literature see: Pearson, Hesketh.
1930. Doctor Darwin (p. 180). London, Dent and Sons; Oswald,
Arthur. 1953. Aston Hall, Warwickshire—I. The property of the
Corporation of Birmingham. County Life, London, 114:552-555, 620-
623, 694-697; Historic Houses and Castles in Great Britain and North-
ern Ireland (1959 Edition), Index Publishers Limited, London, 1959.

Pages 177-178 missing.

Deshayes, Gérard Paul. 1833-1834. Mémoires et communications séance
du 3 février 1834. Bulletin de la Société Géologique de France,
4:200-293.

I have not been able to find any bibliographic reference to Mr. McClay.
Undoubtedly should be Macleay (see note 21).

Dr. Andrew Smith, op. cit., note 14.

Herbert, William. 1822. On the production of hybrid vegetables; with
the result of many experiments made in the investigation of the sub-
ject. Transactions of the Horticultural Society of London, 4:15-50. In
this article Herbert proposes that new species are formed by divergence
and hybridization.

Morton, George. 1821. A singular fact in natural history. (Peculiar-
ities of the progeny of an Arab horse from a mare that had previously
bred with a Quagge.) Philosophical Transactions of the Royal Society
of London: 20-23.

Pages 185 through 190 missing. ‘‘ Folds’’ no doubt should be ‘‘foals.’’

Herbert, William. 1820. Instructions for the treatment of the Amaryllis
longifolia, as a hardy aquatic, with some observations on the produe-
tion of hybrid plants, and the treatment of the bulbs of the genera
Crinum and Amaryllis. Transactions of the Horticultural Society of
London, 3:187-196. An interesting quotation from p. 196 of this article
is the following: ‘‘Considering the wide field that is open for the
creation of new species of plants, by hybrid intermixture, some mode
of naming them must be adopted, or the art of cultivators will break
down all the landmarks of the botanist.’’? See also Herbert, William.
1837. Amaryllidaceae; preceded by an attempt to arrange the mono-
cotyledonous orders, and followed by a treatise on cross-bred vegetables,
etc. London.

294

121.

128.
129.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY
Darwin probably was referring to either George Don or David Don,
and not Donn. See note 42.

Carmichael, Dugald. 1818. Some account of the island of Tristan da
Cunha and its natural productions. Transactions of the Linnean Society
of London, 12:483-513.

Probably Ellis, William. 1829. Polynesian researches, during a resi-
dence of nearly six years in the South Sea Islands, etc. 2 vols., London.

The word ‘‘planet’’ was inserted in brackets by Darwin, Francis, op.
cit., note 6 (p. 370 in Darwin).

Pages 197 through 202 missing.

Lowe, R. T. 1833. Primitiae Faunae et Florae Maderae et Portus-
Sancti; sive species quaedam novae vel hactenus minus rite cognitae
animalium et plantarum in his insulis degentium breviter descriptae.
Transactions of the Cambridge Philosophical Society, 4:1-70.

The most nearly appropriate bibliographic references that I could find
are: Laurillard, Charles Léopold, Valenciennes and Pentland. 1835?
Catalogue des préparations anatomiques laissées dans le Cabinet
d’ Anatomie comparée du Muséum d’Histoire Naturelle, par G. Cuvier.
(Paris?) ; and Darwin, op. cit., note 56 (p. 153 in Darwin).

Pages 209 and 210 missing.
Op. cit., notes 84 and 103.

I believe Darwin means that under his theory two distinct species
would not independently evolve into a single species.

Bell, op. cit., note 11.
See note 44. The correct spelling should be ‘‘Stonesfield.’’

Op. cit., note 123. See also Williams, John. 1837. Narrative of Mis-
stonary enterprises in the South Sea Islands; with remarks upon the
natural history ....etc. London.

See Duperrey, op. cit., note 13.
This sentence crossed out.

Oolite, or Jura Limestone Group. See Lyell, op. cit., note 4 (vol. 4,
p. 303 in Lyell).

Valvata, a genus of freshwater snails; Conus, a genus of marine snails.

Probably the Ribston Pippin apple occasionally produced a crab apple.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.
149.

150.
151.

152.

BARRETT: DARWIN’S FIRST EVOLUTIONARY NOTEBOOK 295

Macrauchenia, an extinct three-toed, long-necked ungulate of the South
American Pleistocene.

Francis Darwin transcribed this word as ‘‘melted’’ rather than
‘“netted,’’ op. cit., note 6 (p. 368 in Darwin).

Pages 233 and 234 missing.

Coulter, Thomas. 1835. Notes on upper California. Journal of the
Royal Geographical Society of London, 5:59-70.

Martens, M. 1838. On hybridity in ferns. Athenaewm Journal of
English and Foreign Literature, Science and the Fine Arts, no. 539:
154. London.

Pages 237 and 238 missing.

Royle, John Forbes. 1835. Illustrations of the botany and other
branches of the natural history of the Himalayan Mountains, and of
the flora of Cashmere. Journal of the Royal Geographical Society of
London, 5:361-655.

Crag, a geological formation of the Pliocene of England. See Lyell,
op. cit., note 4 (vol. 4, pp. 71-72 in Lyell).

Probably Hope, Frederick W., author of: Descriptions of some species
of Carabidae, collected by Charles Darwin, Esq. in his late voyage.
Transactions of the Entomological Society of London, 2:128-130,
1837-40.

Pages 249 and 250 missing.

““M. Jarred’’ is crossed out. Perhaps Darwin confused Jarred with
Gérard of Gérard Paul Deshayes in: Bélanger, Charles. 1834. Voyage
aux Indes-Orientales, ... les iles de Java, de Maurice etc., ... Zoologie,
par M.M. Bélanger, I. Geoffroy Saint-Hilaire, Lesson, Valenciennes,
Deshayes et Guérin. Paris. The reference to Duméril is probably:
Dumeéril, André Marie Constant and Gabriel Bibron. 1834-54. Erpé-
tologie générale; ou, histoire naturelle complete des reptiles. Paris,
Roret.

Pages 258 through 260 missing.

The remainder of this and the next page blank in the notebook, and
pages 266-271 missing.

Falconer, Hugh, author of numerous zoological and botanical papers.
See: Catalogue of scientific papers (1800-1863). Royal Society of
London, 6 vols. London, Eyre, etc. 1867-1872.

160.

iL

162.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Smellie, William. 1790-99. The philosophy of natural history. 2 vols.
Edinburgh.

White, Charles. 1799. An account of the regular gradation in man,
and in different animals and vegetables; and from the former to the
latter. London.

Fleming, John. 1822. The philosophy of zoology; or a general view
of the structure, functions, and classification of animals, ete. 2 vols.
Edinburgh.

Royle, op. cit., note 145.
Pages 276 and 277 missing.

Bell, Thomas. 1828. On Hydrapsis, a new genus of freshwater tor-
toises, of the family Emydidae. Zoological Journal; by Thomas Bell,
ete., London, 3:511-513.

Murchison, Roderick Impey. 1835. On the fossil fox of Oeningen, with
an account of the lacustrine deposit in which it was found. Transac-
tions of the Geological Society of London. 3:277-290.

Candolle, Alphonse de. 1834. Fragment d’un discours sur la Géographie
Botanique. Bibliotheque Universelle des Sciences, Belles Lettres, el
Arts, etc., Genéve. 56:1-29.

Lindley, John. 1830. An introduction to the . .. natural system of
Botany, or a systematic view of the organisation . .. of the whole
vegetable kingdom, etc., London.

Dictionnaire des sciences naturelles, etc. Par plusiers professeurs,
ete., .. . Strasbourg, Levrault; Paris, Le Normant, 1816-30. (M. F.
Cuvier est chargé...)

Candolle, op. cit., note 160.
Horae entomologicae, op. cit., note 21.
Geoffroy Saint-Hilaire, op. cit., note 60.

Waterhouse, op. cit., note 34.

Bulletin of the Museum of Comparative Zoology
ATs Elie AW Ea Vic Ae Eee) © @ Ni Ter Ey Gok

Vor 1226No" 7

SKELETON AND MUSCULATURE OF THE THORAX OF
GELASTOCORIS OCULATUS (FABRICIUS)
(HEMIPTERA-HETEROPTERA )

By Maraaret C. PARSONS
Harvard Biological Laboratories

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM

May, 1960

PUBLICATIONS ISSUED BY OR IN CONNECTION
WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

BULLETIN (octavo) 1863 — The current volume is Vol. 122.

BREVIORA (octavo) 1952 — No. 124 is current.

Memorrs (quarto) 1864-1938 — Publication was terminated with
Vol. 55.

JOHNSONIA (quarto) 1941 — A publication of the Department of
Mollusks. Vol. 3, no. 39 is current.

OcCASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo)
1945 — Vol. 2, no. 25 is current.

PROCEEDINGS OF THE NEW ENGLAND ZooLocicAL CuLuB (octavo)
1899-1948 — Published in connection with the Museum. Publication
terminated with Vol. 24.

The continuing publications are issued at irregular intervals in num-
bers which may be purchased separately. Prices and lists may be
obtained on application to the Director of the Museum of Comparative
Zoology, Cambridge 38, Massachusetts.

Of the Peters ‘‘Cheek List of Birds of the World,’’ volumes 1-3 are
out of print; volumes 4 and 6 may be obtained from the Harvard Uni-
versity Press; volumes 5 and 7 are sold by the Museum, and future
volumes will be published under Museum auspices.

Bulletin of the Museum of Comparative Zoology
ANMY Jel AN Dey WN diy ID) (O}(O) 1D; 1b) 1) (EF 18:

Vorwi22, No.7

SKELETON AND MUSCULATURE OF THE THORAX OF
GELASTOCORIS OCULATUS (FABRICIUS)
(HEMIPTERA-HETEROPTERA )

By MARGARET C. PARSONS
Harvard Biological Laboratories

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM
May, 1960

n ‘i - a - ihe of sy is we ro 7 :

Ht »
ii eat) 1, We +. Sie -— s

: - . WS datrelt a) Ae Del Ve rust a nh Bly:
awe i | ny a) per e ree eas Rot ay | ae

a wa — -
116. cas -

ola ae,

. =~
~ Ae t E - } an - :
cal a R - ee é. j 4
| ' | he > ¢
ar Ve ak
. 7 | . _
ate

No. 7 —WNkeleton and Musculature of the Thorax of
Gelastocoris oculatus (Fabricius )
(Hemiptera-Heteroptera)

By Maraaret C. Parsons

TABLE OF CONTENTS

PAGE
emt.0 GC HO Ne Sas aan RE ee cy AA a aes ee De ern ee a 299
Niatenialswan de sMethod sian. 2.4) ae, ke ERR rahe pat inane 301
Skeleton
Frit OTN wee pera RA and oD EP orf a Ag OS cs I A oe 302
RtenOGhonaee perce te. et Cae ee eat ek ene ers hopes : ; 308
IL SERS a ek a ik ate I 2 eerie peli da ms & ieee RAIN ee ed RIBAS he, te 321
NUNIT Vege MS a tr gee la et ch ede SA or Le Rey hi eR Met 326
Museulature
Muselesmots thelr othonasds 12 hoe sth cseiseie 4 eee senor chee. ae 329
Muscles otetlenmesothonaquy tort t ae ee ees es, Sennen Ae eee ere ar 335
Mars clesrotathemmetathonaxensste ten ey net eae iehcrs certs © eeyaie cate 339
Intrinsic muscles of the legs DA Leek ee eR SNe eS he 343
DT SCUSSTON peas ene ie a Ay oe ey An oe tye a A Se Spee. Mae cee Bae 346
iberaunegciteds <p at ono. egtews ee eae cee Ba oe 302
Ep lame tomo fe OUNeSh sc 1.s hs. 5 eeu ees ss ro SONG Gaeta as 355

INTRODUCTION

The members of the family Gelastocoridae are predaceous,
littoral inseets which live in swamps or along the shores of ponds
and streams. The family is comprised of only two genera,Gela-
stocoris and Nertha. They are commonly called ‘‘toad bugs’’
beeause of their jerky, hopping means of locomotion, their
roughened exoskeleton, and their prominent, laterally projecting
compound eyes.

In a previous paper (Parsons, 1959) the author has described
the cephalic skeleton and musculature of Gelastocoris oculatus.
The study of the anatomy of this insect was undertaken for two
reasons. First, except for those characters which are of taxonomic
interest, almost nothing was known of the morphology of the
Gelastocoridae. The few earlier works on the biology and tax-
onomy of this family are briefly reviewed in my previous paper.

300 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Secondly, the Gelastocoridae, along with the closely related shore-
dwelling families Ochteridae and Saldidae, occupy a basic posi-
tion in previous theories of the evolution of the aquatic and
semi-aquatiec Heteroptera. These families are generally believed
to represent an intermediate stage in the evolution of the water
bugs from terrestrial forms. The various phylogenetic schemes
which have been proposed have been based upon relatively few
anatomical features, most, if not all, of which are external. It
appeared, therefore, that a thorough study of both the external
and internal morphology of one of the littoral Heteroptera might
reveal additional evidence which could clarify the phylogenetic
position of this group. Gelastocoris was used for this study be-
cause it could be obtained in large numbers and because it proved
to be an excellent, easily kept laboratory animal.

The present study was greatly facilitated by the previous work
of Larsén. That author has published many excellent papers on
the thorax of the Heteroptera, and the two works (1945a and b)
in which he described and compared in detail the thoracic skele-
ton and musculature of representatives of a great many different
families are especially helpful. Previous to Larsén’s studies
there were very few works on the heteropteran thorax. Tower
(1913) deseribed briefly the external appearance of the thorax
in Anasa, and Taylor (1918) discussed the thoracic sclerites of
several families of Heteroptera, but included little detail. Ham-
ilton (1931) studied the skeleton and musculature of the thorax
of Nepa rather superficially. A much more extensive description
of the skeleton and musculature was given by Malouf (1933) in
his well illustrated paper on the thorax of Nezara. Rawat (1939)
also briefly described both the muscles and the sclerites of this
region in Naucoris. More recently Griffith (1945) and Esaki and
Miyamoto (1955) have described the thoracic skeletons of Ram-
phocorixa and the Veliidae respectively. Three other studies have
been made which inelude both the skeleton and the musculature ;
these are the works of Sprague (1956) on Hydrometra, Akbar
(1957) on Leptocorisa, and Lauck (1959) on Lethocerus. The
latter study is particularly valuable for its brief but thorough
deseription of the musculature.

The author is indebted to Mr. Edwin P. Marks, of Washburn
University, and to the members of the C. V. Riley Entomological

«

PARSONS: THORAX OF GELASTOCORIS 301

Society, of Columbia, Missouri, who provided the insects used
in the present study. I also wish to thank Professor Frank M.
Carpenter and my husband, Dr. Thomas S. Parsons, both of
Harvard University, for their helpful advice and for their critical
examination of the manuscript. This study was carried out dur-
ing the tenure of the Ellen C. Sabin Fellowship, awarded by the
American Association of University Women.

Figure 1. Dorsal view of Gelastocoris oculatus, legs removed. About 11X.

MATERIALS AND METHODS

Since Gelastocoris is a fairly large insect (approximately 6 to
5 mm. long), the skeleton and musculature could be studied by
means of dissection under a stereoscopic microscope. Most of
the dissections were made upon insects preserved in aleoholic

302 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Bouin’s solution, 10 per cent formalin, or Kahle’s solution, and
stored in 70 per cent alcohol. Since fresh material was some-
times needed, live gelastocorids were kept in the laboratory. The
conditions under which they were kept have been described in a
previous paper (Parsons, 1959).

Figure 2. Ventral view of G. oculatus female, spines and hairs of legs

omitted.

SKELETON

PROTHORAX
Cervical region. The cervical region of Gelastocoris is relatively
simple. A cervical membrane (CM), devoid of cervical sclerites,
connects the postoeciput of the head with the inturned anterior

PARSONS: THORAX OF GELASTOCORIS 303

margin of the prothorax (Fig. 5). The latter forms a tight collar
around the postocciput. Two short tendons extend from the mid-
dorsal region of the cervical membrane into the thorax. From
the posterior margin of the postocciput, two pairs of apodemes
project into the thorax, providing points of attachment for
muscles. The longer of these, the occipital condyles (O), extends
dorsally from the ventrolateral regions of the postocciput. Dorsal
to them are the much shorter lateral apodemes (L), which project
posteroventrally.

Figure 3. Lateral view of G. oculatus female.

Tergum. The prothoracie tergum is a large plate covering
much of the anterior half of the body dorsally (Fig. 1). Much
of its area is due to marginal evaginations, and the actual pro-
thoracic cavity is comparatively small. These evaginations pro-
duce lobes which overlap the postoeciput anteriorly and the
mesonotum posteriorly. They are especially pronounced along
the posterior and posterolateral tergal margins, where they
form a broad posterior protergal lobe (LP) (the ‘‘tergal flap’’
of Malouf, 1933, and the ‘‘Hinterlappen’’ of Larsén, 1945a and
b) shielding the anterior part of the mesonotum. The antero-
lateral lobes of the protergum are concave anteriorly, conforming
to the shape of the head at the bases of the compound eyes. The
two walls of all these marginal evaginations are fused together
and appear as a single layer. A pronounced transverse ridge

304 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

(TR) (‘‘Querleiste’’ of Larsén, 1945a and b) marks the anterior
border of the posterior lobe and the posterior limit of the pro-
thoracic cavity (Figs. 5 and 6). This ridge is a continuation of
the ventral layer of the posterior lobe; attached to it is the inter-
segmental membrane (J) joining the prothorax to the mesothorax.

Pleuron. The boundary between the pleuron and the marginal
tergal evaginations is not clear externally. A short pleural suture
(Fig. 4, PS) separates the small episternum (HS) from the larger
epimeron (EP). The ventral regions of each of these are evagin-
ated to form supracoxal lobes (EPS and ESS) (‘‘epimeral and

p-CC
ES CL PEC
Geese
West pes |
We ; 1 i i
[PGrssssssseeee et
EPs
FP Yeti secre

EL POC XR X
Figure 4. Ventral view of the prothorax, tilted slightly to the left. The
muscles and the right leg have been removed; the left leg is cut off near
the distal end of the coxa.

episternal flaps’’ of Malouf, 1933, and Akbar, 1957; ‘‘Supracox-
allappen’’ of Larsén, 1945a and b) overlapping the bases of the
coxae. The episternal and epimeral supracoxal lobes are sep-
arated by a pronounced coral cleft (CL), which extends dorsally
to the coxal process (to be described below). Externally the coxal
cleft appears to be a ventral extension of the pleural suture; in-
ternally the cleft and the suture are separated by the coxal pro-
cess. The posterior margin of the epimeron forms a large
posterior epimeral lobe (EL) (the ‘“‘lateral epimeral fiap’’ of
Malouf, 1933), which covers much of the mesothoracic epimeron
as well as the bases of the forewings (Figs. 4,5, and 6). Laterally,
this lobe is continuous with the posterior lobe of the protergum,

PARSONS: THORAX OF GELASTOCORIS 305

the point of junction being braced by a U-shaped sclerotized
strut (Fig. 5, 87). The anterior margin of the inner wall of the
epimeral lobe is turned dorsally to form the ventral part of the
transverse ridge, which medially becomes quite low and con-
tinues into the posterior sternal process (to be described below)
(Fig. 6).

Figure 5. Posteromedial view of the left half of the prothorax and of the
postoccipital region of the head. The tendons of the pericoxal membrane
and the muscles have been removed.

The large coral cavity (CC) is bordered laterally by the epim-
eron and episternum and medially by the sternum (Fig. 4). An-
terior to the coxal cavity the episternum meets the sternum,
forming a narrow precoxal bridge (PEC) ; posteriorly the fused
epimeron and a sternum form a broad postcoxal bridge (POC).
There is no clear boundary between the sternal and pleural ele-
ments. A view of the inner surface of the pleuron (Figs. 5 and
6) shows that the short pleural suture produces a distinct pleural
ridge (PR) internally. At the end of the ridge is a strong coxal
process (CP) (‘‘coxal articulation”’ of Griffith, 1945; ‘‘ Huftgel-
enkkopf’’ of Larsén, 1945 a and b; ‘‘pleural articular process’’ of
Akbar, 1957) which projects a short distance into the coxal cavity

306 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

and articulates with the base of the coxa. Just dorsal to the
coxal process is a large, lamellar pleural apophysis (PA)
(‘‘pleural arm’’ of many authors; ‘‘Pleuralhaken’’ of Larsén,
1945a and b) which extends dorsally nearly to the lateral part
of the notum. It does not fuse with the latter, as it does in many
water bugs. A thick membrane (Fig. 5, PL) projects medially
from the pleural apophysis to the furca (to be described below).
A similar connection between the prothoracic pleural apophysis

1
1

LP EP TIS PR —

|
1
1
1
1
1
1
1
t

i PG

Figure 6. Dorsomedial! view of the left half of the prothorax. The muscles
and the pleurosternal bridge have been removed, and much of the medial
part of the protergum has been cut away.

and the furca has been deseribed in Corixa and Salda by Larsén
(1945a and b), and in Ramphocorixa by Griffith (1945); the
latter author termed it the pleurosternal bridge, and that term
will be used here.

Sternum. The anterior part of the sternum is produced into
a large triangular process which hes between the coxae (Fig. 4,
X’). A similar process in the metathorax of Heteroptera has been
termed the x7phus by many authors, and that term wi!l be used
for this prothoracic structure. Along the posteroventral edge of

PARSONS: THORAX OF GELASTOCORIS 307

the xiphus runs a sharp, V-shaped ridge (Fig. 4, Y&) which
produces a groove internally (Figs. 5 and 6, XG). Laterally this
ridge, which will here be ealled the aiphal ridge, continues into
the precoxal bridge and runs along the posterior edge of the
supracoxal lobe of the episternum, ending at the coxal cleft. The
xiphal ridge corresponds to the ‘‘external ridge of the basi-
sternum’’ of Akbar (1957) ; since it does not pass through the
bases of the fureae, it probably does not represent a sternacostal
suture. Akbar (1957) found a sternacostal suture posterior to
this ridge in the prothorax of Nezara; in Gelastocoris there is
no visible prothoracic sternacostal suture.

Just medial to the coxal cavities, on either side of the midline,
are the short, thick sternal apophyses or furcae (‘‘anterior en-
dosternites’’ of Rawat, 1939; ‘‘Furcadste’’ of Larsén, 1945a and
b) (Figs. 5 and 6, #). Externally their positions are marked by
the furcal pits (Fig. 4, FP). As Larsén (1945b) has pointed out,
the term ‘‘fureal branch’’ is preferable to that of ‘‘furca’’ in the
Heteroptera, since the furca 1s a single medial structure in many
higher insects ; the paired apophyses of the Heteroptera probably
represent the branches of the single furca of other insects. For
convenience, however, the term ‘‘furea’’ will be used here. The
prothoracic furcae of Gelastocoris are each subdivided into a
lateral and a medial arm.

The posterior part of the sternum projects into the cavity of
the mesothorax as a broad plate. Lateral to this plate are two
prong-like posterior sternal processes (PSP), which are con-
tinuous with the ventral part of the transverse ridge (Figs. 5 and
6). In some specimens these processes are heavily sclerotized,
while in others they appear to be more or less membranous. They
may represent the ‘‘posterior horn-like processes’? or ‘‘spina’’
of Akbar (1957), and the ‘‘posterior prothoracic endosternites’’
of Rawat (1939). Lauck (1959), like Rawat, believed them to
be part of the furcae. Both Akbar (1957) and Larsén (1945b),
however, considered them to represent the spina of other insects;
the latter author observed these processes in several aquatie and
terrestrial Heteroptera. Just lateral to the posterior sternal
processes, in the intersegmental membrane between the prothorax
and mesothorax, les the large first thoracic spiracle (Fig. 5.

SL)

308 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

PTEROTHORAX

The mesothorax and metathorax, unlike the prothorax and
mesothorax, are bound closely together, the intersegmental mem-
brane being either very harrow or, more commonly, entirely
absent. In the following discussion the two segments will be con-
sidered together.

TERGUM

Mesothorax. On the anteromedial portion of the mesothoracic
tergum is the semicircular first phragma (Figs. 7 and 10, P71)
(‘‘prephragma’’ of Akbar, 1957), an invagination which extends

PM BUSEGRRE
4 Te vellin ul

'

PNII SCI SM—

ad

Figure 7. Dorsal view of the pterothorax. The wings have been removed

from the right side, and the wings of the left side have been extended and
cut off near their bases.

anteroventrally into the thoracic cavity. The two thin layers of
the invagination are closely apposed, the intersegmental mem-
brane from the prothorax attaching alone the posterior edge of
the more dorsal layer (Fig. 10). A narrow sclerotized precosta
(PC) is present where the membrane meets the margin of the
phragma, and medially two slender tendons extend anteriorly
from the ventral surface of the membrane (Fig. 7).

Posterior and lateral to the first phragma lies the mesonotum.
A broadly U-shaped parapsidal suture (Figs. 7 and 10, PAS)
(‘“parapside’’ of Malouf, 1933; ‘‘pre-seutal suture’’ of Rawat,
1939; *‘convergent suture’’ of Lauck, 1959) separates an antero-
medial prescutum (PU) (‘‘seutum’’ of Malouf, 1933, and Lauck,

309

THORAX OF GELASTOCORIS

PARSONS :

‘soSuq IY} TBM JO JMO puv papUo}xo UWadq ALY SBUIM OIL], *poAaowat Woo dALTT SSULM OY, “XvL0yzO.104d

*xB10Y}O10}d OY} JO OPIS JYStA oyZ FO MOTA [eVSIOG °G INST dy} FO Opts FYSta oy} FO MoTA [VSLOG “g oAINSTT

IOV als

1
!
1

| pe

IL dM-

SE
oe ae
at
aan

310 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

1959) from the rest of the mesonotum. The anterolateral margins
of the prescutum are connected anteriorly to the mesothoracie
episternum by the prealar bridges (Fig. 10, PEB), which are
roughly V-shaped. The parapsidal suture ends just posterior to
the prealar bridges; internally it produces a very low parapsidal
ridge (Fig. 12, PAR). Medial to the parapsidal suture is a nar-
row membranous area which is most conspicuous laterally.

The rest of the mesonotum is made up of the scutwm (SC) and
scutellum (SM). No scutoscutellar suture is present to separate
these regions; their boundaries can be determined only by the

Figure 10. View of the anterior surface of the mesothorax.

insertions of the tergal muscles, which, according to Snodgrass
(1927) arise only from the prescutellar area. According to this
criterion, the scutellum is very narrow laterally but becomes
wider medially. The scutum is bordered anteriorly by the parap-
sidal suture and by the prealar membrane (PM), which extends
from the lateral edge of the prealar bridge into the membrane
of the forewing (Figs. 7 and 10). Laterally, the scutum extends
to the bases of the forewings on either side. Where it articulates
with the base of the wing it is cleft by a posteriorly curved tergal
fissure (Fig. 8, TF) (‘‘Tergalspalt’’ of Larsén, 1945a and b).
This fissure separates two small processes which together form the
anterior notal wing process (Fig. 7, AWII) (‘‘anterior wing

PARSONS: THORAX OF GELASTOCORIS Bilal

process’’ of Rawat, 1939). The more anterior process has been
termed the ‘‘vectis dorsualis anterior’’ (Malouf, 1933), the
‘‘vorderen Tergalhebel’’? (larsén, 1945a) and the ‘‘anterior
notal wing process’’? (Akbar, 1957), while the more posterior
process has been called the ‘‘vectis dorsualis posterior’ (Malouf,
1933) and the ‘‘hinteren Tergalhebel’’ (Larsén, 1945a).

The seutellum (‘‘scutoscutellum’’ of Lauck, 1959) is the
posterior region of the mesonotum, which medially covers much
of the metatergum. Since the latter is concealed by the fore-
wings and since the anterior part of the mesoscutum is overlapped
by the posterior lobe of the pronotum, only the mesoscutellum
and the posterior part of the mesoseutum are normally visible
dorsally (Fig. 1). The posterior apex of the scutellum has been
termed the ‘‘scutellar flap’’ (Malouf, 1933) or the ‘‘Scutellar-
lappen’’ (Larsén, 1945 a and b). Along the posterior part of the
seutellum runs a deep wing groove (Fig. 7, W@) which receives
the posterior margin of the forewing when the latter is at rest.
Medially the wing groove is narrow and sclerotized, resembling
a deep suture, but laterally it becomes wider and membranous,
the membrane continuing into the forewing (this membranous
part of the groove corresponds to the region labelled “‘X’’ in Fig.
3B of Rawat, 1939). The posterior wall of the wing groove,
along with the narrow region of the scutellum posterior to it,
forms the scutellar process (SP) (‘‘processus scutellaris tertius
alae’’ of Malouf, 1933: ‘‘narrow sclerotized strip’’ of Rawat,
1939; ‘‘Scutellumfortsatz’’ of Larsén, 1945a and b; ‘‘lateral
seutellar plate’? of Akbar, 1957; probably the ‘‘frenum”’ of
Tower, 1913). The surface of the seutellar process is somewhat
lower than that of the seutellum proper. Laterally it narrows
and bends anteriorly, becoming fused with the subalare (Fig. 8,
SBIT), an irregularly shaped sclerite lying in the posterior part
of the base of the wing. Just behind the subalare is the axillary
cord (ACIT), which comes off from the seutellar process. Both
Rawat (1939) and Akbar (1957) considered the lateral limit of
the scutellar process to represent a posterior notal wing process.
Larsén (1945b), however, claimed that the Heteroptera lack a
posterior notal wing process, the wing instead articulating with
the subalare; this is certainly the case in the mesothorax of
Gelastocoris.

312 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

The postnotum (‘‘pseudonotum’’ or ‘‘postscutellum’’ of many
authors) is externally visible only laterally (Fig. 11, PNIIZ).
This lateral portion (the ‘‘ Lateropostnota’’ of Larsén, 1945a and
b) lies posteroventral to the subalare and axillary cord, and is
continuous with the mesothoracie epimeron, forming the postalar
bridge (Fig. 8, POBIT) (‘‘Postalare’’ of Taylor, 1918). The pos-
terior margin of the postalar bridge is joined to the anterior edge
of the metathoracie episternum, the boundary between the two
being indistinct in many specimens. Medially the postnotum is
concealed by the scutellar process, and the boundary between
these two regions is also indistinct. A view of the inner surface

SP- WPIHT-;PNIT NI ;-WPI

WY

!
|

1
|
{
{
i)
|
It:
|
i
i
|
|

xf

fl

Seay i Os pee a)

Se er ees a

Ease eate cali ie eee

ft aeraceni ite 2 asp gaeesy
ES I-- am Gi

ae Piece raeyle
PP) ePaai! Igitep
PSH) -:ES Ml

Figure 11. Lateral view of the right side of the pterothorax. The wings
and legs have been removed.

of the tergum (Fig. 12) reveals that the medial part of the post-
notum is extremely narrow. Posteriorly it becomes continuous
with the anterior wall of the second phragma (P2) (‘‘mesopost-
phragma’’ of Akbar, 1957), an invagination of the tergum be-
tween the mesothorax and metathorax. This is the largest of the
three thoracic phragmata. Laterally it is a fairly low ridge which
meets the pleuron at the boundary between the mesothoracic
epimeron and the metathoracic episternum; medially it becomes
much higher, and possesses, on either side of the midline, a ventral

PARSONS: THORAX OF GELASTOCORIS a3

process (V) (‘‘ventrale Fortsatz of Larsén, 1945a and b). This
process extends ventrally nearly to the mesosternal furca (FIT)
(to be described below) ; its two walls are not closely apposed and
may be easily separated.

Metathoraxr. The metathoracic tergum is considerably shorter
than the tergum of the preceding segment. This is true of most
Heteroptera, since the main flight muscles are located in the meso-
thorax (Larsén, 1945b). The metathoracic tergum consists mainly
of the notum, the postnotum being visible, as in the mesothorax,
only dorsolaterally. Medially the notum is quite narrow and
partially concealed by the mesoseutellum (Fig. 7), but laterally it
becomes broader. Its anterior wall (the ‘‘preseutum’’ of Tower,
1913) is continuous medially with the posterior wall of the second
phragma. Laterally the notum is separated from the phragma
by a narrow membrane (concealed, in Figures 7-9, by the scu-
tellar process). Larsén (1942) has pointed out that this separa-
tion enables the lateral edges of the metanotum to be bent
downwards, indirectly causing the extension and raising of the
hindwing. The lateral border of the metanotum bears both an
anterior and a posterior notal wing process (Fig. 8, AWIIT and
PWIITI), although Larsén (1945b) reported the latter to be
absent in the Heteroptera. In the wing membrane ventral to the
posterior notal wing process and the third axillary sclerite (to be
described later) lies an extremely minute subalare which is easily
overlooked (Figs. 8 and 11, SBIIZ).

The medial part of the metanotum bears a broadly U-shaped
groove (Fig. 7) which produces a ridge on the inner surface.
This groove appears to turn laterally just before the posterior
notal border and to run to a point just behind the posterior
notal wing process (Fig. 8). It may represent a scutoscutellar
suture (SL) (‘V-Leiste’’ of Larsén, 1945a and b). If so, the
narrow notal region behind it may be termed the scutellum
(SMIII) and the more extensive area anterior to it the scutum
(SCIII). The posterolateral part of the scutellum resembles the
scutellar process of the mesothorax. The azillary cord of the
hindwing (Fig. 8, ACIIZ) comes off from the lateral edge of this
region, but the subalare is not joined to the scutellum as it is in
the mesothorax.

314 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

The medial part of the narrow postnotum (PNIIT) is concealed
by the notum; laterally it becomes somewhat broader and is vis-
ible dorsally (Fig. 7) and laterally (Fig. 11). It joins the meta-
thoracic epimeron to form the postalar bridge (Fig. 8, POBIIT)
just posterior to the axillary cord. The posterior margin of the
postnotum continues into the anterior wall of the third phragma;
the latter, which marks the dorsal boundary between the thorax
and abdomen, is a low ridge without ventral processes (Fig. 12,
P3). It is much less pronounced than the second phragma.

PNIL4
V P2s, ERT (a PATt ERT NIL

ES
CCIL XI
Figure 12. Medial view of the left half of the pterothorax, with the
muscles removed. Only the bases of the coxae are shown.

PLEURON

Mesothorax. The intersegmental membrane (I) from the pro-
thorax meets the pleuron of the mesothorax along the medial
edge of the prealar bridge and the anterior margin of the meso-
thoracic episternum (Fig. 10). Within the membrane, just an-
terior to the most dorsal part of the prealar bridge, lies a very
small selerite (SE) which provides an insertion for one of the
depressor muscles of the thorax. According to Larsén (1945b),
this sclerite is a free lateral part of the first phragma.

PARSONS: THORAX OF GELASTOCORIS SLs

A ventral view of the mesothorax (Fig. 13) shows an extensive
episternum (HSIZ) and a somewhat smaller epimeron (HPIT)
separated by a long and distinet pleural suture (PSIT). Both
these sclerites have supracoxal lobes overlapping the coxal bases
as in the prothorax; in addition, a posterior lobe of the epimeron
(ELIT), which is continuous with the epimeral supracoxal lobe,
overlaps part of the metathoracic episternum, concealing the
lateral boundary between the mesothoracie and metathoracic
pleura (Fig. 14). At the posterolateral corner of this posterior
epimeral lobe is a conspicuous knob (Figs. 7 and 14, K)
(‘‘Hocker’’ of Larsén, 1945a and b) which fits into a depression
on the anterior margin of the forewing and thus holds the resting

XI XRT

1

EST

Figure 13. Ventral view of the pterothorax. The legs, except for the right
coxae, have been removed and the forewings have been cut off near their
bases.

wing firmly in place. Dorsal and considerably anterior to this
knob, the epimeron meets the mesopostnotum forming the postalar
bridge (Fig. 11).

Ventral to the base of the wing, both pleural sclerites are
somewhat evaginated to form a sort of shelf; in the epimeron,
this shelf continues into the posterior epimeral lobe (Fig. 11).
Dorsal to this evaginated portion, at the point where the epimeron
joins the episternum, is a short pleural wing process (‘‘wing
fulerum’’ of Taylor, 1918) (Figs. 8 and 11, WPII). This process
hes ventral and lateral to the anterior notal wing process. The
dorsal margin of the epimeron is somewhat thickened and curved

316 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

medially, which helps to strengthen and support the pleural
wing process.

The coxal cavity is bordered medially by the sternum and
laterally by the pleural sclerites (Fig. 15). Anterior to it hes a
very broad precoral bridge where the episternum is fused in-
distinguishably with the sternum. It is difficult to determine
whether or not a postcoxal bridge is present, since the boundary
between the mesosternum and metasternum is uncertain. Ac-
cording to Larsén (1945b), the boundary between these two
regions is marked by the mesosternal fureae. If this criterion
is used, the mesosternal-metasternal boundary must lie just at
or slightly behind the posterior margin of the mesocoxal cavity.
If the former is true, the postcoxal bridge is absent ; if the latter,
the bridge is very narrow, probably being represented by the
slight ridge bordering the posterior edge of the coxal cavity
(Fig. 15, RG). On this ridge, one of the metathoracic muscles,
M. episterno-coxalis (Muscle 66), originates. Larsén (1945a)
considered the point of origin of this muscle to be the anterior
part of the metathoracie episternum, immediately behind the
intersegmental boundary. If such is the case, the mesothoracic
postcoxal bridge must be absent.

At the lateral margin of the coxal cavity there is a conspicuous
coxal process (Fig. 15, CPIT) from which the distinct pleural
ridge (PRII) runs anterodorsally on the inner surface of the
pleuron (Figs. 12 and 15). There is no basicoxal plate (‘‘freie
Basicoxalplatte’’ of Larsén, 1945d). The entire length of the
pleural ridge is visible. About midway along the pleural ridge
is a rectangular pleural apophysis (PAII) (possibly the ‘‘basal-
are apodeme”’ of Akbar, 1957), which bends somewhat postero-
laterally and extends towards the second phragma.

Metathorax. The metathoraciec pleuron is composed mostly of
the episternum (Fig. 13, ESIIT) with its large supracoxal and
posterior lobes. These lobes, which are continuous with each
other, overlap part of the abdomen. The greatly reduced epim-
cron possesses no supracoxal lobe, and is visible only laterally
(Fig. 11, EPIIT). Anteriorly the episternum is covered by the
posterior lobe of the mesothoracie epimeron. This lobe covers both
the intersegmental boundary between the mesopleura and meta-
pleura and the second thoracic spiracle (Figs. 12, 14, and 15, S2)

PARSONS: THORAX OF GELASTOCORIS a

which is located at the anterior edge of the metathoracic epi-
sternum. The posterior margin of the mesothoracic epimeral
lobe lies very closely upon the metathoracie episternum, and it
seems unhkely that air could reach the spiracle from a posterior
direction. Laterally, however, between the posterior lobe of the
mesothoracie epimeron and the metathoracie episternum, there
is a space which bears some resemblance to the ‘‘air chamber’’
of corixids, as described by Griffith (1945). The air reaching the
second thoracic spiracle probably comes in through this space,
which is located directly beneath the base of the hindwing. This
condition is typical of the aquatic Heteroptera rather than the
terrestrial bugs, as Larsén (1945b) has pointed out.

The coxal cavity is very large and oval in shape, its long axis
running mediolaterally in relation to the body (Fig. 15). An-
teriorly it is bordered by a very broad precoxal bridge which
joins the episternum with the sternum, and anteromedially by
the metasternum. Posteriorly and posteromedially it is bordered
by the abdominal sternites. There is no postcoxal bridge or basi-
coxal plate. A very large coxal process (CPIII) articulates
laterally with the base of the cora. This process has an anterior
extension which articulates with the trochantin (TNIII) (to be
described later). Dorsal to the coxal process a high pleural ridge
(PRITIT) extends anterolaterally for a short distance and then
curves to run directly anteriorly, parallel to the longitudinal axis
of the body (Fig. 15). Externally only the dorsal part of the
pleural suture (PSIIT), which ends in the pleural wing process
(WPIIT), is visible (Fig. 11); the ventral part of the suture is
concealed by the lateral part of the metathoracic posterior epis-
ternal lobe. There is no pleural apophysis in the metathorax.

Across the precoxal bridge runs a deep stink groove (Fig. 14,
SG) which produces a high stink ridge internally (Fig. 15, SR).
The groove extends from the external orifice of the thoracic stink
eland, at the base of the furea, to the point where the posterior
mesothoracie epimeral lobe begins to overlap the metathoracic
episternum. The part of the precoxal bridge, anterior to the
eroove corresponds to the ‘‘anterior laterale’’ plus the ‘‘basi-
sternum’? of Brindley (1934), the ‘‘basisternum’’ of Griffith
(1945), and the ‘‘laterale’’ of Akbar (1957) ; the part posterior
to the groove represents the ‘‘antecoxal laterale’’ of Brindley

318 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

(1934), the ‘‘episternum’’ of Griffith (1945), and the ‘‘antecoxal’’
of Akbar (1957). The anteromedial wall of the groove forms a
narrow flap (Fig. 14, AW) (the ‘‘stink fold’’ of Akbar, 1957,
and the ‘‘anterior laterale’’ of Brindley, 1934), which extends
posteriorly, while the posterolateral wall forms an even larger
flap (PF) in the lateral part of the groove. This second flap
extends anteriorly and overlaps the smaller one. Together the
two flaps form a partial covering over the ventral side of the
groove, helping to hold the secretion in the channel. The form
and position of the stink groove is more like that of the aquatic
Heteroptera, as described by Larsén (1945b), than like that of

Figure 14. Ventral view of the posterior portion of the pterothorax, left
side, showing the stink groove. The legs have been removed. The broken
lines indicate the positions of the anterior metathoracic episternum and of
the second thoracic spiracle, which are overlapped by the posterior lobe of
the mesothoracic epimeron.

the terrestrial bugs; Larsén did not consider the stink grooves
of these two groups to be homologous. There is an especially
strong resemblance between the grooves of Gelastocoris and the
Corixidae, the chief difference being that the corixids lack the
larger of the two flaps.

Just beyond the lateral end of the stink groove, the posterior
edge of the mesothoracic posterior epimeral lobe is modified to
form a very smooth, shiny band (Fig. 14, HV), whose surface
contrasts sharply with the rough texture of the rest of the sclerite.
This band is so placed that the secretion emerging from the
stink groove flows out upon it. It probably serves as an evaporating

PARSONS: THORAX OF GELASTOCORIS 319

surface for the secretion; to the author’s knowledge, such an
evaporating surface has not been reported in any other heterop-
teran.

The epimeron is largest in the anteroposterior direction.
Posterodorsally it joins the postnotum to form the postalar bridge,
and posterolaterally it is somewhat produced to form a posteriorly
projecting lobe (the ‘‘epimeral flap’’ of Akbar, 1957) (Fig. 11).
Just ventral to this lobe the epimeron is more or less fused with
an anterolateral process from the abdomen (Figs. 11 and 15, AP).
This process, the ‘‘Pleura des ersten Abdominalsegments’’ of
Larsén (1945a) and the ‘‘epimeral fold’’ of Akbar (1957), bears
the first abdominal spiracle at its base, and appears to be abdom-
inal in nature; Snodgrass (1909) considered it to be part of the
abdomen in Benacus.

STERNUM

Mesothorax. Like the prosternum, the mesosternum possesses
an evaginated triangular xiphus between the coxae (Fig. 13,
XII). Along the posteroventral edge of the xiphus runs a xiphal
ridge (XRII) similar to that of the prothorax; internally it
appears as a groove (Fig. 15, YG). Unlike the corresponding
ridge of the prosternum, it does not reach the coxal cleft, but
runs instead to the anteromedial border of the coxal cavity.
Since it does not pass through the fureal bases, it probably does
not represent a sternacostal suture.

Just medial to the coxal cavities he the conspicuous, oval furcal
pits (Wigs. 13 and 14, FPIT). They mark the broad bases of
the large furcae (Figs. 12 and 15, FIZ). The latter possess an-
terior and posterior arms (together probably corresponding to
the ‘‘ posterior arm of the endosternite’’ of Rawat, 1939) which
extend dorsally nearly to the ventral processes of the second
phragma. From the more anterior furcal arm a long, slender
furcal apodeme (FA) (‘‘anterior arm of the endosternite’’ of
Rawat, 1939; ‘‘Furcaapodem”’ of Larsén, 1945a and b; ‘‘anterior
process of the furca’’ of Griffith, 1945; ‘‘lateral arm of the meso-
thoracic furea’’ of Laueck, 1959) extends dorsolaterally and
nearly touches the tip of the pleural apophysis.

There is no clear boundary between the mesosternum and the
metasternum. As has been previously mentioned, the mesosternal
fureae probably indicate its approximate position.

320 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Metathorax. The xiphus (‘‘sternellum’’ of Griffith, 1945) of
the metasternum is considerably larger than those of the preced-
ing two segments (Fig. 13, Y///). Along its posteroventral
margin runs a ridge (SS) which resembles the xiphal ridges of
the prothorax and mesothorax. Since it ends at the bases of the
metathoracic furcae, however, it may represent a sternacostal
suture. If so, the part of the sternum anterior to it may be
termed the basisternum, and the short portion posterior and
lateral to it (the posterodorsal wall of the xiphus) may be

PRID-; ;-ESIL
EPI; |! r-WPIl

Rees

wn
By

|
DD
>)

I
4
a8
Ss
>)

Figure 15. Dorsal view of the inner surface of the pterothorax, left side,
showing the sterna and pleura. The pterotergum has been cut off at the
postalar bridges and at the mesothoracic prealar bridge, and the tendons in
the pericoxal membranes have been cut off near their bases. The muscles
have been removed.

called the sternellum or furcasternum. Extending between the
metathoracic furcae is a membrane which separates the meta-
thoracic sternum from the sternum of the first abdominal segment.

The furcal pits are concealed externally by the lateral edges of
the xiphus. Internally the fureae appear as two unbranched
processes (Figs. 12 and 15, FIJI). They extend posterolaterally
and are longer and much more slender than the mesothoracic
fureae.

PARSONS: THORAX OF GELASTOCORIS 321

LEGS

The raptorial forelegs of Gelastocoris are oriented differently,
with respect to the body, than are the walking and jumping
pterothoracice legs. For convenience, however, the descriptive
terms applied to the surfaces of the last two pairs of legs will
be the same as those used for the corresponding surfaces of the
forelegs. The terms ‘“‘anterior’’ and ‘‘posterior’’ are here ap-
plied to the anteromedial and posterolateral sides of the foreleg
respectively ; “‘ventral’’ refers to the inner surfaces (those which

BE Cee Bane 5
! (J msunyu te
(PR-PB ees ie
o//) Hee
AB---A/ | of -----HA TI WA

Figure 16. Medial view of the left prothoracic leg and trochantin.
Figure 17. Medial view of the inner surface of the left prothoracic leg,
showing the tendons.

meet, on the femur and tibia, when the latter are apposed), and
‘‘dorsal’’ refers to the outer surfaces. The numbers used to
designate the various tendons are the same as those of the muscles
which insert on them.

Prothoracic legs (Figs. 6, 16, and 17). The prothoracic coxal
cavity is fairly round. In the anterior part of the pericoxal
membrane (PE) (‘‘coxal corium’’ of Griffith, 1945, and Akbar,
1957) les a small trochantin (TN), which does not appear to
articulate with either the pleuron or the base of the cora. The
coxa is therefore articulated only at the coxal process, and this
single joint allows it to move freely in all directions. Such free-
dom of movement is advantageous in a raptorial lee; Rawat

322 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

(1939) has reported a similar condition in the foreleg of Naw-
coris. Four tendons in the pericoxal membrane provide insertions
for the muscles which move the coxa; Tendon 13 is located just
medial to the medial end of the trochantin, to which it is partially
attached, Tendon 14 just lateral to the lateral end of the tro-
ehantin, Tendon 15 shghtly posterior to the coxal process, and
Tendon 16 in the posteromedial region of the pericoxal membrane.

The coxa (CX) is nearly cylindrical in form, and projects
out farther from the body than do the coxae of the second and
third pairs of legs. A basicostal suture (BS) encircles its proximal
end; posteromedially the suture is very faint and close to the
edge, but laterally it becomes clearer, producing a_ basicostal
ridge (BR) (‘‘basicosta’’ of Snodgrass, 19385) internally. It
separates off an anterior basicoxrite (AB) (‘‘vorderes Basicoxale’’
of Larsén, 1945¢) anterior to the coxal process, and an equally
large posterior basicoxrite (PB) (‘‘hinteres Basicoxale’’ of Larsén,
1945¢c) posterior to that process. Snodgrass (1935) termed the
posterior basicoxite the ‘‘meron’’; Larsén (1945e and d), how-
ever, has shown that the posterior basicoxite and the meron are
two separate elements, and that the latter is absent in the Het-
eroptera. Between the two basicoxites the basal coxal rim is in-
vaginated to form a socket (the ‘‘articular process’’ of Griffith,
1945) into which the coxal process fits.

A dicondyhe joint with anterior and posterior articulations
joins the coxa with a short, curved trochanter (TC). On the
ventral surface of the latter are two irregular rows of short
spines. Two tendons, whose bases are attached to the proximal
rim of the trochanter by tough membranes, extend into the coxa.
The longer of these, Tendon 20, is located in the part of the rim
which is farthest from the femur, and reaches into the thoracic
cavity. A shorter, three-branched Tendon 24 comes from the part
of the rim nearest the femur.

A dicondylie joint with dorsal and ventral articulations joins
the trochanter with the femur (FH). These two segments are
joined so closely together that the condyles are difficult to see.
The femur is greatly thickened to accommodate the powerful
tibial muscles which originate on its inner walls. These muscles
enable the tibia (TI) to open and close upon the femur. The
femur is broadest proximally, the dorsal part of the segment

PARSONS: THORAX OF GELASTOCORIS a3

forming a hump above the articulation with the trochanter. The
ventral surface of the femur is flattened, and an irregular row
of stout spines extends along each side of the flattened area. On
the anterior surface of the femur, just dorsal to the row of spines,
is a comb of long, fine hairs (HA). This meets a similar comb
of hairs on the tibia when the two segments are brought together.
It probably serves, as Weber (1930) has suggested, as a cleaning
organ for the head and antennae; the insects often perform
‘‘orooming’’ movements with their forelegs in the region of the
head.

The femur and tibia are joined by a dicondylie joint with
anterior and posterior articulations. The ventral surface of
the tibia is flattened, and bears two rows of spines similar to
those of the femur. When the tibia and femur are closed upon
each other, prey may be caught between the apposed flattened
areas and held in place by the spines. Two long tendons from
the ventral (Tendon 26) and dorsal (Tendon 27) regions of the
proximal edge of the tibia extend into the femur. The base of
Tendon 26 is expanded into a broad sclerotized plate (the ‘‘gen-
uflexor plate’’ of Akbar, 1957) which is movably bound to the
rim of the tibia.

The tibia and tarsus (7A) are joined primarily by a membrane,
but have a weak anterior and posterior dicondylie joint. From
the ventral region of the proximal edge of the tarsus, Tendon
28 extends into the tibia. There is only one tarsal segment.
Distally the tarsus is Jomed by a membrane to the pretarsus
(PT), which consists of two fairly long, stout claws (CW) and
a ventral plate, the wngwitractor (U) (‘‘flexor plate’’ of Rawat,
1939). The distal end of the unguitractor is narrowed, and
bears two very fine, short spines. Akbar (1957) reported sim-
ilar spines in Leptocorisa and suggested that they may be anal-
agous with the ‘‘empodium”’ of Diptera. From the base of the
unguitractor, a very long Tendon 29 (‘‘depressor apodeme’’ of
Akbar, 1957) extends through the tarsus and tibia and into the
femur.

Mesothoracic legs (Figs. 15 and 18). Unlike the first pair of
legs, the second and third pairs have coxae which articulate with
the pleuron at two points. Their movement is thus more re-
stricted. A small invagination of the lateral rim of the mesocoxa

324 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

forms a socket into which the coxal process fits; in addition, a
rather broad trochantin (TNII) articulates medially with the
anterior margin of the coxa and laterally with an anterior ex-
tension of the coxal process. Three tendons in the pericozal
membrane (PEIT) provide insertions for muscles; Tendon 40
lies just beside the medial end of the trochantin and is partially
attached to the latter, Tendon 41 is located in the posterior region
of the pericoxal membrane, and J'endon 42 lies just anterior and
medial to the coxal process.

Figure 18. Medial view of the left mesothoracic leg and trochantin.

Figure 19. Medial view of the left metathoracic leg and trochantin.

The mesothoracic coxae le closer to the body than do those
of the prothorax. They project posteromedially, nearly touching
each other at the midline (Fig. 2). Distally they are nearly
spherical in shape; proximally the side which contacts the coxal
process is considerably longer than the opposite side. The basv-
costal suture is not as marked as that of the prothorax. It seems
to disappear medially, while laterally it separates off the very
narrow anterior and posterior basicoxites.

The joints between the various segments of the leg are essen-
tially the same as those of the prothoracic leg. Also the tendons
within the segments occupy the same positions as those of the

PARSONS: THORAX OF GELASTOCORIS 325

first pair of legs, their terminology and the corresponding pro-
thoracic tendons being as follows: Tendon 46 (Tendon 20), Ten-
don 50 (Tendon 24), Tendon 52 (Tendon 26), Tendon 53 (Tendon
27), Tendon 54 (Tendon 28), Tendon 55 (Tendon 29).

The mesothoracie femur is longer and not nearly as broad as
that of the prothorax, and it lacks the flattened ventral area.
The tibia is also longer, and the tarsus consists of two segments,
the first one being much reduced. The wnguitractor of the pre-
tarsus resembles that of the forelegs, and possesses similar term-
inal spines; the pretarsal claws are smaller than those of the
first pair of legs. On the ventral surfaces of the trochanter and
femur are rows of short spines. The tibia possesses longer spines
on all its surfaces; these are especially numerous distally. A
few spines are also present on the distal segment of the tarsus.

Metathoracic legs (Figs. 15 and 19). The metathoracic cozae,
like those of the preceding segment, project posteromedially ;
distally they are very spherical, while proximally the side which
contacts the coxal process is much elongated. They are articulated
with the pleuron both directly, at the coxal process, and indi-
rectly, by means of the very long trochantin (TNIII). Unlike
the coxae of the two anterior pairs of legs, the metathoracie coxa
forms a narrow lateral process at its rim, this process fitting into
a socket on the coxal process; in the prothorax and mesothorax,
the socket is on the coxa. The pericoral membrane (PETIT)
possesses only two tendons: Tendon 63, which is partially at-
tached to the medial end of the trochantin, and Tendon 64, in
the posterior part of the membrane. At the proximal end of the
coxa, the basicostal suture separates off a distinct posterior
basicoxite and a very narrow anterior basicoxite.

The form of the various joints and tendons is the same as in
the first pair of legs. The terminology of the different tendons is
as follows: Tendon 70 (Tendon 20), Tendon 74 (Tendon 24),
Tendon 76 (Tendon 26), Tendon 77 (Tendon 27), Tendon 78
(Tendon 28), and Tendon 79 (Tendon 29).

The shapes of the femur, tibia and tarsus are quite different
from those of the corresponding segments of the forelegs. Since
the latter are modified for catching prey, while the former are
adapted for jumping, these differences are not surprising. The

326 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

metathoracie femur is much longer and narrower than the pro-
thoracic one. In the Hemiptera, according to Weber (1930), the
main muscles of the jumping lees are those of the trochanter,
not those of the tibia as in Orthoptera. In the gelastocorid fore-
leg, on the other hand, the tibial muscles are greatly developed
for capturing prey, and therefore the femora, on which these
muscles originate, are much enlarged. The metathoracic tibia
and tarsus are also much longer than those of the foreleg, and
the tarsus is three-segmented, the proximal segment being re-
duced. The great length of the femur, tibia, and tarsus provides
additional leverage for jumping.

On the femur there are a few very fine spines or hairs, but
very stout spines are present only on the tibia and tarsus, where
they are very numerous. In addition, the ventral surfaces of
the tibia and tarsus bear rows of long, fine hairs; there are two
such rows on the tibia and one on the tarsus. Weber (1930) has
suggested that the spines on the last two pairs of lees in gelasto-
eorids help to anchor the legs in the sand and to prevent them
from slipping backwards when the animal is jumping. The meta-
thoracic tibial and tarsal hairs are probably used to clean the
sides of the abdomen; the author has often observed live gelasto-
eorids rubbing their hindlegs over the edges of the abdomen.

WINGS

Forewing. Most of the forewing is coriaceous, and its surface
is covered with tubereles of various sizes, similar to those on the
body. Its tip, the membrane (MB), is smooth-textured and less
eoriaceous. The rest of the wing is divided into clavus (CV),
corium (CO), and embolium (EM), as shown in Figure 20. The
boundaries between these areas are marked by very narrow mem-
branous elefts in the surface of the wing. These probably repre-
sent wing veins, but the author will not attempt to homologize
them. Both Tanaka (1926) and Hoke (1926) studied the veins
of the forewings of a few Heteroptera, but none of the species
studied by them resembles Gelastocoris closely enough to permit
comparison. The boundary between the clavus and the corium is
very clear, and the wing possesses a flexible fold along this line.
The embolium is marked off by a long longitudinal and a short

PARSONS: THORAX OF GELASTOCORIS 327

transverse vein; these two do not meet medially. A fourth vein
runs longitudinally along the middle of the clavus; it is difficult
to see in many specimens.

The anterolateral edge of the embolium is greatly thickened
and folded ventrally. In this thickened, folded region there is
a laree, socket-like depression which receives the knob on the
posterolateral margin of the mesothoracic epimeron, holding the
resting wine securely in place. Similar wing-locking devices
have been reported in a great many Heteroptera by many authors,

Eee

CV oo N
:} ae )

fie oy pele ’ Shem
ue feat
oh lt) vel

Pa sence Ea CO
fee oties |
|
|
)

Mg)
=—=-f-—— MB

Figure 20. Dorsal view of the right forewing.

Figure 21. Dorsal view of the right hindwing.

and appear to be a common feature in this order of insects. As
has been previously mentioned, the pleural sclerites ventral to
the base of the forewing are somewhat evaginated, forming a
shelf-like projection. The thickened edge of the embolium lies
upon this shelf when the wing is at rest.

The axillary selerites by which the forewing articulates with
the mesothorax are shown in Figure 9. The first axillary sclerite
(1AX), which articulates with the anterior notal wing process,
is small and oval in shape; laterally it contacts a large, irregu-
larly shaped second axillary sclerite (2AX). The latter is fused

328 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

anteriorly with the humeral plate (H), the boundary between
the two being indistinct. A U-shaped third azillary sclerite
(3A4X) articulates anterolaterally with the posterior part of the
second sclerite ; posteromedially it is movably joined with a small
fourth axillary sclerite (44AX). A suture divides this fourth
sclerite into a proximal and a distal part, the proximal part
articulating medially with the subalare. Lateral to the antero-
lateral portion of the third axillary sclerite is a small, triangular
median plate (MP). This selerite articulates anteriorly with a
larger process (AJP?) which appears to be the lateral part of the
second axillary sclerite, but which may represent a second median
plate which has become fused with that selerite. A similar situ-
ation is found in the forewings of the belostomatids Benacus and
Lethocerus ; both Snodgrass (1909) and Lauek (1959), who studied
those forms, considered the process in question to be a median
plate.

Hindwing. Figure 21 shows the veins of the hindwing of
Gelastocoris. For convenience, the homologies suggested by Hoke
(1926) are used here. That author, who studied the wing vena-
tion of representatives of 25 families of Heteroptera, figured the
hindwing of Gelastocoris sp.

As shown in Figure 9, the first axillary sclerite (1AX) of the
hindwing articulates with the anterior notal wing process of the
metathorax and is very small. The third axillary sclerite (8AX)
is much larger and articulates with the posterior notal wing
process ; Taylor (1918) mistook it, in Belostoma, for the subalare.
The third axillary sclerite is broad and U-shaped, bearing a
small, knob-like projection laterally. This projection contacts
the base of the second anal vein (Fig. 21, 24). Between the first
and third sclerites lies a small second axillary sclerite (2A4X) ;
an even smaller, triangular median plate (MP) is located just
lateral to the second axillary sclerite.

MUSCULATURE

In general, the names of the following muscles and the numbers
by which they are designated are the same as those used by
Larsen (1945a). A few of the muscles described by Larsén ap-
pear to consist of two parts in Gelastocoris; in such cases they

PARSONS: THORAX OF GELASTOCORIS 329

have been given the name proposed by that author, with the ad-
dition of ““primus’’ or ““secundus’’andsan= 7A’ or B’ ‘has
been added to Larsén’s number. All the thoracic muscles are
paired.

An attempt has been made to list, for each muscle, similar
muscles which have been reported in other Heteroptera. Those
listed are included because both their origins and their insertions,
as described in the literature, are the same or very similar to
those of the corresponding muscle in Gelastocoris. Whether or
not they are actually homologous to the gelastocorid muscle which
they resemble cannot, in most cases, be definitely stated. The
names used by Larsén are given only when they differ from those
employed in the current work.

In Figures 22-31, the muscles are designated by the numbers
given below.

MUSCLES OF THE PROTHORAX

1. M. PRONOTI PRIMUS (Fig. 22)

Origin: Anteromedial region of the pronotum.

Insertion: On the two tendons in the mid-dorsal region of the
cervical membrane.

Action: Raises and retracts the head.

2. M. PRONOTI SECUNDUS (Fig. 22)

Origin: Anterior region of the pronotum, lateral to M. pronoti
primus,

Insertion: Tip of the occipital condyle.

Action: Rotates or depresses the head.

Similar muscles: Muscle 1 and Muscle 2 (?) (Malouf, 1933) ;
cephalic depressor (?) (Rawat, 1939); first and second pairs
of levators of head (Akbar, 1957).

Go

M.

=

PRONOTI TERTIUS (Fig. 22

A well developed longitudinal muscle.

Origin: Ventral part of the first phragma.

Insertion: On the dorsomedial margin of the postoceiput, and on
the two tendons in the cervical membrane.

Action: Rasies and retracts the head.

Similar muscles: Tergal longitudinal muscle (Malouf, 1933) ;
ventral fibers of dorsal muscle (Rawat, 1939) ; indirect levators
of head (Akbar, 1957).

330 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Figure 22. Medial view of the left half of the prothorax showing the more
medial muscles. The left halves of the postocciput and of the anterior part
of the mesothorax are shown in place.

Figure 23. Same view as above. The mesothorax and Muscles 1, 2, 3, 4,
6, 7,9, and 13 have been removed. Tendon 13 has been cut off near its base.

M.

M.

M.

PARSONS: THORAX OF GELASTOCORIS aol

PRONOTI QUARTUS (Fig. 22)

A well developed longitudinal muscle, just dorsal to M. pronoti
tertius.

Origin: On the dorsal part of the first phragma, and on the two
tendons in the intersegmental membrane.

Insertion: On the inturned dorsomedial margin of the pronotum.

Action: Raises the prothorax.

Similar muscles: Muscle rétracteur du prothorax (Poisson, 1924) ;
dorsal fibers of dorsal muscle (Rawat, 1939).

. PRONOTI QUINTUS (Fig. 22)

A slender muscle.

Origin: On the small sclerite in the intersegmental membrane
anterior to the prealar bridge of the mesothorax.

Insertion: Posterior region of the pronotum.

Action: Depresses the prothorax.

Similar Muscles: Indirect protractor of fore legs (Malouf, 1933) ;
depressors of pronotum (?) (Akbar, 1957).

PROSTERNI PRIMUS (Fig. 22)

A broad longitudinal muscle.

Origin: Anterior surface of the medial arm of the prothoracic
furea.

Insertion: On the occipital condyle and on the tip of the hypo-
pharyngeal wing.

Action: Depresses and retracts the head. May also cause some
rotation.

Similar muscles: Sternal longitudinal musele (Malouf, 1933) ;
depresso-extensors of head (Akbar, 1957).

PROSTERNI SECUNDUS (Fig. 22)

A broad longitudinal muscle, just lateral to WM. prosterni primus.

Origin: Anterior surface of the lateral arm of the prothoracic
furea.

Insertion: Occipital condyle.

Action: Same as M. prosterni primus.

. DORSOVENTRALIS (Fig. 22)

A slender muscle.

Origin: Anterior margin of the prealar bridge of the mesothorax,
medial to M. pronoti quintus.

Insertion: Posterior sternal process of the prothorax.

Action: Raises the posterior part of the prosternum, thus de-
pressing the prothorax as a whole.

Similar muscles: Tergo-sternal fureal muscle (Rawat, 1939):
fuy-prsco (Lauck, 1959).

332

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

10A. M. PROEPISTERNO-POSTOCCIPITALIS PRIMUS (Fig. 22)

10B,

15.

M.

A short muscle.

Origin: Anterolateral region of the proepisternum.

Insertion: Lateral apodeme of the postocciput.

Action: Raises the head (contraction of both muscles) or moves it
to one side (contraction of one muscle).

Similar muscles: Part of M. proepisterno-postoccipitalis (Larsén,
1945a) ; promoto-extensors of head (?) (Akbar, 1957).

PROEPISTERNO-POSTOCCIPITALIS SECUNDUS (Figs. 22
and 24)

A short muscle.

Origin: Lateral surface of the prothoracie pleural apophysis.

Insertion: Tip of the lateral apodeme of the postocciput.

Action: Depresses the head (contraction of both muscles) or moves
it to one side (contraction of one muscle).

Similar muscle: Part of M. proepisterno-postoccipitalis (Larsén,
1945a).

. NOTO-TROCHANTINALIS (Fig. 22)

A large, fan-shaped muscle.

Origin: Pronotum, just lateral to MW. pronoti primus.

Insertion: Tendon 13, at the medial end of the trochantin.

Action: Rotates the coxa and promotes the leg.

Similar muscles: Tergal promotor of coxa (?) (Malouf, 1933) ;
tergal promotor (Rawat, 1939); first promotor of coxa (?)
(Akbar, 1957).

NOTO-COXALIS PRIMUS (Fig. 23)

A large, fan-shaped muscle.

Origin: Pronotum, lateral to M. pronoti secundus and M. noto-
trochantinalis.

Insertion: Tendon 14, lateral to the trochantin.

Action: Rotates the coxa and abducts the leg.

Similar muscles: Internal rotator (Rawat, 1939); second pro-
motor of coxa (?) (Akbar, 1957).

. NOTO-COXALIS SECUNDUS (Fig. 25)

A large, fan-shaped muscle.

Origin: Posterolateral region of the pronotum.

Insertion: Tendon 15, just posterior to the coxal process.

Action: Rotates the coxa and remotes the leg.

Similar muscles: External rotator (Rawat, 1939); first remotor
of coxa (?) (Akbar, 1957).

PARSONS: THORAX OF GELASTOCORIS Bao

PG

1
2o

Figure 24. Same view as Fig. 22. Muscles 5, 10A, 14, and 16 have been
removed. Tendon 16 has been cut off near its base, and the pleurosternal
bridge has been cut away.

Figure 25. Posteromedial view of the left half of the prothorax and of
the postocciput (same view as Fig. 5), showing Muscles 15, 20B, and 21.
The posterior lobes of the pronotum and epimeron have been cut away, and
the pleurosternal bridge has been removed.

334

16.

20A. M.

20B. M.

21.

M.

M.

M.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

NOTO-COXALIS TERTIUS (Figs. 22 and 23)

A large, fan-shaped muscle.

Origin: Pronotum, posterior and lateral to M. noto-trochantinalis.

Insertion: Tendon 16, in the posteromedial region of the pericoxal
membrane.

Action: Rotates the coxa and adducts the leg.

Similar muscles: Tergal remotor of coxa (?) (Malouf, 1933) ;
tergal remotor (Rawat, 1939).

PLEURA-COXALIS (Fig. 24)

A short, broad muscle.

Origin: Medial surface of the prothoracie pleural apophysis.
Insertion: Tendon 14.

Action: Same as M. noto-coxalis primus.

NOTO-TROCHANTERALIS PRIMUS (Fig. 24)

A long, well-developed muscle.

Origin: Pronotum, between J/. noto-coralis secundus and M. noto-
coxalis tertius.

Insertion: Tendon 20, from the part of the proximal rim of the
trochanter which is farthest from the femur.

Action: Depresses the trochanter.

Similar muscles: Depressor of trochanter, tergal branch (Malouf,
1933); extra-coxal depressor, branch from tergum (Rawat,
1939); part of M. noto-trochanteralis (Larsén, 19452) ; tergal
depressor of trochanter (Akbar, 1957).

NOTO-TROCHANTERALIS SECUNDUS (Figs. 22, 24, and 25)

A long, slender muscle.

Origin: Anterolateral region of the pronotum, very near the
lateral margin of the episternum.

Insertion: Tendon 20.

Action: Depresses the trochanter.

Similar muscles: Extra-coxal depressor, branch from tergum
(2?) (Rawat, 1939); part of M. noto-trochanteralis (Larsén,
1945a).

PLEURA-TROCHANTERALIS (Fig. 25)

A short, broad muscle.

Origin: Lateral surface of the prothoracic apophysis.

Insertion: Tendon 20.

Action: Depresses the trochanter.

Similar muscles: Depressor of trochanter, pleural branch (Malouf,
1933); extra-coxal depressor, branch from pleural region
(Rawat, 1939); pleural depressor of trochanter (?) (Akhbar,
1957).

PARSONS: THORAX OF GELASTOCORIS 339

MUSCLES OF THE MESOTHORAX

30. M. MESONOTI PRIMUS (Fig. 26)

When developed, this is the largest muscle in the thorax. In the

majority of specimens, however, it, like the other indirect flight

muscles, is degenerate.

Origin: Anterior surfaces of the medial part of the second phragma
and of the ventral processes of the latter.

Insertion: First phragma and prescutum of the mesothorax.

Action: Indirect flight muscle. Depresses the forewing by acting in
antagonism to M. dorsoventralis primus and M. mesonoti se-
cundus.

Similar muscles: Muscle vibrateur dorsal longitudinal (Poisson,
1924); tergal longitudinal muscle (Malouf, 1933); dorsal
muscles of mesothorax (Rawat, 1939); indirect and principal
depressor of fore-wings (Akbar, 1957); 1ph-2ph and sco9-2ph
(Lauck, 1959).

dl. M. MESONOTI SECUNDUS (Figs. 26 and 27)
Quite large when developed; degenerate in the majority of speci-

mens.
Origin: Lateral surface of the ventral process of the second
phragma.

Insertion: Anterolateral region of the mesoscutum.

Action: Indirect flight muscle, raising the forewings. Its con-
traction forces the anterior notal wing process downward upon
the first axillary sclerite. Since the pleural wing process forms
a fulerum upon which the second axillary sclerite pivots, the
rest of the wing is forced upwards.

Similar muscles: Tergal longitudinal oblique muscle (Malouf,
1933); secondary indirect levator of fore-wings (Akhbar,
1957); se-scle-2ph (Lauck, 1959).

32. M. MESOSTERNI PRIMUS (Figs. 22 and 26)

A fairly long, well-developed muscle.

Origin: Anterior surface of the mesothoracic furea.

Insertion: Posterior part of the prosternum, between the posterior
sternal processes.

Action: Depresses the prothorax.

Similar muscles: Sternal longitudinal muscle (?) (Malouf, 1933) ;
ventral muscle of mesothorax (Rawat, 1939); fuy-fus (Lauck,
1959).

336 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Figur; 26. Medial view of the left half of the mesothorax (same view as
Fig. 127, showing the more medial mesothoracic muscles. The middle part
of Muszle 30 has been cut away. Muscles 30, 31, and 34 are fully developed
in this specimen.

Figure 27. Same view as above. Muscles 30, 32, 34, 35, and 40 have been

removed. Tendon 40 is not shown.

2@Q
od.

39:

Ol

M.

PARSONS: THORAX OF GELASTOCORIS 337

DORSOVENTRALIS PRIMUS (Fig. 26)

Large when developed; degenerate in the majority of specimens.

Origin: In a depression on the mesothoracie precoxal bridge, just
anterior to the coxal cavity.

Insertion: Anterior part of the mesoscutum, just lateral to the
parapsidal ridge.

Action: Same as M. mesonoti secundus.

Similar muscles: Muscle vibrateur transversal (sternali-dorsal)
(Poisson, 1924); tergo-sternal muscle (Malouf, 1933) ; indirect
and principal levator of fore-wings (Akbar, 1957); scs-bse
(Lauck, 1959).

. DORSOVENTRALIS SECUNDUS (Fig. 26)

A very short muscle.

Origin: Posterior arm of the mesothoracic furea.

Insertion: Tip of the ventral process of the second phragma,
between the two layers of this process.

Action: Depresses the posterior mesotergum and the anterior
metatergum (?).

Similar muscles: Tergo-sterno-furcal muscle (Malouf, 1933) ;
tergo-sternal furcal muscle of mesothorax (Rawat, 1939) ;
secondary indirect depressor of fore-wings (Akbar, 1957) ;
2ph-fug (Lauck, 1959).

. EPISTERNO-ALARIS (Fig. 30)

Lies beneath M. pleura-trochanteralis primus and M. episterno-
coxalis.

Origin: Anterior region of the mesothoracic episternum, just pos-
terior to the point of origin of VW. episterno-coxalis.

Insertion: On a tendon from the ‘‘elbow’’ of the third axillary
selerite of the forewing.

Action: Direct flight muscle. Contraction causes the third ax-
illary sclerite to flip over, thus flexing a previously extended
forewing.

Similar muscles: First flexor of fore wing (Malouf, 1933; Akbar,
1957); axillary muscle of mesothorax (Rawat, 1939) ; 3ax9-epse
(Lauck, 1959).

. FURCA-PLEURALIS (Figs. 26 and 30)

A very minute muscle.

Origin: Tip of the mesothoracic fureal apodeme.

Insertion: Tip of the mesothoracic pleural apophysis.

Action: Uncertain.

Similar muscles: Sterno-pleuro-apophysal muscle (Malouf, 1933) ;
pleurosternal muscle (Rawat, 1939); promoto-extensor of fore-
wings (?) (Akbar, 1957); plre-fug (Lauck, 1959).

338

40.

41.

46,

Me

M.

M.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

NOTO-TROCHANTINALIS (Fig. 26)

A well-developed, fan-shaped muscle. ’

Origin: Mesoseutum, between M. dorsoventralis primus and M.
noto-trochanteralis.

Insertion: Tendon 40, at the medial end of the mesothoracic
trochantin.

Action: Rotates the coxa and promotes the leg.

Similar muscles: Tergal promotor of coxa (Malouf, 1933); tergal
promotor of mesothorax (Rawat, 1939); se-sclo-exg (Lauck,
1959).

. NOTO-COXALIS (Figs. 26, 27, and 30)

A well-developed, fan-shaped muscle.

Origin: Posterolateral limit of the mesoscutum.

Insertion: Tendon 41, in the posterior region of the pericoxal
membrane.

Action: Rotates the coxa and remotes the leg.

Similar muscles: Tergal remotor of coxa (Malouf, 1933); tergal
remotor of mesothorax (Rawat, 1939); first remotor of coxa
(Akbar, 1957) ; se-selg’-exg’ (Lauck, 1959).

EPISTERNO-COXALIS (Figs. 26, 27, and 30)

A yather small, fan-shaped muscle.

Origin: Anterior part of the mesothoracic episternum, in the
region of the prealar bridge.

Insertion: Tendon 42, just anterior to the coxal process.

Action: Rotates the coxa and promotes the leg.

Similar muscles: Sternal promotor of coxa (2?) (Malouf, 1933) ;
second promotor of coxa (Akbar, 1957); epsg-exy (Lauck,
1959).

NOTO-TROCHANTERALIS (Fig. 27)

A well-developed muscle.

Origin: Mesoscutum, between M. mesonoti secundus and M. noto-
trochantinalis.

Insertion: Tendon 46, from the part of the proximal rim of the
trochanter which is farthest from the femur.

Action: Depresses the trochanter.

Similar muscles: Depressor of telopodite, tergal branch (Malouf,
1933); extra-coxal depressor of the trochanter of the meso-
thorax, tergal branch (Rawat, 1939); tergal depressor of
trochanter (Akbar, 1957); se-selo-tro (luauck, 1959).

PARSONS: THORAX OF GELASTOCORIS 339

47A. M. PLEURA-TROCHANTERALIS PRIMUS (Figs. 27 and 30)

60.

(Uk

M.

M.

M.

A rather slender muscle.

Origin: Anterior part of the mesothoracie episternum, just lateral
to M. episterno-coxalis.

Insertion: Tendon 46.

Action: Depresses the trochanter.

Similar muscles: Depressor of telopodite, pleural branch (Malouf,
1933); extra-coxal depressor of the trochanter of the meso-
thorax, pleural branch (Rawat, 1939); part of M. pleura-
trochanteralis (Larsén, 1945a); pleural depressor of tro-
chanter (Akbar, 1957); epse-trg (Lauck, 1959).

. PLEURA-TROCHANTERALIS SECUNDUS (Figs. 26, 27, and 30)

Origin: Medial surface of the mesothoracic pleural apophysis.

insertion: Tendon 46.

Action: Depresses the trochanter.

Similar muscles: Extra-coxal depressor of the trochanter of the
mesothorax, pleural branch (Rawat, 1939); part of M.
pleura-trochanteralis (Larsén, 1945a).

FURCA-TROCHANTERALIS (Figs. 27 and 30)

A small muscle, rather difficult to see.

Origin: Base of the furecal apodeme of the mesothorax.

Insertion: Tendon 46.

Action: Depresses the trochanter.

Similar muscles: Extra-coxal depressor of the trochanter of the
mesothorax, sternal branch (Rawat, 1939); fus-tro (Lauck,
1959).

MUSCLES OF THE METATHORAX

DORSOVENTRALIS (Fig. 28)

A slender muscle.

Origin: Tip of the metathoracic furea.

Insertion: Third phragma, lateral to the midline.

Action: Depresses the posterior part of the metanotum (?).

Similiar muscles: Tergo-sternal fureal muscle of metathorax
(Rawat, 1939); 8ph-fug (lauck, 1959).

EPISTERNO-ALARIS (Figs. 29 and 30)

A very slender muscle, difficult to see.

Origin: Wateral part of the metathoracie episternum, just lateral
to the point of origin of M. plewra-trochanteralis.

Insertion: On a tendon from the ‘‘elbow’’ of the third axillary
sclerite of the hindwing.

340 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Action: Direct flight muscle. Flexes the hindwing in the same
way that M. episterno-alaris of the mesothorax flexes the
forewing.

Similar muscles: Flexor of hind wing (Malouf, 1933); axillary
muscle of metathorax (Rawat, 1939); first flexor of hind-
wings (Akbar, 1957); 3ax-epsg (Lauck, 1959).

6|

Figure 28. Medial view of the left halves of the metathorax and of the
posterior mesothorax (same view as Fig. 12), showing the more medial
metathoracic and abdominal muscles. The ventral process of the second
phragma has been cut off.

Figure 29. Same view as above. Muscles 60, 63, 70, and 80 have been

removed.

63. M. NOTO-TROCHANTINALIS (Fig. 28)

A well-developed, fan-shaped muscle.

Origin: Anterior part of the metanotum, just lateral to the mid-
line.

Insertion: Tendon 63, at the medial end of the metathoracic tro-
chantin.

Action: Rotates the coxa and promotes the leg.

Sinular muscles: Tergal promotor of coxa (Malouf, 1933); tergal
promotor of metathorax (Rawat, 1939); first promotor of coxa
(?) (Akbar, 1957); ses;-exg (Lauck, 1959).

64.

PARSONS: THORAX OF GELASTOCORIS 341

M. NOTO-COXALIS (Figs. 29 and 30).

A well-developed, fan-shaped muscle.

Origin: Metanotum, lateral and posterior to M. noto-trochantinalis.

Insertion: Tendon 64, in the posterior part of the pericoxal mem-
brane.

Action: Rotates the coxa and remotes the leg.

Similar muscles: Tergal remotor of coxa, first branch (Malouf,
1933); tergal remotor of metathorax (Rawat, 1939); first
remotor of coxa (Akbar, 1957); se3’-exs’ (Lauck, 1959).

Zz, og
es TREE

T6365 4g ‘746

Figure 30. Dorsal view of the inner ventral surface of the pterothorax,

left side (same view as Fig. 15), showing the ventral and lateral muscles.
The middle parts of Muscles 47A and 71 have been cut away, and most of
the tendons of the leg muscles, along with Muscles 41 and 64, have been
cut off near their bases. The third and fourth axillary sclerites are shown
in place.

65.

M. FURCA-TROCHANTINALIS (Figs. 28 and 30)
A slender muscle, rather difficult to see.
Origin: Base of the posterior arm of the mesothoracic furea.
Insertion: On Tendon 63 and on the medial end of the metathoracic
trochantin.
Action: Rotates the coxa and promotes the leg.
Similar muscle: M. episterno-trochantinalis (?) (Larsén, 1945a).

342

66,

M.

M.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

EPISTERNO-COXALIS (Fig. 30)

A flat, broad musele lateral to M. fwrea-trochantinalis.

Origin: On the ridge bordering the posterior margin of the meso-
thoracic coxal cavity (Larsén, 1945a, considered this ridge
to be part of the metathoracie episternum).

Insertion: Anterior margin of the anterior basicoxite of the
metathoracie coxa.

Action: Rotates the coxa and promotes the leg.

Similar muscle: Sternal promotor of metathorax (?) (Rawat,
1939).

. COXA-SUBALARIS (Figs. 29 and 30)

A slender muscle.

Origin: On the basicostal suture of the metathoracic coxa, in the
region of the coxal process.

Insertion: On the very small metathoracie subalare.

Action: Direct flight muscle. Depresses the posterior margin of
the hindwing.

Similar muscles: Depressor of posterior margin of hind wing
(Malouf, 1933); second flexor of hind-wings (Akbar, 1957).

NOTO-TROCHANTERALIS (Fig. 28)

A very well-developed muscle.

Origin: Lateral part of the metanotum.

Insertion: Tendon 70, from the part of the proximal rim of the
trochanter which is farthest from the femur.

Action : Depresses the trochanter.

Similar muscles: Depressor of trochanter, tergal branch (Malouf,
1933); extra-coxal depressor of the trochanter of the meta-
thorax, tergal branch (Rawat, 1939); tergal depressor of
trochanter (Akbar, 1957); ses-trg (lauck, 1959).

. PLEURA-TROCHANTERALIS (Figs. 28 and 30)

A very well-developed muscle.

Origin: lateral and anterolateral region of the metathoracic
episternum.

Insertion: Tendon 70.

Action: Depresses the trochanter.

Similar muscles: Depressor of trochanter, pleural branch (Malouf,
1933); extra-coxal depressor of the trochanter of the meta-
thorax, pleural branch (Rawat, 1939); pleural depressor of
trochanter, (Akbar, 1957).

. FURCA-TROCHANTERALIS (Figs. 28 and 30)

Origin: Base of the metathoracie furea.
Insertion: Tendon 70.

PARSONS: THORAX OF GELASTOCORIS 343

Action: Depresses the trochanter.

Similar muscles: Extra-coxal depressor of the trochanter of the
metathorax, sternal branch (Rawat, 1939); fug-trg (Lauck,
1959).

80. M. VENTRALIS ABDOMINALIS (Fig. 28)
A short, broad abdominal muscle.
Origin: Posterior surface of the metathoracie furca.
Insertion: On a ridge on the ventrolateral part of the second ab-
dominal segment.
Action: Raises the abdomen.
Similar muscles: veM, (Larsén, 1945a) ; fug-2S (Lauck, 1959).

INTRINSIC MUSCLES OF THE LEGS
Prothoracie legs (Fig. 31)

23. M. COXA-TROCHANTERALIS MEDIALIS

A short, broad, well-developed muscle.

Origin: Posteromedial wall of the coxa.

Insertion: Tendon 20.

Action: Depresses the trochanter.

Similar muscles: Coxal branch of depressor of trochanter (Malouf,
1933); depressor of the trochanter (Rawat, 1939); coxal de-
pressor of trochanter (Akbar, 1957).

24. M. COXA-TROCHANTERALIS LATERALIS

A muscle consisting of three bundles.

Origin: Anterior wall of the coxa.

Insertion: On the three-branched Tendon 24, from the part of the
proximal rim of the trochanter which is nearest the femur.

Action: Raises the trochanter.

Similar muscles: Wevator of trochanter (as shown in Pl. XVI,
fig. 1, by Malouf, 1933; Rawat, 1939; Akbar, 1957).

bo
SK

M. REDUCTOR FEMORIS

A short, broad muscle.

Origin: Posteromedial wall of the trochanter.

Insertion: lateral part of the proximal margin of the femur.
Some strands enter the femur and insert on Tendon 26.

Action: Moves femur laterally. Strands entering the femur depress
the tibia.

Similar muscles: Remotor of femur (Malouf, 1933); reductor of
the femur (Rawat, 1939; Akbar, 1957).

344 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

26. M. DEPRESSOR TIBIAE

A very well-developed muscle.

Origin: Walls of the ventral half of the femur.

Insertion: Tendon 26, from the ventral region of the proximal]
margin of the tibia.

Action: Depresses the tibia, closing it upon the femur.

Similar muscles: Depressor of tibia (Malouf, 1933; Rawat, 1939;
Akbar, 1957).

===-24 27

—
lo ow

o,
Cs
—~ ese

Stews

Figure 31. Medial view of the left prothoracie leg, with the medial walls
of the leg removed (same view as Fig. 17), showing the intrinsic leg muscles.

27. M. LEVATOR TIBIAE

Less well-developed than VW. depressor tibiae.

Origin: Walls of the most dorsal part of the femur.

Insertion: Tendon 27, from the dorsal region of the proximal
margin of the tibia.

Action: Raises the tibia.

Similar muscles: Levator of tibia (Malouf, 1933; Rawat, 1957) ;
extensor of tibia (Akbar, 1957).

28. M. DEPRESSOR TARSI

Composed of many short, fine muscle strands.

Origin: Ventrolateral walls of the tibia.

Insertion: Tendon 28, from the ventral region of the proximal
margin of the tarsus.

Action: Depresses the tarsus.

Similar muscles: Depressor of tarsus (Malouf, 1933; Rawat, 1939;
Akbar, 1957).

PARSONS: THORAX OF GELASTOCORIS 345

294A. M. DEPRESSOR PRAETARSI PRIMUS
A well-developed muscle.
Origin: Lateral walls of the dorsal half of the femur, between
M. depressor tibiae and M. levator tibiae.
Insertion: Intrafemoral part of Tendon 29, from the unguitractor
of the pretarsus.
Action: Depresses the pretarsus.
Similar muscles: Depressor of pretarsus, femoral branch (Malouf,
1933; Rawat, 1939); part of M. depressor praetarsi (Larsén,
1945a) ; depressor of pretarsus, proximal muscle (Akbar, 1957).
29B. M. DEPRESSOR PRAETARSI SECUNDUS
A very weak muscle, consisting of only a few strands.
Origin: Proximal region of the dorsal wall of the tibia.
Insertion: Intratibial part of Tendon 29.
Action: Depresses the pretarsus.
Similar muscles: Depressor of pretarsus, tibial branch (Malouf,
1933; Rawat, 1939); part of M. depressor praetarsi (Larsén,
1945a); depressor of pretarsus, distal muscle (Akbar, 1957).

Pterothoracic legs

Kach of the following muscles (Nos. 49-55B and Nos. 73-79B)
corresponds to a similar muscle in the foreleg which bears the
same name. The origins, insertions, and actions of the correspond-
ing muscles are similar, and each muscle, with the exception of
M. coxa-trochanteralis medialis, inserts on a tendon bearing the
same number as the muscle. In the following account, for each
pterothoracic muscle the number of the corresponding prothoracic
muscle will be noted, along with any significant differences in
general appearance.

Mesothoracic leg's
49. M. COXA-TROCHANTERALIS MEDIALIS
Similar to Muscle 23. Inserts on Tendon 46.

50. M. COXA-TROCHANTERALIS LATERALIS
Similar to Muscle 24.

dl. M. REDUCTOR FEMORIS
Similar to Muscle 25.
52. M. DEPRESSOR TIBIAE
Similar to Muscle 26, but less well developed.
53. M. LEVATOR TIBIAE
Similar to Muscle 27, but somewhat less well developed.

346 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

54. M. DEPRESSOR TARSI

Similar to Muscle 28.
55A. M. DEPRESSOR PRAETARSI PRIMUS

Similar to Muscle 29A, but much less well developed.
55B. M. DEPRESSOR PRAETARSI SECUNDUS

Similar to Muscle 29B.

Metathoracic legs

73. M.COXA-TROCHANTERALIS MEDIALIS
Similar to Muscle 23. Inserts on Tendon 70.
74. M.COXA-TROCHANTERALIS LATERALIS
Similar to Musele 24.
75. M. REDUCTOR FEMORIS
Similar to Musele 25.
76. M. DEPRESSOR TIBIAE
Similar to Muscle 26, but much less well developed.
77. M. LEVATOR TIBIAE
Similar to Muscle 27, but less well developed.
78. M. DEPRESSOR TARSI
Similar to Muscle 28; the muscle strands are shorter and weaker.
79A. M. DEPRESSOR PRAETARSI PRIMUS
Similar to Muscle 29A, but much less well developed.
79B. M. DEPRESSOR PRAETARSI SECUNDUS
Similar to Muscle 29B.

DISCUSSION

Among the Heteroptera, degeneration of the flight muscles,
such as has been observed in the large majority of the gelasto-
corids examined, is not uncommon. Many species have individuals
which are unable to fly because of reduction of the wings, of the
muscles, or of both. Polymorphism of the wings is found in some
terrestrial families, such as the Pyrrhocoridae, Aradidae, and
Lygaeidae (Weber, 1930) and in many aquatic and semi-aquatic
families. Poisson (1924), who studied polymorphism in the
aquatie Corixidae, Aphelocheiridae, and Naucoridae, and in the
semi-aquatie Gerridae, Hydrometridae, Veliidae, and Mesoveli-
idae, found that individuals with reduced wings usually showed
degenerate flight muscles, although in a few cases the muscles were
normal. Larsén (1950) found that in Aphelocheirus the degree

PARSONS: THORAX OF GELASTOCORIS 347

of reduction of the flight musculature increased in proportion to
the amount of reduction of the wings.

Degeneration of the flight muscles in normal-winged individ-
uals, as in Gelastocoris, is also quite common in both aquatic and
terrestrial Heteroptera (Larsén, 1950). Among the aquatic
forms, it has been reported in the Nepidae, Naucoridae, and
Aphelocheiridae (Ferriére, 1914; Poisson, 1924; Larsén, 1949
and 1950). In their general appearance, the degenerate muscles
of Gelastocoris closely resemble the reduced muscles of macrop-
terous individuals of Aphelocheirus, as illustrated by Larsén
(1950; his Fig. 8b). The degenerate dorsal longitudinal muscles
of the mesothorax were termed the ‘‘tracheo-parenchymatous
organ’’ by some authors because of the abundance of tracheoles
which penetrate them. Early workers such as Dufour (1833)
and Does (1909) believed this ‘‘organ’’ to be respiratory in fune-
tion. It appears, however, that the tracheoles are only those
which would penetrate a normal muscle, and that the tracheo-
parenchymatous organ has no special respiratory function (Fer-
riere, 1914; Brocher, 1916). A degenerate MW. mesonoti primus
of Gelastocoris, when teased apart and examined under a com-
pound microscope, shows a rich supply of tracheoles similar to
those figured by Ferriére in the tracheo-parenchymatous organ
of Nepa.

One puzzling feature noted in the present investigation is
that although some gelastocorids possess well developed flight
muscles as well as normal wings none of the insects were ever
observed to fly. During nearly a year of captivity they were
constantly given opportunities to do so, but never showed any
inclination towards flight. Larsén (1950) made a similar ob-
servation on a few individuals of Ranatra which never flew even
when strongly stimulated to do so. Examination of their muscu-
lature showed it to be normal. That author proposed that this
might be due to a reduction of the nervous component of the
fight apparatus. Whether or not this is a plausible explanation
for the lack of flight in Gelastocoris may be elucidated by further
anatomical work. Todd (1955), who also observed no flight in
Gelastocoris oculatus, has noted that several other species of
Gelastocoridae have forewings which are fused or which have
reduced membranes, and in some species the hindwings are
reduced.

348 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

The present study offers a few elues to the possible phylogen-
etie position of the Gelastocoridae among the Heteroptera. A
brief review of the literature on this problem has been presented
in a previous paper (Parsons, 1959), and the reader is referred
to that work for a discussion of the theories of earlier authors. It
is generally agreed that the three littoral families Gelastocoridae,
Ochteridae, and Saldidae are closely related to each other, the
first-named family having arisen from the second. It also appears
that these three families represent a stage in the evolution of the
totally aquatic and semi-aquatic bugs (the Hydrocorisae and
Amphibicorisae respectively) from the terrestrial forms (the
Geocorisae). Authorities have disagreed, however, as to which
of the littoral families are related to the Hydrocorisae and which
to the Amphibicorisae. De la Torre-Bueno (1923) believed the
Hydrocorisae to be descended from saldid-like ancestors, with
the ochterids and gelastocorids as intermediate stages. Spooner’s
(1938) work on the head capsule led him to place the latter two
familes with the Amphibicorisae, and the saldids with the Geo-
eorisae. More recently, China (1955) has proposed that the
Amphibicorisae arose from ‘‘Proto-Saldidae’’ and the Hydro-
eorisae from ‘‘Proto-Ochteridae.’’

Larsén (1945b), after studying a large number of heterop-
teran families, found five characteristics of the thoracic skeleton
which seem to be more typical of the Hydrocorisae than of the
other Heteroptera. First, the metanotum of the aquatie bugs is
longer than the metapostnotum; in the Geocorisae the latter is
longer than the former, while in Salda (Saldidae) the two are
equal in leneth. Unfortunately, the boundary between these two
regions is indistinct in the Amphibicorisae studied by Larsén, so
that it is difficult to compare them with the Hydrocorisae and
Geocorisae. The present study has shown the metanotum of
Gelastocoris to be much longer than the metapostnotum, and in
this character it resembles the Hydrocorisae.

A second feature of the Hydrocorisae, according to Larsén,
is the presence, in all three thoracic segments, of a distinet pleural
ridge (except in the mesothorax of Ranatra). Taylor (1918)
also pointed out the distinctness of the pleural ridge in the ptero-
thorax of corixids, belostomatids, and notonectids. In all the
semi-aquatic and terrestrial bugs studied by Larsén, the pleural

PARSONS: THORAX OF GELASTOCORIS 349

ridge is indistinct in at least one segment. A distinct prothoracic
pleural ridge with a pleural apophysis is present in all of the
Hydrocorisae studied by Larsén, but in only three of the Geo-
corisae and in none of the Amphibicorisae. In Gelastocoris, how-
ever, all three segments show distinct pleural ridges, and a pro-
thoracic pleural apophysis is present. A large posterior lobe on
the mesothoracie epimeron is a third character distinguishing
the Hydrocorisae. This lobe is quite extensive in Gelastocoris,
overlapping much of the metathoracic episternum, and its size
is comparable to that of the aquatic bugs Hesperocoriza, Noto-
necta, and Pelocoris. In the Amphibicorisae, in Salda, and in most
of the Geocorisae studied by Larsén the posterior mesothoracic¢
epimeral lobe is more weakly developed. Two other Geocorisae
showing weakly developed mesothoracic epimeral lobes are Nezara
(Malouf, 1933) and Lepiocorisa (Akbar, 1957).

A fourth characteristic of the aquatic bues, as cited by Larsén,
concerns the width of the metathoracie epimeron which is not
as reduced as in many terrestrial bugs. Unfortunately, he did
not compare the width of this selerite in the Amphibicorisae and
the Hydrocorisae, and did not state how many, if any, Geocorisae
are exceptions to this generalization. The metathoracic epimeron
of Gelastocoris appears to be as well developed as that of the
aquatic bugs Hesperocorixa, Notonecta, Pelocoris, Belostoma,
Nepa, and Ranatra. Finally, Larsén stated that the mesothoracic
pleural apophysis in the Hydrocorisae is large and extends dor-
sally. The size of this process in Gelastocoris is comparable to
that of Belostoma and of Naucoris as figured by Larsén (1945a) ;
it appears to be somewhat smaller than that of Notonecta, but
is considerably larger than that of Hesperocorixa and Ranatra.
It extends dorsally, like the pleural apophyses of the aquatic
forms. Among the semi-aquatic bugs, according to Larsén
(1945a and b), the mesothoracic pleural apophysis is absent in
Velia, small in Gerris, and well developed in Hydrometra; among
the Geocorisae it is variable in both size and position (Larsén,
1945b). Malout’s (1933) figure of the mesothoracic pleural
apophysis in the terrestrial bug Nezara shows it to be fairly
small and medially directed.

In a few other features of the thoracic skeleton, Gelastocoris
resembles the Hydrocorisae, Amphibicorisae, or Geocorisae.

350 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Larsén (1945b) found a distinct separation between the meta-
thoracic seutum and scutellum only in the aquatic and semi-
aquatic bugs. Although Akbar (1957) described a clear separa-
tion between these two regions in the metanotum of Leptocorisa,
a terrestrial bug, it seems that his interpretation is open to eriti-
cism; the part termed the ‘‘scutum’’ by him seems to be the
notum, while his ‘‘seutellum’’ (as shown in his Fig. 66) resembles
the postnotum. Malouf (1933) also incorrectly described a
distinct metascutum and metascutellum in Nezara, as Larsén
(1945b) has pointed out. On the gelastocorid metanotum there
is a fairly definite groove which may represent a scutoscutellar
suture; if this interpretation is correct, this character links the
gelastocorids with both the Hydrocorisae and the Amphibicorisae.

Larsén (1945b) also reported the prothoracic postcoxal bridge
to be broader than the precoxal bridge in the majority of Hydro-
corisae; this is also the case in Gelastocoris. In the Geocorisae
and Amphibicorisae, either the precoxal bridge is the broader
of the two or both bridges are equal in size. This does not serve
to distinguish the Hydrocorisae as a whole, however, since in
Notonecta and Corixa, according to Larsén, the postcoxal bridge
is narrower than the precoxal.

Larsén (1945b) found that although the metathoracie sub-
alare is present in most terrestrial bugs it is absent in most
aquatic (with the exception of Notonecta) and semi-aquatie
forms. The presence of a metathoracic subalare in Gelastocoris
is, therefore, a character most commonly found in the Geocorisae ;
this sclerite is, however, much reduced in Gelastocoris.

The thoracic musculature does not shed as much hght on the
phylogenetic problem as does the thoracic skeleton. Larsén’s
comparative study revealed very few differences between the
three major heteropteran groups on the basis of musculature.
Three generalizations can be made, however. First, the two dorsal
longitudinal muscles of the heteropteran metathorax (‘‘Mm.
metanoti primus’’ and ‘‘secundus’’ of Larsén, 1945a) are absent
in all the Hydrocorisae examined by that author, while at least
one of the two is present in Salda, in all the Amphibicorisae,
and in all but two of the Geocorisae. Gelastocoris resembles the
aquatic bugs in this respect, since it lacks both metathoracie
dorsal longitudinal muscles. Secondly, the ventral longitudinal

PARSONS: THORAX OF GELASTOCORIS 351

muscle of the abdomen (M. ventralis abdominalis of the present
study) is well developed in all Larsén’s Hydrocorisae and in
Salda, but is weak or absent in most of the semi-aquatic and
terrestrial forms examined by him. Here again, Gelastocoris
resembles the Hydrocorisae. Thirdly, a M. coxa-subalaris is pres-
ent in both the mesothorax and the metathorax of most of the
terrestrial bugs studied by Larsén, but is absent in the aquatic
and semi-aquatie forms, the only exception being its presence
in the metathorax of Notonecta. The presence of this muscle in
the metathorax of Gelastocoris links this bug with the Geocorisae ;
the link is not very strong, however, since Notonecta also pos-
sesses this muscle in the metathorax, and since the muscle is absent
in the mesothorax of Gelastocoris.

In general, therefore, the skeleton and musculature of the
thorax of Gelastocoris bear more resemblance to those of the
Hydrocorisae than to those of the Amphibicorisae or Geocorisae.
This is in agreement with the conclusions reached in a previous
study of the gelastocorid head (Parsons, 1959), and supports
the phylogenetic theory of China (1955). Similarities to the
aquatic bugs are seen in the structure of the metatergal sclerites,
the presence of distinct pleural ridges in all three segments,
the size of the mesothoracic and metathoracic epimera, the degree
of development of the mesothoracic pleural apophyses, and the
breadth of the prothoracic posteoxal bridge. Further resem-
blances to the aquatic Heteroptera are the absence of meta-
thoracic dorsal longitudinal muscles and the presence of
M. ventralis abdominalis. The gelastocorids also resemble both
the Hydrocorisae and the Amphibicorisae in the separation be-
tween the metascutum and the metascutellum. Only two features
of the gelastocorid thorax are atypical of the Hydrocorisae; in
their possession of a metathoracic subalare and subalar muscle,
they resemble the Geocorisae (although these two characters are
also found in Notonecta, which is definitely one of the Hydro-
corisae). It must be borne in mind, however, that there are ex-
ceptions in the literature to all of the above generalizations, and
that these are not clear-cut distinctions.

Boe BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

LITERATURE CITED

AKBAR, S.S.
1957. The morphology and life-history of Leptocorisa varicornis Fabr.
(Coreidae, Hemiptera) — A pest of paddy crop in India. Part I.
Head and thorax. Aligarh Muslim Uniy. Publ. (Zool. Ser.)
Ind. Ins. Typ., no. 5, 53 pp.

BRINDLEY, M. D. H.
1934. The metasternum and pleuron of Heteroptera. Trans. Roy. Ent.
Soc. London, vol. 82, pp. 43-50.

BrRocHER, F.

1916. La neépe cendrée. Etude anatomique et physiologique du systéme
respiratoire, chez l’imago et chez la larve; suivie de quelques
observations biologiques concernant ces insects. Arch. Zool.
Exp. Gén., vol. 5, pp. 483-514.

CuHina, W. HB.
1955. The evolution of the water bugs. Nat. Inst. Sei. India. Bull.
no. 7, pp. 91-108.

Doas, W.
1909. Metamorphose der Respirationsorgane bei Nepa cinerea. Mitt.
Nat. Ver. Neuvorpommern-Riigen, Jahrg. 40, pp. 1-55.

Durour, L.
1833. Recherches anatomiques et physiologiques sur les hémiptéres, ac-
compagnées de considérations relatives 4 l]’histoire naturelle et
a la classification de ces insectes. Mém. Sav. Etrang. Acad. Sci.
Paris, vol. 4, pp. 129-462.

EsaAkl,T., and S. MiyamMorTo
1955. Veliidae of Japan and adjacent territory (Hemiptera-Heterop-
tera). I. Microvelia Westwood and Pseudovelia Hoberlandt of
Japan. Sieboldia, vol. 1, pp. 169-204.

FERRIERE, C.
1914. L’organe trachéo-parenchymateux de quelques hémiptéres aqua-
tiques. Rev. Suisse Zool., vol. 22, no. 5, pp. 121-145.

GRIFFITH, M. E.
1945. The environment, life history and structure of the water boat-
man, Ramphocorixa acuminata (Uhler) (Hemiptera, Corixidae).
Univ. Kans. Sci. Bull., vol. 30, pt. 2, no. 14, pp. 241-365.

PARSONS: THORAX OF GELASTOCORIS 300

HAMILTON, M. A.
1931. The morphology of the water scorpion, Nepa cinerea Linn.
(Rhynchota, Heteroptera). Proc. Zool. Soe. London, 1931, pp.
1067-1136.

Hoxg, 8.
1926. Preliminary paper on the wing-venation of the Hemiptera
(Heteroptera). Ann. Ent. Soe. Amer., vol. 19, pp. 13-34.

LARSEN, O.

1942. Bisher unbeachtete wichtige Ziige im Bau des Metathorax bei den
Heteropteren. Kungl. Fysiograf. Sallskap. Lund Forhandl., vol.
IPA rio, 18}, 15) joo,

1945a. Der Thorax der Heteropteren. Skelett und Muskulatur. Lunds
Univ. Arsskrift., N.F., Avd. 2, vol. 41, no. 3, 96 pp.

1945b. Das thorakale Skelettmuskelsystem der Heteropteren. Ein Bei-
trag zur vergleichenden Morphologie des Insektenthorax. Lunds
Univ. Arsskrift., N.F., Avd. 2, vol. 41, no. 11, 83 pp.

1945e. Das Meron der Insekten. Kung]. Fysiograf. Sallskap. Lund
Forhandl., vol. 15, no. 11, 9 pp.

1945d. Die hintere Region der Insektenhiifte. Kungl. Fysiograf. Salls-
kap. Lund Forhandl., vol. 15, no. 12, 12 pp.

1949. Die Ortsbewegungen von Ranatra linearis L. Hin Beitrag zur
vergleichenden Physiologie der Lokomotionsorgane der Insekten.
Lunds Univ. Arsskrift, N.F., Avd. 2, vol. 45, no. 6, 82 pp.

1950. Die Verinderungen im Bau der Heteropteren bei der Reduktion
des Flugapparates. Opuse. Ent., vol. 15, pp. 17-51.

Lauck, D. R.
1959. The locomotion of Lethocerus (Hemiptera, Belostomatidae).
Ann. Ent. Soc. Amer., vol. 52, pp. 93-99.

Ma.our, N.S. R.
1933. The skeletal motor mechanism of the thorax of the ‘‘stink bug’’,
Nezara viridula L. Bull. Soc. Roy. Ent. Egypte, n.s., vol. 16,
pp. 161-203.

Parsons, M. C.
1959. Skeleton and musculature of the head of Gelastocoris oculatus
Fabricius (Hemiptera-Heteroptera). Bull. Mus. Comp. Zool.
Harv. Coll., vol. 122, pp. 1-53.

Poisson, R.
1924. Contribution 4 1’étude des hémiptéres aquatiques. Bull. Biol.
France Belg., vol. 58, pp. 49-305.

354 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Rawat, B. L.
1939. Notes on the anatomy of Naucoris cimicoides lL. (Hemiptera-

Heteroptera). Zool. Jahrb., Abt. Anat., vol. 65, pp. 535-600.

SNODGRASS, R. E.
1909. The thorax of insects and the articulation of the wings. Proce.
U.S. Nat. Mus., vol. 36, no. 1687, pp. 511-595.
1927. Morphology and mechanism of the insect thorax. Smithsonian
Mise. Coll., vol. 80, no. 1, 108 pp.
1935. Principles of insect morphology. McGraw-Hill, New York.

ix + 667 pp.

SPOONER, C.S.
1938. The phylogeny of the Hemiptera based on a study of the head
capsule. Illinois Biol. Monog., vol. 16, no. 3, 102 pp.

SPRAGUE, I. B.
1956. The biology and morphology of Hydrometra martini Kirkaldy.
Univ. Kans. Sci. Bull., vol. 38, pt. 1, no. 9, pp. 579-693.

TANAKA, T.
1926. Homologies of the wing veins of the Hemiptera. Annot. Zool.
Jap., vol. 11, pp. 33-57.

TAYLOR, L. A.
1918. The thoracic sclerites of Hemiptera and Heteroptera. With notes
on the relationships indicated. Ann. Ent. Soc. Amer., vol. 11,

pp. 225-254.

Topp, E. L.
1955. A taxonomic revision of the family Gelastocoridae (Hemiptera).
Univ. Kans. Sci. Bull., vol. 37, pt. 1, no. 11, pp. 277-475.

TORRE-BUENO, J. R. DE LA
1923. Family Saldidae. in: W. E. Britton. Guide to the insects of
Connecticut. Part IV. The Hemiptera of sucking insects of
Connecticut. Conn. Geol. Nat. Hist. Surv., Bull. No. 34, pp. 408-

416.

Tower, D. G.
1913. The external anatomy of the squash bug, Anasa tristis deG. Ann.
Ent. Soe. Amer., vol. 6, pp. 427-441.

WEBER, H.
1930. Biologie der Hemipteren. Eine Naturgeschichte der Schnabel-
kerfe. Springer, Berlin. vii + 543 pp.

PARSONS: THORAX OF GELASTOCORIS 355

EXPLANATION OF FIGURES

In the figures, the membranes, the muscles, the tendons, and
the cut edges of the skeleton are unstippled, while the skeletal
surfaces are either stippled or blackened. The muscles are in-
dicated by the numbers given in pages 329-346. Each major ten-
don is indicated by a ‘‘T’’ followed by the number of the muscle
attaching to it; when more than one muscle attaches to a tendon,
the tendon’s number is that of the lowest-numbered muscle. The
numeral II after an abbreviation indicates a mesothoracic struc-
ture, while the numeral III indicates a metathoracie structure.

The abbreviations used in the figures are as follows:

1, 2, or 3 A — first, second or third anal vein

AB — anterior basicoxite

AC — axillary cord

AF — anteromedial flap of stink groove
AM — abdomen

AP — anterolateral abdominal process
AW — anterior notal wing process

1, 2 3, or 4 AX — first, second, third or fourth axillary sclerite
BR — basicostal ridge

BS — basicostal suture

C+CS — costa plus subcosta

CC — coxal cavity

CL — coxal cleft

CM — cervical membrane

CO — corlum

Cle — coxal process

CU — cubitus

CV — clavus

CW — claw

CX — coxa

EL — posterior epimeral lobe

EM — embolium

EP — epimeron

EPS — supracoxal lobe of epimeron
ES — episternum

ESS — supracoxal lobe of episternum
EV — evaporating surface

F — furea

FA — fureal apodeme

FE — femur

396

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

FP — fureal pit

FW — forewing

H — humeral plate

HA — comb of hairs on foreleg

HW # — hypopharyngeal wing

iL —- intersegmental membrane

K — knob, on mesothoracie epimeron, for anchoring forewing
L — lateral apodeme

LP — posterior lobe of protergum

M — media

MB — membrane

MP — median plate

N — notum

O — occipital condyle

P 1, 2 or 3 —first, second, or third phragma
PA — pleural apophysis

PAR — parapsidal ridge

PAS — parapsidal suture

12418} — posterior basicoxite

PC — precosta

PE — pericoxal membrane

PEB — prealar bridge

PEC — precoxal bridge

PF — posterolateral flap of stink groove
PG — protergum

PL — pleurosternal bridge

PM — prealar membrane

PN — postnotum

PO — postoceiput

POB — postalar bridge

POC — postcoxal bridge

PR — pleural ridge

PS — pleural suture

PSP — posterior sternal process

124 — pretarsus

J21O) — prescutum

PW — posterior notal wing process

R — radius

RG — ridge bordering posterior edge of mesocoxal cavity
S 1 or 2 — first or second thoracic spiracle
SA — abdominal tympanal organ

SB — subalare

sc — scutum

~!

PARSONS: THORAX OF GELASTOCORIS aD

sclerite for attachment of IZ. pronoti quintus
stink groove

seutoscutellar suture

seutellum

sternum

scutellar process

stink ridge

sternacostal suture

strut between posterior tergal and epimeral lobes of prothorax
tendon

tarsus

trochanter

tergal fissure

tibia

trochantin

transverse ridge

unguitractor

ventral process of second phragma
wing groove

pleural wing process

xiphus

xiphal groove

xiphal ridge

c

'
a

cs

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Voi. 122, No. 8

THE PALATINE PROCESS OF THE PREMAXILLA IN
THE PASSERES

A study of the variation, function, evolution and
taxonomic value of a single character
throughout an avian order

By Wauter J. Bock
Biological Laboratories, Harvard University

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM

JUNE, 1960

PUBLICATIONS ISSUED BY OR IN CONNECTION
WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

BULLETIN (octavo) 1863 — The current volume is Vol. 122.

BREVIORA (octavo) 1952 — No. 126 is current.

Memorrs (quarto) 1864-1938 — Publication was terminated with
Vol.55:

JOHNSONIA (quarto) 1941 — A publication of the Department of
Mollusks. Vol. 38, no. 39 is current.

OccASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo)
1945 — Vol. 2, no. 25 is current.

PROCEEDINGS OF THE NEW ENGLAND ZOOLOGICAL CLUB (octavo)
1899-1948 — Published in connection with the Museum. Publication
terminated with Vol. 24.

The continuing publications are issued at irregular interva!s in num-
bers which may be purchased separately. Prices and lists may be
obtained on application to the Director of the Museum of Comparative
Zoology, Cambridge 38, Massachusetts.

Of the Peters ‘‘Check List of Birds of the World,’’ volumes 1-3 are
out of print; volumes 4 and 6 may be obtained from the Harvard Uni-
versity Press; volumes 5 and 7 are sold by the Museum, and future
volumes will be published under Museum auspices.

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Vor. 122, No:?s

THE PALATINE PROCESS OF THE PREMAXILLA IN
THE PASSERES

A study of the variation, function, evolution and
taxonomic value of a single character
throughout an avian order

By Wa.rter J. Bock
Biological Laboratories, Harvard University

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM

JUNE, 1960

aa ;
abe mit of _ ey Poni r ai hag’
od ee a) ae

_ he. Yon: bv. > ee! ‘ihe +
a a =) > ayes 1 Tire Ba
>

:
Dood

ee ——
.— >.
ie a
a ay la ae |
> ea
We >, i
’ 7
7
a, vo
.
5 7
6 j 2
¥ =
i
7 i
ial =
on
i
=
BL, rus
. i , :
: —
et
'
coe =
7 ’ r
a7
i a -
.
_ - Vs a

No. 8 — The Palatine Process of the Premazilla in the Passeres

A Study of the variation, function, evolution and taxonomic
value of a single character throughout an avian order

3Y Water J. Bock
Biological Laboratories, Harvard University

CONTENTS
IMETOCUCEL Ont eee ke) Las eee Ss Se Te eee or aa 361
Acknowledgements). 4.5551). st Seton te ed. ite ldo eel aed Oe 365
Description of the palatine process of the premaxilla .............. 366
1B EEO Ao 5 EOL a EN yO DO MOLG SO OP ane] Ord ot REN cas cred NT eS CRM og Gen cho. cso 371
Development of the palatine process of the premaxilla ....... F AMLSILD
Function of the palatine process of the premaxilla ..............- 5 Bxelll
Variation of the palatine process of the premaxilla ............... 429
Evolution of the palatine process of the premaxilla .......... cca ds, 3409
Taxonomic value of the palatine process of the premaxilla.......... 470
CONCIUSIONS ee tee ee ene) Nate meee este este 478
Sumimnrys « asrns, Ae thre coe oe yin serch eels a Pal mes Oe aaa ares 479
hiteraturemerted Peres. epi oie ee ee eee Spine ag a .. 481
INTRODUCTION

Ever since the beginnings of avian taxonomy, ornithologists
have concentrated on the species problem, with the study skin as
the traditional object of study. This was in many ways a for-
tunate choice, and as a result, avian systematics on the species
level is today the most advanced area in the field of taxonomy.
But at the same time, interest in the higher categories of birds
has lagged so far behind that we know virtually nothing about
the affinities of most groups of birds. Even now, most systematic
work on the supergeneric level represents scarcely more than
guesswork, there is little agreement on the limits of the orders or
on their relationships, and within the relatively sharply defined
orders, the arrangements of the families are, at best, obscure.
Most neglected of all the orders are the Passeres which, although
they contain about half of the recent species of birds, have
received less attention than any other group. The lack of interest
in the anatomy as well as in the classification of the perching
birds dates back to the beginnings of ornithology and is reflected
in the attitude of the standard texts (Fiirbringer, 1888; Gadow,

362 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

1891-93 ; and Beddard, 1898). These authors give detailed cover-
age of the families and even subfamilies of the non-passerine
birds, yet they barely distinguish between the suborders of the
perching birds, the tacit assumption being that their highly
uniform morphology precludes the use of comparative anatomy
as a basis for their classification. Unfortunately, this high degree
of morphological similarity has usually been interpreted as uni-
formity, with the conclusion that comparative anatomical studies
are of no use whatsoever in untangling the relationships within
the Passeres.

Recently, there has been a revival of interest in the Passeres,
as is indicated by the publication of a number of papers on
their anatomy (Arvey, 1951; Ashley, 1941; Beecher, 1951a,
1951b, 1953; Berger, 1957; Engels, 1940; Fiedler, 1951; Hudson
and Lanzillotti, 1955; Mayr, Andrew and Hinde, 1956; Moller,
1930, 1931; Nelson, 1954; Sims, 1955; Stalleup, 1954; Stonor,
1937, 1938, 1942; Sushkin, 1924, 1925, 1927, 1929; Swinebroad,
1954; and Tordoff, 1954a, 1954b). These papers have shown that
the passerines are not absolutely uniform in their internal anat-
omy and that comparative anatomical studies may aid in the
understanding of relationships on the familial level. With the
removal of this psychological block and with increasing interest
in the problems of passerine anatomy we may at last be on the
way to understanding the evolution and classification of the
perching birds.

This revival of interest in the Passeres is, however, not without
its problems, of which the most important is the disagreement in
interpretation of the morphological findings and their value in
showing relationships. Stresemann (1959) has presented an
excellent picture of the problems confronting avian systematics
which should be read by every worker interested in this field.
Mayr (1955, 1958) has discussed some of the perplexing evolu-
tionary assumptions pertinent to passerine classification, and
Starck (1959) has commented on some of the anatomical prob-
lems. These authors agree, more or less, that the major problems
stem from the characters used as clues to relationships, and from
uncritical use of the pertinent evolutionary and morphological
principles. But something else is involved. Perhaps the difficulty
arises from the relatively small degree of anatomical difference
between passerine families; perhaps it stems from insufficient
study of the characters or perhaps it is a result of the method by
which the groups and their structures are compared. Undoubt-
edly, the answer is a combination of all three suggestions, but the

BOCK: PALATINE PROCESS OF THE PREMAXILLA 363

last one is probably the most important, and attention will
therefore be focused on it.

The best approach in taxonomic studies is a comparison of as
many characters as possible throughout the entire group. This
ideal method is feasible only with comparatively small orders
and families of birds. It is not practical when dealing with a
large group such as the Passeres; alternate methods must be
employed. These are of two types. The first is a comparison of
as many characters as possible in two or more families. This is
the method used in most of the works cited above. The second
is an analysis of a single character or character-complex through-
out the whole group under consideration. No proper study of
this type has, to my knowledge, been made for the passerines.
Therefore, this paper presents a sample study of a single char-
acter — the palatine process of the premaxilla — in the Passeres,
as a basis on which some of the problems of passerine anatomy
and classification may be explored.

The method of ‘‘single character study’’ is the analysis of all
aspects of the character essential to understanding its evolution
— this being the major goal of these studies. Although certain
specialized aspects must be investigated in some cases, as for ex-
ample, the embryology of the palatine process in this study, the
following steps must be included in every ‘‘single character
study.’’

a) A survey of the occurrence, structure and variation of
the character must be undertaken. In general, the scope of the
survey includes the next higher taxonomic category that con-
tains the group under consideration. For instance, if the affini-
ties of a passerine family are being studied, then the character
must be surveyed throughout the Passeres. The degree of varia-
tion should be ascertained in each taxonomic group down to the
species. All aspects of variation, e.g., sexual, age, geographical,
must be separated and clearly distinguished from one another.
Usually in studies of avian anatomy, it is not necessary to con-
sider infrageneric variation, since most anatomical characters
do not vary among congeneric species. This is especially true in
the Passeres.

b) The functional significance of the character, including the
meaning of its structural changes within the group, must be
established. This is the most important part of the analysis of a
taxonomic character and the one most often omitted or, if in-
cluded, covered only in a superficial way. Because of limitations
and technical difficulties, conclusions concerning the functions

364 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

are usually only deductions based on physical considerations of
the morphology exhibited by the character. It is only rarely pos-
sible to observe the bird alive and to deduct the function from
actual observations, or to conduct the necessary experiments to
prove it conclusively. However, although most of these conclu-
sions are only inferences, they are better than nothing and, with
practice, a worker can infer the function of a structure with
considerable accuracy. Two things must be remembered. First,
such results are usually only rough approximations; we cannot
hope, at this time, to determine the exact meaning of every
minor variation in anatomical features. Second, as deductions,
they are subject to error and hence the resulting conclusions re-
garding evolution and taxonomy are no better than the deduc-
tions on which they are based.

ce) Lastly, the evolution of the character must be investigated.
With a knowledge of its functional significance, one can estimate
the selection forces which were operative during the evolution of
the character. A knowledge of the selection forces is essential
because while it is possible to outline the phylogeny of a structure
without knowing the selection forces, it is impossible to under-
stand its evolution without them. And here the important thing
is the evolution, not the actual phylogeny, of the structure.

Once these aspects of a character are ascertained, it is possible
to judge its taxonomie value. In general, the taxonomic value
varies inversely with (a) the tightness of the control by the
selection forces acting on the character, and (b) the changeability
— independent origin, reversal of direction, ete. — of these selee-
tion forces. For example, if a structure is tightly bound to its
selection forces, and if these selection forces have altered their
direction frequently during the evolution of the group, then that
structure would have little taxonomic value. Statements such as
‘““the taxonomic value of a character depends upon how constant
that character is within the group’’ are misleading and inconclu-
sive. Lastly, I would like to suggest that the importance placed
on the taxonomic value of a character be de-emphasized and that
more stress be placed on studying its evolution. The former has
not produced any really concrete results while the latter holds
much promise for future studies of avian classification.

Before proceeding to the main part of the study, a word must
be said about the classification and linear sequence of the pas-
serine birds. The past lack of interest in the anatomy of the
perching birds has resulted in chaos. In recent years, a number
of conflicting classifications for the passerine families have been

BOCK: PALATINE PROCESS OF THE PREMAXIDLLA 365

proposed (Mayr and Amadon, 1951; Wetmore, 1951; Mayr and
Greenway, 1956; Wetmore, 1957; Amadon, 1957; Delacour and
Vaurie, 1957; and Mayr, 1958), yet these proposals and sugges-
tions represent little more than personal opinion — the necessary
information to verify the relationships suggested in these pro-
posals does not exist. The central problem of passerine classifica-
tion is the lack of factual evidence with which we can determine
the evolution of the Passeres and eventually establish the most
reasonable classification for them. Speculation on these problems
is premature at the present time and it seems probable that it
will be many years before enough information on the anatomy,
behavior and other attributes of the passerines has been gathered
to allow us to speculate on their phylogeny and relationships.
Until that time comes, it is most advantageous to have a standard
sequence of families which everyone knows and can use. For
the purposes of this paper, I shall adopt the sequence agreed
upon by the committee appointed by the XIth International
Ornithological Congress at Basel which is the one to be used in
the coming volumes of ‘‘Peters’ Check-list’’ (Mayr and Green-
way, 1956). This sequence covers only the Oscines. For the
suboscines, I shall follow the sequence suggested by Wetmore
(1951). I must emphasize that I do not believe that these
particular systems are correct or even satisfactory. Nor do my
findings support them better than the others. Nevertheless, it is
strongly urged that workers in passerine anatomy follow the
‘*Peters’’ sequence until enough evidence has been gathered to
establish a classification acceptable to most workers.

ACKNOWLEDGEMENTS

I am deeply indebted to Dr. Ernst Mayr for his advice and aid
throughout the course of this study and for reading the several
drafts of the manuscript which has been vastly improved by his
criticisms and suggestions. Dr. Ernest Williams and Professor
Bryan Patterson have read the manuscript and have offered
many helpful criticisms for which I am most grateful. Special
thanks are extended to Dr. Carl Gans for his comments on the
manuscript and especially for his valuable criticisms of the dis-
cussion on function. Dr. Dean Amadon of the American Museum
of Natural History, Dr. Herbert Friedmann of the United States
National Museum, and Mr. James C. Greenway of the Museum of
Comparative Zoology must be thanked for their help and coopera-
tion in making available collections in their care. I am indebted

366 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

to Mr. D. K. Wetherbee for a loan of several specimens of hatch-
ling cardinals and rose-breasted grosbeaks which proved to be
most valuable in the embryological analysis, to Dr. Lester Short
for his help in collecting young eardinals, to my wife, Kitty, for
preparing histological slides of the free palatine process and for
checking the manuscript for errors, and to my mother for typing
the final draft of the manuscript. I also wish to extend my
thanks and gratitude to the many other people who have aided
me in one way or another during the course of this study. This
study was completed while working under a National Science
Predoctoral Fellowship.

Figure 1. Ventral surface of the skull of a crow (Corvus). Note the
absence of the palatine process of the premaxilla (= fused to the pre-
palatine process of the palatine). The deep-lying bones are stippled for
contrast. The key for the abbreviations used in all figures can be found
on pages 487-488.

DESCRIPTION OF THE PALATINE PROCESS OF
THE PREMAXILLA

The palatine process is a posterior extension of the premaxilla
starting from the medioventral part of that bone. It lies along
the lateral edge of the palatine and fuses to a greater or lesser
degree with that bone. The palatine process has many features
that render it suitable for this study. It is a simple structure
which is easily observable and which exhibits several quite
dissimilar conditions in the Passeres. However, almost any other

BOCK: PALATINE PROCESS OF THR PREMAXILLA 367

anatomical feature that is not uniform throughout the order
would be equally suitable for a study of this type.

It may occur in any one of four conditions — fused, unfused,
free, or lateral flange.

Fused palatine process. In a typical passerine bird, such as
a crow (Corvus, Fig. 1), we find the simplest possible adult
condition of the palatine process of the premaxilla — namely,
that it is lacking as a distinct structure. The anterior bars of
the palatine (prepalatines or prepalatine processes) merge into
the premaxillary mass without the slightest indication of a break.
There are no sutures or processes at the junction of the prepala-
tine process and the premaxilla to reveal the presence of a
palatine process of the premaxilla. The palatine process has fused
completely with the prepalatine process, as will be shown later in
the section on development. On the lateral side of the skull, the
premaxilla merges with the maxilla, which in turn continues
into the jugal bar.t The maxillo-palatines (not to be confused
with the ‘‘palato-maxillaries’’) originate from the maxillae and
pass medially beneath the palatines to approach one another in
the region of the anterior end of the vomer. The distal ends of the
maxillo-palatines expand to form flat plates; these plates partly
cover the tip of the vomer when the ventral aspect of the palate
is examined. Returning to the palatines, these bones run _ pos-
teriorly and then expand medially to approach the midline. The
palatine shelf’ and the posterior extension of the palatines (the
transpalatine process) serve as the point of origin for a large
part of the M. pterygoideus (at least for the lateral parts of this
muscle). The medial parts of the palatine (the interpalatine
process, anteriorly, and the mediopalatine process, posteriorly),

1In his recent paper on the development of the chick skull, Jollie (1957)
suggests that the names for a number of bones in the skull be changed to agree
with their embryological origins and homologues in the reptilian skull. Thus,
for example, the palatine would become the pterygopalatine and the pterygoid
would become the posteropterygoid. These new names are certainly correct tech-
nically, but the change to them would not lead to greater clarity. The technically
correct names are only necessary for comparisons of the avian skull with the
skull of other classes of vertebrates; however, only a very few workers are
interested in this problem. The terminology used currently for the parts of the
avian skull was developed specifically for the adult skull, and in many cases the
term refers to a functional region or unit rather than to an individual bone.
The present system of names is perfectly suitable for studies in which the skull
is compared within birds. Consequently. it is recommended that the standard
terminology for the parts of the avian skull be retained. I do not, however, want
to convey the impression that the embryological origin of the bones of the skull
is unimportant. These studies are very important and indeed, many more
studies similar to Jollie’s investigation of the chick skull are needed.

2The medial projection of the palatine appears to be unnamed although all
of the other parts of this bone have been given special names. Because of the
complex structure of the M. pterygoideus which originates from the palatine bone,
it would be helpful if this projection also had a specific name. The most appropri-
ate name is the medial shelf of the palatine or more simply, the palatine shelf.

368 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

the vomer and the pterygoid are of no interest to us as they are
far removed from the palatine process of the premaxilla and
from the muscles which may originate from it.

The fused condition of the palatine process of the premaxilla
is typical for many families of passerine birds. There may be
considerable variation in the relative lengths of the different
processes of the palatine, but although this is of considerable
importance in studies on the kinetics and functional significance
of the avian skull, it is of no concern to us in this study.

Figure 2. Ventral surface of the skull of a white-throated sparrow
(Zonotrichia). The palatine process of the premaxilla lies along the pre-
palatine process of the palatine and is separated from it by a distinct

suture.

Unfused palatine process. This condition of the process (called
the ‘‘palato-maxillaries’’ by Parker, 1877, and more recently
by Tordoff, (1954a, see page 374) is present in the adult stage of
many genera, such as the white throated sparrow (Zonotrichia,
Fig. 2), as a small posterior extension of the premaxilla which
lies along the lateral edge of the prepalatine process of the
palatine. The palatine and other bones of the skull in the white
throated sparrow are similar to those of the crow and need not
be deseribed again.

There is considerable variation in the length of the palatine
process and in the degree of fusion between it and the palatine
in the genera possessing an unfused palatine process. Some of

BOCK: PALATINE PROCESS OF THE PREMAXILLA 369

this variation is a result of a difference in the age of the speci-
mens and hence in the degree of ossification of the skull; this
feature of age variation will be discussed later. In some genera,
the anterior end of the palatine process degenerates, thereby de-
stroying the connection between it and the main body of the
premaxilla; the final result is an isolated splint of bone lying
along the lateral edge of the prepalatine process. This isolated
splint of bone may appear as if it were a new bone arising from
a distinct center of ossification, but it is actually nothing more
than the posterior end of the palatine process detached from

Figure 3. Ventral surface of the skull of a cardinal (Cardinalis). The
palatine process of the premaxilla is free of the prepalatine process of
the palatine and lies free in the space between the palate and the jugal bar.

the rest of the premaxilla; again, a full discussion of the develop-
ment of this variant will be presented below in the section on
development.

The fused condition of the palatine process may be combined
with the unfused under the heading of the ‘‘normal palatine
process,’’ as found in most passerine birds.

Free palatine process. The palatine process in some groups
of finches, such as the cardinal (Cardinalis, Fig. 3), is free of the
palatine bone and les in the space between the palate and the
jugal bar. The free palatine process originates at the junction
between the palatine bone and the body of the premavxilla.
In some genera, there is a ‘‘suture’’ at the base of the free
palatine process separating it from the rest of the premaxilla;

370 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

in others the palatine process continues into the rest of the
premaxilla without a break (compare Figs. 31B and 31C; see
also Tordoff, 1954a). The palatine and other bones of the skull
are similar to those of the crow and do not require a separate
description.

Lateral flange. The palatine process is lacking as a distinct
structure in several groups of finches, such as the evening gros-
beak (Hesperiphona, Fig. 4). However, in contrast to the last
three types, there is a lateral flange on the anterior end of the
palatine which extends almost to the jugal bar. This lateral

Figure 4. Ventral surface of the skull of an evening grosbeak (Hesperi-
phona). The palatine process of the premaxilla is absent (= fused to the
prepalatine process of the palatine). A lateral flange is present at the site
of the fused palatine process.

flange is the bony boss referred to below in the section on func-
tion and elsewhere in this paper. The lateral flange of the
palatine is usually fused to the premaxillary mass, but it is
sometimes separated from that bone by a suture. The palatine
and other bones of the skull are similar to those described for
the crow except that they are stouter and the transpalatine
process is divided into two subprocesses. There is no evidence
of a strengthening of ‘‘twisted’’ prepalatine bars such as de-
seribed by Tordoff (1954a, p. 18).

BOCK: PALATINE PROCESS OF THE PREMAXILLA SHA

HISTORY

Many of the current problems in understanding the palatine
process of the premaxilla have a historical basis and thus ean be
fully appreciated only after one knows the history of the studies
on this structure. The most important of these problems concerns
the ‘‘distinction’’ between the palatine process and the ‘‘palato-
maxillary’’ of Parker and of Tordoff ; these terms actually refer
to the same structure as will be shown below (see p. 381).

Study of the palatine process in birds began in the 1860’s
with the work of W. K. Parker. No other student of avian
anatomy mentioned the process prior to the late 1880’s. Parker
clearly described and figured the palatine process of the pre-
maxilla in all of his works including those on the palate of the
‘‘aegithognathous birds’’ (1875c, 1877). But, for some inexplic-
able reason, he stated in the description of Tanagra cyanoptera
(1877, pp. 252-253) that: ‘‘the praemaxillary mass is. . .; the
palatine processes are aborted (d., px., ppz.).

‘“Where the latter processes existed in the embryo, a falcate
spicule of bone appears, a separate ‘palato-maxillary (p. mz.).’
This is a character to be found in several families of the Cora-
comorphae, as I shall soon show. Its presence suggests some
delicate bond of affinity between the families where it is found.’’
Parker then described a ‘‘ palato-maxillary’’ instead of a palatine
process of the premaxillary in the members of the New World
nine-primaried oscines. The most puzzling aspect of the ‘‘ palato-
maxillary’’ is that it appears to be identical to the palatine
process of the premaxilla found in other passerine families when
the two structures are compared in the adult. Yet Parker never
stated how one distinguished between the two bones in the adult
passerine bird. Nor did he present in this paper (1877) or in
any other, the evidence supporting his belief that the palatine
process aborts in the embryo of the New World nine-primaried
oscines and that a separate center of ossification — the ‘‘palato-
maxillary’’ — develops to take its place. Although Parker had
studied the development of the palate in many species of passerine
birds, he never investigated fully the embryology of the palate
in any member of the nine-primaried complex. His only mention
of the development of the ‘‘ palato-maxillary”’ is a description of
one stage in the development of the skull of a cardinal. In this
description, Parker said only that the ‘‘ palato-maxillaries’’ grow
in the space between the palate and the jugal bar as additional
wedges (Fig. 8A). However, he did not give the age of this

372 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

specimen, nor did he have a series of specimens of different ages ;
hence there is no direct evidence of the palatine process aborting
and a separate ‘‘palato-maxillary’’ taking its place. Parker did
offer a very important suggestion on a possible origin of the
‘‘pnalato-maxillaries’’ in a footnote (1877, p. 263), although he
did not follow it up: ‘‘The rapid development and early anky-
losis of the bony centres in birds makes the study of their osteol-
ogy very difficult; also the breaking off of a projection of a
primary centre to make a new bone, as in the mesopterygoid. I
am in some doubt whether this lateral piece of the tetramerous
vomer of the type now being described is not formed in this way.
Perhaps, also, in some eases, the distinct ‘palato-maxillaries’ may
be the palatine process of the praemaxillary detached; I have,
however, no proof of this; and that process is very apt to become
absorbed when no palato-maxillary appears. It is sure to be
removed if a new centre came in behind it to take its place.’’
The evidence and reasoning presented here by Parker is as strong
an argument for the ‘‘palato-maxillary’’ being the same as the
palatine process as it is for the two bones being different strue-
tures. Thus it can only be concluded that Parker did not have
any good evidence supporting his belief that the palatine process
of the premaxilla aborts in the New World nine-primaried
oscines and that a separate ‘‘palato-maxillary’’ takes its place.
It is difficult to understand how a worker of Parker’s caliber
could describe a separate bone on such flimsy evidence until one
realizes that he was the first worker to describe the minute
processes found in the passerine skull. In his work on the
development of the palate, he had only the erudest technical aids,
especially stains, and could easily be misled by a poorly preserved
specimen in which the posterior tip of the palatine process had
broken off and resembled a separate center of ossification. The
remarkable thing is that Parker was able to describe as much as
he did with primitive methods and equipment.

At this point, Parker’s work on the skull of the woodpeckers
(1875a) should be mentioned because he described a separate
‘“palato-maxillary’’ in this group, this being the first description
of the structure. According to Parker, the palatine process of
the premaxilla in the woodpeckers lies on the inside of the pala-
tine and becomes fused to the medial side of that bone, not to
the lateral side as in most birds. In some (all?) species of wood-
peckers, there is a separate spicule of bone lying along the
lateral edge of the prepalatine process, which Parker called

BOCK: PALATINE PROCESS OF THE PREMAXILLA 373

the ‘‘palato-maxillary.’’ Thus if Parker’s observations are cor-
rect (I have not been able to check them), there is a separate
‘‘nalato-maxillary’’ in the woodpeckers and the term ‘‘palato-
maxillary’’ should be used only for this structure.

Curiously enough, later workers used only the term ‘‘palato-
maxillary’? even when discussing the non-New World nine-
primaried passerines, and extended its meaning until it became
almost synonymous with the palatine process of the premaxilla.
The reason for the initial confusion is obscure, but the results
are clear enough — today it is impossible to determine what is
meant when the term ‘‘palato-maxillary’’ is used.

In the years following Parker’s work, the palates of a number
of passerine species were described by various workers (Garrod,
1872, 1877; Forbes, 1880, 1881, 1882; and Pyeraft, 1905a, 1905b,
1905e, 1907). Unfortunately, there is no indication whether these
workers knew Parker’s papers so that we can never be certain
if the palatine process of the premaxilla was truly absent in
the adult of the species described when an author failed to men-
tion or to figure it; often the palatine process was overlooked if
present, or otherwise omitted from discussion.

A number of workers did, however, describe the palatine
process under several different names. Thus, Shufeldt (1888),
in describing the osteology of Pheucticus melanocephalus, the
black-headed grosbeak, stated (p. 489): ‘‘. . . the palatines on
either side develop a secondary palatine process (sp. p., Fig. 1),
extending backwards from a point to the outer side of where
the anterior palatine limb fuses with the premaxillary.’’ Later
in the same paper (p. 441), he described the secondary palatine
process in Piranga and claimed that the possession of a secondary
palatine process by these birds (a tanager and a eardinaline
finch) indicated an affinity between them. Apparently, Shufeldt
had not seen Parker’s paper on the palate of ‘‘aegithognathous
birds’’ because his secondary palatine process is the same as
Parker’s ‘‘palato-maxillary.’’ Neveretheless, the two authors
agree as to the taxonomic value of this structure.

Lueas, in a series of papers (1888 to 1895), reported on the
osteology of many groups of American Passeres. He did not men-
tion the palatine process in his studies on the thrushes, the
thrashers and the wrens, families in which the palatine process
is usually lacking in the adult. We can be certain that Lucas had
read Parker’s papers for he described the palatine process (under
the name ‘‘palato-maxillary’’) in some members of the New
World nine-primaried oscines. He was, however, doubtful of its

ce

374 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

embryological origin for he stated in his study on the osteology
of the swallow-tanager (Tersina) (1895, pp. 505-506) that:
‘‘There is a stout palato-maxillary process, whether or not de-
veloped from a separate center is not known.’’ In addition,
Lucas questioned its taxonomic value and stated (1894, p. 304) :
‘‘Tts exact [taxonomic] value remains to be shown, for it appears
in forms which are not related, at least closely and drops out in
some that are nearly allied. It is present in the Swallows, but
not in the Flyeatchers or Thrushes, is well developed in such
stout-billed Finches as Cardinalis and Habia, missing in Cocco-
thraustes. It appears as a slender splint in Plectrophanes and
Calcarius, while it is lacking in Phoenicophilus. None of the
Drepanididae and Meliphagidae examined have a palato-maxil-
lary.’’ These questions posed by Lucas on the embryological
origin and on the taxonomic value of the palatine process of the
premaxilla, or the ‘‘palato-maxillary’’ as he called it, are most
pertinent and have remained unanswered to the present day.

In the years between 1900 and 1950, several workers described
the palatine process in a number of passerine families (Clark,
1912, 1913a, 1913b, 1913¢; Lowe, 1924, 1931, 1938a, 1938b, 1947,
1949; Stonor, 1942; Sushkin, 1924, 1925, 1927, 1929), but no
further contributions were made regarding its origin or taxo-
nomic significance. Amadon, in his monograph on the Drepa-
niidae (1950a), included a long discussion on the ‘‘palato-
maxillaries’’ (pp. 213-216). He stated that they are absent in the
Drepaniidae, but suggested (p. 216) that the flange on the
lateral side of the prepalatine bar may represent the fused
‘‘nalato-maxillary.’’ However, because of the scope of his paper,
Amadon was forced to leave many questions unanswered and
concluded (p. 216) that: ‘‘Little is known of the significance
of the palato-maxillaries.’’

Tordoff’s studies (1954a, 1954b) on the relationships of the
‘‘Fringillidae’’ and the New World nine-primaried oscines are
based almost entirely on the structure and variation of the
‘‘yalato-maxillaries’’ in these families. This work has been, up
to the present, the most extensive study of the ‘‘palato-maxil-
laries’’ in any group of passerine birds and the only one that
bases important taxonomic conclusions on them. Unfortunately,
Tordoff did not examine families outside of the nine-primaried
complex and the ploceids for the presence of ‘‘palato-maxil-
laries,’’ nor did he examine the embryology of this structure.
He was apparently unaware that Parker had described a very
similar structure under the name ‘‘palatine process of the

BOCK: PALATINE PROCESS OF THE PREMAXILLA 375
premaxilla’’ in other passerine families and had even suggested
that the two bones might be the same. In addition, Tordoff’s con-
clusions of the functional significance of the ‘‘ palato-maxillaries’’
are decidedly different from those arrived at in this paper. Due
to the evidence presented in this study, I am unable to accept
Tordoff’s paper.

Recently, Jollie (1958, pp. 27-28) investigated the develop-
ment of the ‘‘palato-maxillaries’’ in connection with his studies
on the embryology of the avian skull. He showed that the
‘‘yalato-maxillary’’ is the remnant (posterior part) of the pala-
tine process of the premaxilla and not a separate center of ossifi-
eation. This is a most important contribution to the clarification
of the origin of the ‘‘palato-maxillary.’’ Unfortunately, Jolhe
neglected to include a clear statement as to whether the term
‘‘yalato-maxillary’’ is or is not synonymous with the palatine
process of the premaxilla.

From this brief history of the past studies on the palatine
process of the premaxilla (or the ‘‘palato-maxillary’’ as it is
usually, but erroneously called), it can be seen that the available
information is very limited in spite of the fact that it was
deseribed many years ago and has been studied by many work-
ers. Therefore, it will be necessary to investigate all facets of
the palatine process before its value in showing relationships
within the Passeres ean be ascertained.

DEVELOPMENT OF THE PALATINE PROCESS
OF THE PREMAXILLA

The development of the palatine process must be known before
several questions on its nature and identity with the ‘‘palato-
maxillary’’ can be answered. Unfortunately, there are too few
studies on the development of the skull in passerine birds and
even fewer which include the development of the minor
processes of the upper jaw and of the palate. The present dis-
cussion, consequently, rests almost entirely on the old but
excellent studies by Parker (1872; 1873a, 1873b, and 1875b) and
on the recent work by Jollie (1957 and 1958). The original con-
tributions of the present study are meager and include only the
development of the palatine process in the cardinal (Cardinalis),
and observations on the ossification of the skull in such post-
fledgling birds as are available in collections.

The following questions should be kept in mind while reading
the descriptions of the development of the palatine process:

376 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

a) Is the palatine process present in the embryo of all pas-
serine birds including those which do not exhibit a distinct pala-
tine process as adults?

b) Is the ‘‘palato-maxillary’
tion ?

c) Is the free palatine process as seen in the adult cardinal
homologous with the palatine process of other passerine birds, or
is it an ossified tendon ?

?

a separate center of ossifica-

Figure 5. Series showing the development of the palatine process of
the premaxilla in the crow (Corvus). Except for figure A which shows
both halves of the skull, the figures illustrate the ventral surface of the
left half of the skull. The ages of the specimens are: (A) Sixth day of
incubation; (B) Ninth day of incubation; (C) Hatchling; (D) Week-old
hatchling; and (E) Fledgling. The figures are redrawn from Parker
(1872).

Fused palatine process. A few specimens with an unfused
or a partly fused palatine process can be found in almost every
large series of birds normally having the palatine process of the
premaxilla completely fused with the prepalatine process in the
adult (e.g., Cyanocitta cristata, Fig. 28F). These specimens
usually show signs of immaturity, such as unossified ‘‘ parietal
windows.’’ This would suggest that the palatine process is pres-
ent in the young bird and becomes increasingly fused with the
prepalatine process until the two bones are completely fused in
the adult.

The typical course of development of the palatine process in
the Passeres can be seen in the crow (Corvus). The following
account and figures have been taken from Parker’s description
of the development of the skull of the crow (1872), which is still

4
=

BOCK: PALATINE PROCESS OF THE PREMAXILLA 3

the most complete one available for any species of passerine
birds.

The premaxilla of the crow appears at about the sixth day of
incubation in the form of three separate nodules of cells (Fig.
5A), the center nodule corresponding to the nasal process of
the premaxilla while the lateral nodules correspond to the two
halves of the main body of the bone. Neither the palatine
process of the premaxilla nor the palatine have appeared by this
time. By two or three days later (Fig. 5B), the nodules have
enlarged and fused together to form a recognizable premaxilla.
The dentary processes of the premaxilla have appeared by this
time and run backwards to meet the maxillae on either side. The
palatines have also appeared and are quite well developed,
although the palatine processes of the premaxilla have not yet
made their appearance. Parker’s next stage (Fig. 5C) is a hatch-
ling bird. The palatine processes have appeared and are small
projections on the medial side of the premaxilla. They overlie
the palatines. By the time the hatchling is a week old (Fig. 5D),
the palatine process has enlarged to cover the lateral half of the
prepalatine process. Up to the time of fledging, the palatine
process continues to grow and to remain distinct from the pre-
palatine process (Fig. 5E). From the time of fledging or shortly
thereafter, the palatine process of the premaxilla starts to fuse
with the prepalatine process of the palatine until the two bones
are completely fused together. There was no sign of a palatine
process of the premaxilla in any of the adult crow skulls that I
examined.

Among other birds possessing a fused palatine process, in-
formation on its development is available for the titmouse (Fig.
6A), the thrushes (Figs. 6B, 6C, and 6D), and the house sparrow
(Fig. 7A). These species agree with the crow in possessing a
distinct palatine process of the premaxilla in the embryo which
becomes fused with the prepalatine process during development.
The presence of a palatine process in the house sparrow (a
ploceid finch) is of interest since Tordoff claimed that this group
lacked ‘‘palato-maxillaries.’’

Unfused palatine process. The unfused condition of the pala-
tine process, or the isolated splint lying along the prepalatine
process, as seen in some of the emberizine finches, is the typical
‘‘nalato-maxillary’’ of Parker and of Tordoff. According to
Parker, the palatine process of the premaxilla aborts in the
New World nine-primaried oscines and a separate center of

278 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

ossification — the ‘‘palato-maxillary’’—takes its place. How-
ever, Jollie’s description of the junco, an emberizine finch (Fig.
7D) shows that the development of the palatine process in this
species is identical to that described for the crow except that the
fusion between the palatine process and the prepalatine process
does not go to completion. I have examined a fledgling towhee
(Pipilo) which has a perfectly normal development of the pala-
tine process similar in all respects to that seen in the junco.
Thus there is no evidence supporting Parker’s hypothesis that a
separate center of ossification takes the place of the aborted
palatine process.

Figure 6. Development of the palatine process of the premaxilla in the
titmouse (Parus) and the thrush (Turdus). Figure A shows the palatine
process in a titmouse at about the tenth day of incubation; redrawn from
Parker (1873a). Figures B, C, and D illustrate the palatine process in a
prehatchling Turdus viscivorus, a day-old T. merula, and a week-old T.
merula respectively; redrawn from Parker (1873b).

The isolated splint lying along the prepalatine, which some
workers might consider to be the true ‘‘palato-maxillary,’’ de-
velops by the degeneration of the anterior end of the palatine
process which thereby destroys the connection between the rest
of the palatine process and the main body of the premaxilla.
Jollie illustrates the development of this splint in the junco and
I have seen good series of this change in Formicarius (Figs.
23D, 23E), Spizizos (Figs. 24D, 24E, 24F), Melospiza (Figs.
25G, 25H, 251) and Paradisaea (Figs. 28G, 28H, 281). These
observations substantiate Parker’s hypothesis that the ‘‘palato-
maxillary’? may be the posterior part of the palatine process

BOCK: PALATINE PROCESS OF THE PREMAXILLA 379

detached from the rest of the premaxillary, and hence not a
separate bone.

Free palatine process. The palatine process of the cardinal is,
in many respects reminiscent of an ossified tendon. The M.
pterygoideus ventralis lateralis originates from this process by
means of a tendon, and consequently, it is possible that the entire
free process seen in the cardinal could be an ossified tendon
which originates from the main body of the premaxilla. My
suspicions of this possibility were increased by the presence in

Figure 7. The palatine process of the premaxilla in: (A) A nestling house
sparrow (Passer, redrawn from Jollie, 1958); (B) A five-day old embryo
linnet (Carduelis, redrawn from Parker, 1875b); (C) A nestling house
finch (Carpodacus, redrawn from Jollie, 1958): and (D) A fledgling junco
(Junco, redrawn from Jollie, 1958).

one specimen of a faint longitudinal suture on the lateral half of
the prepalatine process. This could be the suture between the
semifused palatine process and the palatine if the free process
seen in the adult cardinal was not the true palatine process of the
premaxilla. Histological sections were prepared of the free
process in the hope of ascertaining its identity. No difference
could be detected between the bone of the free process and that
of the premaxilla, but this result is inconclusive. Ossified tendon
and bone are almost identical, if not identical, histologically.
Therefore the only means of solving this problem was to study
the development of the palatine process in the cardinal. Unfor-
tunately, Parker did not give sufficient detail in his treatment on
the embryology of the palate in the cardinal (see Fig. 8A) so that
a series of cardinals ranging in age from hatchling to post-
fledgling were gathered and stained to show the details in the

380 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

development of the palatine process. These specimens (Figs.
8B, 8C, 8D) prove that the free process in the cardinal is the
true palatine process of the premaxilla and not an ossified ten-
don. It is possible that the tendon attaching to the free process
has ossified for a short distance starting at its origin and thus
has elongated the process, but this would be exceedingly difficult
to verify. However, even if the free process was enlarged through
ossification of the attached tendon, it would still be the palatine
process. The position of the palatine process in a hatchling rose-
breasted grosbeak (Pheucticus) also indicates that the free proc-
ess in this species is the true palatine process of the premaxilla.

PPM

Figure 8. Development of the palatine process of the premaxilla in the
cardinal (Cardinalis). Figure A is a bird of unknown age redrawn from
Parker, 1875b. The series B, C, and D are drawn from specimens of a
hatchling cardinal, a fledgling cardinal, and a post-fledgling, half-grown
cardinal, respectively.

Lateral flange. Those birds, such as the cardueline finches,
which possess a lateral flange at the anterior end of the palatine,
also lack a palatine process in the adult. Tordoff stated that the
Carduelinae do not have a ‘‘palato-maxillary’’ (with the tacit
assumption that it is also absent in the embryo) and are therefore
related to the ploceid finches. However, Parker (1875b) shows a
very distinct palatine process in the early embryo (five days)
of the linnet (Fig. 7B) and Jollie (1958, p. 29) shows an
equally distinct process in the house finch (Fig. 7C). Henee, the
palatine process of the premaxilla is present in the embryo of
the ecardueline finches and becomes fused to the palatine during
development. Ossification of the lateral flange starts at the pala-
tine process as can be seen in Jollie’s figure of the house finch
Ghia (©)

BOCK: PALATINE PROCESS OF THE PREMAXILLA 381

Conclusion. The palatine process of the premaxilla is present
in the immature of all studied species of passerine birds and it
is probably present in the immature of all passerines (see also,
Jollie, 1958, p. 27, who concludes that the palatine process is
probably present in the immature of all birds). It is most prob-
able that the palatine process was overlooked in those studies
(e.g., Huggins, et al., 1942) in which it is not mentioned. In
most passerines, the palatine process becomes indistinguishably
fused with the prepalatine process of the palatine during post-
hatching development. There is no indication of the palatine
process aborting and a separate center of ossification taking its
place in the New World nine-primaried oscines. Therefore, all
of the structures in the passerine birds which have been called
the ‘‘palato-maxillary’’ or the ‘‘secondary palatine process’’
are the same as the palatine process of the premaxilla; that is,
these terms are synonymous. None of these structures, e.g., the
free process in the cardinals, are non-homologous structures
which have been misidentified as the palatine process. Lastly, it
is best not to give the isolated splint lying along the prepalatine
process a separate name. This procedure implies that the splint
developed as a separate bone while it is nothing more than the
posterior part of the palatine process detached from the main
mass of the premaxilla.

FUNCTION OF THE PALATINE PROCESS
OF THE PREMAXILLA

Analysis of the functional significance of most morphological
systems is, by necessity, based on deductive reasoning. The
functional conclusions are only hypotheses and must be treated
as such. Only after these hypotheses have been tested by exten-
Sive experiments, can they be relied upon and, even then, there
is a chance that some important factor has been overlooked. The
deductive method of functional anatomy is based partly on a
consideration of the laws of mechanics and partly on a consider-
ation of the relative development of the structure in forms hav-
ing different habits. For instance, if the shape and mass of
certain jaw muscles differ between seed-eating and insect-eating
birds, then the basic assumption would be that this difference is
somehow associated with feeding habits. The details of the partic-
ular functions are, then, worked out using the principles of
mechanics. This simple method has enabled functional anatom-
ists to analyze highly complex systems, even though their

382 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

results are largely hypothetical. It is hoped that experimental
workers will test the conclusions of functional anatomy and
determine which of their conclusions and working hypotheses
are correct. Such work may be most difficult from a technical
standpoint, but the results would be invaluable to students of
anatomy and evolution.

The palatine process of the premaxilla has three major func-
tions which are partly independent of one another. One function
is found in all passerine birds and is complementary to the two
others which appear to be mutually exclusive. The first function
is the insurance of a firm connection between the palate and the
upper jaw, while the two mutually exclusive ones are: a point of
origin for part of the M. pterygoideus, and a bony boss against
which seeds are crushed. The first function may be considered
to be the primary function of the palatine process and the others
to be secondary ones. This division of functions into primary
and secondary ones is not to be confused with original and suc-
cessive functions; it is a division according to relative importance,
not according to the time of appearance. A function may be
defined as primary if it is the most important or the most basic
funetion of the structure. It is present in all species possessing
the structure and thus can be considered as the function respon-
sible for the maintenance or the preservation of the structure.
Secondary functions are subservient to the primary function in
that their action must be in harmony with the action of the
primary function. Usually secondary functions are not found
in all species possessing the structure. An example of primary
and secondary functions may be found in the avian wing. Active
flight is generally the primary function of the wing, while dis-
play, defense, underwater swimming and so forth are secondary
functions. So long as a bird must be able to fly, these secondary
functions are subservient to the primary function of flight. Al-
though the primary function is responsible for maintaining a
structure, it is not necessarily responsible for the origin of that
structure. A former secondary function could have become the
primary function in the course of evolution and thus become
responsible for the preservation of the structure. The original
primary function would then become a secondary function or
drop out entirely. This is the well-known phenomenon of pre-
adaptation or functional change (Functionwechsel of Dohrn, see
Bock, 1959). In the example of the avian wing, active flight is
currently the primary function, but it is not the original function
responsible for the origin of the wing. The original function was

BOCK: PALATINE PROCESS OF THE PREMAXILLA 383

probably gliding which was replaced by active flight when the
fore limb became sufficiently developed as a wing to acquire this
new function. Similarly, underwater swimming was once a
secondary function of the wing in the ancestral penguins, as it is
in the auks and the diving-petrels, but became the primary fune-
tions when penguins no longer needed to fly.

The following analysis will be divided into two parts. The
first will deal with the function responsible for maintaining the
palatine process in birds, while the second will deal with the
functions responsible for the modifications of the process during
the evolution of the Passeres. Throughout the discussion, I
will switch from the function to the selection force associated
with that function and vice versa. In general, there is a major
selection force for each function and that selection force can be
described in the same terms as the function. Thus, if the function
of a bony process is that of a brace to support the bone, then
the selection force is for a brace to support the bone. For those
who are not used to switching from function to selection force,
the best way to keep the two separate is to think of the function
as the static phenomenon and the selection force as the dynamic
phenomenon.

Before proceeding to the discussion of the function of the
palatine process, it is necessary to establish the limits of this
study. The palatine process of the premaxilla is part of the
extensive character complex of bones, muscles, hgaments and
other structures that make up the jaw mechanism. A character
complex may be defined as that group of characters that acts
together as a single functional unit. A structure may belong to
several character complexes, and a large complex, such as the
jaw mechanism, may be divided into a number of smaller com-
ponent complexes. Whenever possible, the entire character com-
plex, not the individual component characters, should be the unit
of study. A complete study of the jaw mechanism in the Pas-
seres is most desirable and must eventually be done in order
to understand the passerine feeding modifications and the rela-
tionships between groups of passerine birds (e.g., the develop-
ment of the seed-eracking bill versus the relationships between
the various groups of finches), but I do not have the knowledge
to undertake such a study at this time. In this paper, I have
restricted myself to the function of the palatine process of the
premaxilla, but have included the function of such other struc-
tures as seemed pertinent to the problem.

384 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Maintenance of the palatine process. The kinetic skull of birds
with its movable upper jaw necessitates a firm connection be-
tween the palate and the upper jaw. The strength of this con-
nection is increased by the palatine process which provides a
larger surface to which the palatine can fuse. Hence, it is
postulated that the primary function of the palatine process is to
insure a firm connection between the palate and the upper jaw,
and that the selection force associated with this function would
be for a stronger connection between the palate and the upper
jaw. A discussion of the mechanics of the skull can provide some

Figure 9. Diagrammatic drawing showing the mechanics of the kinetic
skull in birds. When the quadrate rocks forward, it pushes on the jugal
bar and the palate which in turn push on the base of the upper jaw.
Because it is attached to the brainease at the nasal-frontal hinge, the upper
jaw rotates upward. When the quadrate rocks backwards, the upper jaw
rotates downward. Redrawn from Engels, 1940.

evidence supporting this hypothesis. I will only outline the
salient features of this mechanism and refer the interested reader
to Beecher’s excellent discussion of the mechanics involved in
elevating and depressing the upper jaw in birds (1951la, pp. 412-
416).

The upper jaw of birds is not solidly fused to the brainease as
in mammals and in some reptiles, but can be raised and lowered
by means of a complex mechanism of bones (Fig. 9). Rotation
of the upper jaw is about the nasal-frontal hinge — the connec-
tion between the upper jaw and the braincase. At its ventro-pos-
terior end, the upper jaw is attached to the jugal bars laterally
and to the palate medially. These elements connect the upper

BOCK: PALATINE PROCESS OF THE PREMAXILLA 385

jaw to the quadrate. All parts of this system except for the
nasal-frontal hinge are free of the brainease and can move rela-
tive to it. Thus, as the quadrate rocks forward, it pushes the
base of the upper jaw forward. The upper jaw, being attached
to the brainease at the nasal-frontal hinge, rotates upward (Fig.
10). When the quadrate rocks backwards, it pulls the base of
the upper jaw backwards and thus depresses the upper jaw.
Because the muscles operating this system insert on the quadrate
and the pterygoid, their force must be transmitted to the upper
jaw by means of the palate and the jugal bars. The push that
raises the upper jaw is probably transmitted to it only through
the palate because the thin jugal bars would bend if a push was
exerted on them. The pull could be transmitted through the
palate and the jugal bars; however, it seems likely that most of
the pull is along the palate. Hence, in addition to other factors,
the proper functioning of this kinetic system is dependent upon a
strong connection between the palatines and the premaxilla.
At least two important functions are achieved by the kinetic
skull of birds. First, it permits a wider gape than a stationary
upper jaw; this feature is desirable in such birds as the swallows
and the flycatchers, which need a wide gape. Second, it preserves
the primary orientation of the skull (see Moller, 1931, p. 146;
Beecher, 195la, pp. 414-415) by allowing the bird to open its
bill without shifting the position of the eye with reference to the
prey or ‘‘leading’’ the prey (Fig. 10). If the axis of the skull
shifted when the bird opened its bill to capture its prey, the
entire orientation of the head and neck in respect to the prey
would be destroyed. The bird would have to re-orient completely
in the brief instant between bill-opening and prey-capture. De-
velopment of the elaborate nervous mechanism needed for this
rapid re-orientation would be difficult. It would be far easier to
preserve the orientation of the skull by mechanical means, e.¢.,
a kinetie skull. Evolution in birds has followed the latter course.
The importance of these functions is indicated by the fact that
almost all birds possess a kinetic skull. Hence there would be
a strong selection force favoring all parts of the kinetic skull,
including a firm connection between the upper jaw and the
palate. It was postulated above that the palatine process of the
premaxilla serves to increase the contact and presumably the
degree of fusion between the premaxilla and the prepalatine
process of the palatine; thus the palatine process would be
favored by the selection force for the kinetic skull. Therefore,
it can be concluded that the selection forces favoring the kinetic

386 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

A

NFH
Dp es Kee . ee A

NFH

ee ee eee eee ree.|

BS fos mis ci en En IO S|

Figure 10. Diagrammatic drawings of the avian skull illustrating how
the kinetic upper jaw preserves the primary axis of the skull, i.e., the
position of the eyes in respect to the prey. Line A-A represents the primary
axis of the skull; it lies along the gonys (the junction between the upper and
lower jaws) as shown in figure A. The primary axis remains stationary
and midway between the jaws when the bill opens (Figure B) and closes
on the prey (Figure C). Line B-B represents the secondary axis of the
skull when the bill opens if the upper jaw is not movable. Figures redrawn
from Moller, 1931.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 387

skull are responsible for maintaining the palatine process of the
premaxilla in birds.

It should not be assumed that the palatine process of the
premaxilla appeared as a new structure in the passerine birds.
The palatine process had doubtless originated at the time birds
evolved from reptiles, if not before. The palatine in most rep-
tiles abuts against the maxilla and the vomer anteriorly and
against the pterygoid posteriorly. In birds, the anterior connec-
tion of the palatine is with the premaxilla and perhaps with
the maxilla by means of secondary ossification. The connection
with the vomer is medial and more posterior than in the reptiles
while the connection with the pterygoid is posterior as usual.
The shift of the palatine from the maxilla to the premaxilla
probably required the development of a point of abutment or
anchorage on the premaxilla. This is the palatine process of the
premaxilla. It is not known whether the shift of the palatine
was associated with the development of the kinetic skull in birds
because the palates of neither the pseudosuchians nor Archaeop-
teryx are known. The kinetie skull evolved sometime after the
Archaeopteryx-stage in the evolution of birds. Nevertheless, once
birds possessed a kinetic skull and a palatine process of the pre-
maxilla, the palatine process was preserved because of the selec-
tion forces associated with the kinetie skull. It was thus avail-
able (preadapted) for other selection forces which arose during
the subsequent evolution of the passerine birds.

Modifications of the palatine process. Modifications in the strue-
ture of the palatine process of the premaxilla in the Passeres
have developed under the control of the several secondary
functions of this structure. These will be discussed with three
problems in mind. (a) Changes in the palatine process associ-
ated with modifications in the M. pterygoideus. These changes
arose in connection with the development of a free palatine
process, such as is found in the cardinal. (b) Development of a
bony boss at the anterior end of the palatine. This is associated
with the development of the lateral flange on the anterior end
of the palatine in the cardueline and other finches. (c) The
variation in the degree of fusion between the palatine process of
the premaxilla and the palatine, and the variation in the de-
velopment of the isolated splint lying along the palatines as seen
in the emberizine finches.

These are not sharply separated problems, but are all inter-
related under the general heading of adaptive modifications in
the bill for seed-eating. I shall, therefore, first describe the

388 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

various functional mechanisms of the avian jaw which are
prerequisite for understanding the modifications in the palatine
process of the premaxilla and, then, under the heading of conclu-
sions, return to these questions and answer them as best I ean.
Discussion of the functional mechanisms will be in the following
order : first, the structure, function and variation of the M. ptery-
gvoideus in the Passeres; second, a comparison of the adaptive
pathways through which the strength of the bite can be in-
creased; third and last, a comparison of the jaw muscles and
seed-cracking methods in the several groups of finches.

The M. pterygoideus. Tordoff (1954a, p. 12) assumed that the
origin and evolution of the palatine process of the premaxilla
(his ‘‘palato-maxillary’’) was dependent upon changes in the
mass of the M. pterygoideus. However, he apparently only ex-
amined the jaw muscles of the cardinal (Cardinalis) and extrapo-
lated the correlation between the palatine process and the M.
pterygoideus in the other New World nine-primaried oscines
from the condition seen in the cardinal. To be sure, part of the
M. pterygoideus originates from the palatine process in the
cardinal, yet it is necessary to survey the jaw musculature in the
Passeres and to correlate the changes in the M. pterygoideus with
the modifications in the palatine process before any statements
about the evolution of the palatine process can be made. It can
be stated in advance that the M. pterygoideus is the only jaw
muscle that originates from the palatine process of the pre-
maxilla.

Dissection of the M. pterygoideus is rather difficult because of
the incomplete separation of the muscle into four parts and the
complex arrangement of the muscle fibers. Much care must be
taken to separate the parts correctly and to determine the
direction of the muscle fibers in each part. The M. pterygoideus
is usually divided into a ventral and a dorsal portion and each
portion is, in turn, divided into a lateral and a medial half
(Lakjer, 1926, pp. 65-67). Some workers (e.g. Engels, 1940, pp.
359-361) do not recognize any subdivisions of the M. ptery-
goideus because they cannot separate the parts with complete cer-
tainty. It is true that the subdivisions of the M. pterygoideus are
not clearly defined units, but it is not necessary for the parts of a
muscle to be sharply separated from one another before they are
recognized as separate units. If the parts of a muscle, such as
the M. pterygoideus, have different functions and are unequally
developed in different forms, then they are best recognized even
if they are not sharply separated from one another. It should

BOCK: PALATINE PROCESS OF THE PREMAXILLA 389

be emphasized that the functional unit of a muscle is not the
whole muscle, or a recognizable part thereof, or even a muscle
fiber, but the motor unit—which is the aggregate of muscle
fibers innervated by a single nerve fiber (= motor cell axon). If
we wish to be completely precise in our studies of muscle fune-
tion, then we must separate the muscle into its motor units, which
is an impossible task. Therefore, the degree of analytical pre-
cision is not noticeably reduced if the recognized subdivisions of
a muscle merge into one another.

A

Figure 11. Jaw muscles of the gray jay (Perisoreus canadensis). (A)
Ventral view of the M. pterygoideus. (B) Oblique view into the orbit
showing the dorsal jaw muscles and the dorsal aspect of the M. pterygoideus.
In the ventral view of the jaw muscles of this and all other species, the
posterior end of the palatine process of the premaxilla or a point on the
palatine posterior to the palatine process is indicated by an arrow (PPM).
Thus the reader can note the relationship between the palatine process and
the M. pterygoideus ventralis lateralis. In the gray jay, for example, there
is no connection between the palatine process and the M. p. ventralis
lateralis.

The following description of the M. pterygoideus (Fig. 11)
is for the gray jay (Perisoreus canadensis), a bird having a
medium-sized bill of fairly generalized shape. I shall regard the
arrangement of the M. pterygoideus in this species as ‘‘typical’’
for the Passeres.

a) M. pterygoideus ventralis lateralis. This large segment
comprises almost all of the ventral portion of the M. ptery-
goideus. It originates from the entire ventral surface of the
transpalatine process and from much of the ventral surface of the
medial shelf of the palatine, and inserts on the medial and
ventral surfaces of the mandible and on the medial process of the

390 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

mandible (‘‘internal articular process’’ of some authors). Both
the origin and insertion of this muscle are fleshy. The M. p.
ventralis lateralis is a fan-shaped muscle with some of its medial
fibers inserting on the M. p. ventralis medialis. It is important to
note that in the gray jay, the origin of the M. p. ventralis lat-
eralis is from the transpalatine process and the medial shelf and
not from the region of the fused palatine process of the pre-
maxilla; that is, there is no association between this muscle and
the palatine process. The M. p. ventralis lateralis is, however,
the part of the M. pterygoideus that may take origin from the
palatine process of the premaxilla in some groups of passerine
birds.

b) M. pterygoideus ventralis medialis. This small subdivision
of the M. pterygoideus comprises only a minor part of the ventral
portion of the muscle and is frequently difficult to separate from
the lateral part. It originates from the lateral side and from the
tip of the mediopalatine process, and inserts on the distal end of
the medial process of the mandible. Both the origin and the
insertion are fleshy and the fibers are parallel to one another.
Beeause of its position, the M. p. ventralis medialis is never
associated with the palatine process of the premaxilla.

ce) M. pterygoideus dorsalis lateralis. This large segment of
the dorsal part of the M. pterygoideus lies directly over the
slightly larger M. p. ventralis lateralis; only in some of the
heavy-billed finches is the M. p. ventralis lateralis smaller than
the M. p. dorsalis lateralis. It takes origin from the dorsal
surface of the transpalatine process and the medial shelf of the
palatine, and inserts on the medial side of the mandible just
dorsal to the insertion of the M. p. ventralis lateralis. The muscle
fibers appear to be parallel to one another and to run obliquely
backwards from their origin to their insertion., Except for a
small aponeurosis at the corner between the mandible and its
medial process, the origin and insertion of this muscle are fleshy
in the gray jay; in some birds, they are quite tendinous. In the
gray jay, the origin of the M. p. dorsalis lateralis is limited to the
posterior part of the palatine and is far removed from the fused
palatine process. However, in some passerine birds, the origin

1 Actually these fibers are not parallel, but are pinnate for they insert on a
membrane that runs along the dorsal side of the muscle rather than directly on
the mandible. Pfuhl (1936) stresses this problem of the true pinnate nature of
some apparent parallel-fibered muscles, but for simplicity I shall regard pinnate
muscles of the M. p. dorsalis lateralis type as parallel-fibered. I realize that
this is incorrect and that someday a correct description of these muscles must
be given, but this simplifying assumption will not affect the results of the
present paper.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 391

of the M. p. dorsalis lateralis extends forward along the palatine
as far as the premaxilla. In these groups, the origin of this
anterior extension of the M. p. dorsalis lateralis is usually from
the dorsal surface of the prepalatine process. But in those few
groups where it takes origin from the lateral edge of the prepala-
tine process, the M. p. dorsalis lateralis is not associated with
the palatine process of the premaxilla.

d) M. pterygoideus dorsalis medialis. This is the most sharply
defined part of the M. pterygoideus. It takes origin from both
sides of the pterygoid bone and from a small part of the posterior
tip of the mediopalatine process, and inserts on the distal tip
of the medial process of the mandible. The pterygoid divides this
muscle into an anterior and a posterior part. The anterior fibers
are pinnate, inserting on a tendon that runs along the anterior
edge of the muscle; the posterior fibers are parallel. Except for
the insertion of the anterior fibers, the origin and insertion of
this muscle are fleshy. Because of its medial position, the M.
pterygoideus dorsalis medialis is never associated with the pala-
tine process of the premaxilla.

e) ‘‘M. retractor palatini.’’ The ‘‘M. retractor palatini’’ is
not a separate muscle as listed by some workers, but is part of
the M. pterygoideus. In most passerine birds, some of the medial
fibers of the M. pterygoideus run directly backward and insert on
the basitemporal plate instead of on the distal tip of the medial
process of the mandible. In the gray jay, a few fibers appear to
take origin from the middle of the M. p. ventralis medialis and
insert on the basitemporal plate. These fibers, which form a very
thin layer of tissue in the gray jay, are homologous with a band
of fibers in such groups as the thrashers and the thrushes that
originate on the medial shelf of the palatine next to the M. p. ven-
tralis lateralis and pass over the M. p. ventralis medialis to insert
on the basitemporal plate. These fibers are probably part of the
M. p. ventralis medialis although they appear to be associated
with the M. p. ventralis lateralis. (I shall discuss this group of
fibers in more detail below, p. 396). In addition to these ventral
fibers, a small group of fibers run back from the posterior tip
of the mediopalatine process to insert on the basitemporal plate
dorsal to the insertion of the ventral fibers. The fibers that insert
on the basitemporal plate may be grouped together as a part of
the M. pterygoideus or they may be included as part of the M. p.
dorsalis medialis or the M. p. ventralis medialis according to
their position. I will identify them on the figures as the ‘‘M.
retractor palatini,’’ but will consider them as part of the medial

?

392 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

portion of the M. pterygoideus in functional discussions. The
fibers of the ‘‘M. retractor palatini’’ are never associated with
the palatine process because of their extreme medial position.

Function of the M. pterygoideus. The action of the M. ptery-
goideus is usually described as raising the lower jaw and lowering
the upper jaw; however, this is not precise enough for the pur-
poses of this paper. Presumably, each of the four parts of the
M. pterygoideus has its own innervation and can contract inde-
pendently of the others. Also, only one part may enlarge to meet
the demands of a particular selection force. If all four parts
of the M. pterygoideus had the same function, then one would
expect that the whole muscle would evolve as a unit. Certainly
then, it can be assumed that, although the action of the M.
pterygoideus is to close the bill, the exact role of each of its four
parts in closing the bill differs and must be determined. The
following discussion is an attempt to ascertain the action of each
part of the M. pterygoideus from an analysis of their origins and
insertions, the directions of their muscle fibers, and their rela-
tive development in different types of birds.

The parts of the M. pterygoideus which serve to raise the
mandible must be so oriented that their pull causes the depressed
mandible to swing upward about its quadrate articulation. To
envision the direction of these muscle fibers, one must consider
the mandible and the M. pterygoideus from their lateral side as
well as from their ventral side. Two groups of fibers possess the
qualifications for raising the mandible. The first are those fibers
originating from the lateral side of the palate, the transpalatine
process and the palatine shelf and inserting on the medial side
of the ventral edge of the mandible anterior to its articulation.
These fibers draw the mandible directly upward and would be
effective even with the bill almost closed. The second group
of fibers are those which originate from the transpalatine process
and the palatine shelf and insert on the medial process of the
mandible, usually on its anterior face but sometimes along the
ventral edge of its posterior face. These fibers pull the medial
process of the mandible forward and thereby raise the mandible.
When the mandible is depressed, the medial process is slightly
posterior and ventral to its position when the bill is closed. The
difference between the normal and the depressed position of
the medial process is very*small, perhaps only 149 of the distance
between the posterior tip of the transpalatine process and the
medial process of the mandible when the bill is closed. This
means that a sheht movement of the medial process toward its

BOCK: PALATINE PROCESS OF THE PREMAXILLA 393

normal position results in the mandible being raised over a
considerable distance. Because of their insertion near the quad-
rate hinge, these fibers raise the mandible rapidly, but with little
power. The medial process reaches its final position when the
bill is about half closed. Thus, the fibers of this second group
are effective in raising the mandible only when the bill is wide
open and can no longer serve in this connection after the bill is
half closed. Lastly, it should be mentioned that those fibers which
insert along the ventral edge of the posterior face of the medial
process rotate the process and thereby raise the mandible. These
fibers may be effective in raising the mandible until the bill is
almost closed ; however, I have not studied this point in detail.

Lowering of the upper jaw would depend upon the ability of
the M. pterygoideus to retract the palate. Probably all of the
fibers of this muscle, no matter what their origin and insertion
might be, would draw the palate backward. However, those
fibers which run obliquely from the palate to the medial side of
the mandible exert only a slight backward pull on the palate.
The fibers which retract the palate most effectively are those that
originate on the posterior part of the palate and run directly
back to insert on the medial process of the mandible. Those
particular fibers which insert on the basitemporal plate can only
retract the palate; they cannot have any effect on the mandible.
Lastly, the fibers which originate from the pterygoid probably
have as their only action, the lowering of the upper jaw.

With these background remarks in mind, the following actions
may be suggested for the parts of the M. pterygoideus.

a) M. p. ventralis lateralis. The lateral fibers of this muscle
act mainly to raise the mandible. The medial fibers, which insert
on the medial process of the mandible, retract the palate and
thus depress the mandible during their entire contraction cycle,
but can serve to raise the mandible only when the bill is about
half closed. In the insect-eaters, the M. p. ventralis lateralis is
probably more important as a palatine retractor, but in the
large-billed seed-eaters, this muscle is probably more important
as an adductor of the mandible.

b) M. p. ventralis medialis. This muscle, by virtue of its
origin on the mediopalatine process and its insertion on the distal
tip of the medial process of the mandible and the basitemporal
plate, has as its major and probably only action, the retraction
of the palate. It may be noted that those birds which need large
palate retractors, such as the swallows and the flycatchers, have
a large M. p. ventralis medialis. I include those fibers which

394 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

originate on the palatine next to the M. p. ventralis and pass over
the M. p. ventralis medialis to insert on the basitemporal plate
(see drawing of the thrasher, Figs. 13A and 13B) as part of the
M. p. ventralis medialis; their function is, of course, palatine
retraction.

ce) M. p. dorsalis lateralis. All of the fibers of this muscle run
obliquely from the palatine bone to the medial side of the
mandible. Thus, the action of this muscle is to raise the mandible
with, at most, a very minor part of the force used to retract the
palate. It could be noted that this muscle is small in insect-eaters
and greatly enlarged in seed-eaters.

d) M. p. dorsalis medialis. This muscle takes origin from the
pterygoid and the posterior end of the mediopalatine process and
inserts on the distal end of the medial process of the mandible
and the basitemporal plate; hence its sole action is retraction of
the palate.

To recapitulate, the major function of the medial parts of the
M. pterygoideus is to retract the palate and hence depress the
upper jaw, while the major function of the lateral parts is to raise
the lower jaw. The medial fibers of the M. p. ventralis lateralis ean
raise the mandible only during the early part of their contraction
while they can retract the palate during all of their contraction.
This separation of functions between parts of the M. pterygoideus
is not a sharp one, for it seems likely that each part of the muscle
has at least a small role in both functions. However, this division
of labor between the parts of the M. pterygoideus is clearly re-
flected in their relative sizes in different types of passerine birds,
as for example, insect-eaters as compared to seed-eaters.

Variation of the M. pterygoideus in the Passeres. The results
of a survey of the M. pterygoideus in the Passeres will be re-
ported in this section. This survey is far from complete, but it
does include a number of different types of passerine birds and
is, I believe, adequate for the purposes of this paper. The muscle
will not be described in detail as has been done for the gray
jay ; instead, its ventral aspect will be figured for each species
available. In the figure, the posterior end of the palatine process
will be indicated to allow the reader to determine the relation-
ship between the M. p. ventralis lateralis and the palatine
process. A word of warning should be given. First, my drawings
are crude representations of the very complex system of jaw
muscles. I have tried to show the spatial relationships of the
muscles and the directions of the fibers; however, I cannot vouch
for the accuracy of the proportions or the perspective. These

BOCK: PALATINE PROCESS OF THE PREMAXILLA 395

figures were drawn to illustrate the points discussed in this paper
and should not be used to illustrate any other aspect of the jaw
muscles. Second, the style of each author differs; thus much of
the difference in the jaw muscles in a bird as shown by Engels
or Fiedler or myself is artificial. The significance of this survey
in relation to the functional significance and evolution of the
palatine process will be discussed in the conclusion of this see-
tion.

The method used in dissecting the M. pterygoideus was
simply to remove the hyoid apparatus and associated muscles,
to cut off the mandible just anterior to the insertion of the M.
pterygoideus, and lastly to remove the lining on the roof of the
mouth plus the horny covering of the upper jaw. Usually the eye
was also removed to allow examination of the other jaw muscles
and the dorsal aspect of the M. pterygoideus. The M. ptery-
eoideus is now exposed and after some cleaning up of connective
tissue and blood vessels, it is ready for study. Some care must
be taken when removing the lining of the mouth and the horny
palate to make certain that the tendons and muscle fibers in the
region of the prepalatine process are not damaged or destroyed.

The following species are available for comparison :

Tyrannidae Tyrannus dominicensis Figure 12A
Alaudidae Eremophila alpestris Be 12Crvand 2)
Hirundinidae Tridoprocne bicolor me 12B
Bombyeillidae Bombycilla cedrorum Gt 12H
Troglodytidae ITeleodytes brunneicapillus UE 12F
Mimidae Toxostoma rediviwum 6 13A
Toxostoma rufum ig 13B
Nesomimus macdonaldi oe 13C
Dumetella carolinensis se 13D
Turdinae Turdus philomelos as 13E
Turdus migratorius aM 13F
Hylocichla sp. Oe 14A
Paradoxornithinae Paradoxornis sp. “ 145
Polioptilinae Polioptila caerulea as 14D
Sylviinae Regulus calendula af 14C
Paridae Parus bicolor ai 14B
Sittidae Sitta europaea Bi 14h
Nectariniidae Cinnyris chalybaeus ae 15A
Zosteropidae Zosterops annulosa ne 15B
Meliphagidae Anthornis melanura ae 15C
Emberizinae Emberiza citrinella fue 15D

Passerella iliaca
Melospiza melodia

15E and 15F
17A

396 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Emberizinae (cont ’d) Spizella pusilla Figure 16A and 16B
Pipilo erythrophthalmus os 16E and 16F
Zonotrichia albicollis of 16C and 16D
Cardinalinae Cardinalis cardinalis Oh 17E and 17F
Passerina cyanea oe 17€ and 17D
Tanagrinae Piranga rubra aie 17B
Coerebinae Dacnis cayana oh 19B
Parulidae Seiurus aurocapillus oe 18C and 18D
Vireonidae Vireo olivaceus ae 18A and 18B
Icteridae Molothrus ater oe 18E and 18F
Quiscalus quiscala ee 19A
Fringillinae Fringilla coelebs oe 19C and 19D
Carduelinae Spinus tristis (o> OR vand el9in
Carpodacus purpureus es 20C
Hesperiphona vespertina es 20A and 20B
Carduelis carduelis oe 20D
Pinicola enucleator oe 20H
Loxia curvirostra GG 20F
Coccothraustes coccothraustes ‘‘ 21A
Estrildidae Lonchura oriziwora GE 21B
Ploceidae Passer domesticus uS 21C@ anda
Sturnidae Sturnus vulgaris Ges 28H
Corvidae Corvus crassirostris es 21F
Perisoreus canadensis io ial

Some comparative notes on the structure of the M. ptery-
goideus can be given at this point. It has already been pointed
out that, in the seed-eaters, the medial parts of this muscle are
relatively small while the lateral parts are relatively large. In the
insect-eaters, the medial parts are relatively large, although they
are still smaller in mass than the lateral parts of the M. ptery-
goideus; the M. p. ventralis lateralis makes up a major share of
the total mass of the muscle. For example, the M. p. dorsalis
lateralis is very small in the kinglet (Regulus) and the gnat-
catcher (Polioptila), while the medial parts of the M. ptery-
goideus comprise only about 5 per cent of the total muscle mass
in such heavy-billed finches as the evening grosbeak (Hesperi-
phona). The structure of the ‘‘M. retractor palatini’’ in the
Old World insect-eaters, such as the kinglet, thrushes, thrashers
and wrens, is very characteristic. The dorsal band of fibers
originates along with the M. p. dorsalis medialis from the distal
tip of the mediopalatine process, and is unquestionably part of
that muscle. The ventral band of fibers originates from the
palatine shelf next to the M. p. ventralis lateralis and passes
over the M. p. ventralis medialis before inserting on the basi-
temporal plate. Although these fibers appear to be part of the

BOCK: PALATINE PROCESS OF THE PREMAXILLA 397

B

Figure 12. Jaw muscles of: (A) Tyrannus; (B) Iridoprocne; (C and
D) Eremophila; (E) Bombycilla; and (F) Heleodytes (redrawn from
Engels, 1940).

398 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

M. p. v. lateralis, they are most probably part of the M. p. v.
medialis. They may have at first originated from the ventral
edge of the mediopalatine process along with the rest of the
M. p. v. medialis and then moved laterally to the palatine shelf
as their mass increased. It is possible that these fibers are part
of the M. p. v. lateralis and that their insertion shifted from the
medial process of the mandible to the basitemporal plate, but
this does not seem probable. A more thorough study of the in-
nervation of these fibers is needed before we can be certain of
their origin.

Beecher (195la, 1953) illustrated the dorsal aspect of the M.
pterygoideus and identified four subdivisions — the M. p. dor-
salis anterior, M. p. dorsalis posterior, M. p. ventralis anterior
and M. p. ventralis posterior (Beecher’s terminology). Most of
his figures (1953) show the usual four subdivisions, but some
(sunbird, p. 291; white-eye, p. 291; wood warbler, p. 306; see Fig.
22B; and wren, p. 318) show five subdivisions (the identity of
the fifth subdivision is usually not mentioned) and others show
only three subdivisions. In the case of birds with only three
parts of the M. pterygoideus visible through the orbit, these parts
are always the M. p. d. anterior, M. p. d. posterior and the M. p.
y. anterior, as for example, in the house finch, the cowbird (Fig.
22D) and the song sparrow (Fig. 22C). In the latter species,
there can be no doubt of Beecher’s identification of the M. p. v.
anterior for he states (1953, p. 307) that: ‘‘Large M. 4a [= M.
p. v. anterior] overlying M. 4b [ — M. p. v. posterior].’’ Yet my
dissections of the finches revealed that the origin of the M. p.
dorsalis lateralis extended anteriorly along the prepalatine proc-
ess as far as the premaxilla in some species (see also, Fiedler,
1951; and Sims, 1955, p. 381). In these birds, the M. p.
ventralis would be completely covered by the M. p. dorsalis and
invisible when the jaw muscles are viewed through the orbit.
Dissection of other passerine birds showed that the M. p. dorsalis
lateralis covers much of the M. p. ventralis lateralis and that the
M. p. ventralis medialis is usually not visible when the muscles
are viewed through the orbit. Only in the thin-billed species can
much of the ventral portions of the M. pterygoideus — usually
the M. p. ventralis lateralis— be seen when the dorsal aspect of
this muscle is examined through the orbit. In any case, the
ventral parts of the M. pterygoideus are visible between the two
dorsal parts of this muscle, not lateral to both dorsal segments
as shown by Beecher. Consequently, his identifications of the
parts of the M. pterygoideus would seem to be wrong and should

BOCK: PALATINE PROCESS OF THE PREMAXILLA 399

B

MRP ‘MPVM

MRP

Figure 13. Jaw muscles of: (A) Toxostoma redivivum (redrawn from
Engels, 1940); (B) Towostoma rufuwm; (C) Nesomimus (redrawn from
Engels, 1940); (D) Dumetella; (E) Turdus (redrawn from Fiedler, 1951) ;
and (F) Turdus migratorius.

400 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

be corrected. His M. p. dorsalis anterior and M. p. d. posterior
are the two parts of the M. p. dorsalis medialis which are anterior
and posterior to the pterygoid respectively. His M. p. ventralis
anterior is the M. p. dorsalis lateralis. Thus, his M. p. ventralis
posterior would be part of the M. p. ventralis and most likely
the M. p. ventralis lateralis. The fifth portion shown in some
figures would be the M. p. ventralis medialis. However, there is
some doubt as to the identification of the ventral parts of the
M. pterygoideus shown in Beecher’s figures. In some eases, his
M. p. ventralis posterior may be the M. p. ventralis medialis in-
stead of the M. p. v. lateralis, or, more likely, two muscles should
have been shown instead of just one. I may add, at this point,
that the only way to be certain of the identification of the parts
of the M. pterygoideus is to dissect them from the ventral side of
the muscle. These misidentifications and the tacit assumption
that the M. pterygoideus only retracts the palate invalidate
Beecher’s remarks on the structure and the function of the M.
pterygoideus.

Here may be the best place to interject a few comments on the
factual parts of Beecher’s work as there are a number of dis-
erepancies between his drawings and my dissections of the same
bird or a species within the same family. For example, Beecher
shows the medial slip of the M. adductor mandibulae in the
larks (1953, p. 816) as a parallel-fibered muscle, but in my dis-
section of the same species, this muscle was complexly pinnate.
Again, Beecher shows the same muscle slip in the Icteridae as
pinnate (1953, p. 308), although he showed it as parallel-fibered
in his earlier paper (1951la). My dissections of several genera of
the Icteridae, including the cowbird, agree with his earlier paper.
In both his dissections and drawings, Beecher studied only the
external aspect of the muscles and did not dissect the muscles
themselves, nor did he attempt to ascertain the mass or cross-
sectional area of the muscles. Thus a muscle is considered to be
important if it exhibits a large surface area as shown in his
drawings. All pinnate muscles are lumped together as one type
and automatically considered to be better and more efficient
than parallel-fibered muscles. Aside from these points, there
are many interpretations which do not seem to agree with the
facts presented. For example, the drawings of the adductor
(= medial) slip of the M. adductor mandibulae do not agree
with his separation of the oscine families into two superfamilies
on the basis of a pinnate slip in the one group and a parallel-
fibered slip in the other group. Nor can I understand the evolu-

BOCK: PALATINE PROCESS OF THE PREMAXILLA 401

PPM

MPDM MPVL
MPDL MRP MPVM

MPDL MRP

Figure 14. Jaw muscles of: (A) Hylocichla; (B) Paradoxornis (redrawn
from Fiedler, 1951); (C) Regulus; (D) Polioptila; (E) Parus; and (F)
Sitta (redrawn from Fiedler, 1951).

402 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

tionary shift of the muscle fibers from the M. adductor mandibu-
lae to the M. pseudotemporalis superficialis (Beecher, 1951b, p.
278). However, Beecher’s functional discussions are excellent
and should be read by those interested in the functional anatomy
of the avian skull.

Yet, this critical evaluation of Beecher’s work should not be
interpreted as meaning that the jaw muscles cannot supply good
elues to the relationships and evolution of the passerine birds.
They may well prove to be useful if analyzed with extreme cau-
tion and with the realization that they are subject to the same
evolutionary phenomena, such as convergence, that make the
study of any taxonomic character difficult (see also Starck,
1959).

Comparison of the adaptive pathways for increased force of
the bite in the passerine birds. I have mentioned above that the
major modifications of the palatine process in the passerine birds
appear to be associated with the functional demands of seed-
eating. However, the relationship between seed-eating and the
structure of the palatine process is not a simple one such as the
free palatine process (cardinal condition) becoming more and
more fused as the M. pterygoideus decreases in size (Tordoff,
1954a, p. 12) and vice versa. If this were true, then why do the
heavy-billed cardueline finches lack the free palatine process and
possess lateral flanges on the anterior end of the prepalatine
processes? This question leads to the fundamental question of
the entire problem: What are the basic requirements for seed-
eating, and how have passerine birds evolved the necessary
structural adaptations to meet these demands?

Aside from behavioral traits and such morphological features
as the length of the gut (see Eber, 1956), the necessary digestive
enzymes and so forth, the basic requirement of a seed-eating bird
is to be able to crack the hard shell of a seed without damage to
itself. One way to meet this requirement is to grind the seeds
in the muscular gizzard, as done by gallinaceous birds and
pigeons. Passerine birds have not utilized this method, but crack
seeds by means of a powerful closing of their bill. Thus, seed-
eating passerines must be able to crack seeds in their bill with-
out damaging the structures of the head, especially the brain
and sense organs. Larks are apparently an exception for they
swallow seeds whole and grind them in their gizzard (Meinertz-
hagen, 1951, p. 84). The central problem of this section is,
therefore: What are the ways by which a passerine bird might
inerease the strength of its bite and at the same time protect the

BOCK: PALATINE PROCESS OF THE PREMAXILLA 403

\

oT |

| MPDL
MRP ‘MPDM

Figure 15. Jaw muscles of: (A) Cinnyris (redrawn from Moller, 1930) ;
(B) Zosterops (redrawn from Moller, 1931); (C) Anthornis (redrawn from
Moller, 1931); (D) Emberiza (redrawn from Fiedler, 1951); and (E and
FE) Passerella.

404. BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

other structures of its head against the forces and shocks associ-
ated with seed-cracking? The several methods by which the
strength of the bite can be increased will be described first, and
then it will be shown how different combinations of these methods
have evolved in the several groups of finches.

Muscles that close the bill. Four separate jaw muscles in the
Passeres act to close the bill, either by raising the mandible or by
depressing the upper jaw. Increase in the mass of any of these
muscles would increase the strength of the bite. The usual condi-
tion in the finches is that all of these muscles have increased in
size, but that the relative increase of the several muscles differs
in the different groups. The descriptions of the jaw muscles will
be for the gray jay (Fig. 11), with comparative notes on their
structure in the finches.

a) M. adductor mandibulae. This is usually the largest of the
jaw muscles or is second in mass only to the M. pterygoideus.
Without doubt, it is the most complex of the jaw muscles. The M.
adductor mandibulae is the most posterior of the dorsal ad-
ductors of the mandible and takes origin from the lateral side
of the skull posterior to the orbit, and from the outer rim of the
orbit, and inserts on the dorsal edge and lateral side of the
mandible. The action of the M. adductor mandibulae is to raise
the mandible, but because of the anterior position of its insertion,
it is probably most important when the mandible is more than
half closed. The anterior position of its insertion gives the M.
adductor mandibulae a mechanical advantage through increased
leverage (the farther a force is applied from the fulerum point,
which in this case is the quadrate-articular hinge, the greater
is the resulting force). The anterior insertion also results in a
mechanical disadvantage when the bill is wide open because of
the unfavorable angle of insertion —a very acute angle which
means that most of the streneth of the muscle is lost (see Mollier,
1937; and Dullemeijer, 1951, for a discussion of the ‘‘unprofit-
able’’ angle of insertion). The M. adductor mandibulae usually
does not leave a muscle sear on the roof of the skull in passerine
birds, but in several of the heavy-billed finches, such as the
evening grosbeak and the cardinal, a slight depression can be
seen on the roof of the skull outlining its area of insertion.

b) M. pseudotemporalis superficialis. This tripartite (some-
times bipartite) muscle originates from the posterodorsal wall of
the orbit, just medial to the origin of the M. adductor mandibu-
lae, and inserts on the medial side of the mandible close to the
quadrate hinge. Its action is to raise the mandible, but in

BOCK: PALATINE PROCESS OF THE PREMAXILLA 405

yf a
ie SS

WY

Vv

WUE

MPVL
MRP “MPVM

MPD M
MPDL
Figure 16. Jaw muscles of: (A and B) Spizella; (C and D) Zonotrichia;

and (E and F) Pipilo.

406 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

contrast to the M. adductor mandibulae, the M. pseudotemporalis
superficialis is probably more important as an adductor of the
mandible when the bill is opened widely. Its insertion close to
the quadrate hinge allows it to close the bill rapidly at the price
of a reduction of the exerted force. It is interesting that in the
eardueline finches, the enlarged part of the M. pseudotemporalis
superficialis has the most anterior insertion. The large mass and
importance of this muscle in the finches is indicated by the
several bony processes on the posterodorsal wall of the orbit to
which this muscle attaches. These processes are absent in most
other passerine birds, especially in the thin-billed insect-eaters.

ec) M. pseudotemporalis profundus. This muscle originates
from the orbital process of the quadrate and inserts on the medial
side of the mandible anterior to its insertion of the M. p. super-
ficialis and opposite the insertion of the M. adductor mandibulae.
Like the M. pterygoideus, this muscle has the dual function
of raising the lower jaw and depressing the upper jaw; however,
it is difficult to determine which of these functions is the most
important. The M. pseudotemporalis profundus is a relatively
small muscle as compared to the other jaw muscles, especially the
other adductors of the mandible. Increase in the mass of this
muscle could serve for increased strength of the adductors of
the mandible or for increased strength of the palatine retractors
(= depressors of the upper jaw). The latter function may be
the more important because this muscle is relatively small in the
finches. It is also possible that the M. p. profundus functions to
oppose the outward forces of the M. adductor mandibulae and
to strengthen the quadrate hinge.

d) M. pterygoideus. This muscle is the most anterior of the
jaw muscles and lies ventral and anterior to the M. pseudotem-
poralis profundus. The M. protractor quadrati les dorsal to the
M. pterygoideus and separates it from the M. pseudotemporalis
superficialis. The M. pterygoideus has already been described
and discussed. I need only to emphasize that the medial parts of
the M. pterygoideus —the depressors of the upper jaw — are
relatively more highly developed in groups with a highly kinetic
upper jaw while the lateral subdivisions are more highly de-
veloped in the groups which have only a slightly kinetic upper
jaw. In the finches, only the dorsal portions of the M. ptery-
goideus can be seen through the orbit.

Gross muscle function. The action of a muscle depends not
only upon its size and attachment, but upon the orientation of
its fibers. In some muscles, the fibers run parallel to one another

BOCK: PALATINE PROCESS OF THE PREMAXILLA 407

Figure 17. Jaw muscles of: (A) Melospiza; (B) Piranga; (C and D)
Passerina; and (E and F) Cardinalis.

408 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

and to the longitudinal axis of the muscle, while in others, the
fibers are oblique to the longitudinal axis and insert on a tendon
or an aponeurosis. The former are usually called parallel-fibered
or simple muscles, and the latter, pinnate or complex muscles.
Pinnate muscles are not identical in their internal structure,
but vary greatly in the number of central tendons and in the
directions of their fibers. The action of parallel-fibered muscles
is relatively easy to analyze. Since all of the muscle fibers are
oriented along the longitudinal axis of the muscle, the speed,
streneth and distance of the muscle contraction is proportional
to the number of fibers that have contracted. The angle of inser-
tion of the muscle fibers on the central tendon and the change in
this angle during contraction must be considered in addition to
these factors when one analyzes the action of a pinnate muscle.
Few workers have considered pinnate muscles in detail with the
result that virtually nothing is known about their action. Pfuhl
(1936) is the only worker, to my knowledge, who has attempted
to analyze pinnate muscles with the use of trigonometrical
models. The reader is referred to his paper and those by Mollier
(1937) and Dullemeijer (1951).

Both parallel-fibered and pinnate muscles are found in the jaw
muscles of passerine birds; indeed, some of the jaw muscles, such
as the M. adductor mandibulae, are among the most complex
muscles found in birds. The same muscle may be parallel-fibered
in some species and pinnate in others. Some workers, notably
Beecher, have differentiated between parallel-fibered and pinnate
muscles in their functional discussions. But their basic assump-
tions are so simplified that their results are misleading. In
general, they have assumed that pinnate muscles are one type
and parallel-fibered muscles are another, that pinnate muscles
are universally more efficient (i.e., stronger) than parallel-
fibered muscles and that pinnate muscles have evolved only in
response to a selection force for increased strength. In an at-
tempt to clarify some of these problems, I have started an
analysis of the action of pinnate muscles using trigonometrical
models and hope to present the results in the near future. A
few tentative conclusions will, however, be outlined to illustrate
the major aspects in the action of pinnate muscles.

The angle of insertion of the muscle fibers on the central ten-
don determines the relative number of fibers, the relative amount
of useful force and the relative speed of central tendon. If pin-
nate muscles of equal lengths and diameters are compared, the
number of fibers increases, the amount of useful force decreases

BOCK: PALATINE PROCESS OF THE PREMAXILDLA 409

A
PPM
PDM | MPVL
MPDL 'mRP \mMpvom
C
PPM
/
PDM MPVL

E PPM

MRP ‘MPVM

Figure 18. Jaw muscles of: (A and B) Vireo; (C and D) Seiurus; and
(E and F) Molothrus.

410 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

and the relative speed increases as the angle of insertion of the
fibers increases. Similarly, during contraction, the angle of in-
sertion increases with a corresponding decrease in the amount
of useful force and an increase in speed. Thus it can be seen
that pinnate muscles can be not only strong muscles acting over
a short distance, but also weak, rapid muscles acting over a long
distanee. For example, the M. pseudotemporalis superficialis is
frequently pinnate. This muscle inserts on the mandible close
to the quadrate hinge; hence it serves to raise the mandible
rapidly, but with little force. Most likely, it has become pinnate
in response to a selection force for increased speed. On the
other hand, the M. adductor mandibulae is a ‘‘power muscle.’’
It inserts on the mandible far anterior of the quadrate hinge and
serves to raise the mandible with great force. Also, its action is
frequently over a very short distance as, for example, when a
finch eracks a seed. Thus, this muscle has become pinnate in
response to a selection force for increased strength.

Unless complex pinnate muscles, such as the jaw muscles, are
dissected in great detail and all possible reasons for their becom-
ine pinnate are considered, it is better to omit this factor from
consideration. For this reason, I have not attempted to compare
the pinnateness of the several jaw muscles in the finches. But it
is obvious that the degree of pinnateness cannot be omitted if we
hope to understand the function of the jaw muscles and to com-
pare properly the jaw muscles of different groups of passerine
birds. Thus, in investigations of the jaw muscles, there is really
no choice but to dissect the pinnate muscles in great detail and
to take great care in interpreting their functional significance.

Relationship between the processes of the skull. Another
factor influencing the strength of the bite, but quite apart from
the muscles themselves, is the size, shape, and spatial relation-
ships of the various bony processes to which the muscles attach.
Changes in these processes would modify the leverage of the
jaw muscles. The role of leverage in the action of the jaw
muscles has been studied extensively by Kripp (1935) and more
recently by Fisher (1955). However, most workers completely
overlook the importance of the bony processes in the action of
the jaw muscles. A notable exception is Beecher’s discussion
(195la, p. 420) of the orbital process of the quadrate. He shows
this process to be a lever and discusses the functional significance
of the difference in its length in two genera of blackbirds.

The variation in several bony processes of the skull is directly
correlated with changes in the jaw muscles. Some examples are

BOCK: PALATINE PROCESS OF THE PREMAXILLA 411

MPDM MPVL
MPDL MRP ‘MPVM

Figure 19. Jaw muscles of: (A) Quiscalus; (B) Dacnis (redrawn from
Moller, 1931); (C and D) Fringilla; and (E and F) Spinus.

412 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

the size of the medial process of the mandible, the length of the
transpalatine process (and hence the distance between this bone
and the mandible), the length of the palatines themselves, the
size of the auditory bullae (‘‘inflated squamosal region’’ Tordoff,
1954a, pp. 9-10, which is associated with the mass and length of
the M. depressor mandibulae), and finally, the free palatine
process of the premaxilla as seen in the cardinals. This list could
be easily expanded, but it is sufficient to show that a comparative
study of the musculature of so complex a system as the jaw
muscles must include the detailed mechanics of the underlying
bone-lever system. The converse is also true; a study of the
skull must also include the muscles and other influencing factors.

Comparison of the jaw muscles in the finches. The jaw muscles
of the ‘‘nine-primaried’’ and ‘‘ploceid’’ finches will now be
compared, using the information presented in the preceding sec-
tions. The major question to be answered is: Has the same
morphological adaptation for cracking seeds evolved in the
several groups of finches, or have different adaptations evolved
in these groups (ef. multiple pathways of evolution) ? This ques-
tion may appear to be unrelated to the central problem of this
paper — the evolution of the palatine process of the premaxilla
—especially when the jaw musculature of the finches is com-
pared, but it is essential to study the entire set of jaw muscles
before the differences in the M. pterygoideus can be understood
and its correlation with the palatine process of the premaxilla
clarified. In addition to comparing the jaw muscles of the sev-
eral groups of finches, I shall compare, whenever possible, a
small-billed species with a large-billed species of the same group,
to determine whether there is any variation in the jaw muscles
within familes or subfamilies of passerine birds and more
precisely, whether the jaw muscles have changed within a
group of finches to meet the demands of a stronger bite. I shall
deseribe the small-billed species first and then compare it to the
large-billed species. This procedure is used for convenience only
and not to imply that the small-billed species is primitive in the
group or that evolution in the finches has always been from the
small- to the large-billed size.

The jaw muscles of a warbler (Seiurus aurocapillus, Figs. 18C
and 18D) and a vireo (Vireo olivaceus, Figs. 18A and 18B) —
both insect-eaters with thin bills — and ineluded, in addition to
those of the gray jay, for comparison with the heavy-billed seed-
eaters. These species were chosen because of convenience only
and not because of any special relationship to the finch groups

BOCK: PALATINE PROCESS OF THE PREMAXILLA 413

WA
z

UD:

SX

—

MPVL

Figure 20. Jaw muscles of: (A and B) Hesperiphona; (C) Carpodacus ;
(D) Carduelis (vedrawn from Fiedler, 1951); (E) Pinicola (redrawn from
Fiedler, 1951); and (F) Lomwia (redrawn from Fiedler, 1951).

414 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

discussed here. When examining the jaw muscles of these in-
sectivorous birds, the general impression one receives is that the
muscles are weakly developed, or, to put it in another way, the
muscles are not overly developed and do not show any specializa-
tions for a strong bite. For example, the origin of the M. adductor
mandibulae has not spread over the roof of the skull and the M.
pseudotemporalis superficialis has not expanded to cover the M.
p. profundus. One of the most striking features in the jaw
muscles of these species is the relative weakness of the parts of
the M. pterygoideus that adduct the mandible, especially the M.
p. dorsalis lateralis. In the warbler and the vireo, parts of the
M. p. ventralis may be seen between the two halves of the M. p.
dorsalis when the jaw muscles are viewed through the orbit —
an indication of the weakness of the M. p. dorsalis lateralis. This
muscle is large in the gray jay.

The emberizine finches. The emberizine finches may be con-
sidered as generalized or, better, as unspecialized seed-eating
birds; they feed on smaller seeds and are more insectivorous than
most other groups of finches. In accordance with these feeding
habits, the morphological specializations for seed-eating are less
developed than in other finches. For example, the bill of the
emberizines, although shorter and stouter than the bill of insect-
ivorous birds, is longer and thinner than the bill of other groups
of finches. The palatine process of the premaxilla is essentially
the same as in the insect-eaters; it lies along the prepalatine
process and is more or less fused with that bone. The major
exceptions are Melopyrrha and Tiaris, which have a free palatine
process such as is found in the cardinals, and Oryzoborus, which
has a lateral flange on the prepalatine process similar to that
found in the cardueline finches. These ‘‘aberrant’’ genera will
be discussed in the section on relationships. The emberizine
finches are a useful starting point, for their lack of extreme
specializations in the skull and in the jaw muscles allows us to
analyze the basic modifications in these structures for seed-
cracking. Insectivorous and granivorous birds are not sharply
distinet types, but grade into one another ; hence, it is not always
possible to distinguish insectivorous from granivorous adapta-
tions.

The jaw muscles of a field sparrow (Spizella pusilla, Figs. 16A
and 16B) are similar to those of the warbler and the vireo except
for the increase in mass of the mandible adductors. The M.
adductor mandibulae and the M. pseudotemporalis superficialis
are larger and more pinnate than those in the warbler, but they

BOCK: PALATINE PROCESS OF THE PREMAXILLA 415

A

PPM ZA |

GUGRV AL
MPVM ‘MRP MPDL

PVL

Figure 21. Jaw muscles of: (A) Coccothraustes (redrawn from Fiedler,
1951); (B) Lonchura (redrawn from Fiedler, 1951); (C and D) Passer;
(E) Sturnus; and (F) Corvus (redrawn from Fiedler, 1951).

416 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

are quite similar in other respects. The M. pseudotemporalis pro-
fundus is not hidden by the M. p. superficialis as in the cardue-
line finches. Most striking is the expansion of the M. ptery-
goideus dorsalis lateralis which completely covers the ventral
parts of the M. pterygoideus. Turning to the ventral aspect of
the M. pterygoideus, the increase of the mandible adductors at the
expense of the palate retractors can be readily seen. Yet the
palate retractors are still relatively large. A tendon of the M.
pterygoideus ventralis lateralis appears to extend forward along
the palatine to the position of the semifused palatine process
of the premaxilla. I have shown this tendon running forward
in my drawing of the field sparrow (Fig. 16B), but wish to
emphasize that it is not certain whether this tendon really exists.
The small size of the field sparrow makes it difficult to determine
whether the strip of connective tissue seen along the prepalatine
process is the periosteum of that bone or a tendon of the M.
pterygoideus. This problem may be resolved by histological
examination, but I am not sure whether it can ever be decided
beyond all doubt. Therefore, although there is an indication in
the field sparrow of a direct association between the M. ptery-
goideus and the palatine process by means of a tendon, this must
still be proven.

In such a medium-billed species as the white-throated sparrow
(Zonotrichia albicollis, Figs. 16C and 16D) and the larger rufous-
sided towhee (Pipilo erythrophthalmus, Figs. 16K and 16F), the
jaw muscles increase in mass as the size of the bill increases. In
the towhee, there is a muscle sear on the roof of the skull outhn-
ing the origin of the M. adductor mandibulae, a reflection of the
increase in size of this muscle. The most conspicuous changes
in the muscles are, however, the increase in size of the antero-
medial part of the M. pseudotemporalis superficialis toward the
cardueline condition and the increase in the adductor parts of
the M. pterygoideus. Both the M. p. dorsalis lateralis and the
M. p. ventralis lateralis increased in mass. The change in the
M. p. v. lateralis is of particular interest. The lateralmost fibers
of this muscle converge on a tendon that runs along the lateral
edge of the palatine up to the fused palatine process of the pre-
maxilla. Although this tendon is easily destroyed during dissec-
tion, it ean readily be demonstrated in the towhee. The tendon
is intimately associated with the periosteum of the palatine. In
the smaller species, even if it is present, the tendon is almost
indistinguishable from the periosteum of the palatine, as for
example in the field sparrow. The towhee, the largest of these

BOCK: PALATINE PROCESS OF THE PREMAXILLA 417

Figure 22. Jaw muscles as seen through the orbit of: (A) Phylloscopus;
(B)Oporomis; (C) Melospiza; and (D) Molothrus (redrawn from Beecher,
1953). The M. pterygoideus has been labeled according to my identification.
Beecher’s identifications for Phylloscopus are: M. pterygoideus dorsalis
anterior (8a) = MPDM (anterior half); M. p. dorsalis posterior (3b) =
MPDM (posterior half); M. p. ventralis anterior (4a) = MPDL; M. p. ven-
tralis posterior (4b) = MPVL (usually, but sometimes part or all of the
MPVM is included in this muscle by Beecher).

species, has a well-developed tendon very similar to the tendon
of the M. p. ventralis lateralis that attaches to the free palatine
process of the premaxilla in the tanagers and the cardinals. Thus,
with increase in the size of the bill in these species of emberizine
finches, the M. pseudotemporalis superficialis changes toward the
cardueline condition while the M. pterygoideus changes toward
the cardinal condition.

Not all emberizine finches show these changes in the structure
of the jaw muscles. The fox sparrow (Passerella tliaca, Figs. 15K
and 15F), another heavy-billed species, tends toward the cardue-
line condition not only in the structure of its M. pseudotemporalis
superficialis, but also in the structure of its M. pterygoideus
ventralis lateralis. The lateralmost fibers of this latter muscle do
not send a long tendon forward along the palatine, but rather
insert on the distal tip of the transpalatine process by means of

418 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

a short tendon. In accordance with this condition of the M. p. v.
lateralis, the tip of the transpalatine process is forked in a man-
ner similar to that seen in the cardueline finches. The lateral
branch of the transpalatine process is associated with the fibers
inserting on the medial side of the mandible while the medial
branch is associated with the fibers inserting on the medial process
of the mandible. Thus, the changes in the structure of the jaw
muscles in the fox sparrow are exclusively toward the cardueline
condition. I have dissected a specimen of the Lincoln sparrow
(Melospiza lincolnii) and a specimen of the song sparrow (Melo-
spiza melodea, Fig. 17A). These species are smaller than the fox
sparrow and consequently have smaller (both absolute and rela-
tive) jaw muscles. Nevertheless, the structure of their M. pseudo-
temporalis and their M. pterygoideus ventralis lateralis is similar
to those seen in the fox sparrow. It is, however, not certain
whether these species lack the lateral tendon as seen in the fox
sparrow, or have the tendon which was overlooked because of
the small size of these species.

In the emberizine finches, all of the adductor muscles of the
mandible and the retractors of the palate have increased in size.
This increase is relatively ‘‘even’’ in that one adductor or re-
tractor has not assumed a highly dominant role in closing the
bill. Sims (1955, p. 382) points out that the ‘‘division of labor’’
between the several muscles which close the bill has two import-
ant attributes. First, it spreads the origin of these muscles and
hence the strain on the bones over a larger area of the skull.
Second, the ‘‘harmful’’ components of force are counteracted.
For example, the M. adductor mandibulae tends to pull the
mandible backwards and outwards as well as upwards. These
backward and outward forces are counteracted by the M. ptery-
goideus which pulls the mandible inward and forward as well as
upward and by the M. pseudotemporalis superficialis and the
M. p. profundus, both of which have inward and backward com-
ponents as well as upward components of force. If only one of
these adductors were powerfully developed, as for example, the
M. adductor mandibulae, it might put uneven forces on the
mandible and possibly might even disarticulate it during a par-
ticularly powerful contraction. This would, however, never hap-
pen because the jaw muscles function and evolve as a unit. Con-
sider, for example, a bird which is becoming a seed-eater and
thus subject to a selection force for a larger bill and stronger jaw
muscles. As soon as one adductor begins to become dispropor-
tionally large, it would put an uneven strain on the mandible.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 419

The other adductors must enlarge to counteract its ‘‘harmful’’
components of force or the bird will have a selective disad-
vantage. If the other adductors did not enlarge, the bird would
be selected against long before the one muscle became large
enough to disarticulate the mandible.

Compared to a warbler or a vireo skull, the skull of an emberi-
zine finch is a more substantial structure with a shorter and
heavier bill and a stouter palate. Yet, it cannot be called a
reinforced skull, for the interorbital septum and the anterior
part of the interpalatine space are both unossified. Nevertheless,
there are other skeletal adaptations for seed-eating, such as the
bony processes on the posterior wall of the orbit and on the
lateral side of the skull, which are directly correlated with the
inerease in the mass of the adductor muscles, but these do not
need to be considered separately. One of the most important
features of the skull is the fact that the upper jaw has retained
its kinetic property, which plays a large role in the seed-cracking
method of the emberizines.

In the emberizines, a seed to be cracked is held between the
jaws just anterior to the angle of the mandible. This is approxi-
mately the point where the palatine meets the premaxilla, where
the nasal process of the maxilla meets the body of the maxilla,
and where the horny covering of the upper jaw ends. This point
is just anterior to the insertions of the adductor muscles —
hence as close as possible to the jaw articulation and the point
where the maximum force may be exerted on the seed — and yet
it is still the most reinforced part of the skull. When the
adductor and retractor muscles contract, the upper jaw is de-
pressed and the lower jaw is raised until the seed coat is cracked.
This pincer action can be compared to the cracking of the shell
of a nut by means of a nutcracker or a pair of pliers. The chief
advantage of the ‘‘nutcracker’’ method is that the initial shocks
are borne by the jaws which form a system partially isolated
from the brainease. This system permits the retention of a
lighter skull and eliminates the need for reinforcement of the
brainease. A light skull has a lower inertia which means that
smaller muscles are needed to move it —a distinct advantage for
a flying animal. Such a light skull and a faster-closing bill enables
the emberizine finches to capture insects, but also limits them
to smaller seeds.

The cardinaline finches. The cardinaline finches feed, as a
rule, more exclusively on seeds and perhaps on larger seeds than
do the emberizine finches ; therefore, it is not surprising that they

420 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

possess more specialized modifications for seed-eating. They have
a shorter and more conical bill with a more decurved upper jaw
and a greater angle in their mandible than the emberizines. The
most conspicuous specialization, however, is the palatine process
of the premaxilla, which lies free of the palatine and is situated
in the space between the palate and the jugal bar. The other
osteological features of the cardinaline skull will be described
below.

The jaw muscles of the indigo bunting (Passerina cyanea,
Figs. 17C and 17D), when viewed through the orbit, are very
similar to those of the field sparrow. The M. adductor mandibu-
lae has expanded over a larger area of the roof of the skull, but
otherwise the differences between these species appear to be
minor ones of proportions. However, the M. pterygoideus of
the indigo bunting, when viewed from beneath, is unlike that
seen in the field sparrow. The superficial fibers on the lateral
half of the M. p. ventralis lateralis form a distinct bundle that
originates from the free palatine process of the premaxilla by
means of a separate tendon. These fibers and tendon would cor-
respond to the lateral fibers and tendon seen in the towhee.
Examination of the rest of the M. pterygoideus shows that the
palatine retractors are still well developed.

The cardinal (Cardinalis cardinalis, Figs. 17E and 17F) is
one of the largest-billed members of this group. Most of its jaw
muscles, as seen through the orbit, are similar to those of the
indigo bunting except for the M. adductor mandibulae, which
has become larger and has spread over the roof of the skull. In
fact, this muscle leaves a clearly visible muscle scar outlining
its area of origin on the roof of the skull. The M. pterygoideus,
especially its lateral subdivisions, has also enlarged greatly.
Again, the lateral and superficial part of the M. p. ventralis
lateralis originates by means of a tendon from the palatine
process of the premaxilla. These fibers comprise less than 5 per
cent of the total mass of the M. pterygoideus, not 25 per cent as
Tordoft estimates. The functional significance of this separate
bundle of fibers will be discussed later. It is interesting to note
that there has been no expansion of the medial part of the M.
pseudotemporalis superficialis in the cardinals. Possibly, the
genetic capacity for this structure had never appeared in the
cardinals or perhaps this muscle cannot function in harmony
with the superficial bundle of the M. p. ventralis lateralis.

The fact that the cardinals feed more exclusively on seeds
is reflected in the structure of their skull as well as in the jaw

BOCK: PALATINE PROCESS OF THE PREMAXILLA 421

muscles. The bill is shorter and broader, and the bones of the
palate are stouter than those in the emberizine finches. The
entire skull is reinforced; the nasal septum and the anterior
interpalatine space are ossified, the maxillo-palatines are fused
to the vomer and the nasal process of the maxilla is at right
angles to the body of the mandible and parallel to the force on
the upper jaw. Yet, the upper jaw has retained its kinetic
property —a fact that is reflected in the fusion between the
vomer and the maxillo-palatines and in the medial ossification
at the jugal-maxilla connection, both of which are absent in the
eardueline finches.

The cardinals crack seeds by the nutcracker method as has
just been described for the emberizines ; the mobility of the upper
jaw permits the use of this method. Thus the cardinals do not
need heavy bosses on the upper jaw to protect the braincase. The
function of the separate bundle of fibers of the M. p. ventralis
lateralis is still a problem. Obviously it serves some particular
function, for its structure is relatively constant within the
cardinals — an indication that a selection force responsible for
its maintenance is present. These fibers do not appear to play
a vital part in cracking seeds; the other adductors of the
mandible and retractors of the palate are many times more
massive than this bundle of fibers and are probably able to
erack seeds without any aid from these superficial fibers of the
M. p. ventralis lateralis. Because of their greater length and
their insertion on the mandible near its articulation and on the
ventral rim of the medial process of the mandible, these fibers
appear to have as their chief action, the raising of the mandible.
These fibers would raise the mandible rapidly because of their
insertion on the medial process as has been discussed above (p.
392). I suggest, therefore, that the function of the superficial
fibers of the M. p. ventralis lateralis is to raise the mandible
quickly until the seed or insect is grasped firmly between the
jaws. The more massive adductors and retractors would then
take over the task of cracking the seed. If this assumption is
correct, the origin of the separate bundle of fibers and the free
palatine process is a mystery. It does not appear to be an essen-
tial modification for seed-eating ; indeed, it is somewhat contrary
to what would be expected. Perhaps it is a specialization for a
fast-closing bill to allow the cardinals to feed on insects as well
as on seeds, or perhaps it is a modification of a similar specializa-
tion in the insectivorous ancestors of the cardinals (possibly the
tanagers?).

422 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

The cardueline finches. The ecarduelines differ from the em-
berizines and agree with the cardinalines in feeding more on
seeds (in fact, the carduelines feed almost exclusively on seeds),
and in having a shorter, conical bill with a decurved upper
jaw and a greater angle in the mandible. The chaffinch and
brambling (Fringilla) are exceptions and resemble the emberi-
zines in the structure of their skulls. The similarity between
the cardinals and the ecarduelines is a superficial one, for these
eroups are strikingly different in the structure of the skull and
in the arrangement of the jaw muscles. For example, in the
cardueline finches, the palatine process of the premaxilla is
completely fused with the palatine, and in its stead is a lateral
flange as described above. Again Fringilla differs from the rest
of the cardueline finches in having an unfused palatine process
in the adult and in lacking completely the lateral flange.

The jaw muscles of the goldfinch (Spinus tristis, Figs. 19E
and 19F), as seen through the orbit, are quite different from
those in the emberizine finches or the cardinals. The M. adductor
mandibulae is relatively large for a bird the size of a goldfinch,
with the portion spread over the roof of the skull doubled —
a condition not seen in any other passerine family examined in
this study. The M. pseudotemporalis superficialis has enlarged
unevenly. Only the anteromedial part of this muscle has enlarged
to a great degree; the lateral parts of the M. p. superficialis
appear as a small isolated muscle sandwiched between the larger
medial portion and the M. adductor mandibulae. The large
medial portion of the M. p. superficialis almost completely covers
the M. p. profundus. Those muscles associated with the raising
and lowering of the upper jaw, the M. p. profundus and the M.
protractor quadrati, are relatively small muscles with fleshy
origins and insertions. Turning to the ventral side of the head,
the large M. pterygoideus can be seen. Again, the mandible
adductor parts of this muscle have enlarged while the palatine
retractors have decreased in size. Special note should be taken
of the M. p. ventralis lateralis. It takes origin only from the
transpalatine process; no muscle fibers or tendons run forward
to attach to the palatine in the region of the lateral flange. As
in the fox sparrow, the tip of the transpalatine process is forked ;
the lateral branch is associated with the fibers running to the
ramus of the mandible while the medial fork is associated with
those fibers running backwards to the medial process of the
mandible.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 493

If the jaw muscles of a medium-billed species, such as the
purple finch (Carpodacus purpureus, Fig. 20C), and those of a
large-billed species such as the evening grosbeak (Hesperiphona
vespertina, Figs. 20A and 20B), are examined, two important
changes from the goldfinch condition are discernible. First, the
M. adductor mandibulae has increased in size until, in the
evening grosbeak, its origin spreads over most of the roof of
the skull and leaves a well defined muscle scar. Second, the
medial part of the M. pseudotemporalis superficialis has increased
in mass to dominate the muscles inside the orbit. It completely
obscures the M. p. profundus and almost completely hides the
M. protractor quadrati and the lateral part of the M. pseudo-
temporalis superficialis. The two major dorsal adductors of the
mandible —the M. adductor mandibulae and the medial part
of the M. pseudotemporalis superficialis — converge upon the
mandible from the outside and the inside respectively — an
excellent example of two muscles so placed that their ‘‘harmful’’
effects are counteracted. There are no significant changes other
than increase in mass in the structure of the M. pterygoideus.

Sims (1955) has reported on the jaw muscles of the hawfinch
(Coccothraustes coccothraustes), a species very similar and ap-
parently closely related to the evening grosbeak. Unfortunately,
his excellent analysis of the skull and the jaw muscles is marred
by several misidentifications, such as the M. quadrato-mandibu-
laris in his figure 4B (this is actually the enlarged medial part
of the M. pseudotemporalis superficialis) and the M. p. ventralis
lateralis anterioris in his figure 5B (this is probably the M. p.
dorsalis lateralis).

I was fortunate in being able to examine two specimens of the
chaffinech (Fringilla coelebs, Figs. 19C and 19D). In most re-
spects, the jaw muscles are similar to those of the least special-
ized cardueline finches, although they are not as powerful. The
most significant feature of the dorsal adductors is the enlarged
medial portion of the M. pseudotemporalis superficialis. This
muscle is identical to that in the heavier-billed cardueline finches
and, in fact, it completely covers the M. p. profundus, as in the
evening grosbeak. The M. adductor mandibulae is larger than
expected ; its origin has expanded over as large an area of the
roof of the skull as in the larger towhee. However, the M.
adductor mandibulae of the chaffinch is not as specialized as in
the carduelines, but is similar to that seen in the emberizine
finches (see Fiedler, 1951, pp. 241-242). Turning to the ventral
aspect of the M. pterygoideus, we find that it is almost identical

494 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

to that seen in the goldfinch, the main difference being that the
chaffinch has a less massive muscle. The unfused palatine
process of the premaxilla was clearly visible, but there was no
connection between it and the M. p. ventralis lateralis; in fact,
the M. pterygoideus of the chaffinch is very reminiscent of that
seen in the fox sparrow.

The cardueline finches (with the exception of Fringilla, Eber,
1956) feed almost exclusively on seeds and hence have a massive
reinforced skull, one that is even heavier than the cardinaline
skull. The interorbital septum, the nasal septum, and the inter-
palatine space are more heavily ossified in the carduelines than
in the cardinals. The most conspicuous difference between the
two groups is the lack of a free palatine process and the develop-
ment of a lateral flange on the prepalatine process with an over-
lying horny pad of rhamphotheea in the cardueline finches. The
upper jaw has lost most of its mobility, but the fact that it is
not rigidly fused to the cranium as stated by Sims (1955, p.
373) could be ascertained by boiling skulls of Hesperiphona and
Coccothraustes for a minute or two as suggested by Beecher
(195la, p. 412). This technique softens the dried ligaments and
restores flexibility to the skull. However, Sims’ conclusion is
still correct, for the upper jaw is essentially stationary during
the closing of the bill. The immobile upper jaw plus the presence
of the heavy bosses of bone and rhamphotheca suggest that the
cardueline finches employ a method other than that of a nut-
eracker to crack the seed shell.

A seed to be cracked by a ecardueline finch is placed in the
corner of the mouth, just anterior to the angle of the mandible.
The seed lies between the heavy pads of the upper jaw and those
of the lower jaw and is held in place by the tongue, as shown by
Eber (1956). Upon contraction of the adductor muscles, the
mandible is raised and foreed against the seed until its shell
cracks. In this way, the apparatus resembles the action of a vise.
Since the upper jaw is continuous with the brainease, the crack-
ing shock must be transmitted across the skull without harm
to it or to the contained organs. The heavy bosses on the upper
jaw provide an even distribution of the shock, protecting the
brainecase and the brain from injury. Perhaps the slight amount
of mobility of the upper jaw may partly absorb the shock wave
that accompanies the actual cracking of the seed. The heavy
pads of rhamphotheca may serve to absorb some of the shock
wave, but this is open to question. The vise method is intrinsically

Or

BOCK: PALATINE PROCESS OF THE PREMAXILLA 42t

no more efficient than the nutcracker method, but a heavier seed
ean be cracked with the vise method since the bony elements in-
volved are inherently larger. Not only the palatine complex, but
the entire skull is used to transmit the forees and shocks of seed-
eracking ; hence larger forces are possible for an equal amount
of stress on the bone. Sims has shown that the hawfinch must
exert a force of 100 to 150 pounds when it cracks an olive stone.
However, the powerful vise system is developed at the expense
of the mobility of the upper jaw which limits the cardueline
finches to a rather exclusive diet of seeds (Eber, 1956). The
‘<slender-billed’’ chaffinch does not have such a specialized bill
and feeds extensively on insects during the summer. It prob-
ably uses the ‘‘nutcracker’’ method of cracking seeds rather than
the ‘‘vise’’ method.

The above descriptions of the nutcracker and the vise methods
of seed-cracking, being based completely upon deductive reason-
ing from the structure of the skull and the jaw muscles in several
groups of finches, have raised more problems than they have
solved. First, direct observations or experiments are needed
to verify my conclusions as to function and to establish the
exact morphological and functional differences between the two
methods. It must be emphasized that the important functional
difference between the nutcracker and the vise methods of
cracking seeds is not in terms of applying force on the seed, but
in terms of protecting the braincase from the shocks associated
with the breaking of the seed coat. In the nutcracker method,
these shocks are isolated within the jaw apparatus, while in the
vise method, these shocks are distributed evenly to the braincase
by the heavy bosses of bone and rhamphotheca and thus are pre-
vented from concentrating on one bony element. The method
used in each group of finches and the degree of specialization in
each group must be determined. The difference between the
extreme types, such as the cardinals and the carduelines, is
clear, but that between the less specialized forms is quite fuzzy.
For example, the emberizine finches are best classified as having
a ‘‘primitive’’ nutcracker method of seed-cracking. Where is
the boundary between the primitive nutcracker method and the
specialized nutcracker or the vise methods, both of which are
presumed to have evolved from the emberizine condition? I
have assumed for the purposes of this paper that the presence
of a lateral flange indicates that the vise method is used. There-
fore, the estrildids, the advanced ploceids, Oryzoborus of the

426 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Emberizinae and Psittirostra of the drepaniids would use the
vise method of seed-cracking ; however, there is no available evi-
dence supporting a correlation between the lateral flange and
the vise method. These problems must be solved before we can
obtain a complete picture of the evolution of seed-eating in the
passerines and of the origin of the several finch-like groups.

Other finches. The jaw muscles of three other seed-eating
passerines, the horned lark (Hremophila alpestris, Figs. 12C and
12D), the ecowbird (Molothrus ater, Figs. 18K and 18F), and
the house sparrow (Passer domesticus, Figs. 21C and 21D) were
dissected. Although seeds comprise a large part of the diet of
these species, their jaw muscles are relatively unspecialized
when compared to those of the cardinaline and the cardueline
finches. The jaw muscles of the house sparrow and the cowbird
are similar to one another and to those found in the smaller
emberizine finches and show the usual development of the
adductor muscles. Neither species has any specialization of the
M. pseudotemporalis superficialis or the M. pterygoideus; both
have an expansion of the M. adductor mandibulae over the side
of the skull to about the same extent as in the towhee. The medial
part of the M. adductor mandibulae is expanded in both species
and bulges inward to cover part of the M. pseudotemporalis pro-
fundus. This slip of the M. adductor mandibulae has not become
specialized in either species. The upper jaws of the house spar-
row and of the cowbird are kinetic, suggesting that these birds
probably use the nutcracker method of seed-cracking. In almost
all respects, both species are comparable to the emberizine finches
in the specialization of their seed-cracking modifications.

The jaw muscles of the horned lark are the most interesting
because they have evolved a quite unique specialization (for seed-
eracking ?) not found in any other species of passerine birds I
have examined. The M. adductor mandibulae has not spread over
the side of the skull roof as in other seed-eating birds. Instead,
the medial part of this muscle has enlarged and become pinnate.
This large medial slip of the M. adductor mandibulae may be
able to provide much of the force needed for cracking seeds that
is usually supplied by the lateral parts of this muscle and by
the M. pterygoideus, both of which are poorly developed in
this species. It is also conceivable that this pinnate slip has a
special function that is totally independent of seed-cracking.
Larks (all species ?) swallow seeds whole and thus would not
need strone jaw muscles. More work correlating the feeding

BOCK: PALATINE PROCESS OF THE PREMAXILLA 427

methods with the jaw muscles of the larks is needed before
this problem is solved.

Résumé on the feeding methods in the finches. A brief review
of the methods of seed-cracking in the three groups of ‘‘nine-
primaried finches’’ and the associated adaptations may now be
given. The emberizine sparrows (plus the house sparrow and
the cowbird) can be regarded as ‘‘generalized’’ seed-eaters,
lacking the extreme specializations for seed-cracking. The ad-
ductors of the mandible and the retractors of the palate are
equally well developed, and the upper jaw has not lost its kinetic
property — the result is the nutcracker method of seed-cracking.
Some of the larger-billed emberizines (e.g., the towhee) tend
toward the cardinals in their jaw muscles, especially in the strue-
ture of the M. pterygoideus; others (e.g., the fox sparrow) tend
toward the cardueline finches. The cardinals are heavy-billed
forms that have retained the nutcracker method; hence the skull
is still relatively light. Increase in the mass of the M. ptery-
eoideus ventralis lateralis and its lateral tendon has resulted in
the free palatine process of the premaxilla. The free palatine
process and the movable upper jaw probably preclude the de-
velopment of the lateral flanges on the prepalatines as seen in
the cardueline finches. The second group of heavy-billed finches
—the carduelines — employ another method of cracking seeds,
the vise method. However, a presumed ‘‘primitive’’ genus,
Fringilla, is quite similar to the fox sparrow in the structure
of its jaw muscles and probably uses the nutcracker method. In
the specialized carduelines, the upper jaw has lost its mobility ;
it is a nearly stationary block against which the mandible presses.
Heavy bosses of bone (the lateral flanges) and rhamphotheeca
distribute the shocks associated with the cracking of the seed
evenly to all parts of the brainease. Only the adductor muscles
are well developed in the cardueline finches; in fact, the muscles
associated only with the movement of the upper jaw have become
small and are on the verge of becoming functionless.

Conclusions. The primary function of the palatine process of
the premaxilla is to provide a point of anchorage for the pala-
tines and hence promote a stronger connection between the palate
and the upper jaw. Secondary functions have been superimposed
on this primary function in several groups of passerine birds,
but these secondary functions do not conflict with the operation
of the primary function. The secondary functions have been
responsible for the modifications of the palatine process — the
development of the free process and the lateral flanges — within

428 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

the passerines. Using the data assembled in the preceding pages,
I wish to return to the three problems associated with the modifi-
cations of the palatine process which have been listed in the
beginning of the section on ‘‘modifications.’’

a) Development of the free palatine process: In most groups
of passerine birds, there is no connection, either morphologically
or functionally, between the M. pterygoideus and the palatine
process of the premaxilla. Only in a few groups, such as the
cardinals and the tanagers, does part of the M. p. ventralis
lateralis take origin from the palatine process. These groups
have a free palatine process. Therefore, it may be concluded
that the free palatine process as seen in the cardinals serves as
the point of origin for the lateral superficial bundle of the M. p.
ventralis lateralis. This modification is probably not a special-
ization for seed-eating as commonly believed, but for rapid
rising of the mandible during the early phases of closing the bill,
which appears to be a specialization for catching insects.

b) Development of the lateral flange: In a few groups, lateral
flanges have developed at the site of the fused palatine processes.
These birds have acquired a powerful set of jaw muscles for
seed-cracking, but have lost the mobility of the upper jaw. They
employ the vise method of cracking seeds in which the seed is
cracked by raising the mandible against the stationary upper
jaw. The lateral flange plus the overlying pads of rhamphotheca
serve to distribute the shock wave (associated with cracking the
seed) evenly to all parts of the upper jaw and the braincase.

c) Variation in the degree of fusion between the palatine
process and the palatine: Much variation exists in the degree of
fusion between the palatine process of the premaxilla and the
palatine and in the development and final degeneration of the
isolated splint of bone (= the posterior end of the palatine
process), such as seen in the emberizine finches. No sharp break
exists between a slightly fused palatine process and a free
process, or between a heavily and a completely fused palatine
process. Tordoff claimed that the functional basis for this varia-
tion in fusion from a free process to a completely fused process
is the decrease in size of the M. pterygoideus, starting from the
cardinal condition. However, in most passerine birds, there is
no morphological connection between these two structures; those
groups in which part of the M. pterygoideus originates from the
palatine process have been discussed above. In the emberizine
finches, a tendon from the lateral fibers of the M. pterygoideus
originates on or near the palatine process, but there is not the

BOCK: PALATINE PROCESS OF THE PREMAXILLA 429

simple correlation between the size of this muscle and the fusion
of the palatine process as claimed by Tordoff. In fact, the pala-
tine process of the towhee is usually completely fused although
this species has a larger lateral tendon of the M. pterygoideus
than in all other emberizines studied. Therefore, it can be con-
eluded that the variation in the M. pterygoideus is not respon-
sible for the variation in the amount of fusion between the pala-
tine process and the palatine.

This variation in the degree of fusion is the result of two quite
different factors. Part is doubtlessly the result of a difference
in the strength of the primary selection force. It is obvious that
there could be a need for a firmer connection between the upper
jaw and the palate in some birds and consequently, these birds
would have a more heavily fused palatine process. For example,
the difference in the fusion of the palatine process in the several
genera of woodhewers (Dendrocolaptidae) or between Sapayoa
and the other genera of manakins (Pipridae) is probably the
result of a difference in the strength of the primary selection
force. However, this explanation accounts for only part of the
variation. So long as the palatine process fulfills its primary
function as an anchorage to which the palatine can fuse, then
it does not matter what else happens to it. If the demands of
the primary selection force are fulfilled and if there are no
other selection forces acting on the palatine process (e.g., for a
free process or for a lateral flange), then it would not seem to
matter whether the process fuses completely to the prepalatine
process, remains partly unfused or starts to degenerate into an
isolated splint lying along the prepalatine process. Therefore,
I will conclude that, except for the differences in the demands of
the primary selection force, there seems to be no functional basis
for the variation in the amount of fusion between the palatine
process and the palatine, or in the development of the isolated
splint. Since this variation in the degree of fusion lies within
the limits of the primary selection force, the change from an
unfused to a fused process does not appear to be an adaptive
change. It would be, however, an evolutionary change, perhaps
as pleitotropic expression of a complex gene system.

VARIATION OF THE PALATINE PROCESS OF
THE PREMAXILLA

Having ascertained the embryological and the functional
significance of the major variants of the palatine process of the
premaxilla, it is now necessary to survey its occurrence and

430 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

variation throughout the Passeres. In this section, the structure
of the palatine process will be recorded, genus by genus, for all
but a few families of passerine birds. An attempt was made to
separate the observed variation into its separate components —
adult, individual, functional and so forth —but it is realized
that the conclusions are provisional until good age series and
samples of known adults are obtained. These components of
variation are the most important part of this survey, and I will
summarize them now instead of at the usual place at the end of
the section.

The differences in the palatine process between families and
occasionally between subfamilies and genera of passerine birds
are chiefly functional ones of the type discussed in the preceding
section. These variants with their functional interpretation fol-
low.

a) The palatine process unfused to completely fused with the
palatine. Slight variants of this series include (a) the anterior
end of the process partly degenerate to degenerate, leaving the
rest of the process isolated from the premaxilla, and (b) the
posterior part of the palatine process fused to the dentary proc-
ess of the premaxilla and/or to the maxilla. This series repre-
sents the stages associated with the varying strength of the
primary function of the palatine process which is, of course, to
provide a point of anchorage to which the palatine can fuse.
When the palatine process fuses to the dentary process of the
premaxilla, at least its anterior half fuses to the palatine. The
exact reason for the fusion with the dentary process remains,
however, unclear. The major problem of this variation in fusion
is that it is identical to some aspects of age variation and to non-
functional variation in the degree of fusion in the adult; these
will be discussed below. No satisfactory means of separating
these morphologically similar, but basically different, variants is
available. Thus if a series of genera exhibit differences in the
degree of fusion between the palatine process of the premaxilla
and the prepalatine process, we cannot tell whether the variation
is functional, or age variation. However, the differences in the
degree of fusion between families, such as that between the
heavy-billed icterids and the lighter-billed New World warblers,
are frequently functional.

b) The distal end of the palatine process lying free of the pre-
palatine process and in the space between the palate and the
jugal bar. The free process serves as the point of origin for part

BOCK: PALATINE PROCESS OF THE PREMAXILLA 431

of the M. pterygoideus which is apparently associated with rapid
raising of the mandible.

ce) A lateral flange developing at the site of fusion of the
palatine process. The flange serves to distribute the shocks as-
sociated with cracking of seeds by the vise method employed by
the cardueline and other groups of finches.

The variation of the palatine process within a family of pas-
serine birds is commonly quite complex and must be carefully
separated into its several components before offering any state-
ments on the variation of the process within the family or the
contained genera. Earler authors (Lucas, 1894, p. 304; Ama-
don, 1950a, p. 214; and Tordoff, 1954a, p. 23) differ in their
opinions on the amount of variation exhibited by the palatine
process in a genus of passerine birds. However, their arguments
are ambiguous for they did not separate the observed variation
into its several components. The following types of variation
can be expected when examining the palatine process within a
family or genus of passerines.

a) All types of functional variation described above may oc-
eur between genera of a family, but they are infrequent. Ex-
amples may be: the differences in the fusion of the process in
the several genera of woodhewers (Dendrocolaptidae) or between
Sapayoa and the other genera of the manakins (Pipridae), or
the free process in Melopyrrha and Tiaris and the lateral flange
in Oryzoborus as compared to the normal palatine process seen
in the other genera of emberizine finches. I do not know of any
definite examples of functional differences between congeneric
species.

b) Individual variation between adults of a species (or be-
tween adults of a genus as recorded in this study) is difficult to
ascertain because of the variation in age. There are many ex-
amples in which there is no variation between the adults of
a genus, usually when the process is completely fused in the
adult. If the process is only partly fused in the adult,
then it is impossible to determine the adult individual variation
without a series of known adults. These were not available.
Therefore, I will make the assumption for the purposes of the
present study that there is no significant adult variation within
genera of passerine birds. This assumption is obviously not
correct, but it is very likely not far from the actual condition
and it greatly simplifies the study of the minor variations in the
degree of fusion between the palatine process and the surround-
ing bones.

432 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

¢) The variation resulting from differences in the age of the
specimens is of extreme importance because the final adult con-
dition of the palatine process is not reached until the bird is
between six months to a year old, because the expression of age
variation is very similar to adult individual variation and to
certain aspects of functional variation, and because many speci-
mens in skeleton collections are birds less than a year old.
Karlier workers have ignored age variation completely and as a
consequence, have offered some erroneous conclusions on the
amount of individual (adult) and even of functional variation.
While examining the specimens used in the survey, I was able
to separate some immature birds from adults on the basis of the
ossification of the parietal windows and the general ossification
of the skeleton, and to compare these groups for differences in
the amount of fusion of the palatine process. In general, the
palatine process becomes more fused with increasing age. The
danger of age variation is that many genera are represented in
collections by one or a few specimens so that there is no sure
way of determining whether the structure has reached its final
adult condition in the bird being examined and hence whether
comparisons with other species are proper, i.e., adult with adult
and not adult with immature. Realization of the dangers result-
ing from confusing age variation with adult individual or with
functional variation has made me cautious in my discussions of
the last two aspects of variation. Consequently, because much
of the observed variation in the degree of fusion of the palatine
process is the result of differences in the age of the specimens
and because fewer erroneous conclusions can be based on age var-
iation, I have assumed that all of the variation in fusion with
the exception of obvious cases, such as between the genera of
woodhewers, is the result of differences in the age of the speci-
mens. Again, this assumption is possibly wrong, but it is the
safest one to use until we have enough good series of specimens of
known age.

Methods. The original survey and a major share of the study
of the structure and variation of the palatine process of the pre-
maxilla was carried out in the American Museum of Natural
History — the entire passerine section of this skeleton collection
was utilized. Subsequent study of material in the United States
National Museum filled the gaps left after the original survey
was completed. All in all, some 3300 specimens representing 500
genera of all but a few passerine families were examined. The
missing families are all small groups which do not affect the

€

BOCK: PALATINE PROCESS OF THE PREMAXILLA 433

overall picture in the Passeres. I believe that an adequate sample
of most families, certainly of all the important ones, was ob-
tained. In addition to the Passeres, the Pici and the Coraciae
were examined to acquire an idea of the structure of the palatine
process in these groups that are believed to be the nearest rela-
tives of the Passeres.

Only those skulls in which the rhamphotheca of the upper
jaw and the membranes in the anterior half of the palate have
been removed could be used in the survey. The faint suture
separating the palatine process of the premaxilla and, in many
eases, the palatine process itself is obscured if there is a thin
layer of overlying tissue. A binocular microscope with a mag-
nification of 10x to 15x was used in all examinations.

The structure and variation of the palatine process was de-
termined in each genus available to me. Recording was done on
the generic level because, with few exceptions, little or no varia-
tion was found between congeneric species; indeed, usually little
variation was found between genera of the same family. Also in
many eases, the specific identification was lacking or doubtful,
and many old names were used so that the task of separating
the species and determining the synonymies would have been
far greater than justified by the returns. Particular attention
was paid to age variation. I tried to assemble a large series of
at least one genus in every family, from which the ‘‘typical’’
adult condition of the palatine process and the age variation
was determined. The approximate age of the specimens was
judged by the general degree of ossification of the skull and
especially of the parietal windows (see Chapin, 1949, for a
discussion of the parietal windows as a criterion of age in the
passerines).

In the following synopsis, the structure of the palatine process
is summarized for each family, and in a few cases for the sub-
families, of passerine birds. As mentioned above, the families
will be listed according to the sequence to be used in “‘ Peters’
Check-lst.’’ Under each family (or subfamily) heading, a
general description of the palatine process and of its major
variation in that family is given. Usually one genus is chosen
as the basis for description and comparison in the family. I
have tried to choose for this purpose the genus having the most
typical palatine process, but this genus is not to be regarded as
the type of the family or even as completely typical for that
family. If possible, the probable course of ontogenetic develop-
ment is given. The genera examined are then listed. For each

434 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

genus, the number of specimens examined is given in parenthesis
and a brief statement of the structure of the process is given.
The genera within a family are listed alphabetically rather than
according to supposed relationships. A far more complex
problem is which generic names should be used. The labels on
many specimens bore names no longer used; hence there was
the huge task of synonymizing the discarded names with the
currently accepted ones. The most convenient solution to this
problem was to accept the most recent generic revision of each
family. Peters (1951) was followed for the few groups of the
suboseines treated in the last volume of his ‘‘Check-list.’’ Hell-
mayr (1924-1938) was consulted for the remaining New World
families, and various authors for the Old World groups. In some
cases there has not been a recent revision of the family, and
for these groups I have used the material in Mayr and Amadon
(1951) or have followed the arrangement of genera used in the
study collection in the American Museum of Natural History.
In any event, the authority followed is cited for each group. I
must emphasize that this system is used only because of con-
venience and not because I may advocate the generic limits set
forth by these authors. Nor does it make much difference
whether a broad or a narrow generic concept was employed in
these revisions; the results of this paper would be the same
with either concept.

The schematic drawings accompanying the descriptions (Figs.
23 to 28) show the left half of the upper jaw as seen from
below. The arrow points to the palatine process of the pre-
maxilla (labeled ‘‘P’’) or to the region of the fused palatine
process. The figures are not drawn to scale.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 435

Figure 23. Palatine process of the premaxilla (P) of: (A) Smithornis
(Hurylaimidae); (B) Hypocnemis (Formicariidae); (C) Fumarius (Fur-
nariidae); (D) Formicarius (Formicariidae); (E) Formicarius (Formi-
eariidae); (F) Agriornis (Tyrannidae); (G) Tyrannus (Tyrannidae) ;
(H) Cnipodectes (Tyrannidae); and (1) Euchlornis (Cotingidae).

436 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

SYNOPSIS OF PASSERINE FAMILIES
Eurylaimidae

The process is present with a varying degree of fusion, from unfused
up to the point of being completely lacking (= fused). ‘‘Lacking’’ is a
neutral term and will be used when the process is not present as a visible
structure, instead of the term ‘‘fused’’ which assumes that the process
was present in the embryo and has become fused to the palate. In Corydon,
the process is present and bears some resemblance to an ossified tendon
(of the M. pterygoideus ?). All specimens of Smithornis (Fig. 234A) have
the process, but the amount of fusion between it and the palate varies
greatly. This may, in part, explain the absence of the process in Calypto-
mena, in which I have assumed that the palatine process is fused to the
palatine. Lowe (1924) described and figured the palate of Cymbirhynchus,
Pseudocalyptomena and Smithornis, but makes no mention of the process.
Pycraft (1905a) did not mention the process in Calyptura, Chasmorhynchus,
Corydon, and Cymbirhynchus, but did say that the anterior process of the
palatine is broader where it joins the premaxilla which indicates that the
process has fused to the palatines. Thus, it may be concluded that the
typical adult condition of the palatine process in the broadbills is for the
process to be fused to the palatines and that most, if not all, of the
variation observed is because of differences in age and hence the degree of
ossification.

Calyptomena (1), lacking; Corydon (2), present, somewhat resembles an
ossified tendon; Smithornis (4), present, amount of fusion varies. Checked
with Peters (1951).

Dendrocolaptidae

The process is present in all genera of woodhewers, but the amount of
fusion between the process and the palatine varies greatly in the different
genera. As a broad generalization, the process is only slightly fused in the
long, curved-billed species and heavily fused in those species having a short,
straighter and heavier bill. Some of the observed variation may be age,
but some is almost certainly correlated with the strength of the bill.

Dendrocincla (2), heavily fused to palatine; Dendrocolaptes (1), slightly
fused; Drymornis (1), slightly fused; Glyphorynchus (2), heavily fused;
Lepidocolaptes (8), usually slightly fused, but some more fused; Sittasomus
(2), slightly fused; Xiphorhynchus (22), usually slightly fused, but in the
medium- to heavy-billed species, the process is as heavily fused as in Dendro-
cincla, Checked with Peters (1951).

Furnariidae

In most genera, the process is present, lying along the palatine and not
fused with that bone. In Furnarius (Fig. 23C), the process lies free in
the space between the palate and the dentary process of the premaxilla
and in some ways resembles an ossified tendon. Rarely is the process fused

BOCK: PALATINE PROCESS OF THE PREMAXILLA 437

A

Figure 24. Palatine process of the premaxilla (P) of: (A) Progne
(Hirundinidae); (B) Anthus (Motacillidae) ; (C) Acanthorhynchus (Meli-
phagidae, redrawn from Parker, 1877); (D) Spizixos (Pycnonotidae) ; (E)
Spizizos (Pycnonotidae); (F) Spizizos (Pyenonotidae); (G) Garrulax
(Timaliinae) ; (H) Garrulaz (Timaliinae) ; and (I) Garrulax (Timaliinae).

438 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

or absent as in the case of Leptasthenura or Synallazis. The length of the
process varies somewhat between the genera.

Anabacerthia (2); Anumbius (1); Aphrastura (1), lacking ?; Asthenes
(4); Automolus (5); Chilia (1), similar to Furnarius; Cinclodes (1);
Coryphistera (1); Cranioleuca (4); Furnarius (22), present, usually lying
in the space between the palate and the jugal bars, and reminiscent of an
ossified tendon; Geositta (5); Leptasthenura (5), present, degree of fusion
varies greatly; Lochmias (1), lacking ?; Margarornis (2); Philydor (1);
Phleocryptes (3); Premnoplex (1); Pseudocolaptes (2); Pseudoseisura (2) ;
Sclerurus (2); Schoeniophylaz (1); Synallazis (13), present, varies from
unfused to fused; Upucerthia (4); Xenops (2). Process unfused unless
otherwise noted. Checked with Peters (1951).

Formicariidae

The variation of the process in this family is almost as great as that
seen in the entire order. In most genera, the process is present, lying
along the palatine and searcely fused to that bone, e.g., Taraba. In a few
genera, such as Hypocnemis (Fig. 23B) and Cercomacra, the process lies
in the space between the palate and the jugal bar, and seems to fuse, at
least in part, with the overlying bone. A considerable amount of variation
in the degree of fusion between the process and the palatine exists in
some genera, such as Thamnophilus, Phegopsis and Myrmorchilus, in which
the process varies from unfused to completely fused with the palatines. In
Formicarius (Figs. 23D and 23E) the anterior end of the process degener-
ates until in some individuals only a small isolated splint of bone remains
along the palatine. While much of the observed variation in the size and
the degree of fusion is the result of age difference and hence of ossifica-
tion, some of it probably represents true morphological difference between
genera. However, no indication of relationships between the genera of the
Formicariidae could be determined from the structure of the palatine
process.

Cercomacra (5), present, unfused to fused to the palatine; Chamaeza
(2); Cymbilaimus (2); Dysithamnus (2); Formicarius (5), present,
unfused to degenerating at the anterior end to become an isolated splint
lying along the palatine; Formicivora (7), lacking, no hint in any specimen;
Grallaria (1); Gymnopithys (2); Herpsilochmus (1); Hylophylax (2);
Hypocnemis (5) present, rather broad, unfused to fused, lying in the
space between the palate and the jugal bar; Microrhopias (1) lacking ?;
Myrmeciza (8), present, as in Hypocnemis, or lying next to the palatines;
Myrmorchilus (2); Myrmotherula (6), present, unfused to fused, as in
Hypocnemis; Percnostola (1); Phaenostictus (3), present, unfused to
fused; Phegopsis (1), lacking ?; Pyriglena (2); Sakesphorus (3); Taraba
(9); Thamnophilus (27), present, unfused to fused. Process unfused unless
otherwise noted. Checked with Peters (1951).

Figure 25. Palatine process of the premaxilla (P) of: (A) Turdus
(Turdinae); (B) Yurdus (Turdinae); (C) Melopyrrha (Emberizinae) ;
(D) Tiaris (Emberizinae); (E) Oryzoborus (Emberizinae); (F) Pipilo
(Emberizinae); (G) Melospiza (Emberizinae); (H) Melospiza (Emberi-
zinae) and (1) Melospiza (Emberizinae).

440 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Conopophagidae

Several specimens of this family were available for study, but unfor-
tunately, the palates of all were damaged. In one skull, a very faint suture
on the lateral side of the prepalatine process was seen — an indication that
the process may be present, but partly fused with the palatine. Forbes
(1881) figured the palate of Conopophaga in which the process was not
shown. This, however, cannot be accepted as definite evidence that the
process is lacking in this family because many authors were unaware of
the existence of the process or otherwise failed to mention it.

Rhinocryptidae

The process is universally absent in all specimens examined in this fam-
ily.

Pteroptochos (4); Rhinocrypta (1); Scytalopus (1). Process lacking.
Checked with Peters (1951).

Cotingidae

The process is highly variable in this family; however, it must be
remembered that the shape of the bill is also highly variable in the
ecotingas. The process may be present, but more commonly it is lacking
= fused to the palatine). When present, it usually lies next to the palatine,
but in Euchlornis (Fig. 231), it lies free in the space between the palate
and the jugal bar. In Cotinga, the process may be partly fused to the
dentary process of the premaxilla. How much of this variation in the
fusion of the process to the surrounding bones is true difference between
genera and how much is age variation cannot be determined, but certainly
at least some is age. Thus, it is most likely that the adult condition of the
palatine process for most genera of this family of broad-billed forms is a
completely fused process and that those individuals having an unfused or
a partly fused process are immature.

Ampelion (2); Calyptura (1); Cephalopterus (3); Cotinga (1); Euchlor-
nis (4), present, long, lying free in the space between the palate and the
jugal bar; Gymnoderus (1); Lipaugus (1); Pachyramphus (3), present,
unfused in two specimens, fused in the other; Procnias (7); Rupicola
(5); Tityra (3). Process lacking unless otherwise noted. Checked with
Hellmayr (1929).

Pipridae

The process is present, short and unfused in all genera of this family,
except for Sapayoa in which it is lacking (= fused). This genus is larger
in size than the other manakins and has a broader bill which may account
for its completely fused palatine process. This variation is in agreement
with the generalization that the process is more fused in those forms with
a heavier bill.

BOCK: PALATINE PROCESS OF THE PREMAXILLA

Figure 26. Palatine process of the premaxilla (P) of: (A) Caryothrau

44]

stes (Cardinalinae) ; (B) Spiza (Cardinalinae) ; (C) Piranga (Tanagrinae) ;

(D) Thraupis (Tanagrinae); (E) Buthraupis (Tanagrinae); (F) Tersina
(Tersininae); (G) Psittirostra (Drepaniidae); (H) Himatione (Drepa-
niidae); and (I) Vestiaria (Drepaniidae).

442 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Antilophia (1); Chiroxiphia (11); Corapipo (1); Machaeropterus (1) ;
Manacus (4); Pipra (18); Sapayoa (2), lacking, no hint of the process;
Tyranneutes (1). Process short and unfused unless otherwise noted.
Checked with Hellmayr (1929).

Tyrannidae

The process varies greatly in this family of broad-billed birds. It
may be present as a long process lying along the palatine (as in Agriornis,
Fig. 23F) or as a short process at the base of the palatine (as in
Myiarchus) or appear as an ossified tendon arising from the mass of bone
at the junction between the palatine and the premaxilla or rarely from the
maxillo-palatine (as in Cnipodectes, Fig. 23H) although it is not absolutely
certain whether the short process on the maxillo-palatine is the palatine
process), or may be completely absent (as in Pitangus). Even within a
single genus, such as Tyrannus, the process may vary from being present
to what appears as a tendinous process arising from the general mass of
the premaxilla (Fig. 23G). The commonest condition is for the process
to be lacking (= fused) or to vary in the degree of fusion with the pala-
tine, but always with a tendency toward greater fusion. How much of the
observed variation is age and how much is difference between genera is
impossible to determine at this time. Certainly, at least some of the varia-
tion in fusion is age and some is functional.

Agriornis (6), unfused (long-billed form); Arundinicola (2); Blacicus
(1); Cnemotriccus (7); Cnipodectes (1), present, attached to base of the
maxillopalatine; Colonia (4); Colopteryx (4), varies as in Tyrannus;
Elaenia (16), varies as in Tyrannus; Empidonaz (30), usually lacking, but
present in a few specimens; Empidonomus (3); Entotriccus (1); Euscarth-
mornis (1); Flwvicola (1); Gubernetes (1); Knipolegus (2); Legatus
(1); Lessonia (4); Lichenops (2), lacking in one specimen, present in the
other; Lophotriccus (1); Mecocerculus (2); Megarynchus (5), varies as
in Tyrannus; Muscisaxicola (2); Muscivora (9), varies as in Tyrannus;
Myiarchus (60), varies as in Tyrannus; Myiobius (1); Myiochanes (8),
varies as in Tyrannus; Myiophobus (2); Myiotheretes (1), long as in
Agriornis, Myiozetetes (24), varies as in Tyrannus; Nuttallornis (4);
Ochthoeca (4); Ochthornis (1); Oncostoma (1); Onychorhynchus (2), varies
as in Tyrannus; Phaeomyias (1); Pipromorpha (2) varies as in Tyrannus ;
Pitangus (20), varies as in Tyrannus, but most have a free process;
Pseudocolopteryx (1); Pyrocephalus (23); Pyrrhomyias (2); Rhynchocyclus
(1); Sayornis (11); Serpophaga (3); Sirystes (2), varies as in Tryannus ;
Snethlagea (1); Spizitornis (3); Sublegatus (2); Suirirt (1); Todiro-
strum (3); Tolmarchus (13), varies as in Tyrannus; Tolmomyias (4);
Tyranniscus (5), varies as in Tyrannus; Tyrannulus (1); Tryannus (52),
present, to degenerating to what appears as a tendinous process arising
from the general mass of bone of the premaxilla (this can be easily broken
and is lost in many specimens) ; Xolmis (5), unfused and long in one speci-
men, lacking in the others. Process lacking unless otherwise noted. Checked
with Hellmayr (1927).

BOCK: PALATINE PROCESS OF THE PREMAXILLA 443

Figure 27. Palatine process of the premaxilla (P) of: (A) Vireo
(Vireonidae); (B) Icteria (Parulidae); (C) EHuphagus (Icteridae); (D)
Gymnostinops (Icteridae); (E) Quiscalus (Iecteridae); (F) Fringilla
(Fringillinae) ; (G) Coccothraustes (Carduelinae) ; (H) Carduelis (Cardue-
linae); and (1) Lonchura (Hstrildidae).

444 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Oxyruncidae

One badly preserved specimen was available for study. The process is
apparently lacking, but because of the condition of the specimen, there is
still some question as to the true condition of the process.

Oxyruncus (1), lacking ?. Checked with Hellmayr (1929).

Phytotomidae

The process is absent in all specimens of this monotypic family and
may be fused, at least in part, with the dentary process of the premaxilla
instead of with the palatine. Phytotoma is a heavy-billed bird so that its
fused process provides additional evidence for the generalization that the
process fuses completely with the surrounding bones in heavy-billed birds.

Phytotoma (9), lacking (fused with the dentary process of the pre-
maxilla ?). Checked with Hellmayr (1929).

Pittidae

The process is present, but varies from being partly fused to completely
fused with the palatine. The bill is approximately the same shape and
size in all specimens, therefore this variation is most likely age and not
the result of functional differences. Consequently, the typical adult condi-
tion is assumed to be a completely fused process and those specimens with
a partly fused process are probably immature birds.

Pitta (18), unfused in 11 specimens, lacking in the other 7. Checked
with Peters (1951).

Xenicidae

No specimens of this rare family were available for study. Forbes
(1882) did not mention this process in his paper on the anatomy of
Xenicus and Acanthisitta. Pyeraft (1905b) described and figured the
palate of Acanthisitta, but also did not mention the process.

Philepittidae

No specimens of this family were available for study. Forbes (1880)
reported on the anatomy of Philepitta and figured the palate in which the
process was not shown.

Menuridae
The process is lacking in this group of large birds with relatively heavy
bills.

Menura (2), lacking. Checked with Mayr and Amadon (1951).

Atrichornithidae

No specimens of this rare family were available for study. The only
anatomical study of Atrichornis that I know of is by Garrod (1877), but
unfortunately he does not describe the palate.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 445

Figure 28. Palatine process of the premaxilla (P) of: (A) Spermospiza
(Estrildidae); (B) Ploceus (Ploceidae); (C) Passer (Ploceidae); (D)
Dinemellia (Ploceidae); (E) Bubalornis (Ploceidae); (F) Cyanocitta
(Corvidae); (G) Paradisaea (Paradisaeidae); (H) Paradisaea (Para-
disaeidae); and (1) Paradisaea (Paradisaeidae).

446 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Alaudidae

The process is present in most genera of larks, but there is much varia-
tion in the degree of fusion between it and the palatine. In some genera,
such as Hremophila, the process is completely lacking (= fused). When
present, the process usually lies along the palatine, but in Melanocorypha,
it lies free in the space between the palate and the jugal bar. Certhilauda
has a long, decurved bill and lacks the process. This is the opposite of
the condition found in many groups in which the heavy-billed forms lack
the process and the species with long, thin or decurved bills have a well-
developed process. However, Certhilauda uses its bill to dig in the
ground (Meinertzhagen, 1951, p. 101) and the fused process probably gives
the bill greater strength as in the thrashers (Toxostoma).

Alauda (8), degree of fusion varies; Calandrella (1), process almost
completely fused with the palutine and the jugal bar; Certhilauda (1),
lacking; Hremophila (17), lacking; Hremopterix (3), unfused; Lullula
(1), lacking; Melanocurypha (1), process lying free in space between
the palate and the jugal bar; Mirafra (5), partly fused to the palatine.
Checked with Vaurie (1951) and Meinertzhagen (1951).

Hirundinidae

Most genera of swallows have a well-developed process which lies next
to the palatine and is not fused at all with that bone (Fig. 24A). In
Tridoprocne, the degree of fusion varies from unfused to completely fused
with the palatine. All specimens of Hirundo lacked the process. The bill
of all genera of swallows is equally broad and apparently similar in fune-
tion so that the significance of the variation in Jridoprocne and Hirundo
remains a mystery.

Atticora (2); Delichon (4); Hirundo (22), lacking; Iridoprocne (20),
present, but varies from unfused to completely fused; Lamprochelidon
(1); Orochelidon (1); Petrochelidon (6); Phaeoprogne (2); Progne (14);
Riparia (5); Stelgidopteryx (10); Tachycineta (2). Process present and
unfused unless otherwise noted. Checked with Hellmayr (1935) and with
Mayr and Bond (1943).

Motacillidae

In this family of thin-billed species, the process is lost except for a thin
isolated splint of bone lying along the palatines in some specimens (Fig.
24B). This splint can be easily broken off, so that it is not known how
much of the observed variation is natural and how much is artificial.

Anthus (36), usually lacking, but in a few specimens there is an indica-
tion of an isolated splint lying along the palatine; Macronyx (3), isolated
splint in one specimen, lacking in the others; Motacilla (33), lacking or
present as a splint lying along the palatine. Checked against the
A.M.N.H. collection.

fd

BOCK: PALATINE PROCESS OF THE PREMAXILLA 447

Campephagidae

In a few genera of the cuckoo-shrikes, such as Campephaga, there is a
short process at the anterior end of the palatines. In most genera, however,
the process fuses to the palatines. This seems to be the typical adult
condition in this family and the observed variation in fusion appears to be
age variation.

Campephaga (4), short, unfused; Coracina (8), lacking; Edolisoma (2),
lacking; Hemipus (2), lacking; Lalaga (1), present (?), short; Pericroco-
tus (43), lacking, but in a few specimens there is a hint of a process;
Volvocivora (1), short, partly fused. Checked against the A.M.N.H. col-
lection.

Pycnonotidae

The process varies from being present and unfused to completely fused
in the bulbuls. In general, it fuses partly to the prepalatine process to form
a large part of that bone and partly to the dentary process of the pre-
maxilla. In a few genera, such as Spizixos (Figs. 24D, 24E, and 24F), the
process degenerates at its anterior end to become a thin, isolated splint of
bone lying along the palatine, such as in Paradisaea or in Motacilla. Except
for the difference between the Microscelis-type and the Spizixos-type, the
variation in the degree of fusion between the process and the palatine is
probably age variation; the difference between Spizizos and Microscelis is
probably functional, but more than that I cannot say.

Baeopogon (4); Bleda (5); Calyptocichla (2); Criniger (6); Ixonotus
(1), lacking %; Microscelis (33); Phyllastrephus (4); Pycnonotus (93) ;
Spizizos (26), present, the anterior end generally degenerating to an
isolated splint lying along the palatine; Thescelocichla (2); Trachycomus
(5). Unless otherwise stated, the process varies from unfused to fused
with the prepalatine process and, in part, to the dentary process of the
premaxilla. Checked with Delacour (1943a).

Trenidae

The nature and variation of the process in this family is identical to
that in the closely related Pycnonotidae.

Aegithina (8), partly fused to completely fused with the palatine;
Chloropsis (8), lacking; Irena (8), partly fused to completely fused with
the palatine. Checked with Mayr and Amadon (1951).

Laniidae

The process is lacking (= fused) in most members of this heavy-billed
family. A few specimens, however, have a small, partly fused process at
the anterior end of the palatine, which indicates that the process fuses with
the palatine and that any variation is probably due to age.

Dryoscopus (6), lacking; Laniarius (13), lacking, one specimen had a
faint suture showing that the process had fused to the palatine; Lanius

448 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

(13), partly fused to completely fused; Malaconotus (3) lacking; Pelicinius
(1), long, partly fused to the palatine; Rhodophoneus (1), lacking; Tchagra
(3), one specimen had a long, partly fused process, the other lacked the
process; Urolestes (2), lacking. Checked against the A.M.N.H. collection.

Prionopidae

No specimens of this family were available for study, nor could I find
a deseription of the palate in the literature. It is most likely that they
agree with the Laniidae in the structure of the process. Genera checked
with Mayr (1943).

Vangidae

No specimens of the vanga shrikes were available for study, nor could
I find a description of their palate in the literature. It is most probable
that the structure of the process in the vangids is the same as in the
true shrikes.

Bombyeillidae

The process is lacking in all but a few specimens of this family. These
latter specimens, especially the two individuals of Dulus, are probably
immature birds in which the process is less fused to the palatine. As in the
preceding families, the observed variation is the result of a difference in
age and hence in ossification.

Bombycilla (33), lacking or a faint indication in a few specimens; Dulus
(10), usually lacking, but two specimens had a well-developed process;
Phainopepla (9), lacking; Ptilogonys (5), lacking. Checked with Mayr
spd Amadon (1951) and with Delacour and Amadon (1949).

Cinclidae

As in the closely related thrushes, the process in the dipper is lacking
as a distinct structure in the adult and is assumed to be fused with the
palatine.

Cinclus (7), lacking. Checked with Mayr and Amadon (1951).

Troglodytidae

Most specimens examined lacked a process; however, in a few genera,
such as Heleodytes, Cinnycerthia and Thryothorus, the process varies from
being present, but partly fused, to completely fused with the palatine. It
appears that the adult conditon is complete fusion and that the observed
variation is due to age.

Catherpes (2); Cinnycerthia (8), varies as in Heleodytes; Cistothorus
(4); Heleodytes (15), partly fused to completely fused with the palatine;
Henicorhina (1); Leucolepis (1); Salpinetes (6); Thryomanes (1); Thryo-
thorus (55), varies as in Heleodytes; Troglodytes (31). Process lacking
unless otherwise noted. Checked with Hellmayr (1934).

BOCK: PALATINE PROCESS OF THE PREMAXILLA 449

Mimidae

Except for Margarops, the process is lacking (— fused) or heavily
fused with the palatine in all genera of the mimids. Individuals having a
partly fused process are, with little doubt, immature birds and the
variation in the degree of fusion is thus due to age. Im the case of
Margarops, however, it seems likely that the process is present and not
fused to the palatine in the adult bird (that is, the four birds examined
are not immature), but the significance of this difference between Mar-
garops and the rest of the mimids is not known.

Allenia (2); Cinclocerthia (1); Donacobius (1); Dumetella (13), a
hint of the process in a few specimens, and in one specimen the process
was only slightly fused; Margarops (4), unfused, lying next to the pala-
tine, Melanoptila (1); Melanotis (3); Mimodes (1); Mimus (7); Neso-
mimus (3); Oreoscoptes (2); Toxostoma (8). Process lacking unless
otherwise noted. Checked with Hellmayr (1934).

Prunellidae

The process is lacking in Prunella as in the closely related Turdinae.
Prunella (5), lacking. Checked with Marien (1951).

Turdinae

The process is usually fused to one or more of the surrounding bones
(the palatine or the dentary process of the premaxilla). In a few genera,
such as Hylocichla or Turdus, a partly fused process is present in some
specimens which are assumed to be young birds (Mig. 25A and 25B). Hence,
the fused process is the adult condition and any variation in the degree
of fusion is mostly likely due to differences in age.

Alethe (4); Callene (2); Cichladusa (18); Copsychus (18); Cossypha
(5); Enicurus (12); Erithacus (31); Erythropygia (1); Hylocichla (18),
a faint indication of the process in a few specimens; Monticola (1);
Myadestes (3); Neocossyphus (2); Oenanthe (4); Phaeornis (1); Phoeni-
curus (16); Sawxicola (25); Sialia (15); Thamnolaea (2); Turdus (19),
a faint indication of the process in a few specimens and a partly fused
process in two others; Zoothera (2). Process lacking unless otherwise
noted. Checked with Ripley (1952).

Timaliinae

The variation of the process in either Garrulax (Figs. 24G, 24H, 241)
or Liothrix is typical for this subfamily. In these genera, the process
varies from being present and unfused to fused with any of the surrounding
bones (palatine, or the dentary process of the premaxilla). The adult
condition is thus assumed to be a completely or an almost completely
fused process and the observed variation in the degree of fusion to be due
to age.

450 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Alcippe (19), varies as in Liothriz; Babax (4), varies as in Garrulax;
Chamaea (5), lacking; Chrysomma (1), unfused, lying along the palatine;
Dumetia (1), fused to dentary process of the premaxilla; Gampsorhynchus
(1), fused to dentary process of the premaxilla; Garrulax (55), unfused
to fused with the dentary process of the premaxilla, process also
partly overlies the palatine; Heterophasia (3), fused to the dentary
process; Lioptila (4), varies as in Liothrix; Liothrix (38), long and un-
fused, with the anterior end fusing to the dentary process, and thus leav-
ing an isolated splint of bone (the posterior end of the process) lying
along the palatine; Macronus (4), fused to the dentary process; Malaco-
cincla (5), varies as in Garrulax; Malia (1), varies as in Garrulaz;
Pellorneum (1), fused to the dentary process; Pomatorhinus (7), varies as
in Liothrix; Psophodes (2), varies as in Garrulax; Pteruthius (5), varies
as in Garrulax; Rhinocichla (1), heavily fused to the palatine; Siva (10),
varies as in Liothria; Stachyris (6), varies as in Liothriz; Trochalopteron
(25), varies as in Garrulaz; Turdoides (2), varies as in Garrulax; Yuhina
(26), varies as in Liothriz. Checked with Delacour (1946).

Paradoxornithinae
The process is lacking in the one genus examined in this subfamily and

it is assumed to be fused as in the closely related Timalinae.
Suthora (5), lacking. Checked with Delacour (1946).

Polioptilinae

The process is lacking (= fused) in this group of thin-billed birds and
is similar to that seen in the closely related Sylviinae.

Microbates (4), lacking, but with a hint of a fused process; Polioptila
(10), lacking; Ramphocaenus (1), lacking. Checked with Rand and Tray-
lor (1953).

Sylviinae

The process is lacking (= fused) in almost all genera of this subfamily.
In a few genera, such as Cisticola and Prinia, an unfused process is present
in some specimens. This indicates that the typical adult condition is for the
process to be fused with the palatine and that the observed variation in
the degree of fusion is age variation.

Abroscopus (3); Acrocephalus (9); Apalis (2); Bradornis (5), (may
be a muscicapiine?) ; Camaroptera (2); Cisticola (9), present as a long thin
process that is partly or completely fused to the palatine; Hremomela (2),
present in one specimen, lacking in the other; Franklinia (3); Hylia (1);
Hypergerus (2); Oreopneuste (1); Phragamaticola (1); Phylloscopus (23),
a small splint of bone lying along the palatine in one specimen, signs of a
fused process in a few other specimens; Prinia (10), present, small, at the
very anterior end of the palatine, partly fused in two specimens; Regulus
(27); Seicercus (1); Sylvia (6); Sylvietta (2); Tribura (8); Urosphena
(1). Process lacking unless otherwise noted. Checked against the A.M.N.H.
collection.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 451

Malurinae

The process is lacking in the Malurinae.
Malurus (1), lacking. Checked with Mayr and Amadon (1951).

Muscicapinae

The process is almost universally lacking in the Old World flycatchers,
but there is a hint of a process in a few specimens, e.g., Culicicapa, which
indicates that the broad anterior end of the palatine process is composed
in part of the fused process. Thus the adult condition is for the process
to be fused and any variation in fusion is most likely age variation.

Batis (2); Bias (2); Culicicapa (5), hint of a fused process in one
specimen; Hrranornis (2); Erythocercus (1); Ficedula (11); Fraseria
(2); Hyliota (1); Muscicapa (10); Niltava (4); Platysteira (1). Process
lacking. Checked with Vaurie (1953) and against the A.M.N.H. collection.

Monarchinae

The process is lacking as in the closely related Muscicapinae.

Chasiempis (1); Hypothymis (7); Terpsiphone (11); Trochocercus (2).
Process lacking. Checked with Mayr and Amadon (1951) and against the
A.M.N.H. collection.

Pachycephalinae

The process is lacking in this group of heavy-billed flycatchers.

Colluricincla (2); Eopsaltria (1); Falcunculus (1); Pachycephala (4).
Process lacking. Checked with Mayr and Amadon (1951) and against the
A.M.N.H. collection.

Paridae

A short process is present in some individuals which becomes fused,
in part with the dentary process of the premaxilla, but mainly with the
palatine in the adult. The variation in the degree of fusion is assumed
to be age and hence ossification.

Aegithaliscus (3); Aegithalos (2); Parus, a partly fused process is
present in a few specimens; Psaltriparus (6). Process lacking. Checked with
Mayr and Amadon (1951) and against the A.M.N.H. collection.

Sittidae

The process is lacking in all specimens of this heavy-billed family. How-
ever, in one specimen of Sitta, there is a suggestion of a suture on the
prepalatine process indicating that the process has fused with the palatine.

Rhabdornis (1); Sitta (49); Tichodroma (1). Process lacking. Checked
with Mayr and Amadon (1951).

452 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Certhiidae

The process is lacking (= fused) in most specimens of this thin-billed
genus, this being the typical adult condition. In one specimen, however,
the process is only partly fused to the palatine which indicates that the
process is present and unfused in the immature and becomes increasingly
fused with age.

Certhia (17), lacking, but a slight indication of the process in one
specimen. Checked with Mayr and Amadon (1951).

Dicaeidae

The typical condition of the process in this thin-billed family is a
thin isolated splint lying along the palatine. In some specimens, presum-
ably immatures, the process is still attached to the premaxilla. In others,
the anterior connection to the premaxilla has degenerated toward the typi-
cal adult condition. If the process is lacking, it is most likely the result
of loss during preparation, although the process could degenerate com-
pletely in some specimens.

Dicaewm (7), lacking or as a splint lying along the palatine. Checked
with Mayr and Amadon (1947).

Nectariniidae

The process is usually lacking in this family of long, thin-billed birds.
In a few specimens, there is a slight suggestion that the process has
become fused to the palatine during development.

Aethopyga (3); Anthreptes (2); Arachnothera (2); Nectarinia (23),
a slight indication of the process in a few specimens. Process lacking.
Checked with Delacour (1944).

Zosteropidae

The process is usually lacking in this family of short thin-billed species
with the usual indication of a semifused process in a few specimens.
Zosterops (15), usually lacking. Checked with Mayr and Amadon (1951).

Meliphagidae

The process is lacking (= fused) in most specimens of this family. In
Philemon, however, the process is present, but at least partly fused to the
palatine. Thus it is evident that in the adult of this family, the process
fuses to the prepalatine process and that any variation in fusion is prob-
ably due to age.

Parker (1877) reported that the palatine process in Acanthorhynchus
(Fig. 24C) lies along the medial side of the palatine. This is possible,
although the specimen of the genus that I examined did not show this

BOCK: PALATINE PROCESS OF THE PREMAXILLA 453

condition. I was careful to check specimens of the entire order for addi-
tional examples in which the palatine process lay on the inside rather than
on the outside of the palatine, but could not find any such specimens. Thus,
I would suggest that the palatine process in Acanthorhynchus is normal
and lies on the lateral side of the palatine.

Acanthorhynchus (1); Conopophila (1); Melilestes (1); Meliornis (2);
Meliphaga (2); Melithreptus (2); Myzantha (1); Myzomela (3); Phile-
mon (6), present, but partly fused to the palatine. Process lacking unless
otherwise noted. Checked against the A.M.N.H. collection.

Hmberizinae

The process varies greatly in this group of finches. Typically, it lies
next to the palatines and is more or less fused to that bone, but in a
number of specimens, it is completely fused (Fig. 25F). The figures of
Melospiza melodia (Figs. 25G, 25H, and 251) illustrate what seems to be
the typical course of reduction and fusion of the process with increase
in age; however, the process can also fuse to the palatine without any
reduction in size. In Melopyrrha (Fig. 25C) and Tiaris (Fig. 25D), the
process lies free in the space between the palate and the jugal bar and
appears similar to that seen in the cardinals. In Oryzoborus (Fig. 25E),
a very heavy-billed species, the process is completely fused to the palatine
and in its place is a lateral flange similar to that seen in the Carduelinae.
Much of the observed variation in this family is no doubt age variation,
but aside from Melopyrrha, Tiaris, and Oryzoborus which are special prob-
lems, it is difficult to determine the fully adult condition in the different
genera. Hence the significance, both taxonomically and functionally, of
the variation in fusion cannot be evaluated at this time.

Ammospiza (7), unfused to almost completely fused; Arremonops (1),
lacking; Calcarius (1), lacking ?; Hmberiza (6), slightly fused; Guberna-
trix (4), slightly fused; Junco (8), slightly fused; Melopyrrha (2), lying
free in the space between the palate and the jugal bar; Melospiza (16),
slightly fused to completely fused, specimens show reduction of the an-
terior end of the process; Oryzoborus (3), lacking, anterior end of palatine
has a lateral flange; Paroaria (10) slightly fused to almost completely
fused; Passerculus (9), partly fused to completely fused; Passerella (7),
lacking, some specimens with a faint suture; Phrygilus (1), slightly fused
Pipilo (9), partly fused to completely fused; Plectrophenax (7), slightly
fused; Pooecetes (1), lacking; Sicalis (1), lacking; Spizella (7), slightly to
partly fused; Sporophila (7), lacking, one specimen showed a hint of the
process; Tiaris (1), free as in Melopyrrha; Zonotrichia (15), slightly
fused. Checked with Tordoff (1954a).

Cardinalinae

The process is present in all genera of this heavy-billed group of
finches. Usually it lies free in the space between the palate and the
jugal bar (Figs. 3 and 26B), but in some genera, such as Caryothraustes

454 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

(Fig. 26A) and Saltator, the process lies along the palatines. There may
or may not be a suture between the free process and the rest of the pre-
maxilla, but the process never fuses to the palatine. At the present time,
none of the observed variation can be correlated with any functional or
taxonomic significance.

Cardinalis (14), free; Caryothraustes (1), lying along the palatine;
Guiraca (3), free; Passerina (7), free; Pheucticus (9), free; Saltator (8),
in some specimens it lies free, in others it lies along the palatine; Spiza
(1), free (may be icterid). Checked with Tordoff (1954a).

Tanagrinae

The process is present in all specimens of tanagers that were examined.
It may lie in the space between the palate and the jugal bar, but usually
it lies along the palatine although the two bones apparently do not fuse
(Figs. 260, 26D, and 26E). A suture separating the process from the rest
of the premaxilla may or may not be present. Tordoff reports the process
as being partly fused to the palatine with a suture present; this does not
agree with my findings that the process does not fuse at all with the
palatine. However, because of the difficulty of determining the degree of
fusion between the process and the palatine by examination of the ventral
aspect of the palate, this difference of interpretation is of no importance.

Buthraupis (1); Calospiza (9); Chlorophonia (1); Chlorospingus (1),
lying free; Compsocoma (1); Habia (1); Hemithraupis (1); Orthogonys
(1); Poecilothraupis (2); Pipraeidea (2); Piranga (9), lyimg free; Ram-
phocelus (2); Schistochlamys (3); Spindalis (2); Tachyphonus (3); Tan-
gara (21); Tanagra (8). Process present, unfused or free. Checked with
Hellmayr (1936).

Tersininae

The process is present in this swallow-billed species. It lies somewhat
free in the space between the palate and the jugal bar, but still closely
attached to the palatine (Fig. 26F). In general, the process resembles that
seen in the true swallows.

Tersina (3), unfused, lying somewhat free. Checked with Hellmayr
(1936).

Coerebinae

The process is present in this group of thin-billed birds. It always
lies against the palatine, but its length varies as does the degree of fusion
with the palatine, which varies from slightly fused to completely fused.
How much of the variation in the degree of fusion is age variation and
how much represents a true difference between genera is unknown.

Ateleodacnis (4), heavily to completely fused; Chlorophanes (3), long,
slightly to completely fused; Coereba (12), short, partly to completely
fused; Conirostrum (2), long, slightly fused; Cyanerpes (24), long, slightly

BOCK: PALATINE PROCESS OF THE PREMAXILLA 455

(usually) to heavily fused; Dacnis (4), slightly to strongly fused; Diglossa
(3), long, partly fused; Hwmeornis (2), short, not fused. Checked with
Hellmayr (1935).

Catamblyrhynehinae

The process is lacking (= fused) in this heavy finch-billed genus.
Catamblyrhynchus (2), lacking. Cheeked with Hellmayr (1938).

Parulidae

In this family of thin-billed birds, the process is present as a thin
splint lying along the palatine (Fig. 27B). It degenerates at its anterior
end and becomes fused to the palatine with increasing age. The fully adult
condition is apparently for the process to be heavily fused with the palatine;
hence the observed variation in fusion is most probably age variation.

Dendroica (79); Geothlypis (8); Helmitheros (3); Icteria (4); Miniotilta
(12); Parula (2); Protonotaria (2); Oporornis (10); Seiurus (34); Set-
ophaga (10); Vermivora (13); Wilsonia (10). Process present, lying along
the palatine with the anterior end degenerating to leave a thin splint which
finally fuses with the palatine. Checked with Hellmayr (1935).

Drepaniidae

In the heavy, finch-billed genus, Psittirostra (Fig. 26G), the process and
the anterior half of the palate is exactly like that in the cardueline finches.
In the other genera, the process is lacking (= fused), but the anterior
end of the palatine does not flare out to the side as strongly as in Psitti-
rostra, and in some there is no flaring of the palatine (Figs. 26H, and 261).
In at least one specimen, that of Hemignathus, a faint suture is present on
the lateral half of the prepalatine process, which indicates that the process
is present, but has fused completely to the palatine.

Hemignathus (1), long, curved-bill; Himatione (3), short, medium-bill;
Loxops (1), short, thin-bill; Psittirostra (9), lateral flange present, bill
similar to cardueline bill; Vestiaria (1), medium, curved-bill. Process
lacking. Checked with Amadon (1950a).

Vireonidae

The process is lacking (= fused) in all specimens of vireos (Fig. 27A),
except for one in which the process is visible but heavily fused to the
palatine. This indicates that the process is present in the immature and
becomes fused with the palatine with increasing age. The process is
lacking in Cyclarhis, a heavy-billed species. I was unable to examine any
specimen of Vireolanius, another heavy-billed species, but presumably the
process of this genus would be similar to that seen in Vireo and Cyclarhis.

Cyclarhis (7), lacking; Vireo (16), lacking in all but one specimen which
showed a faint suture. Checked with Hellmayr (1935).

456 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Icteridae

The process is lacking (= fused) in most specimens examined in this
family. In some specimens, however, e.g., Huphagus, Gymnomystax, Icterus,
Xanthornus and Quiscalus, the process is absent, but usually heavily fused
with the palatine (Figs. 27C, 27D, and 27E). This variation is, no doubt,
mainly due to differences in age; the adult condition is presumed to be
complete fusion between the process and the palatine.

Agelaius (6), suture sometimes weakly present; Amblycercus (2);
Cacicus (3); Dolichonyx (4); Euphagus (3), present, slightly fused to
almost completely fused; Gnorimopsar (1); Gymnomystax (2), suture very
faint; Gymmostinops (2); Holoquiscalus (1), with a faint suture; Hy-
popyrrhus (1), present, but heavily fused; Icterus (22); heavily fused in
some specimens; Molothrus (7); Notiopsar (2); Pezites (3); Psomocolax
(1); Quiscalus (15), heavily fused in some specimens; Sturnella (6),
suture sometimes weakly present; Tangavius (1); Xanthocephalus (1);
Xanthornus (1), faint suture; Zarhynchus (1). Process lacking unless
otherwise noted. Checked with Hellmayr (1937).

Fringillinae

In Fringilla, the process is present and lies along the palatine, much
the same as that seen in many genera of the Emberizinae. However, it
may be even less fused with the palatine than in that group (Fig. 27F).

Fringilla (3), present, slightly fused, lying along the palatine. Checked
with Mayr et al. (1956).

Carduelinae

The process is lacking (= fused) in all specimens of this group of
heavy-billed finches. In addition, the anterior end of the prepalatine process
bears a lateral flange (Figs. 4, 27G, and 27H).

Carduelis (ineluding Acanthis and Spinus specimens) (18); Carpodacus
(14); Chloris (3); Coccothraustes (3); Eophona (2); Hesperiphona (19) ;
Leucosticte (6); Loxia (6); Pinicola (3); Poliospiza (2); Pyrrhula (1);
Serinus (12). Process lacking, but lateral flange present. Checked with
Mayr and Amadon (1951) and against the A.M.N.H. collection.

Estrildidae

All specimens examined had a lateral flange similar to that in the
Carduelinae (Figs. 271, and 28A). As in that group, it is assumed that
the process has fused with the palatine to form part of the lateral flange.

Amadina (4); Clytopiza (1); Erythrura (2); Estrilda (14); Lonchura
(16); Poephila (11); Pytilia (3); Pyrenestes (1); Spermophaga (1);
Steganopleura (4). Process lacking, but lateral flange present. Checked
with Delacour (1943b).

BOCK: PALATINE PROCESS OF THE PREMAXILLA 457

Bubalornithinae

The process is present in Bubalornis and Dinemellia (Figs. 28D, and
28E) as reported earlier by Sushkin (1927). In these genera, the process
lies next to the palatine and is partly fused with that bone.

Bubalornis (1), present, but heavily fused; Dinemellia (3), present,
lying next to the palatine, slightly to partly fused.

Passerinae

The process is usually lacking (= fused) in this subfamily (Fig. 28C).
It may be present in some specimens as a heavily fused process, but the
sutures are so faint that it is difficult to determine their true nature.

Passer (15), lacking.

Ploceinae

The process is lacking in this subfamily of weaver finches. The anterior
part of the prepalatine process flares out to the side as in the carduelines.

Anaplectes (1); Coliuspasser (3); Diatropura (2); Euplectes (9);
Malimbus (1); Plocepasser (2); Ploceus (7); Pseudonigrita (1); Sporo-
pipes (1). Process lacking, but lateral flange present.

Viduinae

The process in the widow birds is the same as that in the Ploceinae, but
except for Hypochera, they lack the lateral flange.

Hypochera (1), lacking, flange present; Steganura (1), lacking, flange
absent; Vidua (2), lacking, flange absent. Subfamilies of Ploceidae fol-
low Tordoff (1954a), genera checked with Chapin (1917) and against the
A.M.N.H. collection.

Sturnidae

The process is lacking (= fused) in almost all specimens of starlings
examined. In a few specimens of Stwmus, a faint suture could be seen
which suggests that the process is present in the immature and becomes
fused with increasing age.

Acridotheres (2); Creatophora (1); Lamprotornis (4); Mino (2); Ony-
chognathus (2); Spreo (8); Sturnus (11), hint of process in a few speci-
mens. Process lacking. Checked with Amadon (1943).

Oriolidae

The process is lacking (= fused) in most specimens of this family. In
a few specimens of Oriolus, a faint suture can be seen which indicates that
the process is present in the immature and becomes fused with age.

Oriolus (23), present, but partly fused in five specimens, lacking in the
rest; Sphecotheres (3), lacking. Checked against the A.M.N.H. collection.

458 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Dieruridae

The process is lacking (= fused) in all specimens of drongos.
Dicrurus (39), lacking. Cheeked with Vaurie (1949).

Callaeidae

The process is lacking (= fused) in all specimens of this family.

Callaeas (1), lacking, heavy-billed form; Neomorpha (1, female), lack-
ing, long, eurved-billed species (the palatine process is not shown in the
figure of the palate of the male [Garrod, 1872], which has a much shorter
and stouter bill); Philesturnus (1), lacking, short slender bill. Checked
with Stonor (1942).

Grallinidae

The process is lacking (= fused) in all specimens of Corcoraz and
Grallina, but present in Struthidea. This apparently represents a true dif-
ference between these genera which may be a reflection of the possible
artificial nature of this family.

Corcorax (1), lacking; Grallina (8), lacking; Struthidea (4), present,
lying along the palatine. Cheeked with Amadon (1950b).

Artamidae

The process is lacking (= fused) in this broad-billed, swallow-like genus,
but in some specimens there is an indication that the process has fused
with the palatine.

Artamus (7), lacking, but a faint indication in a few specimens. Checked
with Mayr and Amadon (1951).

Cracticidae
The process is lacking (= fused) in all specimens examined in this
family.
Cracticus (1); Gymnorhina (7); Strepera (3). Process lacking. Checked
with Amadon (1951).

Ptilonorhynchidae

The process is lacking (= fused) in almost all specimens of the bower-
birds. Its presence in one specimen of Ailuroedus indicates that the process
has fused to the palatine and that any variation in fusion is age variation.

Ailuroedus (5), present in one specimen; Chlamydera (2); Ptilono-
rhynchus (9); Xanthomelus (1). Process lacking. Checked against the
A.M.N.H. collections.

Paradisaeidae

The process varies greatly in the birds of paradise. In most genera,
it is lacking or heavily fused to the palatine. In a few genera, such as
Manucodia, a quite distinct process is present in some specimens. In

BOCK: PALATINE PROCESS OF THE PREMAXILLA 459

Paradisaea (Figs. 28G, 28H, and 281), the anterior connection between
the process and the main body of the premaxilla begins to degenerate until
only an isolated splint of bone remains. How much of the variation is true
difference between genera is not known, but it is evident that a large amount
of the observed variation is age.

Astrapia (3), lacking; Cicinnurus (6), lacking, a hint of the process
in one specimen; Craspedophora (1), process partly fused; Diphyllodes
(8), lacking; Epimachus (2), partly fused; Lophorina (6), partly fused
to completely fused; Loria (1), lacking; Manucodia (5), present in two
specimens, lacking in the others; Paradisaea (25), present in 23 specimens,
degree of fusion varies, lacking in the other two; Parotia (5), partly fused
to completely fused; Phonygammus (2) slightly fused; Ptiloris (1) partly
fused; Seleucides (5), lacking; Semioptera (1), lacking. Checked with
Mayr (MS for Peters’ Check-list).

Corvidae

The process is lacking (= fused) in the adult crow; the observed varia-
tion in the degree of fusion is most likely due to age.

Calocitta (1); Corvus (34); Crypsirina (2); Cyanocitta (13), faint
indication of a process in one specimen; Cyanocorax (10); Garrulus (2) ;
Kitta (10); Nucifraga (3); Perisoreus (6); Pica (3); Ptilostomus (4);
Pyrrhocorax (1). Process lacking. Checked with Amadon (1944).

EVOLUTION OF THE PALATINE PROCESS
OF THE PREMAXILLA

The discussion of the evolution of the palatine process will, by
necessity, be divided into two parts: (a) the general evolution-
ary principles involved; and (b) the evolutionary pathways tra-
versed by the process (Fig. 29). I am interested here in deter-
mining how the palatine process has evolved— how it has
changed from one condition to another —not in its actual
phylogeny. It is, nevertheless, necessary to consider certain ques-
tions that are closely associated with problems of phylogeny,
namely whether independent origin plus parallel evolution and
reversal of evolution have played important roles in the evolu-
tionary history of the palatine process. Throughout the discus-
sion, the palatine process will be considered as a morphological
structure independent of taxonomy, but it will always be in-
tegrated with the other structures of the character complex to
which it belongs. These evolutionary conclusions are, of course,
speculative with the weakest link in the chain of evidence being
the functional conclusions. If future empirical studies on the
funetion of the jaw mechanism in birds prove the functional
deductions to be wrong, then these evolutionary conclusions
also may well be wrong.

460 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Evolutionary principles. A discussion of some basic evolution-
ary principles may seem to be out of place in this paper and
even to be totally unnecessary. Yet in recent papers on avian
anatomy and classification, there is a distressing lack of under-
standing and appreciation of evolutionary principles underlying
morphological change. Mayr (1955, 1958) alludes to this problem
and discusses some of the basic principles. The reader should
consult these papers and the general treatises on major evolu-
tionary change, such as Sewertzoff (1931), Simpson (1953), and
Remane (1956), for a more complete coverage of this topie.
Yet no single book or paper includes a discussion of all of the
principles pertinent to the evolution of the palatine process.
These principles will, therefore, be discussed to avoid any possible
confusion in their meaning and use in the present paper. I shall
restrict my remarks to the evolution of single characters and
shall omit those principles, such as ‘‘mosaic evolution’’ and
‘‘key innovations,’’ which apply to the evolution of species or
groups.

a) Genetie potential: The members of a taxonomic group are
basically similar in their genetic compositions which tend to
change in a similar fashion. The similarity of the genetic com-
position and change depends upon the degree of relationship ;
the closer the relationship, the closer is the degree of genetic
potential. This bipartite principle is one of the most important
underlying the evolution of both single structures and taxonomic
groups — in fact, it is basic to many other evolutionary concepts,
including those to be discussed below — yet it is one of the least
understood and appreciated concepts of evolution. Although
brief, one of the best discussions of this concept can be found
in Mayr and Vaurie (1948, p. 246) who conclude that : ‘‘ Parallel-
ism in evolution must be accounted for by such basie similarity
of the germ plasm.’’ Simpson (1953, pp. 348-349) alludes to
this principle when he considers the problem of how related
groups can acquire the same adaptation independently (.e.,
the development of the mammalian jaw articulation in five dif-
ferent lines of therapsid reptiles). Again, Amadon (1956, pp.
15-16) suggests that some of the parallelism in the starlings has
been facilitated by the genetic potential of this family.

The evidence supporting the concept of genetic potential is
relatively simple and well known to most biologists. First, if
there is a direct correlation between the genotype and the pheno-
type, then a group of related and structually similar forms
would have to be similar genetically. This statement should be

BOCK: PALATINE PROCESS OF THE PREMAXILLA 461

reversed to read that a group of related and genetically similar
forms would be similar in their phenotypical expression. How-
ever, since the genetics of most groups has not yet been worked
out, we must use the more awkward converse. Second, it is well
known that, although the occurrence of gene mutations is ran-
dom, their direction is not random. A gene can mutate only to
one of the several alleles found at that chromosomal site; the
limitation is most probably imposed by the physical and chemical
structure of the gene. Third, the mutant gene must act in
harmony with the rest of the genotype if it is to have a selective
advantage. Therefore, one can conclude that all of the Passeres
are similar genetically and that a structure could change only in
a rather limited number of ways. The palatine process of the
premaxilla would, then, have a similar genetic basis in all of
the Passeres and could change, both genetically and phenotypic-
ally, in a limited number of ways.

b) Developmental potential: Many structures are present at
some time during development and disappear as visible struc-
tures in the adult, a phenomenon especially characteristic of the
avian skeleton. Although these structures are ‘‘absent’’ in the
adult, they must be regarded as being present in the species
because their genetic basis is present and they appear in every
generation in the developing animal. Why these structures are
present only in the embryo— whether they have a role in
development, or are the remnants of degenerating structures —
is a most important question, but it is not pertinent to this
study. The important thing is that the potential for the structure
is present and that the structure could be restored or retained
in the adult whenever selection would demand it. The evolution
of a structure involving developmental potential may exhibit
certain features which are more apparent than real. First, the
sudden appearance of the structure may appear to be the result
of a single large mutation and second, this mutation may appear
to have occurred independently in several groups. Neither of
these are correct. The structure, along with its genetical basis,
is already present in the embryo and ean be retained in the adult
by inhibiting or suppressing the factors that cause its normal
breakdown and/or disappearance. In the case of an osteological
feature, the fusion of a bony process to another bone or its
breakdown can be prevented if this process becomes the point
of attachment of a muscle or a ligament. This is exactly what
has happened in the evolution of the free palatine process. The
process is present in all Passeres although it is absent as a

462 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

visible structure in the adult, but can ‘‘appear’’ in any family as
a free palatine process if part of the M. pterygoideus takes
origin from it.

ec) Preadaptation: A preadapted structure is one whose form
will allow it to assume a new function whenever need for that
funetion should arise (Bock, 1959). The new function does
not have to be related in any way to the original function of the
preadapted structure. Its position along the palatine and the
fact that it fuses to that bone relatively late in ontogeny pre-
adapts the palatine process for the function as the origin for part
of the M. pterygoideus as seen in the cardinals and several other
families of passerine birds.

d) Multiple pathways of evolution: The same selection force
acting on several groups of animals may elicit different adaptive
responses depending upon the genotypic and phenotypic differ-
ences between these groups. These different adaptive responses
would be the multiple pathways of adaptation or evolution. The
basis for multiple pathways hes in the fact that there are usually
several ways to achieve the same functional goal (see Simpson,
1953, pp. 179-181; and Bock, 1959). For example, at least two
different methods for cracking seeds have evolved in the Passeres
and each has its own special anatomical modifications including
the structure of the palatine process of the premaxilla. Within
this primary dichotomy of multiple pathways for seed-cracking,
are secondary and probably even tertiary or lower levels of
multiple pathways. If, for instance, a passerine bird has
‘‘ehosen’’ the nuteracker method of seed-cracking, then there
are a number of different ways of evolving the necessary adapta-
tions, e.g., differential development of the several adductor and
protractor muscles. The most important consequence of multiple
pathways for the functional anatomist and the evolutionist is
that morphological differences between the adaptations for the
same funetion are non-functional and hence non-adaptive in
terms of that particular function. Only differences between
the structure in the ancestral and descendent species are
functional and thus adaptive. Therefore, one cannot assume
that all of the morphological differences between contem-
porary groups are adaptive differences. It must be empha-
sized, however, at this point, that multiple pathways can
be discussed in terms of only one function, or in terms of
only one selection foree. The differences between the adap-
tive responses for one selection foree may be adaptive in
terms of other functions. Thus one condition may evolve to

9

BOCK: PALATINE PROCESS OF THE PREMAXILLA 463

another under the influence of these other selection forces. For
example, although the differences between the adaptive responses
— the nutcracker method and the vise method — for seed erack-
ing in the Passeres are non-functional in terms of seed cracking,
they are probably functional and adaptive in terms of insect
catching. Thus the ‘‘vise structure’’ could change to the ‘‘nut-
cracker structure’’ and vice versa under the influence of selection
forces associated with insect eating, but could not change under
the influence of the selection force associated with seed cracking.

e) Independent origin and parallel evolution: These two
evolutionary principles are independent in that one does not
depend upon the other; however, because both are dependent
upon the principle of genetic potential and because they fre-
quently occur together, they are best discussed together. The
term ‘‘independent origin’’ refers to the separate evolutionary
origin of the same structure in different taxonomic groups, as
for example the free palatine process in the flycatchers and the
cardinals. This may be the result of de novo appearance of the
structure because of similar genetical changes or it may be the
retention of an embryological structure in the adult. Separation
of these possibilities is impossible without detailed genetical and
embryological investigations. ‘‘Parallel evolution’’ is similar
evolution of a structure in different taxonomic groups, i.e., the
structure undergoes the same series of changes. Clearly then, if
the families of passerine birds have a basically similar genetical
potential for a certain structure, that structure could arise inde-
pendently and evolve in a parallel fashion if the same selection
forces acted on the members of several families. Yet structures
that have arisen independently do not have to undergo a sub-
sequent period of parallel evolution and vice versa. Many ex-
amples of independent origin and of parallel evolution of the
palatine process can be cited. Some notable ones could be the
free palatine process in the cardinals. Purnarius and some of
the tyrant-flycatchers, and the lateral flange in the carduelines,
the ploceids, Oryzoborus and Psittirostra.

f) Convergence: Convergence occurs when formally dissimi-
lar structures or structures in unrelated taxonomic groups be-
come similar to one another. There is no sharp break between
parallelism and convergence, and indeed some workers do not
distinguish between the two concepts (see, however, Haas and
Simpson, 1946, for a discussion of this problem). However, as
there are several important differences between typical converg-
ence and typical parallelism, I shall distinguish between them.

‘

464 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Cases of convergence in the palatine process are somewhat diffi-
cult to separate from parallelism because we know so little about
the relationships of the passerine families. One example may
be the evolution of the lateral flange in the ecarduelines and the
ploceids if these groups are unrelated.

o) Reversal: Reversal is evolutionary change of a structure
to a condition that is similar to the primitive or original struc-
ture. The sequence would be primitive — specialized — see-
ondarily primitive. The question of whether reverse evolution
means an exact backwards change involving the same steps and
genetic changes is a moot if not a ridiculous problem. Most
likely, reversal does not involve the same genetical changes, but
this is of little importance. The important consequence of re-
versal is that a secondarily primitive structure frequently can-
not be distinguished morphologically from an originally primi-
tive structure! These morphologically similar, but evolutionary-
wise different conditions usually can be separated from one
another only after the phylogeny of the taxonomic groups has
been determined. Reversal in the evolution of the palatine
process had probably occurred many times, as for example, when
an insectivorous group has evolved from a heavy-billed group.
An example may be the evolution of the Drepaniidae from the
Carduelinae with the loss of the lateral flange in the thin-billed
members of the Hawaiian honey-creepers.

The most important feature of these several evolutionary
phenomena, especially independent origin, convergence and re-
versal, is the frequency of their occurrence. If they were of rare
occurrence during the evolution of a taxonomic group, then they
would be of only mimor importance. As their frequency in-
creases, so does their importance in evolutionary and taxonomic
studies as well as the difficulty in untangling the phylogeny of
a group, but this increase in importance and difficulty is a rapid
geometric increase. It is my belief that these phenomena were of
common occurrence during the evolution of birds because birds
share a very similar genetic potential. This appears to be
especially true in the Passeres in view of their great similarity
in structure. It is difficult, however, to cite indisputable ex-
amples of these phenomena in the case of the palatine process.
The problem stems from the fact that we know so little about the
phylogeny of the Passeres that we can never be certain, in many
cases, whether similarity in the palatine processes of two fami-
lies is the result of common ancestry, parallelism or convergence.
Nor can we determine at this time the frequency of independent

BOCK: PALATINE PROCESS OF THE PREMAXILLA 465

origin and reversal. It is always better to be cautious in taxo-
nomie studies, hence I shall stress the need to consider these
phenomena at all times.

Evolutionary history of the palatine process.

a) Primitive form: The primitive condition of the palatine
process of the premaxilla is that of the process lying along the
lateral edge of the palatine and more or less fused with that
bone — the normal condition of the palatine process as seen in
the crow or the white-throated sparrow. It is impossible to
determine whether the palatine process of the premaxilla was
completely or only partly fused with the prepalatine process of
the palatine in the ancestral passerine bird, but this is of little
importance. Change from an incompletely fused process to a
completely fused one or vice versa is exceedingly simple and
has probably occurred repeatedly during the evolution of the
Passeres. Two bits of evidence support the conclusion that the
normal palatine process is the primitive condition. First, this
is the condition found in most passerine birds as well as in most
other birds. If the normal palatine process was not the original
condition in the Passeres, one would have some difficulty ex-
plaining its present-day distribution in the passerine families.
Second, all other conditions of the palatine process in the passer-
ines appear to be specializations for some particular mode of
feeding. This is certainly true for the free palatine process and
the lateral flange on the prepalatine process.

b) Selection forces: The factor most frequenly omitted from
discussions of the evolution of a structure is the set of selection
forees acting on that structure. In the ease of the palatine
process, the primary function and hence the primary selection
force is insurance of a firm connection between the palate and
the upper jaw. This selection force may not have been responsi-
ble for the origin of the palatine process which probably took
place some time in the early history of the birds, if not earher.
However, this selection force is responsible for the maintenance
of the palatine process in birds and for some of the variation in
the degree of fusion with the prepalatine process. All modifi-
cations of the palatine process in the Passeres must always fulfill
the demands of this primary selection force. The secondary
functions and hence the secondary selection forces acting on
the palatine process are independent of each other, and indeed
they apparently cannot act together in harmony. One of the
secondary selection forces is associated with the vise method
of cracking seeds and selects for the fusion of the palatine process

BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

466

‘aoreuLBldxe Loy}iny 10f 4x0} oY} oVG ‘“AoYJZOUR 03 UOLFIPUOD 9UO WOLF SULpVo, SOUL OY} Yoostq soaov10F
UOTpPVIJOS oY, “ssad0rd ourzpeyed oy} JO wornpoas oy, ur Skvayyed reu0TYNToOAe oy} SUIMOYS UeIsOIpuog “eZ dINSTT

Q4SN4NN

t

3SV34u930
i
NOILOSNNOD MVE Y3ddN

-3LV1Vd 3HL 40 HLON3BYLS

GQASV3YON!I YO4 G33N
!

Tare 3asvauoga
G3snd4 1118 AAV3H LYOHS V HLIM

Q3LVINOSSVY) 3J7ISIGNVW
SNISON19 LSV4 V YO4 G33N

XN
ASVAYONI
3svayo03d \
7 33yYJ4
$Q33S ONINDVYD 4O

SHIOHS GNV $394O04 BHI

ISNIVOV NOILO3L0OYNd YO4 O33N
¢

ASV3YONI

#
JONVIS WHY3LVI

BOCK: PALATINE PROCESS OF THE PREMAXILLA 467

with the prepalatine process and the development of a lateral
flange at the site of fusion. The other secondary selection force
is associated with a fast closing mandible during the early phases
of closing the bill and selects for a free palatine process from
which a portion of the M. pterygoideus originates.

ec) Maintenance of the palatine process: During most of the
evolution of the Passeres, there was little or no change in the
structure of the palatine process. Shght changes in the strength
of the primary selection force had no doubt occurred which could
account for some of the variation in the degree of fusion between
the palatine process and the palatine, as for example, in the Den-
drocolaptidae and the Pipridae. Parallelism in the amount of
fusion of the palatine process or in its degeneration as an isolated
splint of bone lying along the palatine are probably common, as
are reversals in the trends of fusion or breakdown. It is impossible
to determine whether the normal condition of the palatine
process in any group of passerine birds is originally primitive or
secondarily primitive or even, as will be shown later, if the
group had passed through an earlier stage having a more special-
ized type of palatine process.

d) Evolution of the free palatine process: The evolution of
the free condition of the palatine process is dependent upon
changes in the M. pterygoideus ventralis lateralis. This muscle
becomes larger in some passerine birds with the lateral super-
ficial fibers extending forward along the palatine. It has been
shown that the forward extension of these fibers has several im-
portant functional consequences in addition to the increased
strength of the muscle. These fibers would be longer than all
other fibers of this muscle; hence it would appear reasonable to
suggest that their action is faster and over a longer distance than
the shorter fibers of the M. p. ventralis lateralis. It has been
pointed out above that the insertion of these superficial fibers is
such that their main function is to raise the mandible during the
early phases of closing the bill. This is a rapid, but not a very
powerful closing of the bill. Once the origin of the superficial
fibers reaches the anterior end of the palate, the palatine process
can assume the role of the point of origin. The detailed steps
leading up to this are outlined below.

The tendon of this superficial bundle runs forward along
the palatine to attach somewhere along that bone or at the
junction between the palatine and the premaxilla in the more
advanced forms. This tendon is closely associated with the
periosteum of the palatine, especially where the tendon takes

468 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

origin from the palatine. In the more advanced forms, the ten-
don takes origin from the periosteum of the palatine process.
As the superficial bundle of the M. pterygoideus increases in
size, its tendon shifts away from its original position along the
palatine until it is free of the palatine for most of its length.
Finally, the tendon meets the palatine bone at a slight angle. At
this time, the tendon usually originates from the fused or the
partly fused palatine process, although it has not yet influenced
the position or the fusion of the palatine process which is still
only under the control of the primary selection force. But, as
has been pointed out above, the palatine process fuses to the
palatine bone relatively late in ontogeny. Thus, if the tendon
of the superficial bundle originates on it and if this tendon
approaches the palate at a slight angle, then the distal end of
the palatine process would be ‘‘pulled’’ away from its original
position along the palatine to a new position in the space between
the palate and the jugal bar. The distal end of the palatine
process, then, could not fuse to the prepalatine process. How-
ever, the anterior part of the process still fuses with the palatine,
which apparently fulfills the demands of the primary selection
foree. In its new position, the palatine process would be free
of the palatine and at a slight angle as well as a little ventral
to it. The process would, thus, lie exactly in line with the longi-
tudinal axis of the superficial bundle of the M. pterygoideus and
hence would be in the most effective position as the point of
origin for that muscle. Once the palatine process is free of the
palate, some ossification of the tendon may occur, thereby length-
ening the process. Proof of this point is exceedingly difficult
to obtain because ossified tendons are practically indistinguish-
able from normal bone, both macroscopically and microscopically.
However, this is a minor point with no direct bearing on the
evolution of the free palatine process.

The development of the superficial bundle of the M. p. ven-
tralis lateralis is an exceedingly simple evolutionary change and
has probably occurred independently many times during the
evolution of the Passeres. Here is the basis for the independent
origin, parallel evolution and even convergence of a free palatine
process. The normal palatine process, no matter whether it is
incompletely or completely fused to the prepalatine process, is
preadapted to evolve into the free process with the appearance
of the selection forces favoring this condition (i.e., the develop-
ment of the superficial bundle of the M. pterygoideus). The
appearance of a free process in a group that does not have a

BOCK: PALATINE PROCESS OF THE PREMAXILLA 469

visible palatine process in the adult is facilitated by the genetic
and developmental potential alluded to above.

Loss of the free palatine process and reverse evolution back
to the normal condition will occur whenever the M. pterygoideus
changes back to its original condition, i.e., by losing the super-
ficial bundle of fibers. The loss of this separate bundle is not
associated with decrease in the strength of the bite, but would
occur if the bird no longer needed a fast closing bill or had de-
veloped a more efficient way to close the bill rapidly, as, for
example, by lengthening the bill and palate which automatically
lengthens the M. pterygoideus. In fact, the only birds with a
free palatine process or with tendencies toward this condition
are those which must combine a short bill with rapid closing.
If a group had lost the free palatine process and returned to the
normal condition, there would be no way of distinguishing the
secondary normal condition from the primary or primitive nor-
mal condition. Thus we cannot determine which of the tanagers
possessing a normal palatine process has evolved from ancestors
possessing a free palatine process. Nor can we determine whether
or not the ancestors of the emberizine finches had a free palatine
process. Such determinations must wait until the true phylogeny
of these groups has been established.

e) Evolution of the lateral flange: The evolution of the lateral
flange is far simpler than the evolution of the free palatine
process. Most, if not all, birds have a membrane stretching be-
tween the prepalatine process and the jugal bar or the lateral
edge of the upper jaw. This membrane is part of the lining of
the roof of the mouth. Ossification of this membrane is the
easiest and probably the most efficient way of achieving addi-
tional protection in this region of the skull against the shocks
that may accompany cracking seeds by means of the vise method.
Such ossification would result in an automatic fusion of the pala-
tine process with the palatine if it was not already fused, and
in the subsequent development of the lateral flange as seen in
the cardueline and the ploceid finches. Additional protection
such as the development of overlying pads of rhamphotheca can
evolve after the development of the lateral flanges. I need
scarcely point out the ease with which the lateral flange could
arise independently in different groups and evolve in a parallel
fashion or converge depending upon how distantly related the
eroups were. Also, reversal can occur when the selection force
for protection against the forces and shocks of feeding is no
longer in existence. The secondary normal palatine process, such

470 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

as may occur in some of the thin-billed Drepaniidae, is indis-
tinguishable from the primitive normal palatine process.

TAXONOMIC VALUE OF THE PALATINE PROCESS
OF THE PREMAXILLA

With the aid of the information on the evolution of the pala-
tine process gathered in the preceding section, we can proceed
to the evaluation of its taxonomie significance — this being the
ultimate aim of any study of a taxonomic character. The taxo-
nomic value of the palatine process in relating families of pas-
serine birds will be evaluated first, followed by a discussion of
the relationships of certain families and genera based on the
evidence of the palatine process and, in part, the jaw muscles.
I wish to emphasize that these discussions on the relationships of
the various families and genera are only suggestions based on the
available data; they are not hard and fast opinions or conelu-
sions on the affinities of these groups. Here again, I might
reiterate my earlier statement that it is my belief that we do
not have and probably will not have for many years, the necessary
body of evidence on which to base conclusions on the relationships
within the Passeres.

Taxonomic value. The taxonomic value of a structure depends,
as has been mentioned above, upon the nature of the controlling
selection forces. If these selection forces have a tight control
over the structure and if they have arisen repeatedly and
changed direction often during the evolution of the group, then
that structure would have very little taxonomic value in that
eroup. One could argue about the degree of control on the
palatine process by the several selection forces guiding its evolu-
tion, but it is reasonable to assume that these selection forces
exert a fairly tight control on the palatine process. There is little
question, however, about the fact that the several controlling
selection forces have arisen repeatedly and have reversed their
direction numerous times during the evolution of the perching
birds. Therefore, I would conclude that the palatine process of
the premaxilla has little or no value in showing relationships
between families of passerine birds or in placing problem genera
into the correct family. Here, I must disagree with Tordoff,
who concluded (1954a, p. 33) that the palatine process provides
a good clue to relationships in the New World nine-primaried
oscines. The taxonomic changes advocated by Tordoff are con-
sequently not justified on his evidence and should not be ac-
cepted unless supported by other evidence. A detailed discus-
sion of these changes will be given below.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 471

The Passeres and the New World nine-primaried oscines. The
palatine process of the premaxilla does not offer any clues which
might help solve the problem of the relationship of the Passeres
to other orders, or the problem of the major groupings within
the Passeres. I should mention again that the results of this
study neither support nor refute any of the arrangements pro-
posed in recent years of the Passeres or of the Oscines. There
are, however, parts of each system that are not in agreement with
the available evidence provided by the palatine process and the
jaw muscles. For example, the separation of the Fringillidae
(Carduelinae + Fringilla) from the Emberizinae in both Mayr
and Greenway’s, and Amadon’s lists does not agree with the
similarity in the configuration of the jaw muscles of these
groups. The separation of some of the New World nine-primaried
families in Wetmore’s arrangement does not appear to be justi-
fied in view of the great similarity of these families. In fact,
all of the available morphological evidence indicates that the
nine-primaried oscines are a monophyletie group. The structure
of the palatine process of the premaxilla, although it does not
provide any strong proof for a monophyletic origin of the nine-
primaried group, at least does not argue against it. However,
the relationships within this assemblage are exceedingly complex
and for all practical purposes, are completely unknown. The
probable ancestral group and the directions of evolution within
the group are anyone’s guess. In fact, the family limits cannot
be defined with any degree of certainty; hence many problem
genera cannot be allocated to the proper family or subfamily.
Although much work has been done toward the clarification of
the relationships between the nine-primaried families, much
more must be done before the problem is solved; however, it is
hoped that in the future, more attention will be paid to the
group as a whole and not to whether this family is allied to that
family, or where a certain problem genus should be placed.

Vireonidae. The vireos are, in the opinion of many workers
including myself, the most likely representatives of the ancestral
nine-primaried stock. The reasons for this choice are negative
rather than positive, namely that the vireos are relatively gen-
eralized insect-eaters while most of the other nine-primaried fami-
lies are apparently specialized in one way or another. Rejection
of Tordoff’s hypothesis that the free palatine process as found
in the cardinals is the primitive condition eliminates one of the
serious objections to the vireos being the ancestral nine-primaried
oscines. The vireos have a normal palatine process, a relatively

AT2 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

unspecialized set of jaw muscles, and a tenth primary (only
some species), all of which support the hypothesis that the vireos
represent the ancestral nine-primaried stock. Yet, this evidence
is not very conclusive and much more is needed to verify this
hypothesis.

Vireolanius and Cyclarhis have been included in the Vireo-
nidae. They appear to be heavy-billed vireos with a more
heavily fused palatine process. It should be pointed out, however,
that we do not have a single, thorough, study evaluating the
morphological differences between these genera and the vireos.

Parulidae. The arguments for the origin of the wood-warblers
either from the vireos (Beecher, 1953, p. 307) or from the emberi-
zine finches (Tordoff, 1954a, 1954b, p. 278) are based on
rather weak evidence. Tordoff’s argument stems from his in-
sistence that the free palatine process is the primitive condition
and that evolution proceeded mainly or only in the direction of
reduction and loss of the palatine process. If the wood-warblers
arose from the vireos, which is not an unlikely hypothesis, then
Beecher may be correct in stating that the pinnate muscles of the
warblers (as for example, the M. pseudotemporalis superficialis
which serves to raise the mandible during the early phases of
closing the bill) may have given them an advantage over the
vireos which have only parallel-fibered muscles. Pinnate muscles
do not develop only in response to selection for increased
strength; they may have other functions, perhaps speed, which
we do not suspect and thus could develop in response to selection
for these functions. However, it must be emphasized that there
is very little evidence available that may shed light on the
origin and affinities of the Parulidae.

Icteridae. Again the arguments advanced by Beecher (1951a,
1953) and Tordoff (1954a) for the origin of the Icteridae from
the Emberizinae. are inconclusive; the necessary evidence is
sunply not available. Beecher’s statement that the cowbirds,
Molothrus, represent the ancestral stock of the icterids is pure
speculation. The dickcissel, Spiza, will be discussed below with
the cardinals.

Coerebinae. The New World honey-creepers (Coerebinae) pro-
vide an excellent example of the hodgepodge nature so character-
istic of the nine-primaried families. Beecher (1951b) is prob-
ably completely correct in pointing out the polyphyletie nature of
the honey-creepers, but I am not convinced that he has solved the
convergence problem and clarified the affinities of the coerebid
genera. It is likely that most of the honey-creepers have evolved

BOCK: PALATINE PROCESS OF THE PREMAXILLA 473

from the wood-warblers or from the tanagers, but whether there
are two clear-cut groups, the Coerebini (nectar-feeding warblers)
and the Dacnini (nectar-feeding tanagers), is another question.
Beecher’s interpretations of the similarities in the jaw muscles
between the Coerebini and the warblers and between the Dacnini
and the tanagers are not convincing (see above, p. 400, for a
general discussion of Beecher’s work). Of the other evidence —
the horny palate relief, plumage pattern and differences in the
manner of feeding — presented by Beecher, only the nature
of the ridges on the horny palate appear to have any value in
showing affinities. However since I have not studied this feature,
I do not feel qualified to evaluate it. It is interesting that
Tordoff (1954a, pp. 30-31) has found the palatine process to
be absent in the Coerebini, as in the warblers, but present and
more or less fused with the palatine in the Dacnini, as in the
tanagers, thereby supporting Beecher’s earlier conclusion. I
have found much variation in the degree of fusion between the
palatine process and the palatine bone in the honey-creepers, but
was not able to ascertain any differences in the degree of fusion
between the genera assigned to the Coerebini and those assigned
to the Daenini by Beecher. In view of this inconclusive evidence,
I would agree with Mayr and Greenway (1956) who recognize
the old group ‘‘Coerebinae’’ even though they realize that it
is of polyphyletic origin. The only alternative solution at the
present time is to list the genera of honey-creepers as genera
incertae sedis. However, since both alternatives accomplish the
same end result, it is best to retain the old ‘‘Coerebinae’’ until
we have discovered the true affinities of its genera.
Cardinalinae and Tanagrinae. The cardinals and the tanagers
are so similar to one another in many respects and seemingly
grade so imperceptibly into one another that they are best dis-
cussed together. Beecher (1953) and Tordoff (1954a) agree
that the cardinals are related to the tanagers, but disagree on the
direction of evolution. Nelson (1954) agrees that the two groups
are related on the basis of similarity in the pneumatic fossa of
the humerus, but believes that these families are not related to the
Emberizinae (see Berger, 1957, pp. 266-267, for a discussion of
the pneumatic fossa of the humerus). There cannot be any
question that the cardinals and the tanagers are very similar
morphologically, but I doubt that the evidence advanced by these
authors proves that they are closely related within the nine-
primaried assemblage. Nor can the evidence be used to dis-
tinguish between the cardinals and the tanagers, or between

44. BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

either or both families and the emberizine finches. Mayr (1955,
p. 34) pointed out that the direction of evolution in the nine-
primaried oscines, and especialy in these three families, is in
doubt. This is still true. It seems certain that the free palatine
process-of the cardinals is not the primitive condition in the
New World nine-primaried oscines, but we cannot tell whether
the emberizine finches gave rise to the tanagers or vice versa, or
whether the cardinals gave rise to the tanagers or vice versa, or
whether these groups are more distantly related. We cannot even
be sure that the cardinals and the tanagers are good taxonomic
groups. For these reasons, I do not see the point of worrying
about the correct position of the numerous problem genera —
especialy those from tropical regions —of the nine-primaried
finches. What difference does it make whether Saltator is a
tanager or a cardinal, or whether Chlorospingus and Oreothrau-
pis are tanagers or emberizine finches when we have no idea of
the limits of the familial groups.

One genus, Spiza, must be discussed because of the importance
placed on it by Tordoff. In both of his papers (1954a, 1954hb, p.
218), Tordoff placed Spiza in the Cardinalinae and claimed that
it represents the ancestral fringillid and consequently the an-
cestral nine-primaried stem stock better than any other living
genus. This opinion is based on his belief that the free palatine
process is primitive; his other reasons supporting this opinion
are incidental. On the other hand, Beecher (1951a, 1953, p. 309)
claimed that Spiza is an icterid on the basis of the configuration
of the jaw muscles, although he did not figure these muscles
in either paper. There is no question about the structure of the
palatine process in the dickcissel; it is free as in the cardinals,
but this is far from proof that it is a member of the Cardinalinae.
Unfortunately, I have not been able to dissect the jaw muscles of
the dickeissel, but I would be most surprised if a superficial
bundle of the M. pterygoideus ventralis lateralis did not originate
from the free palatine process. It is of interest that in some
icterids, such as the grackle, the M. p. ventralis lateralis sends
a lateral tendon forward for a short distance along the palatine.
This indicates that the potential for developing a separate super-
ficial bundle of the M. pterygoideus and a free palatine process
may be present in the icterids. Most of the other evidence cited
by both Tordoff and Beecher is likewise inconclusive. For ex-
ample, the similarity in color pattern between the dickcissel and
the meadowlark may well be convergence because of their living
in the same habitat, as has been pointed out by Friedmann

BOCK: PALATINE PROCESS OF THE PREMAXILLA 475

(1946). New and different evidence is needed before the affini-
ties of Spiza are clarified.

The swallow-tanager, Tersina, scarcely deserves a separate dis-
cussion. In the structure of the palatine process, this genus
appears to be a tanager, but whether it should be placed in that
eroup or in a group of its own is another problem. It is
interesting that the true swallows usually have an unfused and
quite distinet palatine process. Although I have not dissected
the jaw muscles of these groups (the one swallow I examined
was a damaged specimen), I would suspect that a part of the M.
pterygoideus took origin from the palatine process and had the
function of raising the mandible rapidly, as in the cardinals and
the tanagers. I also suspect that such an arrangement of the M.
pterygoideus will be found in some of the tyrant-flycatchers.

Fringillidae and Emberizinae. The Fringillidae of older
authors are, with little doubt, one of the most controversial fami-
lies in the nine-primaried complex. Most of the recent arguments
about relationships in the nine-primaried oscines center about
the Fringillidae or start with them. This family according to the
older classifications, e.g., Wetmore, is composed of the Emberi-
zinae, the Cardinalinae and the Fringillinae (including the ear-
dueline finches). Many recent workers suspected, however, that
this family is polyphyletic in origin and have set out to uncover
the true affinities of the several subfamilies. The conclusions
of the various authors differ radically, but they need not be
summed up here. The only one that will be considered is
Tordoff’s conclusion that the cardueline finches (excluding
Fringilla) do not belong to the nine-primaried group, but are
related to the ploceids. Fringilla, according to Tordoff, is a
primitive emberizine finch. The only recent modification of
Tordoff’s conclusions has been that of Mayr, et al. (1956) who
concluded that Fringilla is closely related to the carduelines.
Because I believe that Tordoff’s reasons for separating the
eardueline finches from the nine-primaried assemblage are
baseless, it is necessary to re-examine the entire problem of the
‘‘Wringillidae.’’? In doing so, I will use only the morphological
evidence, especially that supplied by the cranial anatomy.

There seems to be little disagreement about the conclusion that
the cardinals are closely related to the emberizine finches. If the
eardinals are included in the ‘‘Fringillidae,’’ then the tanagers
will most likely also have to be included. However, in view of
the inconclusive evidence pertaining to the affinities between
the eardinals and the emberizine finches, this aspect of the

476 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

problem will be dropped and attention will be concentrated on
the relationships between the emberizine finches and the cardue-
lines plus Fringilla.

Mayr, et al. (1956) have presented much evidence pointing to a
close affinity between Fringilla and the cardueline finches; one of
the most serious disagreements was the unfused palatine process in
Fringilla as compared to the fused process and the lateral flange
in the carduelines. The results of the present study have shown
that the presence of an unfused palatine process in Fringilla and
its apparent absence in the earduelines does not necessarily mean
that the two groups are unrelated, as supposed by Tordoff. Al-
though Fringilla does not possess a lateral flange, its jaw muscles
are almost identical to those in the cardueline finches as has been
reported earlier by Fiedler (1951) and Beecher (1953). The
most distinctive feature of the cardueline jaw musculature is
the enlarged medial portion of the M. pseudotemporalis super-
ficialis. Fringilla agrees completely with the carduelines in this
feature. The major difference between the jaw muscles of the
two groups is the poorly developed M. adductor mandibulae in
Fringilla which is similar to that found in the emberizine finches
(see Fiedler, 1951, p. 242). This muscle is highly developed in
the carduelines with two parts expanding over the side of the
skull. In addition, some of the emberizine finches, e.g., the fox
sparrow, tend toward the cardueline finches in the structure
of the M. pseudotemporalis superficialis and the M. pterygoideus.
Thus in several aspects of the bony palate and the jaw muscula-
ture, Fringilla is intermediate between the emberizine and the
eardueline finches, but it is closer to the carduelines. On the
other hand, none of the ploceid or estrildid finches, to my knowl-
edge, tend toward the cardueline arrangement of the jaw muscles,
especially in the very distinctive condition of the M. pseudotem-
poralis superficialis. I must emphasize, however, that I have
dissected only the house sparrow (Passer) and the heavy-billed
Coliuspasser and examined Beecher’s plates of the jaw muscles
in the ploceid and estrildid finches so that my sample of these
finches is quite small.

Although no definite conclusions on the relationships between
the fringillid, emberizid and ploceid finches can be based on
the available evidence from the cranial morphology, several
hypotheses may be advanced. I cannot see any indications of a
relationship between the carduelines plus Fringilla and the
ploceids or the estrildids, as advanced by Tordoff. Because of the
intermediate position of Fringilla in many features of the cranial

BOCK: PALATINE PROCESS OF THE PREMAXILLA ATT

anatomy, it does seem reasonable to suggest that the emberizine
finches gave rise to the carduelines through a Pringilla-like group.
The eardueline radiation could have eliminated an earlier Frin-
gilla radiation except for Pringilla itself. If this hypothesis is
correct, then the Fringillidae of Wetmore is a monophyletic
group and the similarity between the Carduelinae and the
Kstrildidae and the Ploceidae in the structure of the lateral
flange would be convergence. It is entirely possible, however,
that the carduelines plus Fringilla are related to the ploceids or
the estrildids and that the similarity between the fringilline
finches and the emberizines in the structure of the jaw muscles
is due to convergence.

Drepanudae. The origin of the Hawaiian honey-creepers re-
mains a deep mystery even after many workers have devoted
much time to this problem. One of the main reasons why the
origin of the Drepaniidae has eluded ornithologists is because the
structures so far studied were those which could never supply
conclusive clues to their affinities. Although Drepaniidae con-
stitute one of the best examples of adaptive radiation in feeding
methods in the Passeres, the investigations of most workers on
the origin of this group are based on characters associated with
feeding methods. The present study of the palatine process has
shed some light on this problem and I wish to offer the following
ideas as a suggestion.

The group that gave rise to the Hawaiian honey-creepers had
to have three characteristics. They had to be birds that wander
in flocks, preferably erratically, over long distances, and breed at
the place to which they have wandered. Of all the possible an-
cestral groups, the cardueline finches are the only ones that
possess all three attributes. Here again, it must be pointed out
that the objection to the carduelines as the ancestral drepaniids,
presented by Tordoff, has been removed. It is quite possible that
a cardueline finch or a member of its ancestral group had reached
the Hawaiian Islands and gave rise to the Drepaniidae, a hypoth-
esis already advanced by Sushkin (1929). The possession of a
finch bill, even a heavy ecardueline bill with its lateral flanges
on the prepalatine process, does not present any problem for this
hypothesis. The lateral flanges can be lost whenever the selection
forces for them are eliminated or reversed. It is interesting that
the bones of the palate and upper jaw of the heavy-billed genus
Psittirostra are identical to those found in the large-billed ecar-
duelines. According to Beecher (1953, p. 311), the jaw muscles,
including the expanded medial part of the M. pseudotemporalis

478 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

superficialis, of Psittirostra are almost identical to those of the
eardueline finches. This suggested origin of the Drepaniidae is
speculation and must be verified or rejected on additional evi-
dence. I have offered it mainly to counteract the trend of think-
ing that the Hawaiian honey-creepers had to originate from a
thin-billed member of the nine-primaried oscines.

CONCLUSIONS

A sharp distinction was made in the beginning of this paper
between comparative studies of many characters in two or a few
passerine families and comparative studies of a single character
or character complex in the entire order Passeres, with the tacit,
but clear implication that the former rarely led to conclusive
taxonomic results. It is now necessary to evaluate the conelu-
sions which may be gained in the latter type of study.

The first thing that probably comes to the mind of most readers
is the highly speculative and inconclusive taxonomic results pre-
sented above. In fact, the only definitive conclusions were nega-
tive ones. These inconclusive results are scarcely surprising in
view of the fact that they are based on the evidence supplied by
a single character complex. Nevertheless, these are not very
encouraging results for a taxonomic paper. Let us look, however,
on the positive side of the ledger. This study has provided de-
tailed knowledge of the structure and variation of the palatine
process along with some information on its embryological origin
and a fairly good idea of its functional significance. With this
information, it was possible to gain a reasonably accurate pic-
ture of the evolution, although not of the actual phylogeny, of
the palatine process. Such data, especially those on the evolution,
are only rarely supphed by comparative studies of two families.
Yet one can ask, what can be done with this information? If
only a few inconclusive taxonomic results can be reached after
such an extensive study as the present one, then why continue
with these studies if one is primarily concerned with the classifi-
cation of the perching birds? With comparative studies of
families, we at least know how different these families are.
Would it not be best to return to these studies?

The answer, or at least my answer, to the last question
would be no. Still there is no reason to continue these extensive
studies of single characters for the sake of the taxonomic dedue-
tions that may be reached in individual works. But then, why do
taxonomic conclusions have to be offered at the end of every

BOCK: PALATINE PROCESS OF THE PREMAXILLA 479

taxonomic study? Is it not enough to present the data in such
a way that they can be used to help formulate sound judgments
of affinities when enough data have been gathered? With only
a single study, such as the present one, very little if anything
ean be learned about the evolution and relationships of the Pas-
seres. Perhaps a bit more can be learned with two or three
studies. However, with several dozen such works, there would be
a good chance of unraveling the entire evolutionary history of
the recent Passeres. The evolution of the individual characters
could be pieced together to form an overall evolution of the
Passeres. Use of a large number of different characters will
reduce the chance of error to an insignificant amount. This over-
all picture of the evolution of the Passeres will then serve as the
basis for a classification of the order. It is my firm belief that
a sound classification of the passerine birds can best be reached
through a series of single character studies. This is a long and
tedious task with relatively few rewards until the central prob-
lem is finally solved. But there are no short cuts to the goal,
so that the sooner thorough analyses of single characters are
begun, the sooner we will have a suitable classification of the
Passeres.

SUMMARY

A comprehensive analysis of a single character — the palatine
process of the premaxilla—in the Passeres is presented as a
basis on which to judge the merits of the ‘‘single character
study’’ approach to passerine classification.

The four major conditions — fused, unfused, free, and with a
lateral flange —of the palatine process are described.

A history of the past studies is presented in which the origin
of the erroneous term ‘‘palato-maxillary’’ for the palatine proc-
ess is shown.

The embryology of the palatine process in birds possessing
the process and in birds supposedly possessing the ‘‘palato-
maxillary’’ is compared. It is shown that these terms refer to
the same structure in the passerines and hence ‘‘palato-maxil-
lary’’ should be dropped from use since it is the younger term.

The functional significance of the palatine process is discussed
with two separate problems in mind. The first is the main-
tenance of the process in birds. It is shown that the primary
function of the palatine process is insurance of a firm palate-
upper jaw connection which is associated with the kinetic skull.
The second problem concerns the modifications of the palatine

480 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

process in the finch-like Passeres. These modifications are associ-
ated with certain aspects of seed-eating. Therefore, it is neces-
sary to investigate the correlation between the process and the
M. pterygoideus (the only muscle attached to the palate), the
different ways to increase the force of the bite and the methods by
which the several groups of finches crack seeds. It is shown that
the free palatine process serves as the point of origin for part
of the M. pterygoideus and is associated with the nutcracker
method of cracking seeds. This is an adaptation for a rapidly
closing mandible. The lateral flange is associated with the vise
method of cracking seeds, and is an adaptation for protecting the
brain and sense organs from the forces and shocks that accom-
pany the cracking of seeds.

Finally, the results of a survey of the palatine process in the
Passeres are presented with a résumé of the types of inter- and
intrafamilial variation.

The evolution of the palatine process is outlined with a short
discussion of the several evolutionary principles that figure in
its evolution. The normal condition of the palatine process —
lying along the palatine and fused or unfused to it —is assumed
to be primitive. Change from the fused to the unfused condition
and vice versa has doubtless occurred repeatedly during the
evolution of the Passeres. Evolution of the free process from
the unfused process is under the control of selection forces for a
fast-closing bill, while the evolution of the lateral flange is under
the control of selection forces for protection of the braincase.
These evolutionary changes can be reversed with a reversal of
the selection forces. Apparently, the free process cannot evolve
directly into the lateral flange; it must first pass through the
normal condition.

It is concluded that the palatine process has little value in
showing relationships between families of passerine birds. Some
of the problems of relationships within the nine-primaried
oscines are discussed, but no definite conclusions are reached.

Although individual studies of single characters, such as the
present one, do not lead to definite taxonomic conclusions, it is
suggested that comprehensive studies of single characters
throughout the Passeres will eventually provide the best basis
for understanding relationships within the order and it is
urged that more studies of this type be undertaken.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 481

LITERATURE CITED

AMADON, D.

1943. The genera of starlings and their relationships. Amer. Mus.
Novitates, No. 1247:1-16.

1944. The genera of Corvidae and their relationships. [bid., No. 1951:
1-21.

1950a. The Hawaiian honeycreepers (Aves, Drepaniidae). Bull. Amer.
Mus. Nat. Hist., 95:151-262.

1950b. Australian mud-nest builders. Emu, 51:123-127.

1951. Taxonomic notes on the Australian butcher-birds (family Crac-
ticidae). Amer. Mus. Novitates, No. 1504:1-33.

1956. Remarks on the starlings, family Sturnidae. Jbid., No. 1803:
1-41.

1957. Remarks on the classification of the perching birds (order
Passeriformes). Proc. Zool. Soc., Calcutta, Mookerjee Memor.
Vol.; 259-268.

ARVEY, M. D.
1951. Phylogeny of the waxwings and allied birds. Uniy. Kansas Publ.
Mus. Nat. Hist., 3:437-530.

ASHLEY, J. F.
1941. A study of the structure of the humerus in the Corvidae. Con-
dor, 43:184-195.

BEDDARD, F. E.
1898. The structure and classification of birds. London, xx + 548 pp.

BEECHER, W. J.
1951a. Adaptations for food-getting in the American blackbirds. Auk,
68 : 411-440.
1951b. Convergence in the Coerebidae. Wilson Bull., 63:274-287.
1953. A phylogeny of the oscines. Auk, 70:270-333.

BERGER, A. J.
1957. On the anatomy and relationships of Fregilupus varius, an ex-
tinct starling from the Mascarene Islands. Bull. Amer. Mus.
Nat. Hist., 113:225-272.

Bock, W. J.
1959. Preadaptation and multiple evolutionary pathways. Evolution,
13:194-211.

CHAPIN, J. P.
1917. The classification of the weaver birds. Bull. Amer. Mus. Nat.
Hist., 37:243-280.
1949. Pneumatization of the skull in birds. Ibis, 91: 691.

482 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

CuLaRK, H. L.
1912. Note on the Laysan finch. Auk, 29:166-168.
1913a. Notes on the Panama thrush-warbler. Ibid., 30:11-15.
1913b. Anatomical notes on some genera of passerine birds. Tbid.,
30 : 262-267.
1913c. Anatomical notes on Todus, Oxyruncus and Spindalis. Ibid.,
30: 402-406.

DELACOUR, J.

1943a. A revision of the genera and species of the family Pyenonotidae
(bulbuls). Zoologica, 28:17-28.

1943b. A revision of the subfamily Estrildinae of the family Ploceidae.
Ibid., 28:69-86.

1944. A revision of the family Nectariniidae (sunbirds). Ibid., 29:
17-38.

1946. Les Timaliines. L’Oiseau, 16:7-36.

DELACOUR, J. AND D. AMADON
1949. The relationships of Hypocolius. Ibis, 91:427-429.

DELACOUR, J. AND C. VAURIE
1957. A classification of the oscines (Aves). Contributions in Science
(Los Angeles County Museum), 16:1-6.

DULLEMEIJER, P.
1951. The correlation between muscle system and skull structure in
Phalacrocorax carbo sinensis (Shaw and Nodder). Proc. Kon.
Akad. Wet. Amsterdam, 54:247-259, 400-404, 533-536 ; 55:95-102.

EBER, G.
1956. Vergleichende Untersuchungen iiber die Ernahrung einiger Fink-
envogel. Biologische Abhandlungen, No. 13/14:1-60.

ENGELS, W. L.
1940. Structural adaptations in thrashers (Mimidae; genus Toxostoma)
with comments on interspecific relationships. Univ. Calif. Publ.
Zool., 42:341-400.

FIEDLER, W.
1951. Beitrige zur Morphologie der Kiefermuskulatur der Oscines.
Zool. Jahrb., 71:235-288.

FIsHEr, H. I.
1955. Some aspects of the kinetics in the jaws of birds. Wilson Bull.,
67:175-188.

ForBes, W. A.
1880. Contributions to the anatomy of passerine birds. — Part III.
On some points in the structure of Philepitta and its position
amongst the Passeres. Proc. Zool. Soe. London: 387-391.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 483

1881. (Same title) Part IV. On some points in the anatomy of the
genus Conopophaga and its systematic position. Lbid.:435-438.

1882. (Same title) Part VI. On Xenicus and Acanthisitta as types of
a new family (Xenicidae) of Mesomyodian Passeres from New
Zealand. Ibid.:569-571.

FRIEDMANN, H.
1946. Ecological counterparts in birds. Sci. Monthly, 63:395-398.

FURBRINGER, M.
1888. Untersuchungen zur Morphologie und Systematik der Vogel.
Jena and Amsterdam, 2 vols., 1751 pp.

Gapow, H. (and E. SELENKA)
1891, Vogel. Jn Bronn, H. G., Die Klassen und Ordnungen des Thier-
1893. reichs. Leipzig, vol. 6, div. 4, pt. 1 (1891), 1008 pp.; pt. 2
(1893), vii + 303 pp.

GarRrRop, A. H.
1872. Notes on the anatomy of the huia bird (Heteralocha gouldi).
Proce. Zool. Soe. London: 643-647.
1877. Notes on the anatomy of passerine birds. Part Il. [bid.:447-452.

Haas, O. AND G. G. SIMPSON
1946. Analysis of some phylogenetic terms, with attempts at redefini-
tion. Proc. Amer. Philos. Soc., 90:319-349.

HELLMAYR, C. E.
1924- Catalogue of birds of the Americas. Field Mus. Nat. Hist.,
1938. Zool. Ser., 13, pts. 3-11.

Hupson, G. E. anp P. J. LANZILLOTTI
1955. Gross anatomy of the wing muscles in the family Corvidae.
Amer. Midl. Nat., 53:1-44.

Huaeins, R. A., e¢ al.
1942. Ossification in the nestling house wren. Auk, 59:532-543.

JOLLIE, M.
1957. The head skeleton of the chicken and remarks on the anatomy
of this region in other birds. Jour. Morph., 100:389-436.
1958. Comments on the phylogeny and skull of the Passeriformes. Auk,
75 26-35.

Kripp, D. VON
1935. Die mechanische Analyse des Schnabels. Krimmung und ihre
Bedeutung fiir die Anpassungsforschung. Morph. Jahrb., 76:
448-494.

LAKJER, T.
1926. Studien tiber die Trigeminus-versorgte Kaumuskulatur der Sau-
ropsiden. Copenhagen, 154 pp.

484 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Lows, P. BR.
1924. On the presence of broadbills (Eurylaemidae) in Africa. Proc.
Zool. Soe. London, pt. 1:279-291.
1931. On the anatomy of Pseudocalyptomena and the occurrence of
broadbills (KHurylaemidae) in Africa. Ibid., pt. 1:445-461.
1938a. Some anatomical and other notes on the systematic position of
the genus Picathartes together with some remarks on the
families Sturnidae and Eulabetidae. Ibis, (14)2:254-269.

1938b. Some anatomical notes on the genus Pseudochelidon Hartlaub
with reference to its taxonomic position. [bid.:429-437.

1947. Note on the systematic position of Hypocolius. Proc. Zool. Soc.
London, 117:109-114.

1949. On the position of the genus Zavattariornis. Ibis, 91:102-104.

Lucas, F. A.

1888. Notes on the osteology of the thrushes, Miminae and wrens.
Proc. U.S. Nat. Mus., 11:173-180.

1890. Notes on the osteology of the Paridae, Sitta and Chamaea.
Ibid., 13:337-345.

1894. Notes on the anatomy and affinities of the Coerebidae and other
American birds. Jbid., 17:299-312.

1895. Osteological and pterylographical characters of the Procniatidae.
Ibid., 18:505-517.

MARIEN, D.
1951. Notes on the bird family Prunellidae in southern Eurasia. Amer.
Mus. Novitates, No. 1482:1-28.

Mayr, E.
1943. What genera belong to the family Prionopidae? Ibis, 85:216-
218.
1955. Comments on some recent studies of song bird phylogeny. Wil-
son Bull., 67:33-44.
1958. The sequence of songbird families. Condor, 60:194-195.

Mayr, E. anD D. AMADON
1947. A review of the Dicaeidae. Amer. Mus. Novitates, No. 1360:
1-32.
1951. A classification of recent birds. Ibid., No. 1496:1-42.

Mayr, E. anp J. BOND
1943. Notes on the generic classification of the swallows, Hirundini-
dae. Ibis, 85:334-341.

Mayr, E. anp J. C. GREENWAY, JR.
1956. Sequence of passerine families (Aves). Breviora, Mus. Comp.
Zool., No. 58:1-11.

Mayr, E. AND C. VAURIE
1948. Evolution in the family Dieruridae (birds). Evolution, 2:238-
265.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 485

Mayr, E., R. J. ANDREW AND R, A. HINDE
1956. Die systematische Stellung der Gattung Fringilla. Jour. Orn.,
97 : 258-273.

MEINERTZHAGEN, R.
1951. Review of the Alaudidae. Proce. Zool. Soc. London, 121:81-1382.

Mouuer, W.
1930. Uber die Schnabel- und Zungenmechanik bliitenbesuchender
Vogel. I. Biologia Generalis, 6:651-726.
1931. (Same title) II. Ibid., 7:99-154.

MOLLIER, G.
1937. Beziehungen zwischen Form and Funktion der Sehnen im
Muskel-Sehnen-Knochen-System. Morph. Jahrb., 79:161-199.

NELSON, G. E.
1954. Relationships of the avian family Fringillidae. Unpublished
Thesis, Univ. of Florida: 1-89.

PARKER, W. K.

1866. On the structure and development of the skull of the ostrich
tribe. Philos. Trans. Roy. Soc. London, 156:113-183.

1869. On the structure and development of the skull of the common
fowl (Gallas domesticus). Ibid., 159:755-807.

1872. On the structure and development of the crow’s skull. Monthly
Micr. Jour., 8:217-226, 253.

1873a. On the development of the skull of the tit and the sparrow hawk.
Tbid., 9:6-11, 45-50.

1873b. On the development of the skull in the genus Turdus — the
thrushes. Ibid., 9:102-107.

1875a. On the morphology of the skull in the woodpeckers (Picidae)
and the wrynecks (Yungidae). Trans. Linn. Soe. London, (2)
1:1-22.

1875b. On the structure and development of the bird’s skull. Ibid., (2)
1:99-154.

1875c. On aegithognathous birds (part I). Trans. Zool. Soe. London,
9:289-352.

1877. On the skull of the aegithognathous birds — part II, Jbid., 10:
251-314.

PETERS, J. L.
1951. Check-list of birds of the world. Vol. VII.

PFUHL, W.
1936. Die gefiederten Muskeln, ihre Form und ihre Wirkungsweise.
Zeitschr. Anat. Entwickl., 106:749-769.

PYCRAFT, W. P.
1905a. Contributions to the osteology of birds — Part VII. Eurylaemi-
dae; with remarks on the systematic position of the group. Proc.
Zool. Soe. London, pt. 2:30-56.

486 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

1905b. On the systematic position of Zeledonia coronata with some
observations on the position of Turdidae. Ibis, (8)5:1-24.

1905e. Some points on the anatomy of Acanthidositta chloris, with some
remarks on the systematic position of the genera Acanthidositta
and Xenicus. Ibid.: 603-621.

1907. Contributions to the osteology of birds— Part IX. Tyranni;
Hirundines; Lanii; and Gymnorhines. Proc. Zool. Soc. London,
pt. 1:352-379.

Ranp, A. L. AND M. A. TRAYLOR, JR.
1953. The systematic position of the genera Ramphocaenus and
Microbates. Auk, 70:334-3387.

REMANE, A.
1956. Die Grundlagen des natiirlichen Systems, der vergleichenden
Anatomie und der Phylogenetik. Leipzig, 2nd. ed., vi + 364 pp.

Ripiey, S. D.
1952. The thrushes. Postilla, No. 13:1-48.

SEWERTZOFF, A. N.
1931. Morphologische Gesetzmissigkeiten der Evolution. Jena, xiv
-+ 371 pp.

SHUFELDT, R. W.
1888. The osteology of Habia melanocephala with comparative notes
upon the skeleton of certain conirostral birds, and of tanagers.
Auk, 5:488-444,

SIMPSON, G. G.
1953. The major features of evolution. New York, xx + 434 pp.

Sims, R. W.
1955. The morphology of the head of the hawfinch (Coccothraustes
coccothraustes). Bull. Brit. Mus. (Nat. Hist.), 2:369-393.

STALLCuP, W. B.
1954. Myology and serology of the avian family Fringillidae. Univ.
Kansas Publ., Mus. Nat. Hist., 8:157-211.

STARCK, D.
1959. Neuere Ergebnisse der vergleichenden Anatomie und ihre Bedeu-
tung fiir die Taxonomie erlautert an der Trigeminus-Muskulatur
der Vogel. Jour. Orn., 100:47-59.

Stonor, C. R.
1937. On the systematic position of the Ptilonorhynchidae. Proc.
Zool. Soe. London, (B) 107:475-490.
1938. Some features on the variation of the birds of paradise. Ibid.,
(B) 108:417-481.
1942. Anatomical notes on the New Zealand wattled crow (Callaeas)
with especial reference to its powers of flight. Ibis, (14)6:1-18.

BOCK: PALATINE PROCESS OF THE PREMAXILLA 487

STRESEMANN, E.

1959.

The status of avian systematics and its unsolved problems. Auk,
76 : 269-280.

SUSHKIN, P. P.

1924.

1929.

(On the Fringillidae and allied groups.) Bull. Brit. Orn. Club.,
45 :256-261.

The evening grosbeak (Hesperiphona), the only American
genus of a Palaearctic group. Auk, 42:256-261.

On the anatomy and classification of the weaver-birds. Bull.
Amer. Mus. Nat. Hist., 57:1-32.

On the systematic status of the Drepanidae. Proc. 6th Internat.
Ornith. Cong., 1926:379-381.

SWINEBROAD, J.

1954.

A comparative study of the wing myology of certain passerines.
Amer. Mid]. Nat., 51:488-514.

Torporr, H. B.

1954a.

1954b.

VAURIBE, C.
1949.

1951.

1953.

A systematic study of the avian family Fringillidae based on
the structure of the skull. Mise. Publ., Mus. Zool. Univ. Mich.,
No. 81:1-42.

Relationships in the New World nine-primaried oscines. Auk,
71 :273-284.

A revision of the bird family Dicruridae. Bull. Amer. Mus. Nat.
Hist., 93:199-341.

A study of Asiatic larks. Ibid., 97:431-526.

A generic revision of flyeatchers of the tribe Muscicapini. [bid.,
100: 453-538.

WETMORE, A.

1951.

1957.

Muscles:

A revised classification for the birds of the world. Smithsonian
Mise. Coll., 117:1-22.

The classification of the Oscines, Passeriformes. Condor,
59: 207-209.

ABBREVIATIONS USED IN THE FIGURES

MAM = M. adductor mandibulae

MDM = M. depressor mandibulae

MPDL = M. pterygoideus dorsalis lateralis
MPDM = M. pterygoideus dorsalis medialis
MPQ = M. protractor quadrati

MPTP = M. pseudotemporalis profundus
MPTS = M. pseudotemporalis superficialis

488 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Muscles:

‘¢MRP’’ = ‘‘M. retractor palatini’’

MPVL — M. pterygoideus ventralis lateralis

MPVM = M. pterygoideus ventralis medialis

TT — tendon of the superficial bundle of the M. pterygoideus ventralis
lateralis

3ones:

BT = basitemporal plate

DPM = dentary process of the premaxilla

IPP = interpalatine process of the palatine

JB =jugal bar

LF = lateral flange of the prepalatine process

M = maxilla

MP = maxillo-palatine

MPP = mediopalatine process of the palatine

MSP = medial shelf of the palatine

NFH = nasal-frontal hinge

P = palatine (except in figures 23 to 28)

PM = premaxilla

PP = prepalatine process of the palatine

PPM = palatine process of the premaxilla (in figures 23 to 28, ab-
breviated as P)

PT = pterygoid

Q = quadrate

TPP = transpalatine process of the palatine

V =vomer

eo

¥

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE
Won. (122, .No:'9

THE SNAKES OF ECUADOR
A Check List and Key

By JAMEs A. PETERS

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM

JUNE, 1960

PUBLICATIONS ISSUED BY OR IN CONNECTION
WITH THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

BULLETIN (octavo) 1863 — The current volume is Vol. 122.

BREVIORA (octavo) 1952 — No. 126 is current.

Memorrs (quarto) 1864-19388 — Publication was terminated with
Vol. 55.

JOHNSONIA (quarto) 1941— A pubheation of the Department of
Mollusks. Vol. 3, no.39 is current.

OCCASIONAL PAPERS OF THE DEPARTMENT OF MOLLUSKS (octavo)
1945 — Vol. 2, no. 25 is current.

PROCEEDINGS OF THE NEw ENGLAND ZooLoaicaAL CLUB (octavo)
1899-1948 — Published in connection with the Museum. Publication
terminated with Vol. 24.

The continuing publications are issued at irregular intervals in num-
bers which may be purchased separately. Prices and lists may be
obtained on application to the Director of the Museum of Comparative
Zoology, Cambridge 38, Massachusetts.

Of the Peters ‘‘Check List of Birds of the World,’’ volumes 1-3 are
out of print; volumes 4 and 6 may be obtained from the Harvard Uni-
versity Press; volumes 5 and 7 are sold by the Museum, and future
volumes will be published under Museum auspices.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE
Vor. 122, No. 9

THE SNAKES OF ECUADOR
A Check List and Key

By JAMES A. PETERS

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM

JUNE, 1960

No. 9 —The Snakes of Ecuador. A Check List and Key.

By JAMES A. PETERS

INTRODUCTION

As my work on the herpetofauna of Ecuador progressed, it
became obvious that a complete key to the genera and species to
be found in that country had to be prepared. Also, an adequate
check list of the fauna including known and expected taxa was
needed, for work on the geographical and ecological problems
was halted until such a time as the actual fauna was better
known. This check list and key to the snakes of Ecuador is, I
hope, the first of a series to cover the entire reptilian and am-
phibian species of the country. In the course of its preparation,
it has been necessary to make gestures in the direction of solution
of several problems that were and still remain of considerable
magnitude in South American herpetology. I have tried to avoid
attacking these problems on the basis of artificial, political
boundaries, particularly on the generic and familial level. The
use of political units as the basis for analysis of a fauna is an
expedient one, but it should not be extrapolated to form the
basis of a taxonomic review or revision. If work is to be done
on a particular genus, it should, if possible, cover the genus
throughout its range, not just that genus within a political unit.
Much of the difficulty encountered today in South American
herpetology results from earlier use of this restricted approach.

METHODS

This check list has been drawn up with an eye toward making
it of maximum usefulness to anyone interested in identifying a
snake from Eeuador. It is, in fact, left in the same arrangement
in which I have worked with it over the past ten years, and which
I found most useful. The genera are arranged alphabetically
throughout, rather than phylogenetically (for the phylogeny of
Ecuadorian snakes, see page 493). The species are also arranged
alphabetically within their genera. The phylogeny of many if
not most South American snake genera is so poorly known that
any attempt at an overall phylogenetic arrangement would be
presumptuous, at best. Under each taxon I have included, first,
a citation to the original description of that taxon plus its type
locality and its holotype; second, citations for any taxon with its
type locality in Eeuador that has been synonymized with the
species, its type locality and its holotype; and third, as accurate

2

492 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

a statement of the entire range of the taxon as possible. I have
made no changes in status of any taxon solely within the pages
of this check list. Although no documentation as to the authority
for changes of generic position has been presented in the species
synonymy, this is deliberately omitted, and its absence is not to
be construed to mean that the combination appearing here is a
new one. On the contrary, there are no new combinations in this
list, whatsoever. The generic position of every species has been
previously documented in the literature, with one exception: I
have followed Bogert’s suggestion (in litt.) that Synophis and
Diaphorolepis be considered distinet genera, and in this act an-
ticipate his published documentation.

The presence in the keys of a genus or species name followed
by an asterisk indicates that there is no valid record of the
occurrence of that taxon in Ecuador, but that the likelihood of
its being taken there in the future is sufficiently good that its
presence should be anticipated. It should not be construed from
this that I have been so perspicacious that I have successfully
anticipated and ineluded all likely new members of the fauna.
Any specimen that does not fit in the key is not necessarily a
representative of a species new to science, and must needs be
checked against the considerable literature on South American
snakes. This is particularly true of specimens from the Ama-
zonian slope.

Under each species, the reference to the holotype in the species
citation includes first an abbreviation of the name of the museum
which acts as custodian for that specimen, and the catalogue
number assigned to the specimen by the museum.

The abbreviations used are as follows:

ANSP — Academy of Natural Sciences, Philadelphia
BerM — Berlin Museum

BM — British Museum (Natural History )

CNHM — Chicago Natural History Museum

DroM — Drottningholm Museum

GothM — Gothenberg Museum

GottM — Gottingen Museum

HM — Hamburg Museum

IB — Instituto Butantan, Sao Paulo, Brazil

Id1S — Instituto de la Salle, Bogota, Colombia

IRB — Institut Royale d’Histoire Naturelle de Belge, Brussels
JAP — James A. Peters Collection

LeyM — Leyden Museum

LunM — Lund Museum

MAF — Museum Adolphus Frederici

PETERS: SNAKES OF ECUADOR 493

MCZ — Museum of Comparative Zoology, Harvard

MilM — Milan Museum

MonM — Monaco Museum

MPB — Musée des Pays-Bas

MunM — Munich Museum

MVZ — Museum of Vertebrate Zoology, Berkeley

PM — Museum National d’Histoire Naturelle de Paris

RMS — Royal Museum, Stockholm

Senck — Senckenberg Museum, Frankfort, Germany

SU — Natural History Museum, Stanford University, California
TurM — Turin Museum, Italy

UMMZ — University of Michigan Museum of Zoology, Ann Arbor
UNZM — University of Naples Zoological Museum, Italy

USNM — United States National Museum, Washington, D. C.
VM — Vienna Museum

Z1U — Zoological Institute, Uppsala, Sweden

ZSBS — Zoologischen Sammlung des Bayerischen Staates, Munich

CLASSIFICATION

As pointed out elsewhere, the arrangement in the check list is
entirely alphabetical. This facilitates use of the list, but tells
nothing about the phylogenetic relationships of the genera in-
eluded. I have listed below the eight families of snakes that are
represented in the Ecuadorian fauna, with the genera placed in
their proper categories. I have not recognized the Boiginae, for
the polyphyletic nature of this supposed subfamily of the Colu-
bridae is obvious. I have followed Dunn and Dowling (1957 :260)
in assigning Nothopsis to the Colubrinae rather than to a sub-
family of its own.

Typhlopidae
Typhlopinae
Typhlops
Anomalepinae
Anomalepis, Liotyphlops
Leptotyphlopidae
Leptotyphlops
Boidae
Boinae
Boa, Corallus, Epicrates, Eunectes
Tropidophinae
Trachyboa, Tropidophis
Anilidae
Anilinae
Anilius

494 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

Colubridae
Colubrinae
Atractus, Chironius, Clelia, Coniophanes, Dendrophidion,
Diaphorolepis, Drepanoides, Dryadophis, Drymarchon,
Drymobius, Drymoluber, Erythrolamprus, Helicops, Lnantodes,
Lampropeltis, Leimadophis, Leptophis, Liophis, Lygophis,
Ninia, Nothopsis, Oxybelis, Philodryas, Pliocercus, Pseudoboa,
Pseustes, Rhadinaea, Rhinobothryum, Siphlophis, Spilotes,
Stenorhina, Synophis, Tantilla, Thamnodynastes,
Tretanorhinus, Tripanurgos, Xenodon
Dipsadinae
Dipsas, Sibon
Xenoderminae
Xenopholis
Hydrophiidae
Hydrophiinae
Pelamis
Viperidae
Crotalinae
Bothrops, Lachesis

ZOOGHEOGRAPHY

It has been written many times about Ecuador that it, like
ancient Gaul, ‘‘in tres partes divisa est.’’ These three parts are
the Pacific lowlands, the Andes, and the Amazonian slope. Ac-
tually, this is a broad generalization that tends to obscure more
than it clarifies. It is true that each of the areas possesses a
reptilian and amphibian faunule distinct from the others, but it is
equally true that there are faunal units within each area almost
as distinct one from the other. Thus, on the Pacifie coast, there
is a distinct and major faunal break which corresponds in gen-
eral with the northern limit of the influence on the coast of the
Humboldt Current. North of this line is a fauna in which Central
American species and genera are strongly predominant, and, in
fact, the Chocéan rain forest is to all intents and purposes an
extension of the Caribbean rain forest of Panama and Costa Rica.
South of the limit of current influence there is a somewhat im-
poverished fauna, which has definite affinity with coastal Peru
and perhaps northern Chile.

There are other such distinct units on the Pacifie slope of the
Andes. There is a distinct, ‘‘cliff-hanging’’ fauna, autochthonous
between about 1000 and 3000 meters, quite distinct in species
and even in a few genera from the lowlands of the Pacific. In
addition, there appears to be an as yet inadequately explored

PETERS : SNAKES OF ECUADOR 495

cloud forest on the west slope, investigation of which will almost
certainly explain some of the difficulties currently encountered
in working with the Pacific slope species.

In the interandean valleys the situation is similar. Although
quite depauperate in snake species, what few there are in the
region show that the fauna of the area is not uniform, but is
zonally divided from north to south. There are many more
species of snakes in the Loja hoya, for example, than in the
Quito valley. Many of the taxa in the Loja region are invaders
from the lowlands into an area that is, in part, at least, well
watered and timbered, a situation not true in the north. Each
valley is likely to show distinet changes in its reptilian and am-
phibian fauna from its neighbors, because of the alternation of
drainage directions from east to west, and the correspondingly
divergent faunas available for invasion. The only species that
range from valley to valley are the hardy spirits such as Atelopus
or some Eleutherodactylus, which I have found on the grassy
plains well above treeline. These forms are able to transgress
the horizontal steps of the ladder formed by the north-south
parallels of the Andean chain in this country. The fauna of the
northern part of the interandean area of Ecuador shows strong
and distinct relationships with the Colombian highlanders, while
the Peruvian fauna is predominant in the south, with genera
such as Telmatobius and Stenocercus reaching their northern
periphery there.

On the Amazonian slope, there are two and probably three
different groups of snake species, differentiated not on their
exclusive relationships in Ecuador, but on their more general
distribution. All are found in Amazonian Eeuador, a compara-
tively miniscule region, and are, for the most part, sympatric.
But one group is limited almost entirely to the Amazonian slope
of the Andes in Colombia, Ecuador, and Peru, as well as immedi-
ately adjacent Brazil, while the second is composed of extremely
wide-ranging forms which cover all of the Amazonian Basin to
the Atlantic, reach in many eases to the Caribbean, and are found
in some instances to northern Argentina and coastal southeastern
Brazil. The third element, as yet poorly defined, appears to be
similar to the first in its restriction to upper Amazonian areas,
but is a more southerly group, ranging from Bolivia into Peru,
and only reaching Ecuador on its southern borders. It is pos-
sible, in addition, that a cloud forest fauna may be found on the
eastern slope, as well as on the western. It is an interesting
anomaly of herpetological research in Ecuador that the entryway

496 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

to the Amazonian slope is through the Andes rather than over
them, as it is on the Pacific slope. The valley of the Rio Pastaza
drops directly down from Ambato after the most modest of
climbs, and plunges one directly into the upper reaches of the
tropics, at Banos. As a consequence, one never sees the upper
slopes of the eastern side under normal circumstances — certain-
ly no one has ever worked them thoroughly.

This brief summary hardly does justice to the zoogeographie
patterns in Ecuador, but the picture is much more enlightening
when all reptiles and amphibians are included. I hope to prepare
a more graphic and inclusive analysis when my knowledge of all
groups is more complete.

OMISSIONS

It will be noted by anyone familiar with South American
herpetology that some species long considered to include Ecuador
in their range are omitted from this list. A case in point is
Crotalus durissus terrificus Laurenti. The rattlesnake has long
been considered Ecuadorian (Oreés V., 1942: 147), but I have
been unable to locate any documented records for it. Klauber,
in his recent review of the rattlesnakes (1956: 123) felt the same
way, for he wrote, ‘‘General statements have been made to the
effect that C. d. terrificus occurs in Eastern Ecuador, but no
authentic records are available. Little is known concerning the
presence of rattlesnakes in the western part of the Amazon
Basin.”’ I eliminated Hypsiglena from the fauna of Ecuador and
of South America as a whole (Peters, 1956), and with it the un-
assignable name Pscudodipsas fallax K. Peters. Tachymenis
peruviana Wiegmann has been recorded from the Pacific low-
lands at Guayaquil (Amaral, 1925: 14), but Walker (1945: 14)
examined the specimen, and assigned it to Dryophylax natterert.
I have seen the specimen, and agree with Walker, including it
here in its current generic position under Thamnodynastes. Dry-
adophis boddaerti heathi Cope was indicated as Ecuadorian by
Stuart (1941: 78), but no records are known, and the’ reason
seems to be solely the availability in Ecuador of suitable habitat.
Drymarchon corais corais Boie is certainly to be expected in
Amazonian Ecuador, and previously has been included in that
fauna, but I cannot document that inclusion with specimens.

ACKNOWLEDGEMENTS

It is my pleasant duty to record my indebtedness to several
friends and institutions whose aid in the preparation of this list

PETERS: SNAKES OF ECUADOR 497

has been invaluable. Several of them have protected me from
gross stupidities, for which I am grateful, while retaining full
responsibility for any which I have perpetuated. Others have
made available specimens for purposes of checking identifications
or verifying the keys.

Dr. Gustavo Orcés-Villagomez, of the Escuela Polytecnica
Nacional in Quito, has been a constant help. He has checked the
keys against the Escuela’s snake collection, and found that they
worked successfully on 96 per cent of the 1500 specimens. He
has loaned me specimens that did not fit, with a consequent re-
adjustment of the keys and the addition of several species to the
fauna. He has commented at length on my check list, and has
supplemented my data in many ways. I am unable to exaggerate
my indebtedness to him. Robert Copping, also of Quito, has
helped me in my work, both during my visit in Quito and after-
wards. Dr. Roberto Levi-Castillo, of Guayaquil, has taken me on
field trips on the Pacific Coast, demonstrating his complete fa-
miliarity with his country, and has shown me areas I could not
have known without him. He has sent me snake collections from
the Pacifie slope as well. The late E. R. Dunn reviewed the check
list and gave me the benefit of his criticisms, as has Benjamin
Shreve. Jay Savage and William Duellman gave me keys to the
genera Atractus and Leptodeira in Keuador. Joseph Bailey has
provided information on the genus Oxyrhopus and its relatives.

J.C. Battersby, Charles M. Bogert, Doris M. Cochran, Norman
Hartweg, Robert Inger, Alan Leviton, Arthur Loveridge, George
S. Myers, Neil D. Richmond, the late Joseph R. Slevin, and
Ernest EH. Williams have loaned me specimens from or furnished
me with information concerning the Ecuadorian collections in
their respective institutions.

My field work in Ecuador in 1954 was supported by a grant
from the Penrose Fund of the American Philosophical Society,
and a grant from the Bache Fund of the National Academy of
Sciences made possible a summer’s work in the Stanford Univer-
sity Natural History Museum in 1956. I am grateful to the
authorities in charge of each for providing me with these oppor-
tunities.

BIBLIOGRAPHY

AMARAL, AFRANIO DO
1925. South American Snakes in the Collection of the United States
National Museum. Proc. U. S. Nat. Mus., vol. 67, art. 24, pp.
1-30.

498 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

DuNN, EMMETT REID and HERNDON DOWLING
1957. The Neotropical Snake Genus Nothopsis Cope. Copeia, 1957,
no. 4, pp. 255-261.

KLAUBER, LAURENCE M.
1956. Rattlesnakes, their Habits, Life Histories, and Influence on
Mankind. U. Calif. Press, vol. 1.

OrckS- VILLAGOMEZ, GUSTAVO
1942. Los ofidios venenosos del Eeuador. Flora, vol. 2, nos. 5-6, pp.
147-155.

PETERS, JAMES A.
1956. ‘The Occurrence of the Snake Genus Hypsiglena in Ecuador.
Copeia, 1956, no. 1, pp. 57-58.

SruaRT, LAURENCE C.
1941. Studies on Neotropical Colubrinae. VIII. A Revision of the
senus Dryadophis Stuart, 1939. Mise. Publ. Mus. Zool. Univ.
Mich., no. 49, pp. 1-106.

WALKER, WARREN F.
1945. <A Study of the Snake Tachymenis peruviana Wiegmann and its
Allies. Bull. Mus. Comp. Zool. Harvard, vol. 96, no. 1, pp. 1-55.

Key to the Genera of Snakes Known or Expected in Ecuador
“indicates a genus not yet known from Ecuador, but expected

1. Scales of uniform size around body, or ventrals only feebly enlarged. .2

Ventral seales distinctly larger than dorsals ........... ee Peso)
2. Tail flattened laterally... .Pelamis

shavlemoreror lessuroundede sy, eyes scares cine ae ee eee oes 3
3.) Vientrals teebly enlarged teeth in botht jaws eae eee ere 4

Ventrals not enlarged; teeth in upper or lower jaw, not both....... 5
4. Nostrils dorsal; body unicolor....Huwnectes

Nostrils lateral; ringed pattern... . Anilius

Nasal bordering lip; 14 scale rows on body... .Leptotyphlops

Nasal not bordering lip; more than 14 scale rows................-- 6
Gu hosial ands trontals nou imecoltache ieee ere ene ieee eee ne 7

Rostraleim! proadscontacheswalblyetst 0 Nite) | eee eee eee 8
7. Frontal pentagonal....Anomalepis

Frontal semicircular... .Helminthophis*

8. Small, divided nasal; prefrontals present... .Liotyphlops
Very large nasal (may be divided or not) ; no prefrontals.... Typhlops

10.

TEV

ire

14.

16.

Wie

Oe

20:

PETERS: SNAKES OF ECUADOR 499

Pit presente between eye an demos trillent teenie serene asta 10

INO jorae SO, Grey DING! NOTIN, - eee ow scaccaaneeoccrdootaceonnne 12

Tail with rattle on tip... .Crotalus*

Hua jyahEle loa e THO, AMO) WANE. os ocacauoegboedodpopuecovandoounge inl

Posterior subcaudals replaced by small scales; upper head scales granu-
lar, smooth or obtusely keeled... .Lachesis

No small scales under tail; upper head scales imbricate or subimbricate
.... Bothrops.

Vestiges of pelvis and hind limbs present (not always visible ex-
ternally) ; maximum body scale rows 30 or more................ 13

No vestiges of pelvis or hind limbs, externally or internally; maximum
bodys scalennowse2 Se OrMmlesste ccs: eases ile ore oma nateren oe ners ten: 16

Rostral reduced or absent, several small scales on vertical aspect of
tip of snout... . Trachyboa
Rostral large, extends from lip to dorsal surface of head........... 14

Dorsum and sides of head between eyes and tip of snout with many

small, sbaresilleidhy Guanaeel SORES, 5.2.5 00acecs cane san ysaandonoce 15
Dorsum and sides of head between eyes and tip of snout with large,
regularly placed seales, often in pairs... .Hpicrates

Dark stripe from eye across temporal region; usually another dark
stripe from tip of snout between eyes to mid-dorsum; no deep pits in
labials....Boa

Temporal region unicolor, lacking dark stripe; no stripe from snout
passing between eyes; if either stripe is present, deep pits in labials

_..Corallus

Top of head with small scales. ...Nothopsis

MO Cie IncaGl ywarda lene, mesuiewe jolene. oo.0cosec acces oanscc coco We/
Tip of snout raised into a sharp point... .Phimophis*

Tip of snout rounded, not with hard, sharp point................ 18

Number of scale rows one head length anterior to anus equal to or only

Gin WSN Wa IniUATN|oYere ANE AMC OOCN,. .- 5 ca eeacaccouesnaceseane 5 lf)
Number of scale rows one head length anterior to anus at least two less

than. numiberttat: smidbodiysse 54> ac oe oon. wisn me eee ree eet as 43
Scalesian) 19. orsmoremrows: see oo eee: eee ante SO Bi cei Pee 20
Salles ma less Wham IS) mos) Ain WMER OMEN, joo cco aoe so aaanccconope 22
Prefrontals united into a single scale... .Synophis
Pretrontalsed oul ewes sa eee see hea a kPa ee one ep cco ee a 21
Ventrals less than 200... .Ninta

Ventrals more than 200... .Clelia

ol.

Biase

37.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

(Acialcuping let. crs a. ce ye meaner tet ence Rene aa a 23

Amal divided) se an eee Hao eelatere crete nasties ce acres 35

Seale Powell yes oct eae SiC oa aks Sey ere ee, eon 24

Geale rows hovOreless) ferme ace se cee Noten: elena aoe ete Pi

Prefrontals united into single scale, may show partial suture....
Xenopholis

JORIS) ONE WHO 2. ccesncnponsesevaad mUGaE MODS OG Godras sob 25

Subcaudals single... .Pseudoboa

Subeaudals! divided \..-e ests ceo bes site chen oe fea ieetewe (th eNce keen 26

Ventrals less than 200... . Atractus

Ventrals more than 200... .Clelia

Vertebral scale row equal in size to other dorsals............--..-- 28

Vertebral scale row larger than other dorsals........-..--.------> 33

Internasals fused into single scale.............---------:----:-->: 29

(aa) quierneicell Reales) 4. onccenacccesesorecanccsaudggo0onoesan0008 30

Two prefrontal scales... Hydrops*

One prefrontal scale... . Apostolepis*

Scale pits present... .Drymoluber

Seale: pitsmabsent: aire. teeny. cc ac eaten sete ee age re ae 31

Poison fangs in upper jaw... . Micrurus

INO poison fangs me Upper JAW ce. a) ee oe eo nl 32

No loreal; preocular present....Drepanoides (which includes Pseudo-
clelia)

Loreal present; or, if absent, preocular absent. . . _Atractus

Labial beneath primary temporal greatly enlarged, and in contact with

postocular, primary and secondary temporal. . . Sibon
Labial beneath primary temporals not enlarged, no single labial in
contact with postocular, primary and secondary fFemporalss sss ee 34
Mental groove present... . Imantodes
Mental groove absent... .Dipsas
Seale er OWs ely en ee eA ee GSE aha ee ene mc 36
SVorik endo); (cael Oe ee Ak OA A SOROS asian Oboe Cant oases oo.c 39
Internasals fused with anterior section of nasal... .Stenorrhina
lbmraRNeCAN) Ghisnamver tena WACENI, oo o6enaacaanccccoodkoss eed as bo sos 37

Pattern of stripes, which may be so vague as to appear unicolor....
Rhadinaea

Pattern not of stripes, but mever unicolor..............-.-------- 38

38.

41.

42.

43.

44.

46,

52.

OO.

o4.

PETERS: SNAKES OF ECUADOR 501

Head sharply distinet from neck... . /mantodes

Head and neck about same size... . Pliocercus

Patterns otmedsslackewanGsyelll Osi re Sewer oe A()
Pattennenoteannulla tem eerie ee ey caeem Oy eevee wae a ett 41
Loreal present... .Hrythrolamprus

Loreal absent... . 1 Micrurus

Mental widely separating first lower labials and in contact wtih anterior

pair of chin shields... . Leptomicrurus
Mental not in contact with anterior pair of chin shields; first lower
lnlovralle: ti Common ra tamil. 5.4 obs sce edn sar eesoesseaaoseoeuoe 42
Internasals fused; nasals in contact... .Pseudoeryx*
Two internasals; nasal scales not in contact... . Tantilla
Seales in even number of rows ....... etree ae or sae aca:
Seales in odd number of rows. ........ . EE ad Sea
Seales at midbody in 10 or 12 rows |... Chironius
Seales at midbody in 14, 16 or 18 rows... . Spilotes
Some or all scale rows keeled................... af st We plethora 46
All scales smooth ... Fo apr sas f 4 Sep ee Ay ).. $6'L
Anal divided ....... i areas eS tie MLO oy ba Dt tem on td eee de _ AT
Ae SIM SLO wey cy sce koe. Bes Se were. Y oe ais LR, LO
MASAMI Ss Callen Owis el 9) OTOL: Cheer eee eye ener PERRI 3)
Maximum scale rows 17 or less... . a: tesyeck: ete es
NCaleqpiisrpEesenitwae ay Howee Lote oe ee eee pale) cy ancl geen See Se ee
Scalenpiisralsenteren ce ae aa ces Ae Tne ELL
-attern of red and black rings... .khinobothryum
Patrerne noteotetedeande blacks nillegne seen: St eS bone ROP Roe 50
Pupil round; dorsal scales oblique; green dorsum... Philodryas
Pupil elliptic; dorsal scales not oblique; grayish in color... .Thamnody-
nastes
Usually three prefrontals; maxillary teeth subequal... .Tretanorhinus
Two prefrontals; posterior maxillary teeth gradually increasing in size
....Helicops
Seale rows 15... . Leptophis
SCaIEr LOWS STs blake sonatas PB tle auto nso fe, ae Re ene ee teen 53
No loreal, prefrontal contacts labials; head elongate... Oxybelis

Loreal present, prefrontal does not contact labials; head not elongate 59

Pretnrontalsmumibe dst Oe esine)) Cts call Cee serene nn reer 55
Prefrontals paired, not fused into a single seale................... 56

502

60.

61.

66.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

A single keel on individual scales in the vertebral row... .Synophis
Double keels on individual seales in the vertebral row... . Diaphorolepis
SOHNE thn Ty HOWE) Ghee bohNpN 5 pour de oommeraudnceod an edness oe: 57
Scales! li oremlone sO wise ait eta) Xa0 Oc eee ee ely
Upper labials 6; lower labials 7... . Oxybelis

Upper labials may be 7, usually 8 or more; lower Jabials 8 or more....

Leptophis

Seale rows medially 17 or less... IS. 5. Ree ON. Bee eae 59
Scalesrowsmmediallly-smore shame Ose eee eee ree 60

Subeaudals more than 110; maxillary teeth 32 or more, subequal
Dendrophidion
Subeaudals less than 110; maxillary teeth 34 or less, posteriormost

greatly increased in size |. Drymobius
Less than 50 single subeaudals.... Tropidophis
More than 100 double subeaudals... Pseustes
Anal single be Ewe ARNG AONE eee 62
Anal divided EARN LE Aha ERO 8 le eae Ae Aha Pa JT
No loreal; prefrontal in contact with labials ..... Te Lae + 63
Loreal present; or, if absent, prefrontals do not contact labials.... 64
Two pairs of prefrontals, more than 21 seale rows at maximum
Tropidophis
One pair of prefrontals, less than 21 scale rows at maximum...
Oxybelis
Maximum seale rows 15 or less... . Dipsas
Scalemeoweedl (MOTRIN OLE: cr oar. ke cases Meee hare ee 65
Maximum’ -Scalerst Ows, (7: cps ae ee ee eee 66
Scale-rows U9Lor more... ee ee a ee ee 70
Body striped with light and dark brown, at least posteriorly
Thamnodynastes
Bodysnot striped: ciy4....38 fe seg eee ee ne ee 67
Body unicolor, no dark eross bands... .Drymarchon
Body light with darker cross bands or spots....................... 68

Head much larger than the very slim, vine-like body; eye very large. .
Imantodes
Head not or very slightly larger than body; eye not greatly enlarged 69

Ventrals less than 200... . Atractus
Ventrals more than 200... Clelia
Anterior temporal single, =o: occu eo. ee eee ar 71

Anterior temporals two or more

74,

84.

PETERS: SNAKES OF ECUADOR 503

Seales in diagonal rows... .Xenodon
Seales in regular, horizontal rows... .Oxyrhopus
Dorsal pattern of complete yellow and red rings, separated by black

rings... .Lampropeltis
Dorsal pattern not of complete yellow and red rings............... 73
Labials excluded from eye by subocular row of seales....Dugandia*
Atleast one slabialventersvorbiten 15 cee ee ne eee 74

Dorsum with dark bands, which are very narrow; interspaces 4-5 times
as wide as blotches... .Tripanurgos

Dorsum without bands, or, if bands are present, interspaces are not
4D atIMese AS Wwid Geass sane v2.5: 20 epee ie rey ee een ere oe eer 75

Pattern of spots or numerous cross bands with very irregular zigzag
borders; head shields variegated with light and dark; third to fifth

anterior mandibular teeth very much enlarged... Siphlophis*
Pattern not as above; anterior mandibular teeth may be enlarged with
gradual decrease in tooth length toward posterior............... 76
Anterior mandibular teeth much the largest... . Oxyrhopus
All mandibular teeth approximately the same size... .Clelia
No loreal, prefrontal in contact with labials.... Oxzybelis
A loreal, or if absent, prefrontal not as deseribed.................78
DWOROLPMoreranverior scl Olas see een re tae) tae 4 een rend)
One; anterior temporal) se ae oe Bea ee a Be eS ee ees 82
Posterior maxillary teeth grooved... .Thamnodynastes
iPostericr maxillary ateeth) wathout eLoovess +s. 4242-65 4 -) 2a 80

Posterior maxillary teeth enlarged, separated from other maxillary teeth

by diastema....Rhadinaea
Posterior maxillary teeth not enlarged, no diastema............... 81
19 scale rows... .Drymobius reissii Peters 1868 1
17 or less scale rows... .Dryadoplis
Ma xaTmumMes Cale ar Owise lis Ole lOSS meee ee eee ee Botte Cae Oe 83
Maxamumescalestows Os orsmone see see ae Pe eee OO
Maximum scale rows 15... . Leptophis
Maximum scale rows 17......... SE EERE ote. OS SS Raa pe uate 84

Head short, much broader than slender neck; eye very large with
vertical pupil... .Imantodes
Head not or but slightly larger than neck; eye normal............. 85

1 Described in Monatsb. Akad. Wiss. Berlin, 1868, p. 640 (Type Berlin Mus.

4507), from Guayaquil. I do not know to what genus it belongs, but it apparently
does not fit in Drymobius as defined by Stuart (Occ. Pap. Mus. Zool. Univ. Mich.,
no. 236, 1952, p. 6). The type must be re-examined to assign the species to its
proper genus.

88.

89.

90.

Sle

92.

93.

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Seale pits present... . Leimadophis

Scalexpits: absent ™.qaccbakies- os Soc is eas oh ae eae wee eee .. 86
No stripes on body... . Liophis

Stripes present on body, often on posterior part only............... 87

Chin black with white spotting; dorsal dark brown color extends well
onto tips of ventrals; rest of venter clear yellow with no spotting...
Rhadinaea

Chin white with black spotting; dorsal ground color does not extend
onto ventrals; all of venter spotted and blotched with black....
Lygophis

Scailespits present}. 22.5... 2 S555 son + Shae ae OE ee en Gea 89
Scalleritsrabsemt 6.0 a2 5.eoe eae soe se Ete coe ne ea 92
Sealenpitisdoublésaa vs: dae eens fat ote Ee cere 90

Seale pits single

Vertical pupil; rounded ventrals; no clear line marking end of dorsal
color on ventrals.... Leptodeira

Round pupil; angulate ventrals; a clear light line on ventrals at edge
of dorsal color... .Philodryas

Body banded in young; bands may disappear in adults, which are then

unicolor; seales disposed obliquely... .Xenodon
30dy striped at least posteriorly or some trace of spots or bars, at least
anteriorly; scales in longitudinal rows on body... . Leimadophis
Dorsum with bands... . Liophis
Dorsmmiystrupedmthrou eh out eno thee eee ae 93

First few dorsal scale rows very light, same color as venter, dark stripe
above them; posteriormost maxillary teeth without grooves
Lygophis

First few dorsal seale rows dark, darker than and contrasting with
venter, light stripe above them; posteriormost maxillary teeth grooved

Coniophanes

ANILIUS

Anilius scytale Linnaeus

Anguis scytale Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 228—
““Tndiis’’? (RMS 38 cotypes).

Range: Guianas; Northern Brazil; Colombia; Ecuador;
Peru. Amazonian drainage.

ANOMALEPIS

Anomalepis flavapices Peters

Anomalepis flavapices Peters, 1957, Amer. Mus. Novitates, no. 1851,
p. 3.—Esmeraldas, Esmeraldas Province, Ecuador (JAP 2613).
Range: Lowlands of Northwestern Ecuador.

J

Ile

PETERS: SNAKES OF ECUADOR 505

ATRACTUS

Loreal shield reduced to a minute seale or absent; prefrontals meeting
supralabials... carriont

Loreal shield well developed; prefrontals not meeting supralabials 2
Dorsal seales in 15 rows........ : 4:5, aN eA ae 3
Dorsal scales in 17 rows ... oh Beet) ri ee eR ER Crees, 3 Maes 5

Loreal relatively short, about equal in length to length of postnasal;
prefrontals broader than long, equal to or smaller than rostral; color
pattern including red and black rings... .elaps

Loreal long, between two and three times as long as postnasal; pre-
frontals longer than broad, much larger than rostral; no rings in

COXON (Cpe OFT FLEET 1 AN legen ERE reine eD re NennAiemt tence best ils de ue Ate SN ert nate 4
Maxillary teeth 10-11; supralabials 6 ...roulei
Maxillary teeth 7-8; supralabials 8 (sometimes 7) occipitoalbus
Dorsal color pattern of bands, blotches or stripes..................- 6
Dorsal coloration uniformly dark brown or gray................... 12
Dorsal color pattern of longitudinal stripes . Ee ae 7
Dorsal pattern of bands or blotches........... sae. poe Lk
Ventrals plus caudals less than 210 (range 164-196) ........ ate:
Ventrals plus caudals 210 or more (range 222-240)... gaigeae

No stripes along edges of ventrals; no dorsolateral blotches, although
dorsolateral stripes present .. Pe eed Se ciete: seis ea ae 9
Stripes along edges of ventrals present, as are douolateral blotches. .10

No vertebral stripe; maxillary teeth 8; hemipenes in males extending
to level of 12th caudal... . ecuadorensis

A vertebral stripe; maxillary teeth 6; hemipenes in males extending to
level of 18th caudal... . occidentalis

A vertebral stripe present; ventrals plus caudals 164... .dunni
No vertebral stripe; ventrals plus caudals 194-196... . collaris

Ventrals in males 148-172, in females 157-181, venter yellow spotted
with dark... .majer
Ventrals in males 168-173, in females 177-184, venter unspotted yellow
multicinctus

Venter pale brown with faint line along margins of ventrals
microrhynchus
Venter unicolor or with irregular dark brown markings medially ... 13

Ventrals in males more than 150, in females more than 160 are le
Ventrals in males 142-144, in females 148-153... . .lehmanni

506 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY

14. Loreal moderate, about 1% times as long as postnasal....modestus
Loreal long, well over twice as long as postnasal..... ae ae tor ae eee 155

15. Ventrals plus caudals 184-203; supralabials usually 8... .resplendens
Ventrals plus caudals 215-223; supralabials usually 7... . paucidens

Atractus carrioni Parker
Atractus carrioni Parker, 1930, Ann. Mag. Nat. Hist., ser. 10, vol. 5,
p. 208.—Loja, Ecuador, 2200 m. (BM 1929.10.30.1, female).
Range: Intermontane valley of Loja, Ecuador.

Atractus collaris Peracea
Atractus collaris Peracea, 1897, Bol. Mus. Zool. Univ. Torino, vol. 12,
p. 4.—Rio Cononaco, Napo-Pastaza Province, Ecuador (TurM).
Range: Amazonian Ecuador and Peru.

Atractus dunn Savage
Rhabdosoma maculatum Bocourt, 1883, Miss. Sei. Mex., pt. 3, p. 539,
pl. 34.—‘‘ Equateur.’’ (PM 5986, female).
Atractus dunni Savage, 1955, Proe. Biol. Soe. Washington, vol. 68,
p. 14.—Eeuador (substitute name for Rhabdosoma maculatum Bocourt,
preoceupied).
Range: ‘‘Ecuador’’.

Atractus ecuadorensis Savage

Atractus ecuadorensis Savage, 1955, Proe. Biol. Soe. Washington, vol.
68, p. 15—‘‘Llangate Area,’’ probably Llanganate Range, Tunguruhua
Province, Ecuador (CNHM 23529, male).

Range : Known only from the type locality.

Atractus elaps Ginther
Rhabdosoma elaps Giinther, 1858, Cat. Coll. Snakes Brit. Mus., p. 241.—
Guayaquil, Ecuador, probably in error (BM).
Range: The oriente of Ecuador, northern Peru, Bolivia,
eastern Colombia, Venezuela, and western Brazil.

Atractus gaigeae Savage
Rhabdosoma maculatum Bocourt, 1883, (part), Miss. Sei. Mex., pt. 3,
195 DEAD, jell, aiay, ates, dL
Atractus gaigeae Savage, 1955, Proc. Biol. Soc. Washington, vol. 68,
p. 12.—Santiago-Zamora Province, Ecuador (UMMZ 82887, male).
Range: Amazonian lowlands of Ecuador.

Atractus lehmanni Boettger
Atractus lehmanni Boettger, 1898, Katal. Rept. Mus. Senckenberg, vol.
2, p. 80.—Cuenca, Azuay Province, Ecuador (Senck 8310a, 6 cotypes;
MCZ 33513, cotype).
Range : Known only from the type locality.

PETERS: SNAKES OF ECUADOR 507

Atractus major Boulenger
Atractus major Boulenger, 1894, Cat. Snakes Brit. Mus., vol. 2, p.
307.—Intae (BM 1946. 9.7.56) ; Pallatanga (BM 1946. 9.7.60) ; Canelos
(BM 1946. 9.7.27, designated lectotype by Savage, Mise. Publ. UMMZ,
No. 112, p. 50, 1960); and ‘‘W. Eeuador’’? (BM 1946. 9.7.57-59).
Range: Ecuador and Colombia on Amazonian slopes; Vene-
zuela.

Atractus microrhynchus Cope
Atractus microrhynchus Cope, 1868, Proc. Acad. Nat. Sei. Philadelphia,
p. 102.—Guayaquil, Ecuador (USNM 6693).
Range: Known only from type specimen.

Atractus modestus Boulenger
Atractus modestus Boulenger, 1894, Cat. Snakes Brit. Mus., vol. 2,
p. 304, pl. 15, fig. 1—‘‘W. Eeuador’’ (BM 1946. 1.6.30, male).
Range: Western Ecuador.

Atractus multicinctus Jan
Rhabdosoma badium multicinctum Jan, 1865, in Jan and Sordelli, Icon.
Ophidiens, vol. 10, pl. 4, fig. 5——Lima, Peru, in error (Type unknown).
Range: Northwestern Eeuador into the Choeé of Colombia.

Atractus occidentalis Savage
Atractus occidentalis Savage, 1955, Proe. Biol. Soc. Washington, vol.
68, p. 16.—Mindo, Pichincha Province, Eeuador (BM 1916. 5.23.5).
Range: Higher Pacifie slopes of the Andes in northwestern
Keuador.

Atractus occipitoalbus Jan
Rhabdosoma occipitoaibum Jan, 1862, Arch. per la Zool. Anat. Fisiol.,
vol. 2 (1), p. 16—Andes, Eeuador, 4000 feet (MonM).
Rhabdosoma duboisi Boulenger, 1880, Bull. Soe. Zool. France, p. 44.—
‘¢ Andes of Heuador’’? (IRB, 2 cotypes).
Atractus orcesi Savage, 1955, Proc. Biol. Soc. Washington, vol. 68,
p. 17.—Loreto, Napo-Pastaza Province, Ecuador (SU 15622).
Range: Eastern slopes of the Andes in Eeuador.

Atractus paucidens Despax
Atractus (Atractopsis) paucidens Despax, 1910, Bull. Mus. Hist. Nat.
Paris, 1910, p. 372.—Santo Domingo de los Colorados, Ecuador (PM
06-245).
Range: Slopes of Andes in northwestern Keuador.

Atractus resplendens Werner
Atractus torquatus resplendens Werner, 1901, Ver. Zool.-Bot. Ges. Wien,

508

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

vol. 51, p. 598.—Ecuador (Location of type unknown).
Range: Amazonian slopes of Ecuador.

Atractus rouler Despax

Atractus roulei Despax, 1910, Bull. Mus. Hist. Nat. Paris, p. 370.—
Alausi, Eeuador, 2350 m. (PM 06-243).
Range: Southwestern Ecuador.

BOA

Seales in 81-95 rows; red areas on tail; body with 15-20 dark eross
bars; black median line on head without lateral processes between
eyes... .constrictor constrictor

Seales in 61-79 rows; no red areas on tail; body with 22-30 dark cross
bars; black median line on head with lateral processes between eyes
_...constrictor imperator

Boa constrictor constrictor Linnaeus

Boa constrictor Linnaeus, 1758, Syst. Nat., 10th Ed., vol. 1, p. 215.—
‘<Tndiis’’ (RMS, 2 cotypes).
Range: Amazonian South America to Argentina and Para-
oeuay ; Trinidad ; Tobago.

Boa constrictor imperator Daudin

2

we

1

Boa imperator Daudin, 1803, Hist. Nat. des Rept., vol. 5, p. 150.—
““Mexico,’’? restricted to Cérdoba, Veracruz, Mexico, by Smith and
Taylor, U. Kans. Sei. Bull., vol. 33 (8), p. 347, 1950; E. R. Dunn
indicated in correspondence that he believed the type came from the
Colombian Chocé (Type no longer in PM, probably destroyed).
Range: Mexico south to northwestern South America west of
the Andes to Ecuador and Peru.

BOTHROPS

INon=prehensile stanly sce eeeeriee 2
iPrebensil em tanll eee eee Re CAPO et ee Coe as cit he Wh eet)

Subcaudals all or greater part in two rows
Subeaudals all or greater part single

Seales on the vertex and occiput more or less strongly keeled; dorsal

scales strongly keeled).:..ta9¢) seo) oie ee ee ee tes
Upper head scales smooth; dorsal scales not strongly keeled xantho-
gramma
Second upper labial forming anterior border of loreal pit 5
Loreal pit separated from labials m. micropthalma
Scales in 25-29 rows... a. atrox

Scalesuiny2il-23erows! ae eee : acheter 6

~l

10.

Wile

PETERS: SNAKES OF ECUADOR 509

Keels on dorsal scales much shorter than seale itself; ventrals 156-174;
subeaudals 47-64... pulchra

Keels on dorsal scales reach extremity of scale; ventrals 144-155; sub-
caudals 38-44... lojana

Second upper labial forming anterior border of loreal pit... .castelnaudi
Lonel jos Seal wir wee) JENN 5 ooo se dose sco e noob ace 8
Double internasals; ventrals 125-131; subcaudals 44-50... .hyoprora
Single internasal; ventrals 1380-145; subcaudals 24-35. |. nasuta
Subcaudalsvallmonenecarer parcel twiOnhOwSrieir nei ere nani nrenn ree 10
Subcaudalsrallwonseneaters panbasin 0) Copii ase eis aera ee een aee 12
IDroneseul: KeAMIa: TRON) Tee, IOS) GROWS see cb ose oles de oeceebaachaueuodae: U1
Dorsalescale sows Nec O-ooOmLOWSi ees eet eee ee eee 13
Scales in 19 rows... .alticola

Seales in 21 rows... .albocarinata

Keels on dorsal, dorsal head, and dorsal caudal scales all whitish,
giving a finely striated appearance albocarinata
Keels on dorsal scales same color as scale itself... .schlegeli

Green, uniform or speckled with black; a yellow lateral streak or series
of spots on first row of dorsal scales... . bilineata

Pallid greenish-gray, with rounded, yellow-bordered black spots
pwnetata

Bothrops albocarinata Shreve

Bothrops albocarinata Shreve, 1934, Occ. Pap. Boston Soc. Nat. Hist.,
vol. 8, p. 1380.—Pastaza River between Canelos and Marafion River,
Ecuador (MCZ 36989).

Range: Known only from the Pastaza River drainage in
Eeuador.

Bothrops alticola Parker

Bothrops alticola Parker, 1934, Ann. Mag. Nat. Hist., ser. 10, vol. 14,
p. 272.—5 km. east of Loja, Ecuador, 9200 feet (BM 1933. 6.24.115,
male).

Range: Known only from the vicinity of the type locality.

Bothrops atrox atrox Linnaeus

Coluber atroz Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 222.—
““Asia,’’ Restricted to Surinam by Schmidt and Walker, Zool. Ser.
Field Mus. Nat. Hist., vol. 24, p. 295, 1943 (Type unknown).

Range: Amazonian Brazil and northern Bolivia north to
Mexico. Occurs on Pacific and Amazonian slopes in Eeuador.

510 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Bothrops bilineata Wied
Cophias bilineata Wied, 1825, Beitr. Naturgesch. Brasil, vol. 1, p. 483.—
No type locality given (Type unknown).
Range: Bolivia; Peru; Ecuador ; Brazil; in their Amazonian
parts.

Bothrops castelnaudi Duméril, Bibron and Duméril
Bothrops castelnaudi Duméril, Bibron and Duméril, 1854, Erp. Gen.,
vol. 7 (2), p. 1511.—South America (PM 1582).
Bothrops quadriscutatus Peters, 1861, Monatsb. Akad. Wiss. Berlin,
p. 358.—Quito, Ecuador (BerM).

Range: Northern Brazil; Colombia; Ecuador; and eastern
Peru.

Bothrops hyoprora Amaral
Bothrops hyoprova Amaral, 1935, Mem. Inst. Butantan, vol. 9, p. 222.—
La Pedrera, Colombia (IB 9199, male).
Range: Southern Colombia; eastern Ecuador; Peru; and the
state of Amazonas in Brazil.

Bothrops lojana Parker
Bothrops lojana Parker, 1930, Ann. Mag. Nat. Hist., ser. 10, vol. 5,
p. 568.—-Loja City, Ecuador, 2200 m. (BM 1930. 1.30.1).
Range: Vicinity of type locality.

Bothrops micropthalma micropthalma Cope
Bothrops micropthalmus Cope, 1876, Jour. Acad. Nat. Sci. Philadelphia,
vol. 8 (2), p. 182.—From between Balsas Puerto and Moyabamba,
Peru (ANSP 11515).
Lachesis pleuroxanthus Boulenger, 1912, Ann. Mag. Nat. Hist., ser. 8,
vol. 10, p. 423.—Alpayaea, Ecuador (BM 1946. 1.19.89).
Range: Amazonian Peru and EKeuador.

Bothrops nasuta Boeourt
Bothrops nasutus Bocourt, 1868, Ann. Sci. Nat., ser. 5, vol. 10, p. 202.—
Unknown.
Range: Pacific slope of Ecuador and Colombia; east coast of
Central America to Mexico.

Bothrops pulchra Peters

Trigonocephalus pulcher Peters, 1862, Monatsb. Akad. Wiss. Berlin,
p. 672.—Quito, Ecuador (BerM 3868, 3 syntypes).
Range: Peru and Keuador in the Amazonian lowlands.

Bothrops punctata Garcia

Lachesis punctata Garcia, 1896, Ofid. Venen. del Cauca, Cali, Colombia,

PETERS: SNAKES OF ECUADOR a0

p. 31, pl. 8—‘‘Las montafias del Dagua,’’ Colombia—according to
E. R. Dunn, this is on the Pacific Coast (Type non-existent).

Lachesis monticellii Peracca, 1910, Annu. Mus. Napoli, vol. 3 (12),
p. 2.—‘‘ America tropicale?’’ (UNZM).

Range: Panama; Colombia; Ecuador ; in the Choco region.

Bothrops schlegelu schlegelu Berthold
Trigonocephalus schlegelii Berthold, 1846, Abh. Ges. Wiss. Gottingen,
vol. 3, p. 13, tab. b, figs. 5-6——‘‘New Grenada, Provinz Popayan,’’
which Dunn and Stuart, Copeia, 1951, p. 56, say is same as Popayan,
Colombia (GottM).
Lachesis nitidus Giinther, 1859, Proc. Zool. Soc. London, p. 414.—
‘¢ Western Andes of Ecuador’’ (BM 60.6.16.83).
Range: Chocé Region of Panama, Colombia, and Ecuador.

Bothrops xanthogramma Cope
Trigonocephalus xanthogrammus Cope, 1868, Proc. Acad. Nat. Sei.
Philadelphia, p. 110—Pallatanga, Ecuador (ANSP 9978).
Range: Apparently to be found in the highlands of Ecuador
and Colombia.

CHIRONIUS
le ScalerowselO)onwanberionrspaxtlOl 500 iyi -paeer erie irr 2
Scalesrows lZaonganterl Or p amin Ole Od yen- eis cee ie ei) eee 3

2. All dorsal scale rows keeled to level of anus except paravertebrals and
lowermost row... . grandisquamis
Only vertebral rows (one on each side of body) with keels, other dorsal
rows smooth. . schliiteri

wo

Dark brown or blackish above, vertebral line pale brown, ventral and
subeaudal scales yellow with a fine black edge... .carinatus

Keeled scales with large spots at their base, which gives the appearance
of a yellow vertebral line; lateral scales with diagonal yellow streaks
_... flavopictus

Chironius carinatus Linnaeus
Coluber carinatus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 223 —
‘<Tndiis’’ (RMS).
Range: Central America; Tropical South America; Trini-
dad ; Guadelupe ; San Vicente.

Chironius flavopictus Werner
Herpetodryas carinatus flavopicta Werner, 1909, Mitt. Naturh. Mus.
Hamburg, vol. 26, p. 220.—Guayaquil, Ecuador (HM-V118, 2 cotypes).
Range: Known from the type locality and the headwaters of
the Rio Congo.

aie, BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Chironius grandisquamis Peters
Spilotes grandisquamis Peters, 1868, Monatsb. Akad. Wiss. Berlin, p.
451.—Costa Rica (BerM).
Range: Costa Rica; Panama (MVZ 35559) ; northwestern
Ecuador.

Chironius schliter: Werner
Herpetodryas schliiteri Werner, 1899, Zool. Anz., vol. 22, p. 115.—
Napo, Ecuador (VM).
Range: Amazonian slopes of Ecuador.

CLELIA

1. 19 scale rows... .clelia clelia
17 scale rows... .clelia scytalina

Clelia clelia clelia Daudin
Coluber clelia Daudin, 1803, Hist. Nat. Rept., vol. 6, p. 330, pl. 78.—
‘“Surinam’’ (Type no longer in PM, probably destroyed).
Range: Central America to Argentina and Uruguay.

Clelia clelia scytalina Cope
Scolecophis scytalina Cope, 1866, Proc. Acad. Nat. Sci. Philadelphia,
vol. 18, p. 320, (publ. 1867).—San Juan Bautista, Tabasco, Mexico
(USNM 6581).
Barbourina equatoriana Amaral, 1924, Jour. Washington Acad. Sei.,
vol. 14 (9), p. 201.—Guayaquil, Ecuador (USNM 62790).
Range: Apparently very disjunct: Parts of Mexico; interior
highlands of Colombia (Dunn, Caldasia, vol. 3 (12), p. 201,
1944) ; western Eeuador.

CONIOPHANKS 2

1. Dorsals in 19 rows; 3-5 black lines on top of head between eyes
dromicif ormis
Dorsals in 21 rows, no black lines between eyes... .f. fissidens

Comophanes dromicifornus Peters
Tachymenis dromiciformis Peters, 1863, Monatsb. Akad. Wiss. Berlin,
p. 273.—Guayaquil, Ecuador (BerM 3729-2 syntypes, 3730-3 syntypes,
and 4550-3 syntypes).
Coniophanes signatus Garman, 1892, Bull. Essex Inst., vol. 24, p. 91.—
Guayaquil, Ecuador (CNHM 16941).
Range: Ecuador and Peru, on the Pacific coastal areas.

”

2 Coniophanes brevifrons Bailey, 1937, described from “Ecuador,” is actually
i species known from San Andres Island, according to Dunn and Saxe, Proc.
Acad. Nat. Sci. Phila., vol. 102, p. 162, 1950.

PETERS: SNAKES OF ECUADOR 513

Coniophanes fissidens fissidens Giinther

Coronella fissidens Ginther, 1858, Cat. Snakes Brit. Mus., p. 36.—
““Mexico,’’ restricted to San Andres Tuxtla, Vera Cruz, by Smith and
Taylor, 1950 (6 cotypes, BM 56.3.17.35-36, 57.7.31.48, 57.7.31.54,
60.6.17.17, and one unnumbered specimen).

Range: Central Vera Cruz south on Atlantie Coast of Cen-
tral America to northwestern Ecuador, avoiding high moun-
tains and the Yucatan Peninsula.

CORALLUS
le Nasalsinotemucontach sub caidallsm@ledgutha nies eral at eels 2
Nasals in contact; subcaudals more than 100... .e. enydris

2. Dorsum green (dark blue in preservative) with white line down spine,
lateral branches down sides; which may break into row of spots
... .caninus

Dorsum light, with large, pale-reddish blotches, margined by darker

COLOTaNC oh Bre PoE aI ie corset Been ek WA Cece tok Rk der me ne ede 3

SS)

Two medium-sized lateral internasals, separated by two medial inter-
nasals, arranged one behind the other; four supraloreals .. annu-
latus blombergi

Large pair of lateral internasals, which are in contact anteriorly but
separated posteriorly by a single medial internasal; three supra-
loreals....annulatus colombianus

Corallus annulatus blombergi Rendahl and Vestergren

Boa annulata blombergi Rendahl and Vestergren, 1941, Ark. fiir Zool.,
vol. 383A (5), p. 1, figs. 6-7—Rio Zamora, E. Ecuador (RMS 3141).
Range: Known only from the type locality.

Corallus annulatus colombianus Rendahl and Vestergren
boa annulata colombiana Rendahl and Vestergren, 1940, Ark. fiir Zool.,
vol. 838A (1), p. 2.—Caheceras, Chocé, Colombia (RMS).
Range: Choeé of Colombia and Ecuador.

Corallus caninus Linnaeus
Boa canina Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 215.—
“« America.’? (Type unknown).
Range: Colombia; Venezuela; Guianas; Brazil; Bolivia;
Ecuador.

Corallus enydris enydris Linnaeus
Boa enydris Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 215.—
‘“ America’’ (RMS 1 specimen).
Range: N. Brazil; Bolivia; Ecuador; Peru; Guianas.

014 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

DENDROPHIDION

1 Subcaudalsmlessistham ol (tars ee, eee scat sey caer ter en oe rerercote 2
Subcaudals more than 175... .dendrophis

2. Dorsum striped; a black streak through eye... . bivittatum
Dorsum unicolor; no black streak through eye... . brunnewm

Dendrophidion bivittatum Duméril, Bibron and Dumeéril
Leptophis bivittatus Duméril, Bibron and Duméril, 1854, Erp. Gen.,
vol. 7 (1), p. 540.—‘‘New Grenada’’ (PM 517 and PM 6.3460, 2 co-

types).
Range: Ecuador; Colombia; and Panama. A Chocéan dis-
tribution.

Dendrophidion brunneum Gunther

Herpetodryas brunneus Gimther, 1858, Cat. Snakes Brit. Mus., p. 116.—
‘¢Guayaquil’’ (BM 1946.1.12.98).

Range: Interandean valleys of Ecuador.

Dendrophidion dendrophis Schlegel
Herpetodryas dendrophis Schlegel, 1837, Physiog. Serpents, vol. 2,
p. 196.—‘‘Cayenne’’ (Types not in PM, probably non-existent).
Range: Amazonian region of Ecuador and Guianas.

DIAPHOROLEPIS

Diaphorolepis wagnert Jan
Diaphorolepis wagneri Jan, 1863, Elenco Sist., p. 94.—Andes of Ecuador
(MonM).
Synophis bicolor Peracca, 1896, Bol. Mus. Zool. Anat. Comp. Torino,
no. 266, p. 1—‘‘ Central America’’ (TurM).
Range: Western Ecuador in the lowlands, presumably into
Central America.

DIPSAS

1. Dorsal pattern of broad, dark brown or black bands that are much
wider than interspaces and have vertical edges, interspaces pink or red
in life (arliculata group) 2

Pattern motvas above. Ae «i Shecas ork Re a ee ee eee 3
2. At least one pair of labials in contact behind mental... . gracilis
Mental in contact with paired chin shields... . temporalis

3. Dorsal pattern of rounded, dark brown or black blotches or saddles
on sides, interspaces tawny brown (catesbyi group)...
Dorsalspatbern mot cassalboveraa ee ee eee ee eee 6

“I

10.

ite

12

PETERS: SNAKES OF ECUADOR 516

‘Two pretrontals’s dorsum ‘of head) unicolori sa..9 1. ae ein 5
Prefrontals usually fused; dorsum of head variegated and streaked with
white... .vermiculata

Blotches narrower at vertebral row than laterally; loreal does not enter
eye... .catesbyt

Blotches saddle-shaped, wider at vertebral row than laterally; loreal
enters eye... .pavonina

Dorsal blotches triangular or lozenge-shaped in appearance, usually
widest at ventrals, with yellow spot between corners of blotches at
ventrals'=(endicas Loup) a), ss oe ees echt ee Ue eee 7
Workalop waiter: NOGdas: VOOVEs cies: oe oc, eee Pere cas lee Maen eon Rane mean 8

Occipital region not streaked, may be spotted; first dorsal blotch broad-

ly fused along middorsal line... .indica indica
Occipital region longitudinally streaked; first dorsal blotch pair sepa-
rated by light line at vertebrals.... indica ecuadorensis

Dorsal ground color of light browns and tans, with narrow blotches that
are higher than wide, and much narrower than interspaces (at least
posteriorly), latter streaked, spotted or stippled throughout (variegata

(Eat NDY OV) Or eno Re Lp Ba ech ROE Wine One mney aig tare Rove MnP. come pment ais 9
Dorsalepattennenot: Askabovemacyeerit:. cli ede eee Teer 10

Dark brown spots on head unite on frontals to form U-shaped mark,

posterior tips of ‘‘U’’ often fused to first dorsal blotch... .variegata
nicholsi

Dark brown spots on head not fused on frontal, do not extend to first
dorsal blotch. ...variegata variegata

Dorsal blotches wider than interspaces, little contrast between them;
centers of blotches often considerably lightened, so that blotch resembles
Daireducllipsesm (elu pSsuieEnaaetOlp) eee Peel are eee 11
Dorsale parhernenoteassavovesr(oraiiy 2 LOUD) lems eee tee 12

Centers of dorsal blotches very light, so that blotch resembles paired
ellipses, on all individuals; chin heavily spotted; venter with two
parallel dark streaks; ventrals often less than 170, subeaudals less
than 80... .ellipsifera

Centers of dorsal blotches lightened only in adults, not in juveniles, and
never so light that the blotch resembles paired ellipses; chin not or
sparsely spotted; venter with large rectangular blotches between ends
of neighboring dorsal blotches; ventrals more than 175; subeaudals more
than 79... .oreas

Dorsal ground color yellowish-brown; first few dorsal blotches fused

ventrally; usually two pairs of lower labials in contact behind mental
latifasciata

Dorsal ground color reddish-brown, or, in old adults, interspaces same

color as blotches; usually all dorsal blotches fail to meet on venter;

usually less than two pairs in contact behind mental... .latifrontalis

516 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Dipsas catesbyi Sentzen
Coluber catesbeji Sentzen, 1796, Meyer’s Zool., Arch. 2, p. 66.—Un-
known.
Range: Amazonian Basin, from the Andean slopes of Bolivia,
Peru, Eeuador, and Colombia to coast of Venezuela and Brit-
ish Guiana, and through northern half of Brazil.

Dipsas ellipsifera Boulenger
Leptognathus ellipsifera Boulenger, 1898, Proc. Zool. Soc, London,
p. 117, pl. 12, fig. 2—Ibarra, Ecuador (BM 1946.1.21.26-29).
Range: Western slopes of Eeuadorian Andes.

Dipsas gracilis Boulenger
Leptognathus gracilis Boulenger, 1902, Ann. Mag. Nat. Hist., ser. 7,
vol. 9, p. 57.—St. Javier, Heuador (BM 1946.1.21.24-25, 2 males).
Leptognathus hammondi Boulenger, 1920, Ann. Mag. Nat. Hist., ser. 9,
vol. 6, p. 110.—Guatea (=Gualea?), Ecuador (BM 1946.1.20.95, fe-
male).
Sibynomorphus macrostomus Amaral, 1924, Jour. Washington Acad.
Sci., vol. 14, p. 9—Eeuador (USNM 14047).
Range: Humid forests of coastal northwestern Ecuador.

Dipsas indica indica Laurenti
Dipsas indica Laurenti, 1768, Syn. Rept., p. 90.—‘‘Ceylon,’’ restricted
to Amazonian region of South America by Peters, Mise. Publ. Univ.
Mich. Mus. Zool. no. 114, p. 68, 1960. (Original type unknown, icono-
type designated as pl. 43, fig. 5, of Seba, 1734, Peters, loc. cit., p. 68).
Range: Amazonian basin, in Brazil, Colombia, British
Guiana, Ecuador, and Peru.

Dipsas indica ecuadorensis Peters
Dipsas indica ecuadorensis Peters, 1960, Misc. Publ. Mus. Zool. Univ.
Mich., no. 114, p. 81.—Rio Solis, Caveceras del Rio Bobonaza, 14 km.
ESE of Puyo, Napo-Pastaza Province, Eeuador (UMMZ 118064, male).
Range: Known only from the Amazonian drainage of Eeua-
dor.

Dipsas latifasciata Boulenger
Leptognathus latifasciata Boulenger, 1913, Ann. Mag. Nat. Hist., ser.
8, vol. 12, p. 72.—Upper Marafion River, eastern Peru (BM 1946.1.20.
HD
Range: Amazonian slopes of northern Peru and extreme
southern Ecuador.

Dipsas latifrontalis Boulenger
Leptognathus latifrontalis Boulenger, 1905, Ann. Mag. Nat. Hist., ser.

PETERS: SNAKES OF ECUADOR 517

7, vol. 15, p. 561.—Aricaqua, Venezuela, 1000 meters (BM 1946.1.20.98,
female).

Leptognathus palmeri Boulenger, 1912, Ann. Mag. Nat. Hist., ser. 8,
vol. 10, p. 422.—El]1 Topo, Rio Pastaza, Ecuador (BM 1946.1.20.86,

male).
Range: Lower Amazonian slopes from Venezuela to southern
Keuador.

Dipsas oreas Cope
Leptognathus oreas Cope, 1868, Proc. Acad. Nat. Sei. Philadelphia,
p. 109.—Elevated valley of Quito, Ecuador, probably in error (ANSP
10115, male).
Leptognathus andiana Boulenger, 1896, Cat. Snakes Brit. Mus., vol. 3,
p. 452, pl. 23, fig. 2—Quito, Ecuador, probably in error (BM 1946.
1.20.12).
Range: Higher parts of the western slope of the Andes in
EKeuador.

Dipsas pavonina Schlegel
Dipsas pavonina Schlegel, 1837, Physion. Serp., vol. 2, p. 280.—
“¢Guyanes’’ (PM, cotype not locatable, and MPB, 4 cotypes).
Range: Guianas to Para, Brazil and to Amazonian slopes of
the Andes ; Colombia to Bolivia.

Dipsas temporalis Werner
Leptognathus temporalis Werner, 1909, Mitt. Naturh. Mus. Hamburg,
vol. 26, p. 241.—Esmeraldas, Eeuador (formerly HM, now destroyed).
Range: Chocé area of northwestern Ecuador, Colombia, and
Atlantic coast of Panama.

Dipsas variegata variegata Duméril, Bibron, and Duméril
Leptognathus variegatus Duméril, Bibron, and Duméril, 1854, Erp. Gen.,
vol. 7, p. 477.—‘‘Surinam’’ (PM 7299, and LeyM, one cotype).
Leptognathus robusta Miiller, 1923, Zool. Anz., vol. 57, p. 155.—Kast
Ecuador (ZSBS 622/1920).

Range: Venezuela, Guianas, Ecuador, and Peru.

Dipsas variegata nicholst Dunn
Sibynomorphus nicholsi Dunn, 1933, Copeia, p. 193.—Mid-basin of
Chagres River and mouth of Pequeni River, Panama (MCZ 37884, head
only, sex unknown).
Range: Atlantic slope of Panama to northwestern Ecuador.

Dipsas vernuculata Peters
Dispsas vermiculata Peters, 1960, Misc. Publ. Mus. Zool. Univ. Mich.,
no. 114, p. 65.—Chichirota, Lower Bobonaza River, Napo-Pastaza
Province, Ecuador (UMMZ 118063, male).

518 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Range: Amazonian slopes at lower elevations in Eeuador and
Peru.

DREPANOIDES

Drepanoides anomalus Jan

Clelia anomala Jan, 1863, Elenco Sist., p. 92.—‘‘Brasile’’ (type in
Neuchatel).

Pseudoclelia guitata Rendahl and Vestergren, 1941, Ark. fiir Zool., vol.
33A (5), p. 10.—Rio Pastaza between Rio Puyo and Rio Copataza,
Eeuador (RMS 3172).

Range: Amazonian slopes of Peru and Ecuador.

DRYADOPHIS *

1. Supralabials usually 8; body pattern of alternating dorsal and lateral

dark blotches... . pulchriceps

Supralabials normally 9; body pattern striped, unicolor or narrowly

aN eds eee ek oad Gace t c'eal am acs 3 Ok Fre = Pai ee 2
2. Single light lateral stripe on seale rows 4, 5, and 6... .heathi*

Jither no dorsal striping or a single light lateral stripe involving only

rows 4, 5 boddaerti

Dryadophis boddaerti boddaerti Sentzen
Coluber boddaerti Sentzen, 1796, Ophiol. Frag., Meyer’s Zool. Arch.,
vol. 2, p. 59.—Type and type locality unknown.
Coluber fasciatus Rosen, 1905, Ann. Mag. Nat. Hist., ser. 7, vol. 15,
p. 172, pl. 11, fig. 2—Balao, Ecuador (LunM).
Herpetodryas reticulata Peters, 1863, Monatsb. Akad. Wiss. Berlin, p.
285.—Region of Guayaquil, Ecuador (BerM 4504).
Range: Humid areas of northern South America to the limits
of the Amazon Basin.

Dryadophis pulchriceps Cope
Masticophis pulchriceps Cope, 1868, Proc. Acad. Nat. Sci. Philadelphia,
vol. 2, p. 105.—Plateau Valley of Quito, Eeuador (ANSP 5710, former-
ly USNM 6704).
Range: Guaymas Basin and more humid habitats of West
Central Ecuador.

DRYMARCHON

1. Belly and tail light throughout their length, no distinctive dark marks

° The range for Dryadophis boddaerti heathi Cope is given by Stuart, Misc.
Publ. Univ. Mich. Mus. Zool. no. 49, p. 78, 1941, as “arid coastal desert from
Ecuador possibly as far south as Chile,’ but no specimens are known from
Ecuador, and the range as given is based upon the distribution of apparently
suitable habitat.

PETERS: SNAKES OF ECUADOR 519

on edges of subocular labials; in adults, anterior portion of body
darker than posterior part and tail c. corais*

Posterior part of body and all of tail black above and below, light
brown anteriorly, black marks on subocular labials sharply defined
corais melanurus

Drymarchon corais melanurus Duméril, Bibron and Duméril
Spilotes melanurus Duméril, Bibron and Duméril, 1854, Erp. Gen., vol.
7 (1), p. 224.—‘‘ Mexico’’ (PM 638; 3185; 3354; 3 cotypes).
Drymarchon corais melanocercus Smith, 1941, Jour. Washington Acad.
Sci., vol. 31, p. 473.—(A substitute name for D. c. melanurus, a sec-
ondary homonym).
Range: Pacific slopes of Colombia and Eeuador; Central
America to Veracruz on the Atlantic slope and to Nicaragua
on the Pacifie slope.

DRYMOBIUS

Drymobius rhombifer Ginther
Coryphodon rhombifer Giinther, 1860, Proc. Zool. Soc. London, p. 236.—
Esmeraldas, Ecuador (BM 1946.1.12.94).
Range: Panama; Nicaragua; Costa Rica; Colombia; Eeua-
dor; Peru.

DRYMOLUBER

Drymoluber dichrous Peters
Herpetodryas dichroa Peters, 1863, Monatsb. Akad. Wiss. Berlin, p.
284.—Brazil; Surinam (BerM 1661-62, 2 syntypes).
Herpetodryas occipitalis Giinther, 1868, Ann. Mag. Nat. Hist., ser. 4,
vol. 1, p. 430.—Pebas, ‘‘ Ecuador’? (BM 1946.1.14.61).
Spilotes piceus Cope, 1868, Proc. Acad. Nat. Sci. Philadelphia, p. 105.—
Napo or Upper Maraiion, Ecuador (Formerly USNM 6660, now lost).
Range: Colombia; Ecuador; eastern Peru; northern Brazil.

EPICRATES

Epicrates cenchria cenchria Linnaeus
Boa cenchria Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 215.—
‘“Surinami’’ (RMS).
Range: Amazonian Basin and coastal Guianas.

ERYTHROLAMPRUS

1. Black annuli arranged in diads aesculapui aesculapti
Black annuli not arranged in diads, regularly distributed. ... Jee

520 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

bo

Dorsal blotches and interspaces approximately equal in width...
guentheri
Dorsal blotches much narrower than interspaces .................. 3

Black collar covering posterior tips of parietals and several scales on
midline of neck. ...mimus micrurus

Black collar absent; represented by spots; or only about one scale row
wide on midline, diverging to about three scales on sides of neck... .
mimus MiMUS

(a)

Erythrolamprus aesculapu aesculapw Linnaeus

Coluber aesculapii Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 220.—
‘¢Tndiis.’’ (RMS).
Range: Amazonian South America.

Erythrolamprus guenthert Garman

Erythrolamprus guentheri Garman, 1883, Mem. Mus. Comp. Zool. Har-
vard, vol. 8 (3), p. 154.—‘‘ Mexico (?).’’ (BM 1946.1.8.5-6, 1946.1.8.
38).

Range: Amazonian slopes of Ecuador.

Erythrolamprus mimus mamus Cope
Opheomorphus mimus Cope, 1868, Proce. Acad. Nat. Sei. Philadelphia,
p. 307.—‘‘ High regions of Ecuador or New Grenada.’’? (ANSP 3689).
Range: Eastern Peru and Ecuador.

Erythrolamprus mimus micrurus Dunn and Bailey

Hrythrolamprus mimus micrurus Dunn and Bailey, 1939, Bull. Mus.
Comp. Zool. Harvard, vol. 86 (1), p. 12.—Mine at Santa Cruz de Cana
in Darien, Panama, 2000 ft. (MCZ 31828, female).

Range: Choed of Panama, Colombia, and northwestern
Keuador.

KUNECTES

Hunectes murinus Linnaeus

Boa murina Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 215.—
“¢ America’? (GothM).

Range: Amazonian South America.

HELICOPS

1. Seales in 19 rows; dorsum with more or less regular dark brown, black-
edged cross bars... .angulata

S)

2. Seales in 23-25 rows; dorsum with four. series of confluent spots, upper
two rows may merge anteriorly... pastazae

PETERS: SNAKES OF ECUADOR 521

Helicops angulata Linnaeus

Coluber angulatus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 21.—
“¢ Asia’? (RMS).

Range: Colombia; Venezuela; Guianas; Northern Brazil;
Bolivia; Peru; Ecuador.

Helicops pastazae Shreve

Helicops pastazae Shreve, 1934, Oce. Pap. Boston Soc, Nat. Hist., vol. 8,
p. 129.—Pastaza River, from Canelos to Marafion River, Ecuador (MCZ
36993).

Range: Amazonian Ecuador.

IMANTODES

eS Ales T OWS tyes isc ec aera tr te sede iors Suds ibeenai pear ae ute See 2
Seale rows 15... .lentiferus

2. Vertebral scales not twice as wide as other dorsals; frontal twice as

long as wide; dorsum with small, indistinct spots... .inornatus

Vertebral scales more than twice as wide as other dorsals; frontal not
twice as long as wide; dorsum with large brown saddles... .cenchoa
cenchoa

Imantodes cenchoa cenchoa Linnaeus
Coluber cenchoa Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 226.—
America (type unknown).
Range: South America including Paraguay, Bolivia and
Argentina; Trinidad.

Imantodes inornatus Boulenger

Himantodes inornatus Boulenger, 1896, Cat. Snakes Brit. Mus., vol. 3,
p. 88, pl. 5, fig. 1—Hacienda Rosa de Jerico, Nicaragua (BM 1946.
12:63).

Range: Nicaragua, Costa Rica, Panama, and NW Ecuador
(presumably also Chocé Coast of Colombia).

Imantodes lentiferus Cope

Himantodes lentiferus Cope, 1894, Amer. Nat., vol. 28, p. 613.—Pebas
‘‘BHeuador’’ (Peru), and ‘‘E. Eeuador’’ (2 eotypes, one ANSP 11459,
other unknown).

Range: Amazonas, Brazil; and Amazonian EKeuador.

LACHESIS

Lachesis muta muta Linnaeus

Crotalus mutus Linnaeus, 1766, Syst. Nat., 12th ed., vol. 1, p. 373.—
‘¢Surinami’’ (Type unknown).

BaD BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Range: Panama; Venezuela; Colombia; Guianas; Eeuador ;
Peru; Brazil; Bolivia; Trinidad. Found in the lowlands on
both the Pacific and Amazonian slopes in EKeuador.

LAMPROPELTIS

Lampropeltis doliata micropholis Cope
Lampropeltis micropholis Cope, 1861, Proc. Acad. Nat. Sci. Philadelphia,
1860 (1861), p. 257.—Panama (ANSP 3427).
Range: Panama; Chocé of Colombia and Ecuador.

LEIMADOPHIS
Hee Niaxamumuscal esr ows One Oy, elie sso oa oleae eee eee 2
Maximum scale rows on body 19............. » aa et, Ee if
OPE Sirbcaudalsmmone at harie4: Opes ie eee te aie ee ee es toe 3

Subeaudals less than 46... . pygmaeus

SPNOmblacksnapenspOuggd Orsailliyaa nase sy nee ieee ere ee eee 4
A pair of black nape spots dorsally... .bimaculatus lamonae

4. Two preoculars....epinephalus ecuadorensis
Onewpreoculan> sym 7 dete ee Re een Re nec LE halo Ce Eee 5

5. Belly unspotted, unicolor yellowish white... .albiventris
Belly lotehed! cn 4 2 se Bae ono © 6 eR ee we ee 6

6. Ventrals less than 150... .reginae
Ventrals more than 155... . fraseri

Uo kale spelllone \yaldn, Clima Kyoinabe Sy uc cosmos acduons cee ecocvodone: 8
Belly immaculate... .typhlus

8. Dark streak on side of head, passing through eye.................. 9
No dark streak passing through eye... .poecilogyrus

9. Two preoculars; venter of tail yellow with square black blotches, may be
entirely black... .taeniurus
One preocular; venter of tail yellowish, speckled with olive... .simonsi

Leimadophis albiventris Jan

Liophis reginae albiventris Jan, 1863, Arch. per la Zool., vol. 2 (2),
p. 83.—‘‘ Western Andes’? (MonM) and ‘‘ Fra Lacatunga e Guayaquil’’
(Type no longer in PM).

Ophiomorphus alticolus Cope, 1868, Proc. Acad. Nat. Sci. Philadelphia,
p. 102.—Valley of Quito, Ecuador (Formerly USNM 6703, now lost).
Range: Found in the lowlands on both sides of the Andes
in Keuador. Its range elsewhere in South America is obscured
at present by erroneous use of the name in the literature.

PETERS: SNAKES OF ECUADOR 523

Leimadophis bimaculatus lamonae Dunn
Leimadophis bimaculatus lamonae Dunn, 1944, Caldasia, vol. 10, p. 486.
—Sonson, Antioquia, Colombia, 2410 m. (Idl1S).
Range: Known from the type locality and ‘‘Equateur’
(Laurent, Bull. Inst. Roy. Sci. nat. Belg., vol. 25 (9), p. 8,
1949).

Leimadophis epinephalus ecuadorensis Laurent
Leimadophis epinephalus ecuadorensis Laurent, 1949, Bull. Inst. Roy.
Sei. nat. Belg., vol. 25 (9), p. 8—‘‘Equateur’’ (IRB 5028—I.G. no.
3267).
Range: Known only from the type locality.

?

Leimadophis fraseri Boulenger
Liophis fraseri Boulenger, 1894, Cat. Snakes Brit. Mus., vol. 2, p. 131,
pl. 6, fig. 2—‘‘W. Eeuador’’ (BM 1946.1.6.63, female).
Range: Western Eeuador.

Leimadophis poecilogyrus Wied
Liophis poecilogyrus Wied, 1825, Beitr. Naturgesch. Brazil, vol. 1. p.
371.—Unknown.
Range: Argentina and Uruguay north to Amazonian Brazil
and Keuador.

Leimadophis pygmaeus Cope
Liophis pygmaeus Cope, 1868, Proce. Acad. Nat. Sei., Philadelphia, p.
103.—From Napo or neighboring part of Maranon, Ecuador (Formerly
USNM 6668, no longer there).
Range: Upper Amazonian region of Colombia and Eeuador.

Leimadophis reginae reginae Linnaeus
Coluber reginae Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 219.—
‘“Tndiis’’ (RMS, 2 cotypes).
Liophis reginae quadrilineata Jan, 1863, Arch. per la Zool., vol. 2 (2),
p. 84.—Ecuador; Central America; Colombia (MonM and VM; PM
cotype no longer existent. )

Range: Northern South America, east of the Andes.

Leimadophis simonsi Boulenger
Philodryas simonsti Boulenger, 1900, Ann. Mag. Nat. Hist., ser. 7,
vol. 6, p. 185.—Cajamarea, Peru, 9000 ft. (BM 1946.1.4.98).
Range: Peru; Ecuador.

Leimadophis taeniurus taenurus Tschudi
Liophis taeniurus Tschudi, 1845, Arch. fiir Naturg., vol. 11 (1), p. 166.
—‘‘Tn der heissen Waldregion,’’? Peru (VM?).
Range: Ecuador; Peru.

524

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Leimadophis typhlus Linnaeus

Coluber typhlus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 218.—
‘<Tndiis’’? (RMS).

Xenodon isolepis Cope, 1869, Proce. Amer. Philos. Soe., vol. 11, p. 155.—
Pebas, ‘‘ Ecuador’ ’—actually in Peru (Type not located at ANSP).
Range: South America from Argentina and Uruguay north
through Guianas and Colombia.

LEPTODEIRA
Usually one preocular; no dark, well-defined lateral spots lying between
the ovoid dorsal blotches... .annulata annulata
Usually two preoculars; distinct lateral spots that lie between the
lateralvends vot the dorsal spotseees. sos. «he oe oe ae oe eee 2
No nape stripe... septentrionalis ornata
A dark nape stripe that expands anteriorly to form a butterfly-shaped
mark on the posttemporals and the postparietals....septentrionalis
larcorum

Leptodeira annulata annulata Linnaeus

Coluber annulatus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 224.—
‘“ America,’’ restricted to the Lower Amazon, Para, Brazil (ZIU 9).
Eteirodipsas wieneri Savage, 1884, Bull. Soe. Philom. Paris, ser. 7, vol.
8, p. 146.—Ecuador (type unknown).

Leptodeira nycthemera Werner, 1901, Verh. Zool.-Bot. Ges. Wien, vol.
51, p. 598 Ecuador (BerM 16596).

Range: Amazonian Basin up to 1000 meters on eastern slopes
of Keuador, Peru and Bolivia; along Atlantic coast to Sao
Paulo, Brazil.

Leptodeira septentrionalis ornata Bocourt

Comastes ornatus Bocourt, 1884, Bull. Soe. Philom. Paris, ser. 7, vol. 8,
p. 141.—Isthmus of Darien, Panama (PM 6201, 2 syntypes).

Range: Chocéd region of Central and northwestern South
America, in Costa Rica, Panama, Colombia and Ecuador.

Leptodeira septentrionalis larcorum Schmidt and Walker

Leptodeira larcorum Schmidt and Walker, 1943, Zool. Ser. Field Mus.
Nat. Hist., vol. 24, p. 311.—Chiclin, Libertad, Peru (CNHM 24302,
male).

Range: Coastal regions of northern Peru and southern Ecua-
dor; also recorded from the Upper Maranon Valley, Peru.
Intergrades with ornata in a broad zone in the arid Pacific
coastal zone of Ecuador.

PETERS: SNAKES OF ECUADOR 525

LEPTOMICRURUS

Leptonuicrurus narduccu Jan

Elaps narducci Jan, 1863, Arch. per la Zool., Anat., Phys., vol. 2, p.
222.—Unknown.

Elaps melanotus Peters, 1881, Sitzber. Gesell. Naturf. Freunde, Berlin,
p. 51.—_Sarayacu, Ecuador (BerM 9812, 2 syntypes).

Elaps scutiventris Cope, 1869, Proc. Amer. Philos. Soe., vol. 9 (82),
p. 156.—Pebas, ‘‘ Eeuador’’, actually in Peru (ANSP 6801, location of
other types unknown).

Range: Amazonian slopes of the Andes of Ecuador, Peru,
and Bolivia; Amazonian Brazil.

LEPTOPHIS
Loreal present... .depressirostris
LOYOveey BRET] ofS YE1 slp arena ARGU, at Me atatplcsalt nbato tect peta RR iE ae A Saito, ote he 2
Ventrals 133-149; adults with a color pattern of dark oblique bands;
keels present on the scales of all dorsal rows... .riveti
Ventrals usually more than 149; adults not with a color pattern of dark
oblique bands; no keels on scales of first row of dorsals............. 3
Dorsum with strong coppery tint, ventrals also coppery, but slightly
darker with dark brown and white streaking... .cupreus
Dorsum not as above, usually green or bluish, ventral color contrasts
Tin Clomsal Color, . . nanan SWOO=, 5 oss00-0ahesacesso0rcascess 4

Head plates rarely margined with black; if narrow black margin is
present, head plates are never marked with numerous small black spots
or with a prominent large black spot on each parietal and supraocular
shield... .a. occidentalis

Head plates margined with black and with numerous small, black spots
or with a prominent large black spot on each parietal and supraocular
ey Fs, gn dees a ei a ey ee a gee etre mh he IR Ys Rae oe Wl a a ihe 5)

Head plates and dorsal scales with numerous small, irregularly shaped
black spots, also present on extreme outer edge of ventrals anteriorly

....a, bocourti

Head plates and dorsal scales not marked as above, but with a promi-
nent large spot on each parietal and supraocular plate....a. nigro-
marginatus

Leptophis cupreus Cope

Thrasops cupreus Cope, 1868, Proc. Acad. Nat. Sci. Philadelphia, p.
106.—Napo and Maranon, Ecuador (formerly USNM 6666, now lost).
Range: Lower Amazonian slopes in Ecuador.

4 This key is ineffective on many Ecuadorian specimens, and it is suggested

that identification be made on the basis of the stated ranges of the individual
subspecies until such a time as sufficient material is available to analyze more
completely the situation in Ecuador (see Oliver, Bull. Amer. Mus. Nat. Hist.,
vol. 92, pp. 157-280, 1948).

526 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Leptophis depressirostris Cope
Philothamnus depressirostris Cope, 1861, Proce. Acad. Nat. Sci. Phila-
delphia, 1860 (1861), p. 557—Coeuyas de Veraguas, New Grenada;
actually in Panama (ANSP 5207).
Range: Atlantic slopes of Nicaragua, Costa Rica and Pana-
ma; Pacific slopes of Colombia and Eeuador. A questionable
record from Peru.

Leptophis ahaetulla bocourtt Boulenger
Leptophis bocourti Boulenger, 1898, Proe. Zool. Soe. London, p. 116.—
Paramba (BM 1946.1.6.67-68); and Cachabé, Ecuador (BM 1946.1.6.
Wo)
Range: Northwestern Ecuador and Gorgona Island, Co-
lombia.

Leptophis ahaetulla nigromarginatus Giinther
Ahaetulla nigromarginata Giinther, 1866, Ann. Mag. Nat. Hist., ser. 3,
vol. 18, p. 28.—Upper Amazons (BM 1946.1.4.7).
Range: Extreme southeastern Colombia, western Brazil,
eastern Ecuador, and eastern Peru.

Leptophis ahaetulla occidentalis Ginther
Ahaetulla occidentalis Giinther, 1859, Proc. Zool. Soe. London, p. 412.—
Guayaquil (BM 1946.1.4.48, male); and western Ecuador (BM 1946.
1.6.62, male).
Range: Costa Rica to western Colombia and Eeuador.

Leptophis rweti Despax
Leptophis riveti Despax, 1910, Bull. Mus. Hist. Nat. Paris, 1910, p.
368.—Gualaquiza, Ecuador, 730 m. (PM 06.259).
Range: Higher altitudes up to 5000 feet on both sides of the
Andes in Ecuador; Amazonian Peru; western and north-
central Colombia; Panama. A single record from Trinidad.

LEPTOTYPHLOPS

1. Brown, lighter beneath, each scale with lighter outer edges, forming
more or less distinct longitudinal lines; forehead, lips, and end of tail
usually white... albifrons albifrons
Violet black, except for naso-labials and lower half of rostral, which
are dark brown, and the first 3 pairs of lower labials, which are
yellowish white. ...anthracinus

Leptotyphlops albifrons albifrons Wagler
Stenostoma albifrons Wagler, 1824, in Spix, Serp. Brazil, ssp. nov.,
p. 68, pl. 25, fig. 3—mnear Para, Brazil (MunM).

PETERS: SNAKES OF ECUADOR 527

Range: Argentina and Paraguay to Venezuela and the
Guianas.

Leptotyphlops anthracinus Bailey

Leptotyphlops anthracinus Bailey, 1946, Oce. Pap. Mus. Zool. Univ.
Mich., no. 492, p. 1—mnear Bafios, Ecuador, 1800 meters (UMMZ 90816,
male).

Range: Known only from the type locality and Abitagua,
Ecuador.

LIOPHIS
Sealessan vows ie sc ccc ye a as eis ocd Soe RO he eis Ren =n: ST ee 2
Seales in 19 rows... . festae
Dorsum crossbanded, no striping on body or tail................... 3
Dorsum unicolor, or with striping on part of body and tail.......... +
Six or seven upper labials, third and fourth in eye... . breviceps

Hight upper labials, fourth and fifth in eye... .cobella

Lateral dark streak on posterior portion of body and tail only; no

lateral light streak, subcaudals less than 75... .purpurans
A Jateral light streak on body from head on 5th or 6th scale row, sub-
candals=more phan, SU ov. eee] ae Sem ris ie Sete nedsr nar eue ieee 5

Whitish stripe on first scale row, head whitish to occiput, sides of
ventrals brown, no dark streak through eye... .albiceps

No whitish stripe on first scale row, dorsum of head brown, ends of
ventrals yellowish white, dark streak through eye... .subocularis

Tiophis albiceps Amaral

Rhadinaea albiceps Amaral, 1924, Jour. Washington Acad. Sci., vol. 14
(9), p. 200.—‘‘ Probably from Ecuador’? (USNM 22446).
Range: Known only from the type specimen.

Liophis breviceps Cope

Liophis breviceps Cope, 1861, Proc. Acad. Nat. Sci. Philadelphia, 1860
(1861), p. 252.—Surinam (ANSP 3967).
Range: Dutch Guiana and eastern Ecuador.

Liophis cobella Linnaeus

Coluber cobella Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 2i18.—
‘¢ America’’? (ZIU, no number, 2 cotypes and MAF, 1 cotype, no
number).

Range: Northern South America east of the Andes.

Inophis festae Peracca

Rhadinaca festae Peracea, 1897, Bol. Mus. Torino, vol. 12 (300), p. 16.
—vValley of Rio Santiago, Ecuador (TurM, male).

Range: Known only from the type locality.

528 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Liophis purpurans Duméril, Bibron, and Duméril
Ablabes purpurans Duméril, Bibron and Duméril, 1854, Erp. Gen., vol.
7 (1), p. 312.—Mana, Cayenne (PM 518).
Rhadinaea chrysostoma Cope, 1868, Proc. Acad. Nat. Sei. Philadelphia,
p. 104.—From the Napo or Marafion, Ecuador (USNM 6665).

Range: Guianas; Upper Amazons of Peru, Ecuador, and
Colombia.

Inophis subocularis Boulenger

Rhadinaea subocularis Boulenger, 1902, Ann. Mag. Nat. Hist., ser. 7,
vol. 9, p. 56.—Paramba, Eeuador, 3500 feet (BM 1946.1.5.81-82).
Range: Western Ecuador.

LIOTYPHLOPRS

Inotyphlops peters Boulenger

Helminthophis petersii Boulenger, 1889, Ann. Mag. Nat. Hist., ser. 6,
vol. 4, p. 360.—Guayaquil, Ecuador (BM 1946.1.11.26).
Range: Northwestern Ecuador.

LYGOPHIS
Hee Sealles, sineli/) LOWS emt GD OGY A . sueqee- ty were estes Gey aie cute ee 2
Scales in 19 rows at midbody... .lineatus

bo

Dorsum almost unicolor light brown, with a vague lateral light line
which is poorly distinguished from other dorsolateral color; vertebral
black line never prominent; often with a light, dark-edged ocellus on
each side of the nape... .whymperi

Dorsal ground color brown, with the light lateral line rather strongly
contrasted with dorsal color; vertebral black line prominent, at least
posteriorly ; usually no clearly marked light, dark-edged ocellus on each
side of the nape... . boursieri

Lygophis boursieri Jan

Dromicus boursieri Jan, 1867, Icon. Gen., livr. 25, pl. II, fig. 2.—
Quito, Ecuador (PM?).
Range: Rio Pastaza region on Amazonian slopes of Ecuador.

Lygophis lineatus lineatus Linnaeus
Coluber lineatus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 221.—
““Asia.’’ (DroM)..
Range: Northern South America.

Lygophis whymperi Boulenger
Coronella whymperi Boulenger, 1882, Ann. Mag. Nat. Hist., ser. 5,
vol. 9, p. 460.—Milligalli, Ecuador (BM 1946.1.4.4-5).

=~]

10.

Wh.

PETERS : SNAKES OF ECUADOR 529

Liophis atahuallpae Steindachner, 1901, Anz. Akad. Wiss. Wien, vol.
38, p. 105, pl. 1, figs. 4, 4a.—lLas Palmas, W. Spur of the Andes between
Babahoyo and Guaranda, W. Ecuador, 2500 meters (Type no. 68 in
Prince Bayern’s collection, now in VM?2).

Range: Western slopes of the Andes in Eeuador, above 1000
meters.

MICRURUS ®
Black rinesenOtoineen ad sheewee sce tr sree ee canine ere en eee eee en sae 2
Black rings in well-defined groups of three, these ‘‘triads’’ separated
Vie BEOE ZONES ier, Petterson cette ES ee CaN ne Raney et see Pisa suet aR ase 8
FAC Contrashinespalid Macross al e bells eye. ctsie renner tein ene nae eee er 3
Top of head black, usually covering most of parietals, with or without
amarnrowsliehtaband sacrossysnolte ae ame reeeeeeene eece 4

Very narrow white (yellow?) band across parietals, does not touch
frontal; dorsal pattern without red in adults, but with black rings of

alternating widths....a. annellatus

A broad red band across parietals, usually across posterior tip of
frontal; dorsal pattern with red rings... .mipartitus

Red zones usually much wider than black... .c. transandinus

Red zones, if distinguishable, about equal in length to black......... 5
Yellow rings not distinguishable from red rings; or, yellow rings
reduced to spots (i.e., to half-scales adjacent to the black rings)..... 6
Yellow rings continuous... .mertensi

More than 40 black bands on the body... .ornatissimus

ILS Treen SXS loleyelle lop rncls; ain. ine loyshy, 25 os occa vs ecusenocca coo cuse 7
Subeaudals more than 30... .langsdorffi

Subcaudals less than 25... .spixi obscurus

Anal plate entire... .hemprichi ortoni

Amalisplatesdivaded” «isc di erytia cb: G Ber ecks o Meads ge eds al eebot 9
Head with a narrow light crossband on snout which may not be com-
plete across dorsum of snout, but is indicated by yellow on lips and
post-nasal area; a broader light band across parietals............. 10
No crossband on snout; parietal crossband present or absent........ 12
Ventrals less than 215... .tschudvi olssoni

Wentrals, more; than 22006 se. cuse yo oes io Oe ce eae Seeks hit Acree 11
A moderately slender species; ventrals 220-263; triads of black rings

8-14... .lemniscatus
Body very slender; ventrals 266-321; triads 11-20... . filiformis

5 Elaps alienus Werner, described from ‘‘Venezuela or Ecuador,” has been
shown by Schmidt (Zool. Ser. Field Mus. Nat. Hist., vol. 20, p. 212, 1936) to
be a subspecies of the Central American species Mierurus affinis Jan.

530 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

12. Head with black pileus; a narrow light post-parietal band broadened on

the temporals and posterior labials. .. .e. ecuadorianus

Head variously colored; if black the scales are sharply outlined with

MERE cer. cxsto, <5, 5 SS. s, Diath BOS oy ee Ae en 13
13. Top of head light, the upper head scales all sharply outlined with black;

frontal shield very narrow... .s. surinamensis

Top of head not light, with dark bordered head shields; frontal normal

TMSPAP EM aed cece cke ce ek SES ROH erie, SBR oun Oona ee ey eek eee 14

14. Yellow rings very wide, as wide as outer black rings of the triads; top
of head dark, the shields with light outlines... .spixi obscurus
Yellow rings narrower above than outer black rings of triads; head light
anteriorly, with black spots... .ancoralis ancoralis

Micrurus ancoralis ancoralis Jan
Elaps marcgravii ancoralis Jan, 1872, Icon. Gen. Ophidiens, vol. 42,
pl. 4, fig. 2—Eeuador (MunM).
Elaps rosenbergii Boulenger, 1898, Proce. Zool. Soe. London, p. 117,
pl. 13—Paramba, Ecuador (BM 1946.1.23.74, female).

Range: Pacific drainage of Ecuador.

Micrurus annellatus annellatus Peters

Elaps annellatus Peters, 1871, Monatsb. Akad. Wiss. Berlin, p. 402.—
Pozuzo, Peru (BerM 7185).

Range: Amazonian lowlands of Ecuador, Peru, and Bolivia.

Micrurus carinicaudus transandinus Schmidt
Micrurus transandinus Schmidt, 1936, Zool. Ser. Field Mus. Nat. Hist.,
vol. 20 (19), p. 195—Andagoya, Choeé6, Colombia (MCZ 32744, male).
Range: The Chocé region west of the Andes from the Gulf of
Uraba to Ecuador.

Micrurus ecuadorianus ecuadorianus Schmidt
Micrurus ecuadorianus Schmidt, 1936, Zool. Ser. Field Mus. Nat. Hist.,
vol. 20 (19), p. 196.—Rio Daule, Eeuador (MCZ 3559, male).
Range: Pacific lowlands of Eeuador.

Micrurus fiiformis Giinther
Elaps filiformis Gimther, 1859, Proc. Zool. Soc. London, p. 86, pl. 18,
fig. B—Para, Brazil (BM 1946.1.20.13).
Range: Amazonian basin of Brazil, Ecuador, and Colombia.

Micrurus hemprichi ortoni Schmidt
Micrurus hemprichi ortoni Schmidt, 1953, Fieldiana-Zoology, vol. 34,
p. 166.—Pebas, Peru (MCZ 12423, male).
Range: Upper Amazon Basin in Peru and Ecuador; a single

PETERS: SNAKES OF ECUADOR 531

specimen from Belem which may have been transported from
the upper river.

Micrurus langsdorffii Wagler
Micrurus langsdorfii Wagler, 1824, in Spix, Serp. Brazil, p. 10, pl. 2,
fig. 2—Rio Japura, Amazonas (Type unknown).
Elaps batesii Giinther, 1868, Ann. Mag. Nat. Hist., ser. 4, vol. 1, p.
428.—Pebas, ‘‘Heuador.’’ This type locality is not actually Hcua-
dorian, as Pebas is in Peru (BM 1946.1.17.21).
Elaps wmperator Cope, 1868, Proc. Acad. Nat. Sci. Philadelphia, p.
110.—Napo and Marafion, Ecuador (ANSP 6793).
Elaps steindachneri Werner, 1901, Verh. Zool.-Bot. Ges. Wien, vol. 51,
p. 599.—Eeuador (VM).
Elaps fasslii Werner, 1927, Sitzber. Akad. Wiss. Wien, vol. 135, p. 191.
—KEcuador (VM—this is the same specimen that served as the type of
Elaps steindachneri).
Range: Amazon Basin.

Micrurus lemnscatus Linnaeus
Coluber lemniscatus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 224.
—‘Asia.’? (RMS, 2 cotypes). The type locality has been restricted
to Belem, Para, Brazil (Schmidt and Walker, Zool. Ser., Field Mus.
Nat. Hist., vol. 24, p. 294, 1943).
Range: Amazon Basin.

Micrurus mertensi Schmidt
Micrurus mertensi Schmidt, 1936, Zool. Ser. Field Mus. Nat. Hist., vol.
20 (19), p. 192.—Pacasmayo, Peru (Senck 94206, male).
Range: Coastal areas of northwestern Peru and southwestern
Keuador.

Micrurus mipartitus nipartitus Duméril, Bibron, and Duméril
Elaps mipartitus Duméril, Bibron, and Duméril, 1854, Erp. Gen., vol. 7
(2), p. 1220.—Rio Sucio, Colombia (The type is no longer in PM).
Elaps semipartitus Jan, 1858, Rev. Mag. Zool., p. 516.—Cayenne (MilM;
the specimen in the PM is now missing).

Elaps fraseri Boulenger, 1896, Cat. Snakes Brit. Mus., vol. 3, p. 432,
pl. 22.—West Eeuador (BM 1946.1.17.34).

Elaps mentalis Boulenger, 1896, Cat. Snakes Brit. Mus., vol. 3, p. 432,
pl. 22.—Pallatanga, Eeuador (BM 1946.1.17.16), and Cali, Colombia
(BM 1946.1.17.27).

Elaps calamus Boulenger, 1902, Ann. Mag. Nat. Hist., ser. 7, vol. 9,
p. 57._San Javier, Ecuador (BM 1946.1.20.25).

Elaps aequicinctus Werner, 1903, Zool. Anz., vol. 26, p. 249.—Venezuela
or Keuador (IRB).

Range: Darien, Panama to western EKeuador.

532, BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Micrurus ornatissimus Jan
Elaps ornatissimus Jan, 1858, Rev. Mag. Zool., p. 521.—Mexico (MilM).
Type locality in error, restricted to Rio Putomayo, Colombia, by
Schmidt (Fieldiana-Zoology, vol. 34, p. 345, 1955).
Elaps buckleyi Boulenger, 1896, Cat. Snakes Brit. Mus., vol. 3, p. 416,
pl. 22, fig. 1—Para, Brazil (BM 1946.1.17.18) ; and Canelos, Ecuador
(BM 1946.1.17.17).
Range: Region of headwaters of the Amazon, from the Rio
Caqueta in Colombia to Southern Peru.

Micrurus spixt obscurus Jan
Elaps corallinus var. obscura Jan, 1872, Icon. Gen. Ophidiens, livr.
41, pl. 6, fig. 3—Lima, Peru (Type unknown). Type locality in error,
designated as Iquitos, Peru, by Schmidt (Fieldiana-Zoology, vol. 34,
p. 175, 1953).
Elaps heterozonus Peters, 1881, Sitzber. Ges. Naturf. Freunde, Berlin,
p. 52.—Sarayacu, Ecuador (BerM 9813).
Range: Amazonian drainage of Colombia, Ecuador, Peru,
and Bolivia; head waters of Orinoco in Venezuela.

Micrurus surinamensis surinamensis Cuvier
Elaps surinamensis Cuvier, 1817, Régne Animal, vol. 2, p. 84.—Surinam
(PM 4629).
Range: Guianas, periphery of Amazon Basin, Colombia,
Keuador, Peru, Bolivia and Brazil.

Micrurus tschudui olssoni Schmidt and Schmidt
Micrurus olssoni Schmidt and Schmidt, 1925, Zool. Ser. Field Mus, Nat.
Hist., vol. 12 (10), p. 130, pl. 11—Negritos, Piura, Peru (CNHM 5724,
male).
Range: Pacifie coastal region of NW Peru and SW Eeuador.

NINIA

1. 19 scale rows, head scales smooth... .atrata
21 seale rows, striated head seales, short snout with high loreal....
hudsom

Ninia atrata Hallowell
Coluber atratus Hallowell, 1845, Proce. Acad. Nat. Sci. Philadelphia,
p. 245.—Within 200 miles of Caracas, Venezuela (ANSP 3410-12).
Streptophorus sebae schmidti Jan, 1862, Arch. Zool. Anat., vol. 2, fasc.
1, p. 27.—Guayaquil, Ecuador (HM).
Ninia spilogaster Peters, 1881, Sitzber. Ges. Naturf. Freunde, Berlin,
p. 49.—Ecuador (Type not located in BerM).
Range: Venezuela; Colombia; Ecuador; Panama; Costa
Rica ; Trinidad.

PETERS: SNAKES OF ECUADOR 533

Nima hudsoni Parker
Ninia hudsoni Parker, 1940, Ann. Mag. Nat. Hist., ser. 11, vol. 5, p. 270.
—New River, British Guiana (BM 1939.1.19, male).
Range: British Guiana; Ecuador.

NOTHOPSIS

Nothopsis rugosus Cope
Nothopsis rugosus Cope, 1871, Proe. Acad. Nat. Sci, Philadelphia, vol.
2, p. 201, pl. 13, figs. 1-7.—Isthmus of Darien, Panamé (USNM 12427).
Nothopsis affinis Boulenger, 1905, Ann. Mag. Nat. Hist., ser. 7, vol. 15,
p. 453.—Salidero, Ecuador (BM 1946.1.15.62, male).
Range: Atlantic coast of Nicaragua, Costa Rica, and Pana-
ma; Pacifie coast of Colombia and Ecuador.

OXYBELIS
1. Snout three times as long as eye; anal divided... . Re. os caa SR 2
Snout less than three times as long as eye; anal entire............. 3
2. Distinct white stripe on ends of ventrals.... fulgidus
No white stripe on ends of ventrals... a. aeneus

Broad black stripe on outer ends of ventrals; lateral and vertebral
brown stripe, chin very dark, spotted with black and brown...
argenteus

Unicolor brown, chin lighter brown... . brevirostris

ce

Oxrybelis aeneus aeneus Wagler
Dryinus acneus Wagler, 1824, in Spix, Serp. Brasil, p. 12, pl. 3.—Teffe,
Amazonas, Brazil, according to Bogert and Oliver (Bull, Amer, Mus.
Nat. Hist., vol. 83, p. 389, 1945) (MunM).
Coluber acuminatus Wied, 1885, Beitr. Naturg. Brazil, vol. 1, p. 322.—
Unknown.
Range: Guatemala to southeastern Brazil.

Oxybelis argenteus Daudin
Coluber argenteus Daudin, 18038, Hist. Nat. Rept., vol. 6, p. 336.—
No type locality given in description (Type not located in PM).
Range: Amazonian South America.

Oxybelis brevirostris Cope
Dryophis brevirostris Cope, 1861, Proc. Acad. Nat. Sei. Philadelphia,
1860 (1861), p. 555.—Veraguas, New Grenada, which is the same as
Cocuyas de Veragua, Panamé (USNM 31349).
Range: Atlantic lowlands of Central America; Pacific low-
lands of Colombia and Ecuador.

534 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Oxybels fulgidus Daudin
Coluber fulgidus Daudin, 1803, Hist. Nat. Rept., vol. 6, p. 352, pl. 80.—
‘<Port au Prince, Santo Domingo’’—probably in error (PM).
Range: Northern South America east of the Andes; Central
America and Mexico.

OXYRHOPUS
1. Preocular either in contact with frontal or narrowly separated from it
....petola
Preocular fairly widely separated from frontal by supraocular...... 2
or eHead dark or amottledsabover. a). ace a tea ne ok eo eee 3
Head uniform yellow or orange... . formosus

oy

Red or pale reddish brown, dotted with blackish brown, anterior part

of body may or may not have a few black eross bands disposed in threes
.. melanogenys

Dorsum yellow with irregular dark brown markings which may cover

only a single scale or be confluent into blotches or zig-zag lines....

jfitzingeri frizzelli

Oxyrhopus fitzingeri frizzelli Schmidt and Walker
Oxyrhopus fitzingeri frizzelli Schmidt and Walker, 1943, Zool. Ser. Field
Mus. Nat. Hist., vol. 24, p. 313.—Negritos, Piura, Peru (CNHM 35997,
male).

Range: Dry Pacific lowlands of Keuador and Peru.

Oxyrhopus formosus Wied

Coluber formosus Wied, 1820, Reise in Brazil, part I, p. 257, footnote—
Morro d’Arana, Brazil (type?).

Range: Amazonian South America.

Oxyrhopus melanogenys Tschudi
Sphenocephalus melanogenys Tschudi, 1845, Arch. fiir Naturg., vol. 11
(1), p. 163.—Peru (type?).
Range: Upper Amazonian parts of Bolivia; Peru; Brazil;
and Ecuador.

Oxyrhopus petola Linnaeus

Coluber petola, Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 225.—
““ in Africa’? (ZIU).

Range: Amazonian South America; west coast of Eeuador
and Colombia ; Central America. Roze (Bol. Mus. Cient. Nat.,
Caracas, vol. 1, p. 190, 1957) has revived the name semifascia-
tus 'T'schudi for the Amazonian population of this species,
and it is certainly applicable to the Ecuadorian material on

PETERS: SNAKES OF ECUADOR 535

the Amazonian slope. Use of this name would require action
concerning names for other populations, however, elsewhere
in Keuador, and these names are not currently available.

PELAMIS

Pelamis platurus Linnaeus
Anguis platura Linnaeus 1766, Syst. Nat., 12th ed., vol. 1, p. 391.—
‘¢Surinami’’ (type unknown).
Range: Pacific and Indian oceans.

PHILODRYAS
1. Smooth seales; uniform green above, yellowish green below; scale rows
15 or 13 near anus... . viridissima

Keeled seales; typically with darker dorsal band bounded below by
lighter one, and with darker sides, yellow spotted with brown or gray
ventrally; may occasionally be uniform in coloration; scale rows 17 near
anus... .elegans rufidorsata

Philodryas elegans rufidorsata Gunther

Dromicus rufidorsatus Giinther, 1858, Cat. Colubrid. Snakes British
Museum, p. 130.—‘‘ Ameriea,’’ restricted to northern coastal Peru by
Schmidt and Walker (BM 1946.1.9.28 and 1946.1.9.31, cotypes).
Tachymenis canilatus Cope, 1868, Proce. Acad. Nat. Sci. Philadelphia,
p. 104.—Guayaquil, Eeuador (USNM 12351).

Range: Pacific Lowlands of Eeuador and Peru, in drier
parts.

Philodryas viridissima Linnaeus
Coluber viridissimus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 226.
—‘‘Surinami’’ (Unknown).
Range: Amazonian South America from Argentina to
Guianas.

PLIOCERCUS

Pliocercus euryzona euryzona Cope
Pliocercus euryzonus Cope, 1862, Proc. Acad. Nat. Sci. Philadelphia,
p. 72.—Region of the Truando, New Grenada, which is now Colombia
(Type formerly USNM 4303, now apparently lost).
Range: Colombia; Ecuador; Equatorial Brazil; Panama.

PSEUDOBOA

Pseudoboa coronata Schneider
Pseudoboa coronatus Schneider, 1801, Hist. Amphib., vol. 2, p. 286.—
‘¢ America.’’ (Type unknown).
Range: Guianas; Brazil; Colombia; Ecuador; Peru; Bolivia.

536 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

PSEUSTES
Ie iwo postoculars'tworanltenor emp oral sin sil clr ies eect 2
Three postoculars; one anterior temporal....sulphureus sulphureus
2. Dorsum unicolor, a dull brown in adults; a series of brown bands on
light brown ground color, in juveniles... . poecilonotus polylepis
Dorsum irregularly barred with yellow or each seale has a yellow center
with a black edge... . shropshirei

Pseustes poecilonotus polylepis Peters

Ahaetulla polylepis Peters, 1867, Monatsb. Akad. Wiss. Berlin, p. 709.—
‘<Surinam’’ (BerM 5899).

Range: Guianas; Colombia; Ecuador; Peru; Bolivia; Brazil ;
and Trinidad.

Pseustes shropshire: Barbour and Amaral

Phrynonaxz shropshirei Barbour and Amaral, 1924, Occ. Pap. Boston
Soe. Nat. Hist., vol. 5, p. 131—Fort Sherman, Canal Zone, Panama
(MCZ 18819).

Range: Costa Rica; Panama; Pacifie Colombia and Ecuador.

Pseustes sulphureus sulphureus Wagler

Natrix sulphurea Wagler, 1824, in Spix, Serp. Brasil, p. 26, pl. 9.—
Rio Japura, Brazil (MonM?).
Range: Peru; Ecuador; Brazil; Guianas; Trinidad.

RHADINAEA
1. Dorsals reduce to 15 posteriorly; subcaudals less than 75... . brevirostris
DorsalsaVeateanus subcaudal simone ubiene (oe eee aete ne ee 2

2. Head and nape reddish; body practically unicolor, usually with a vague
lighter yellow area on first seale rows above a black stripe on tips of

ventrals and first scale row pachyura fulviceps
Body with two or more white lines, head and nape blackish, a light
spot on parietal... 1. lateristriga

Rhadinaea brevirostris Peters
Dromicus brevirostris Peters, 1863, Monatsb. Akad. Wiss. Berlin, p.
280.—Apparently from Quito, Ecuador, purchased—probably erroneous
(BerM 8869).
Dromicus viperinus Giinther, 1868, Ann. Mag. Nat. Hist., ser. 4, vol. 1,
p. 418.—Pebas, ‘‘ Eeuador,’’ actually Peru (BM 1946.1.1.9-10).
Range: Guianas to Bolivia.

Rhadinaea lateristriga lateristriga Berthold
Liophis lateristriga Berthold, 1859, Gottingen Gelehrte Anz., vol. 3,
p. 180.—Popayan, Colombia (GottM).
Dromicus frenatus Peters, 1863, Monatsb. Akad. Wiss. Berlin, p. 218
Guayaquil, Ecuador (BerM 2126).

PETERS : SNAKES OF ECUADOR 5ST

Urotheca coronata Steindachner, 1901, Anz. Akad. Wiss. Vienna, p. 106,
pl. 1, figs. 3, 3a—Region of Babahoyo, Ecuador (Prince Bayern Coll.
no. 58).

Erythrolamprus labialis Werner, 1909, Mitt. Naturh. Mus. Hamburg,
vol. 26, p. 237.—Ecuador (HM 117).

Range: Central Colombia to Ecuador on Pacifie slope.

Rhadinaea pachyura fulviceps Cope
Rhadinaea fulviceps Cope, 1885, Proc. Amer. Philos. Soe., vol. 23, p.
279.—__‘‘Panama’’ (USNM 14118).
Range: Panama; Pacific Colombia and Eeuador.

RHINOBOTHRYUM

Rhinobothryum bovallu Andersson
Rhinobothrium bovallii Andersson, 1916, Meddel. Goteborg’s Mus. Zool.,
Afd. 9, p. 32, fig. 4—Siquirres, Costa Rica (GothM 1221).
Range: Costa Rica, Panama, Colombia, and northwestern
Ecuador, in the Choco area.

SIBON

1. Ventrals less than 150... .dunni
Wentralssmoresthan WS: . ces sabre pees te Ss ik ok. Bye ee eae ee aa 2
2. Dorsal pattern often obscured by heavy deposition of black pigment,
belly heavily spotted with dark brown, and may be completely black
... .nebulata leucomelas
Dorsal pattern of chocolate or reddish-brown blotches, contrasting
strongly with the light brown or grayish interblotch areas... .nebulata
nebulata

Sibon dunni Peters
Sibon dunni Peters, 1957, Copeia, p. 110.—Pimanpiro, San Nicholas,
Imbabura, Ecuador (UMMZ 92068, male).
Range: Known only from the type locality.

Sibon nebulata nebulata Linnaeus
Coluber nebulatus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 464.—
‘< America’’ (RMS).
Range: Mexico to northern Brazil and Amazonian Eeuador ;
Trinidad; Tobago. An isolated population exists in Pacific
Ecuador, south of the range of S. n. lewcomelas.

Sibon nebulata leucomelas Boulenger
Leptognathus leucomelas Boulenger, 1896, Ann. Mag. Nat. Hist., ser. 6,
vol. 17, p. 18.—Buenaventura, Colombia (BM 1895.11.16.16, female).
Range: Panama-Colombia border through Chocé of Colombia
to northwestern Ecuador.

038

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

SPILOTES

Spilotes pullatus pullatus Linnaeus

Coluber pullatus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 225.—
‘Agia’? (RMS).

Spilotes megalolepis Giinther, 1865, Ann. Mag. Nat. Hist., ser. 3, vol. 15,
p. 93.—South America-purchased (BM 1946.1.13.2).

Range: Central America south of Nicaragua; northern South
America; Trinidad ; Tobago.

STENORHINA

Stenorhina degenhardtu degenhardtu Berthold

ie

Calamaria degenhardtii Berthold, 1846, Abh. Ges. Wiss. Gottingen,
vol. 3, p. 8, tab. 1, figs. 3-4—‘‘ New Grenada’’ (GottM).
Range: Central America; Pacifie Colombia and Eeuador.

SYNOPHIS
Seales in 19 rows... .miops
Scales in 21 rows... .lasallei

Synophis lasallet Niceforo-Maria

Diaphorolepis lasallei Niceforo-Maria, 1950, Rev. Acad. Colomb. Cien-
cias Exactas, ete., vol. 7 (28), p. 517, 3 figs—Paraje situado al norte
de Alban, Cundinamarca, Cordillera Oriental (vertiente occidental),
2200 m., about 60 km. NW of Bogoté, Colombia (Id1S, male).

Range: Known from type locality and from Rio Sandalias,
Napo-Pastaza Provinee, Eeuador.

Synophis miops Boulenger

ew

Synophis miops Boulenger, 1898, Proc. Zool. Soc. London, p. 115.—
Paramba, Ecuador (BM 1946.1.12.30, female).
Range: Known from type only.

TANTILLA
Subcaudals more than 75... longifrontalis
Subcaudals less than! 75) Secs 26 aes pe ace oe a eee 2
Body with narrow (two seale rows) black half rings... . supracincta
Body.stripeduor unicolor: 3 s..0228 oo Se ee ee 3
hip ot snout swhites pretrontals contacts lalbicls eae ee eee 4

Tip of snout dark brown, prefrontals and labials not in contact....
melanocephala melanocephala

PETERS: SNAKES OF ECUADOR 539

4. All scales above second row brownish, below second row entirely white
melanocephala capistrata
Dorsum striped with dark brown, light brown and white, first row of
dorsal scales white on lower half... . fraseri

Tantilla fraseri Giimther
Homalocranium melanocephalum fraseri Giinther, 1895, Biol. Centr.-
Am., Rept., fase. 19, 1895, p. 148.—Quito, Ecuador (BM 1946.1.9.43-44),
and ‘‘w. Eeuador’’ (BM 1946.1.8.77; 1946.1.8.80; 1946.1.9.84).
Range: High western slopes of Andes, perhaps in Quito
valley.

Tantilla longifrontalis Boulenger
Homalocranium longifrontale Boulenger, 1896, Ann. Mag. Nat. Hist.,
ser. 6, vol. 17, p. 17.—Cali, Colombia (BM 1946.1.8.84).
Range: Eastern slopes of Andes, at lower altitudes, in Co-
lombia and Keuador.

Tantilla melanocephala melanocephala Linnaeus
Coluber melanocephalus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1,
p. 218.—‘‘ America’’ (RMS).
Range: Central America throughout South America north of
Argentina and east of the Andes.

Tantilla melanocephala capistrata Cope
Tantilla capistrata Cope, 1876, Jour. Aead. Nat. Sci. Philadelphia,
(n. s.) vol. 8 (2), p. 181.—Valley of Jequetepeque, (Libertad), Peru
(Type formerly in ANSP, now apparently lost).
Range: Northern coastal Peru to arid valley of the Maranon
and the Catamayo and Malacatos Valley in southern Eeuador.

Tantilla supracincta Peters
Homalocranion supracinctum Peters, 1863, Monatsh. Akad. Wiss. Berlin,
1863, p. 272.—Guayaquil, Ecuador (BerM 4791).
Range: Ecuador, probably on coastal plain from Guayaquil
northward.

THAMNODYNASTES

*Thamnodynastes pallidus Linnaeus
Coluber pallidus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 221.—
‘¢in Indiis’’? (DroM).
Range: Guianas; Brazil; Peru.

Thamnodynastes natterert Mikan
Coluber nattereri Mikan, 1820, Delect. Faun. Flor. Braz., fig. 1.—
‘<Lectus prope Sebastianopolim.’’

540 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Range: Coastal Ecuador and Peru.

Remarks: This species has been recorded from Guayaquil on
the basis of USNM 12277, which Walker (Bull. Mus. Comp.
Zool. Harvard, vol. 96 (1), p. 14, 1945) called Dryophylax
natterer.

TRACHYBOA

1. Top of head without horns... . gularis
Top of head, canthus rostralis with horns... . boulengeri

Trachyboa boulengeri Peracea

Trachyboa boulengeri Peracea, 1910, Annular. Mus. Zool. Univ. Napoli,
vol. 3 (12), p. 1—Type locality unknown (UNZM).
Range: Chocé region of Ecuador, Colombia and Panama.

Trachyboa gularis Peters
Trachyboa gularis Peters, 1860, Monatsb. Akad. Wiss. Berlin, p. 200,
fig. 1—Guayaquil, Ecuador (BerM 3770-71, 2 syntypes).
Trachyboa gularis multimaculata Rosen, 1905, Ann. Mag. Nat. Hist.,
ser. 7, vol. 15, p. 169.—Balao, Ecuador (LunM).
Range: Dry parts of western coastal Eeuador (Brazil?).

TRETANORHINUS

Tretanorhinus taeniatus Boulenger
Tretanorhinus taeniatus Boulenger, 1903, Ann. Mag. Nat. Hist., ser. 7,
vol. 12, p. 350.—Rio Sapayo, N. W. Ecuador, 450 feet (BM 1946.1.15.
40).
Range: Pacific coast in Colombia and Eeuador.

TRIPANURGOS

Tripanurgos compressus Daudin
Coluber compressus Daudin, 1808, Hist. Nat. Rept., vol. 6, p. 247.—
Surinam (PM 3730).
Range: Brazil; Paraguay; Bolivia; Colombia; Ecuador ;
Guianas ; Trinidad ; Panama.

TROPIDOPHIS

1. Smooth scales; 200 ventrals, 42 subcaudals, 11 upper labials with 3
entering eye.... battersbyt
Keeled scales, 150-160 ventrals, 25-31 subeaudals, 8-9 upper labials with 2
entering eye... .taczanowskyi

PETERS: SNAKES OF ECUADOR 541

Tropidophis battersbyi Laurent
Tropidophis battersbyi Laurent, 1949, Bull. Inst. Roy. Sci. Nat. Belg.,
vol. 25 (9), p. 6.—‘‘ Equateur’’ (IB 739-I.G. no. 3701).
Range: Known only from the type specimen.

Tromdophis taczanowskyi Steindachner
Ungalia taczanowskyi Steindachner, 1880, Sitzber. Akad. Wiss. Wien,
vol. 80, Abt. 1, p. 522.—Tambillo, Peru (VM 2 cotypes).
Range: Amazonian Peru and Keuador; Brazil?.

TYPHLOPS

Typhlops reticulata Linnaeus
Anguis reticulatus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 228.—
“¢ America’? (Type unknown).
Range: Guianas; Venezuela; Brazil; Ecuador; Colombia ;
Peru; Argentina; Trinidad.

XENODON

1. Anal entire; scales in 19 rows (rarely 21) _rhabdocephalus
Anal divided; scales in 21 rows... . severus

Xenodon rhabdocephalus Wied
Coluber rhabdocephalus Wied, 1825, Beitr. Naturges. Brasil, vol. 1,
p. 351.—Sertao von Bahia (Type unknown).
Range: Central America; Colombia; Ecuador; Bolivia;
Brazil.

X enodon severus Linnaeus
Coluber severus Linnaeus, 1758, Syst. Nat., 10th ed., vol. 1, p. 219.—
‘“ Agia’? (RMS).
Range: Amazonian South America.

XENOPHOLIS

Xenopholis scalaris Wucherer
Elapomorphus scalaris Wucherer, 1861, Proce. Zool. Soe. London, p. 235.
—Cafiaviera, Matta de S. Joao, Bahia, Brazil (BM 1946.1.8.60, cotype;
other cotypes?).
Range: Brazil; Bolivia; Ecuador.

re

aiaaned
: v brian oar?
#05 ‘enue, ae outs

; .. je yy a
. o * “es i ufugt ab #4 ry tewinagt i} rt A
Ota is
ctesk tent JF hae epreaul ah
“a ene BY Dieta fl ee ts a4 ~

7) a)

See ee Pena, PA Ete] & i4eS and agile

.. | oF - eo” aa

iz oe . . ’ : a y a
; y = ; IZ =

i _ a
ro z hs a ¥ y “A i ae a
4 = ; Fl fs ~*~
' 1 7
>,
}
‘ a
i i , = t se
5
ve
mS fut
: i : cae J 7 “we
et =
a :
4 ig :
; ae, f %
x :
% f 7 i =A
fi _ ' =
a a a9 § | 7 i ; { 5
a ( ite
- a" abe .
caer Ps es : fA) ; th oe ‘ei, has YF i
a" ney See’ 4 7 MPeZ Pee ae
7 i y 4 r : " niga a VY
' -= 7 fe ~~ oa ih co "
: o ats os, at hg Nant i ij
re 5
, dt oe
; , U oon iy
~ Pe Be Ww
te
s a
a

Harvard MCZ Libra

~ iin nn

4 066 304 072