OF THE.

UNIVERSITY
Of ILLINOIS

53O.5
FI

fflou

Return this book on W Before the
Latest Date stamped below. A
charge is made on all overdue
books.

University of Illinois Library

-FEB44-W

M32

ZOOLOGICAL SERIES

OF
FIELD MUSEUM OF NATURAL HISTORY

Volume XX CHICAGO, DECEMBER 11, 1933 Pages 41-66

NOTES ON FISHES OF THE FAMILY
HEMIRHAMPHIDAE

BY ALFRED C. WEED

ASSISTANT CURATOR OF FISHES

In attempting to locate properly a remarkable genus, Odonto-
rhamphus gen. nov., of the family Hemirhamphidae, collected by
the Philip M. Chancellor-Field Museum Expedition to Aitutaki
Island, Cook Archipelago, central Pacific, it has been necessary to
examine critically all the available generic descriptions in the
family. These descriptions show such grave discrepancies and
omissions that it has seemed advisable to bring together all the
information in such form that it may be available as a basis for
future work. The entire group is urgently in need of a revision
based on a series including at least a majority of the nominal species.
Such a revision will probably require reexamination of most of the
specimens now in museums.

The great need for a revision of this group may be illustrated
by some examples taken from the collections of Field Museum.
There are four groups of specimens that were identified as Hypo-
rhamphus roberti. First, a group from the Atlantic end of the Panama
Canal, since named Hyporhamphus hildebrandi, is very slender, with
the ventral fins almost midway between the hypural joint and the
gill opening and with the beak considerably longer than the rest
of the head. A group from Mazatlan, Mexico, is somewhat stouter,
with the ventrals much nearer the gill opening than the hypural
joint and with the beak as long as the rest of the head. Those of
a group from Trinidad are slender, with the ventrals almost midway
between the hypural joint and the gill opening and with the beak
just longer than the rest of the head. Those of a fourth group
were collected in Bermuda, by T. H. Bean. They are much stouter,
with the ventrals much nearer the hypural joint than the gill opening
and with the beak much longer than the rest of the head. There-
fore it appears that four distinct species are labeled Hyporhamphus
roberti.

NO. 326 ;n/S?IY Gr TiJE

FES 26 i

42 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XX

Four groups labeled Hemirhamphus brasiliensis, from Key West,
Bermuda, Jamaica, and the Atlantic end of the Panama Canal,
differ slightly in the relative length of the beak, in the position of
the ventrals and in stoutness of the body. Some of these local
variations might possibly be proved to be specific by combination
with other characters.

Groups of specimens labeled Hyporhamphus unifasciatus from
many localities in the Atlantic and Pacific oceans show a much
greater general uniformity of characters. In each ocean they show
several local groups but the general average is quite uniform, except
that the specimens from the Pacific seem to have somewhat longer
beaks. It seems likely that several local races might be determined
from either the Atlantic or the Pacific and that the groups from the
two oceans might be proved to be distinct species.

The family Hemirhamphidae, as at present understood, is readily
divisible into two major sections. In one section, containing only
a single genus and species, Chnodorus atherinoides Goode and Bean,
the lower jaw is not at all prolonged, not extending as far forward
as the tip of the upper. The teeth are also much larger than in
other members of the family and are arranged in two distinct rows
instead of in more or less broad bands. The remaining genera and
species resemble each other in having at least some indication of
a prolongation of the lower jaw and in having the teeth irregularly
arranged in bands in the jaws.

Hemirhamphidae, other than Chriodorus, may be readily divided
into two groups that show a surprising similarity of parallel develop-
ment. In each group there is at least one genus with practically
no beak. There are genera with beaks of varying lengths and with
various degrees of elongation of the body and, in each group, there
is a genus with bands of teeth extending to a considerable distance
from the rim of the mouth, along the upper surface of the beak.

A group of these fishes, which we may call the river halfbeaks,
has the caudal fin rounded, truncate, or sigmoid with an upper or
lower point, never deeply forked; the dorsal and anal fins rounded,
never with a concave distal margin; the anal fin of the male usually
modified; and the nasal barbels rather long, tapering, and not
concealed in the nasal fossa. Fishes of this group are frequently
(perhaps usually) viviparous. They are found in fresh waters of
the East Indian and central Pacific regions, where they sometimes
enter brackish estuaries and, perhaps, go down into the sea.

1933 FAMILY HEMIRHAMPHIDAE — WEED 43

In the other group of halfbeaks the tail is forked, usually very
deeply, with the lower lobe generally much larger than the upper;
the distal margin of dorsal and anal fins usually more or less concave,
often falcate; the anal fin of the male never modified and the species
probably never viviparous; the nasal barbels short, expanded at
tip and usually more or less concealed in a nasal fossa. These
fishes are inhabitants mainly of shore waters in all warm seas. A
few of them are pelagic, some enter rivers and a very few are per-
manent residents of fresh-water lakes.

Both of these groups have been divided into genera and sub-
genera, often on trivial characters and all too often without any
direct indication of the structural differences on which the separation
is based. In an endeavor to furnish a valid basis for an understanding
of the group I have selected twelve groups of characters that have
been used in some or all of the descriptions and have tried to fit
the published diagnoses into the list. These characters, arranged
somewhat in the order of their taxonomic value, are: (1) nasal barbel;
(2) anal fin modified or not; (3) form of caudal fin; (4) length of
beak; (5) airbladder simple or cellular; (6) position of ventrals;
(7) form of body; (8) length and position of pectoral fins; (9) position
and form of dorsal fin; (10) form and arrangement of the teeth;

(11) form of the premaxillary plate (properly called maxillary-
premaxillary plate as it is formed by the fusion of both elements);

(12) size of the scales. Each of these characters, alone or in connec-
tion with others, has been used in defining some group of Hemi-
rhamphidae but there are very few generic diagnoses in which even
a majority of them are mentioned.

Two nominal genera have been included in the family because
of the prolonged lower jaw. These are Oxyporhamphus Gill and
Hemiexocoetus Fowler. Gill's name was based on Hemirhamphus
cuspidatus Cuvier and Valenciennes (1846), a name given to two
small (4 inches long) specimens taken by Dussumier from the
stomach of a bonito in the Indian Ocean. The published figure of
this fish (Cuvier and Valenciennes, 19, pi. 557, 1846) strongly
resembles various pictures of Evolantia microptera (Cuvier and
Valenciennes) but the fin count, D. 18, A. 12, is quite different.
Weber and De Beaufort (1922) give the fin counts of the latter
species as D. 13-16, A. 14-16. It is possible, although hardly
probable, that cuspidatus represents an extreme variation of microp-
tera, but it does seem likely that both belong to the same genus, in
which case Gill's name, Oxyporhamphus, would have precedence

Tl ' r O rt ^ »/ r"~ —•»»»
-itth/wi ur hii

i.

44 FIELD MUSEUM OF NATURAL HISTORY— ZOOLOGY, VOL. XX

over Evolantia. On the other hand, Oxyporhamphus may be the
young of some species of Parexocoetus. In the absence of specimens,
this question must remain theoretical. This much is sure, with the
information at hand, there is about as much reason for including
Oxyporhamphus in the Exocoetidae as in the Hemirhamphidae.

The genus Hemiexocoetus was based by Fowler on a small
specimen with a very long beak. This has since been shown by
Nichols and Breder (1928) and accepted by Fowler (1929) to be a
young individual of a species of Fodiator Jordan and Meek.

DEFINITIONS

In order to avoid confusion and wrong interpretations it is
necessary to give some definition and discussion of certain terms.

Nostrils and nasal barbels. — The nostrils are openings in a mem-
brane that is stretched over a rather large nasal fossa on each side
of the head a short distance in front of the eye. In some forms
(or in some states of preservation?) this membrane is stretched level
with the surface of the head, while in others it is more or less depressed
into the nasal fossa, in some cases apparently along its bottom.
The nasal barbel arises as a fleshy structure about the center of this
membrane. In the river halfbeaks the barbel is approximately as
long as the pupil, tapering, and extending well above the surface of
the head ("Bin Nasalbartel jederseits," Mohr, 1926, p. 240). In
the genera of the marine group, the barbel is shorter and thickened
at the tip, being mushroom-shaped, leaf-shaped, spatulate or club-
shaped, its basal membrane usually well toward the bottom of the
nasal fossa and the barbel not extending much above the surface of
the head ("Keine Nasalbartel," Mohr, 1926, p. 240). The reason
for considering this structure a barbel in one case and not in the other
does not seem quite clear. It seems to be the same type of structure
in either case and stands near the center of the nasal fossa, directly
behind the opening of the nostril.

Anal fin modified. — Many (perhaps most or all?) of the river
halfbeaks have certain rays of the anal fin more or less enlarged
and variously modified in adult males. The change may be any-
thing from a slight lengthening to a very complex structure. In
some species of Zenarchopterus one (or more) of the rays is enormously
thickened, greatly elongated and provided with secondary rays that
make it look like a portion of the shaft and vane of a feather.

Beak. — The beak, as here used, is that part of the lower jaw in
front of the tip of the premaxillary plate. It is to be measured from

1933 FAMILY HEMIRHAMPHIDAE — WEED 45

the forward end of the depression that forms the mouth, just inside
the band of teeth on the rim of the mouth.

Airbladder. — The structure of the airbladder is somewhat difficult
to observe as the walls are thin and the cells sometimes quite large.

Premaxillary plate. — The "premaxillary" or "intermaxillary"
plate should, properly, be called the maxillary-premaxillary plate
as it is formed by the nearly complete fusion of those elements. In
form it is more or less triangular, with a wing projecting from each
lateral corner. The triangular portion lies on top of the head, forming
the roof of the mouth, while the wings bend down on the sides of
the head and are attached to the lower jaw in such a manner that
depressing the lower jaw raises the plate like a door. The shape and
proportions of this plate are used constantly in generic and specific
diagnoses in this family but, so far as I know, no one has described
just how they are applied. When the mouth is open this plate is
tilted upward and there is a fairly definite joint between it and the
top of the head that forms one logical starting point for a measure-
ment. The tip of the plate forms another definite place for measuring
and the length is easily obtained. Judging from measurements
given in the literature, various authors have used different ways of
determining the width of this structure. If we measure its entire
length, it would seem logical to measure, also, its entire width,
across the wings, a measurement comparatively easy to make with
dividers. Apparently, many authors have tried to measure the width
just in front of the wings. This would also be easy to do if there
were a definite angle between the wing and the edge of the flat
portion of the plate. In very many cases the two come together on
a very gentle curve and it is not possible for one person to tell just
what point another may have considered the junction. The result
has been very serious discrepancies between descriptions of the same
fish by different authors. In original descriptions in this paper,
the measurement across the wings has been used. The parts that
have been compiled from the literature retain the authors' state-
ments and some differences will be found.

Scales. — The size of the scales in this group is not easy to express
and yet it has a definite relationship to the various sections.
Euleptorhamphus has rather large scales, yet, owing to its very long
body, there are about a hundred or more in a lateral series. In
Eulepidorhamphus, which has a body about as elongate as is usual
in the marine halfbeaks, the scales may be considered to be very
small with a count of 90 to 100 or more. They are also small in

46 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XX

Nomorhamphus, which has from 50 to 60 scales in a lateral series.
The scales in the entire group are often closely and irregularly
imbricated, which makes it difficult to count them exactly, especially
when many of them have been lost, as is often the case.

Form of body. — There are three rather distinct types of form of
body in this group. The river half beaks seem to have a tendency
for the body to be somewhat wedge-shaped in cross section, with a
broad back and the straight sides slanting toward a narrow belly.
The body is usually comparatively short. The marine forms, except
Euleptorhamphus, have the body more elongate, with a distinct
tendency for an approach to a rectangular cross section. The back
is fairly broad, with a tendency to be flattish or gently rounded,
the sides are often noticeably flattened, and the belly is broad and
quite flat, with its edges defined by the lateral line; that is, defined
by a ridge of scales that are much less deciduous than those on the
rest of the side. Euleptorhamphus has the body very long, narrow
and flat-sided, so it is properly described as band-like.

KEY TO GENERA OF HEMIRHAMPHIDAE

This key and the generic descriptions following have been
compiled mostly from the literature. They are submitted in the hope
that they may help in bringing some order out of this chaotic group.
Because of the manner in which the key is prepared there will be
found many errors and discrepancies that cannot be corrected
without an amount of work for which the present author has neither
the time nor the material. No attempt has been made to distinguish
between genera and subgenera.

1. Lower jaw not elongated, not extending as far forward as tip of snout.

Chriodorus

Lower jaw more or less elongated, extending at least as far forward as tip
of snout 2

2. Nasal barbel evident, extending above surface of head, about as long as pupil,

not concealed in nasal fossa, tapering; caudal fin usually convex; anal fin

of males usually modified; mostly fresh-water species 3

Nasal barbel more or less concealed in nasal fossa, shorter than pupil, expanded
at tip; caudal fin more or less deeply forked, when deeply forked with
lower lobe much longer than upper; anal fin of males not modified;
mostly marine species 8

3. Band of teeth of lower jaw extending some distance forward beyond tip of

premaxillary plate Hemirhamphodon

Band of teeth of lower jaw not extending on beak beyond rim of mouth .... 4

4. Beak very short, barely extending beyond premaxillary plate. .Nomorhamphus
Beak longer, at least half as long as rest of head 5

5. Beak usually shorter than rest of head; dorsal shorter than anal, inserted

behind anal; ventrals inserted about midway between gill opening and
hypural joint; scales moderate 6

1933 FAMILY HEMIRHAMPHIDAE— WEED 47

Beak usually longer than rest of head; dorsal longer than anal, inserted
before anal; ventrals inserted nearer hypural joint than gill opening; scales
large 7

6. Premaxillary plate longer than broad Dermogenys

Premaxillary plate much longer than broad Rhamphodermogenys

7. Premaxillary plate varying from broader than long to longer than broad.

Zenar chapter us
Premaxillary plate twice as long as broad Labidorhamphus

8. Teeth of lower jaw extending on beak some distance beyond tip of pre-

maxillary plate Odontorhamphus gen. nov.

Teeth of lower jaw not extending on beak beyond rim of mouth 9

9. Beak very short, barely extending beyond premaxillary plate .... Arrhamphus
Beak longer, at least half as long as rest of head 10

10. Body very long, strongly compressed, bandlike, back subcarinate; pectoral

fins very long; scales moderate Euleptorhamphus

Body not extremely elongate, more or less compressed, not bandlike, back
more or less rounded or flattened; pectoral fins moderate, usually about
two-thirds as long as head without beak; scales various 11

11. Body much deeper than wide Farhians

Body not extremely compressed, its width usually about two-thirds of the

depth 12

12. Dorsal inserted in advance of anal, its base longer than that of anal; ventral

nearly as long as pectoral, the inner ray produced Ardeapiscis

Dorsal insertion nearly over that of anal, its base about as long as that of
anal; ventral about half as long as pectoral, the inner ray not produced . . 13

13. Ventral insertion nearer hypural joint than gill opening; airbladder cellular;

teeth simple 14

Ventral insertion about midway between hypural joint and gill opening,
frequently nearer gill opening; airbladder simple; teeth tricuspid 15

14. Beak not much longer than rest of head Hemirhamphus

Beak much longer than rest of head Rhynchorhamphus

15. Scales moderate to large Hyporhamphus

Scales small Eulepidorhamphus

The group to which Whitley (1931) gave the name Reporhamphus
is not included in the above key because the four species referred to
it, as described in the literature, exhibit a bewildering array of con-
tradictory characters. The type of the genus is Hemirhamphus
australis Gunther, a marine form with rather deeply forked caudal,
ventrals inserted nearer hypural joint than gill opening, falcate
dorsal fin, tricuspid teeth and airbladder not described. The
other three are river forms with slightly forked tail, ventrals inserted
nearer gill opening than hypural joint (in ardelio and regularis) or
nearer hypural joint than gill opening (in melanochir), dorsal not
falcate, tricuspid teeth (in melanochir, not described in the other
three), and airbladder simple (in melanochir, not described in the
other three). This genus, then, comprises two species that, on the
basis of generally accepted descriptions, should be placed in

48 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XX

Hemirhamphus and two that, on a similar basis, should be in
Hyporhamphus. Whitley, himself, published no description of his
genus, simply designating a genotype and indicating the species
that he considered should be included in the group. In the absence
of specimens of any members of the group and in consideration of
the very meager descriptions that have been published, it is im-
possible to find any grounds for combining these four species, except
that they all occur in Australian waters (see p. 59).

DESCRIPTIONS
Chriodorus Goode and Bean (1882).

Type Chriodorus atherinoides Goode and Bean 1882.

(1) Nasal barbel leaf-shaped, nearly concealed in nasal fossa,
distal portion almost reniform in outline (specimens). (2) Anal fin
not modified in males. (3) Caudal fin forked, lower lobe larger
than upper. (4) No beak, lower jaw shorter than upper. (5)
Airbladder cellular (specimen). (6) Ventrals inserted nearer gill
opening than hypural joint, small (specimens). (7) Body elongate,
moderately compressed, back rather narrow, rounded (specimens).
(8) Pectoral moderate, inserted high on side (specimens). (9)
Dorsal fin about opposite anal and about equal to it in length, last
rays not produced, distal margin concave (specimens). (10) Teeth
in two rows in each jaw, large for the family, tricuspid. (11) Pre-
maxillary plate very short. (12) Scales large.

Hemirhamphodon Bleeker (1866).

Type Hemirhamphus phaiosoma Bleeker 1852.

(1) Nasal barbel tapering, not expanded at tip, not concealed in
nasal fossa. (2) Anal fin modified in male. (3) Caudal fin rounded
or acuminate. (4) Beak about as long as rest of head. (5) Airbladder
not described. (6) Ventrals inserted about midway between gill
opening and hypural joint, longer or shorter than pectorals. (7)
Body elongate, compressed, back rounded. (8) Pectorals rather
small, inserted low on side. (9) Dorsal fin inserted before or behind
ventrals, much longer than anal, distal margin rounded. (10)
Teeth conical, in bands in both jaws, a band of teeth on each side
of upper surface of beak, represented in the figure of Weber and
De Beaufort (1922) as in three rows in each band and extending to
the tip of the bone. (11) Premaxillary plate longer than broad.
(12) Scales small, various in size and arrangement, enlarged on nape,
much enlarged in parietal region.

1933 FAMILY HEMIRHAMPHIDAE — WEED 49

Nomorhamphus Weber and De Beaufort (1922).

Type Nomorhamphus celebensis Weber and De Beaufort 1922.

(1) Nasal barbel tapering, not concealed in nasal fossa. (2) Anal
fin subtruncate, not otherwise described, perhaps modified in male.
(3) Caudal fin subtruncate, slightly rounded. (4) Beak barely extend-
ing beyond premaxillary plate. (5) Airbladder not described. (6)
Ventrals nearer to gill opening than to hypural joint, moderate.

(7) Body rather short for the group, compressed, back probably
rounded, fairly broad. (8) Pectoral moderate, inserted high on side.
(9) Dorsal inserted behind anal origin, shorter than anal, its distal
margin convex. (10) Teeth simple, in bands in both jaws. (11)
Premaxillary plate as long as or longer than broad. (12) Scales
moderate to small.

Dermogenys van Hasselt (1823).

Type Dermogenys pusillus van Hasselt 1823.

(1) Nasal barbel tapering, not concealed in nasal fossa. (2) Anal
fin not modified in male (Weber and De Beaufort, 1922), distal
margin convex. (3) Caudal fin rounded. (4) Beak usually shorter
than rest of head. (5) Airbladder not described. (6) Ventral fins
inserted about midway between gill opening and hypural joint,
or a little nearer to either. (7) Body moderately elongate, some-
what compressed, back not described, probably somewhat flattened.

(8) Pectorals moderate, inserted high on side. (9) Dorsal fin inserted
behind anal, its base shorter than anal base, last ray not produced,
distal margin convex. (10) Teeth simple, in bands in both jaws,
none on beak beyond rim of mouth. (11) Premaxillary plate
longer than broad. (12) Scales moderate.

I am indebted to Dr. C. L. Hubbs for the following quotation
from Algemeene Konst- en Letter-bode (Haarlem), 2, 1823, giving
the original description of the genus Dermogenys van Hasselt:
"Een Hemirhamphus, welke in de beken rondom Buitenzorg en bijna
overal op Java gevonden wordt, is reeds gedurende het leven van
Kuhl, van dit geslacht, door ons afgescheiden onder den naam
Dermogenys, en wel om de membraneuse uitbreitling, die zich aan
iedere zijde de beide maxillen bevindt; de species draagt den naam
van Pusillus, wordt door de Javanen joelong — joelong genoemd en
is door ons afgebeeld geworden."

The description above is contained in a portion of a letter from
van Hasselt to C. J. Temminck. Apparently the same portion of
the letter was translated to French and printed the next year in the

50 FIELD MUSEUM OF NATURAL HISTORY— ZOOLOGY, VOL. XX

"Bulletin des sciences naturelles." It is not clear from this paragraph
or from the translation whether Kuhl and van Hasselt actually
published the genus and species above described or only prepared
it for publication.

Rhamphodermogenys Fowler and Bean (1922).

Type Dermogenys (Rhamphodermogenys} bakeri Fowler and Bean
1922.

(1) Nasal barbel not described, probably not different from
Dermogenys. (2) Anal fin not described. The figure suggests that
it is considerably modified in the male. (3) Distal margin of caudal
fin sigmoid, with a small point at top. (4) Beak shorter than rest
of head. (5) Airbladder not described. (6) Ventral fins inserted
nearer to gill opening than to hypural joint, small. (7) Body moder-
ately elongate, back broad, sides flattened and converging toward
the narrow belly. (8) Pectoral moderate, inserted high on side.
(9) Dorsal fin inserted behind front of anal, its base shorter than anal
base, distal margin convex. (10) Teeth conical, in bands in each jaw,
none on beak in front of rim of mouth. (11) Premaxillary much
longer than broad. (12) Scales large.

Described as a subgenus of Dermogenys, based on the great length
of the premaxillary plate and the forward position of the ventral
fins. If the form of the anal fin as shown by Fowler and Bean is a
sexual character and there are no intergradations with the condition
mentioned for Dermogenys by Weber and De Beaufort (1922, p.
137), this form probably should stand as a distinct genus.

Zenarchopterus Gill (1863).

Type Hemirhamphus dispar Cuvier and Valenciennes 1846.

(1) Nasal barbel tapering, not concealed in nasal fossa. (2)
Anal fin modified in the male. (3) Caudal fin variously rounded,
sometimes with a secondary point below. (4) Beak various, fre-
quently longer than the rest of the head. (5) Airbladder simple
(Cuvier and Valenciennes). (6) Ventrals inserted nearer to the
hypural joint than to the gill opening, small. (7) Body elongate,
more or less compressed, back flattened. (8) Pectorals moderate,
inserted high on side. (9) Dorsal inserted before anal, longer than
anal, the distal margin convex, often with some of the rays produced
in the males. (10) Teeth conical, those of the mandible confined
to the rim of the mouth. (11) Premaxillary plate varying from
broader than long to longer than broad. (12) Scales large.

1933 FAMILY HEMIRHAMPHIDAE — WEED 51

Labidorhamphus Fowler (1905).

Type Hemirhamphus amblyurus Bleeker 1849.

(1) Nasal barbel tapering, not concealed in nasal fossa. (2) Anal
fin modified in males. (3) Distal margin of caudal fin somewhat
sigmoid, not forked, lower point not produced. (4) Beak much longer
than rest of head. (5) Airbladder not described. (6) Ventrals
inserted much nearer to hypural joint than to gill opening. (7)
Body elongate, somewhat compressed, back rounded. (8) Pectoral
moderate, inserted high on side. (9) Dorsal fin inserted in front of
anal, base longer than that of anal, distal margin somewhat sigmoid
in outline, the last rays not produced. (10) Teeth conical, the
mandibular band on rim of mouth. (11) Premaxillary plate twice
as long as broad. (12) Scales large.

Described as a subgenus of Zenar chapter us, apparently on the
basis of the single character of the very long premaxillary plate.

Odontorhamphus gen. nov.

Type Odontorhamphus chancellor! sp. nov.

(1) Nasal barbel very broadly spatulate, the expanded portion
irregularly subtriangular, reflexed, nearly covering the nasal fossa
but not rising appreciably above it. (2) Anal fin not modified in
the male. (3) Caudal fin moderately forked, the lower lobe much
longer and wider than the upper. (4) Beak as long as rest of head.
(5) Airbladder cellular. (6) Ventrals inserted nearer to hypural joint
than to gill opening, the fin about half as long as pectoral. (7) Body
elongate, compressed, the sides and belly flat, the back gently
rounded. (8) Pectoral moderate, inserted high on side. (9) Dorsal
fin inserted very slightly in advance of anal; its base about as long
as that of anal; the distal margin concave, the first rays somewhat
elevated, not falcate, the last ray somewhat produced. (10) Teeth
simple, in bands on premaxillary plate and on upper surface of beak
nearly to tip. The band of teeth on premaxillary plate extends down
on the lateral wings about to the level of the mandible, with a few
scattered teeth on the edge of the plate below the band on one side.
On the lateral wings the band of teeth at its lower end occupies a
considerable area on the outside of the plate. Careful dissection
would be needed in order to show whether these teeth are on maxillary
or premaxillary. There is a band of teeth on each side of mandible
from near the tip of beak to just inside the corner of the mouth.
These bands are widened so that they cover the entire upper surface
of the mandible from the rim of the mouth outward. They nearly

52 FIELD MUSEUM OF NATURAL HISTORY— ZOOLOGY, VOL. XX

meet in front of the mouth. Just in front of the mouth each band
is narrowed abruptly and continues forward along each side of the
beak, leaving a narrow toothless space the entire length of the center
of the upper surface. There are no teeth on the sides or the lower sur-
face of the beak. The teetty are conical, with a varying degree of slight
curvature, variously pointed; some taper evenly from base to tip,
while others are abruptly pointed at the tip, suggesting the possibility
of descent from, or variation toward, a tricuspid condition. (11)
Premaxillary plate (measured from top of head, with the mouth
open) as long as the width across the lateral wings, including the
lateral bands of teeth. (12) Scales moderate, closely imbricated
forward.

FIG. 1. Odontorhamphus chancellors sp. nov. Dorsal view of head.

Odontorhamphus chancellor! sp. nov.

Type from Aitutaki Island, Cook Archipelago, Pacific Ocean.
No. 16,208 Field Museum of Natural History. Length, total 240
mm. (from tip of premaxillary plate to hypural joint 170 mm.).
Collected, 1930, by Philip M. Chancellor.

D. 15; A. 17; P. 11; V. 6; scales about 83, 5 of them beyond the
hypural joint, rather large but closely imbricated, somewhat decidu-
ous and very difficult to count. Gill rakers 27-29 on the lower
arm of the first arch, 9-7 on the upper, total 36, the longest about half
as long as pupil. No clear division between the upper and lower
arms of the gill arch. Branchiostegals 12. Beak .23,1 barely longer
than rest of head. Head from tip of premaxillary plate .22. Entire
head .45 (.32 of total length). Eye .06 (3.7 in head). Length of
premaxillary plate .053 (4.2 in head), its width .047 (4.7 in head).
Interorbital width .05 (4.4 in head), flat or barely concave, its general
surface slightly below upper edge of orbital rim. Depth .11 (2 in
head). Width of body .08 (2.8 in head, 1.4 in depth). Pectoral
.14 (1.6 in head). Ventral .08 (2.8 in head, 1.8 in pectoral). Distance
from ventral base to hypural joint .35, to gill opening .44. Dorsal
origin about one ray ahead of anal, both fins scaled at base in front,

1 These linear measurements are given as decimals of the standard length, tip
of upper jaw to hypural joint.

1933 FAMILY HEMIRHAMPHIDAE — WEED 53

the scales disappearing at about eighth dorsal and ninth anal rays,
forming a sheath in front. Distal margins of both dorsal and anal
fins concave, the first rays moderately elevated, not falcate, the last
rays slightly produced. Caudal fin moderately forked, the lower
lobe much longer and wider than the upper; the shape of the fin
about as in T. C. Roughley's (1916, pi. 4) plate of " Hemirhamphus
intermedius."

The back is broad and rather flat. The lateral line does not seem
to form so well marked a ridge as in some other species, although
it marks the separation between the flat belly and the rather flat
side. In cross section the body would be rather notably subrec-
tangular in outline.

The premaxillary plate, measured from the top of the head to
the sharply pointed tip, with the mouth open, is slightly longer than
the width across the wings. The teeth are in broad bands, so close
to the edge of the plate that the outer row shows from above. The
lateral bands of teeth do not quite meet in front, leaving a triangular
space that is filled by a patch of smaller teeth. A slight groove
separates this central patch of teeth from the lateral bands. On
the lateral wings of the premaxillary plate the bands cross over the
edge to the outer surface, forming, on each side, a comparatively
broad patch that is just above the level of the upper surface of the
mandible when the mouth is fully open. In the type specimen
the lateral wings of the premaxillary plate are not quite symmetrical.
On the left side, the lower end is noticeably truncate, there seems to
be a small supplementary bone in the lip at the angle of the mouth
and there are several teeth on the edge of the plate below the broad
band. On the right side, the lower end of the wing is distinctly
rounded, the patch of teeth does not extend so far down, and
there are no teeth below it. The patch of teeth on the outside of
the lateral wings of the premaxillary plate, mentioned above, does
not occur in any others of the marine halfbeaks available in Field
Museum. Its place is taken by a rather thick and fleshy lip that
completely hides the teeth on the edge of the plate.

The premaxillary plate is scaled over practically its entire surface,
only the edges and the extreme tip being naked.

The plate where the imbrication of the dorsal scales changes is
directly over the gill opening.

Nasal barbel (see p. 44) very broadly spatulate, the expanded
distal portion irregularly subtriangular, reflexed, nearly covering
the nasal fossa but not rising appreciably above it. The form of

54 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XX

the barbel is very strikingly like that of certain woody fungi that
have a short basal stem attached at one edge of a broadly expanded
and rather thin distal portion.

The airbladder is cellular, with the separate cells about as long
as the width of the entire structure.

Described from a single specimen collected by the Philip M.
Chancellor-Field Museum Expedition to Aitutaki Island, Cook
Archipelago, 1930, and now in Field Museum. Chancellori, for Mr.
Philip M. Chancellor.

Arrhamphus Giinther (1866).

Type Arrhamphus sclerolepis Giinther 1866.

(1) Nasal barbel concealed in nasal fossa, form not described.
(2) Anal fin not described, probably not modified in male. (3)
Caudal fin forked, lower lobe much longer and wider than upper.

(4) Beak very short, scarcely projecting beyond premaxillary plate.

(5) Airbladder not described. (6) Ventrals nearer gill opening than
hypural joint, small. (7) Body elongate, compressed, back not
described, probably rounded. (8) Pectoral moderate, inserted high
on body. (9) Dorsal fin inserted about over front of anal, about
as long as anal, last rays not produced, distal margin concave.
(10) Teeth in bands in both jaws, often tricuspid. (11) Premaxillary
plate broader than long. (12) Scales large.

Euleptorhamphus Gill (1859).

Type Euleptorhamphus brevoortii Gill 1859.

(1) Nasal barbel concealed in nasal fossa, stump-shaped, tapering
from a base that is nearly circular in cross section to a blunt top
that is about as wide and three times as long as the base. The
nasal barbel is not recurved (described from No. 4146 F.M.N.H.).
In most or all of the other genera of this group the tip of the nasal
barbel is considerably flattened and the distal portion is recurved
so as to lie over the top of the nasal fossa. (2) Anal fin not modified
in the male. (3) Caudal deeply forked, the lower lobe much wider
and longer than the upper. (4) Beak very long, much longer than
rest of head. (5) Airbladder simple (Cuvier and Valenciennes, 1846,
"Hemirhamphus macrorhynchus."). (6) Ventrals inserted much
nearer to hypural joint than to gill opening, the fins small, the inner
ray somewhat produced. (7) Body very long and compressed,
band-like, the back thin and subcarinate. (8) Pectoral fins very long,
inserted high on the body, very similar in structure to the pectorals

1933 FAMILY HEMIRHAMPHIDAE — WEED 55

of Exocoetidae. (9) Dorsal fin inserted slightly in advance of anal,
slightly longer than anal, last ray not produced, distal margin more
or less convex. (10) Teeth in few series in each jaw, subconical above
and tricuspid below. In a specimen, No. 4146 F.M.N.H., collected
in the Hawaiian Islands, the teeth are strongly tricuspid in both
jaws. The bands of teeth on the sides of the premaxillary plate
do not quite meet in front, leaving a small interspace that is occupied
by a triangular patch of teeth in other genera, but is toothless, at
least in this specimen. The teeth on the lateral wings of the pre-
maxillary plate do not extend on the outer surface of the plate but
are seen on its forward edge down almost to the level of the top of
the mandible when the mouth is fully opened. It is not possible to
determine whether the teeth on the edge of the wings are concealed
by fleshy lips as in Hemirhamphus and Hyporhamphus. In the
lower jaw the teeth are on the rim of the mouth as in most other
genera of the family. (11) Premaxillary plate about as long as broad
(about two-thirds as long as broad, when measured across the wings,
in specimen No. 4146 F.M.N.H.). (12) Scales moderate.

Farhians Whitley (1930).

Type Hemirhamphus commersonii Cuvier 1817, which is a
redescription and renaming of Esox far Forskal 1775.

(1) Nasal barbel concealed in nasal fossa. In three specimens
in Field Museum the nasal barbel is leaf -shaped, with a thick base
and a reflexed, subtriangular distal portion that covers less than half
of the nasal fossa. (2) Anal fin not modified in males. (3) Caudal
deeply forked, the lower lobe much longer and wider than the upper.
(4) Beak longer than rest of head. (5) Airbladder cellular. (6)
Ventrals inserted nearer to hypural joint than to gill opening.
(7) Body much deeper than wide, back rather broad, rounded. (8)
Pectoral moderate, inserted high on side. (9) Dorsal fin inserted in
advance of anal and longer than anal, distal margin concave, last
ray slightly produced. (10) Teeth tricuspid, in bands on the jaws,
none on the beak. (11) Premaxillary plate broader than long. In
two specimens in Field Museum the premaxillary plate, as measured
across the lateral wings is almost exactly twice as broad as long.
In a very large specimen, 300 mm. in standard length, the width,
measured in the same way, is about two and one-half times the length.
(12) Scales large.

In speaking of the specific name of the genotype, Whitley (1930)
says: "This species is generally known as Hemirhamphus far
(Forskal), but the name far was only used in a vernacular sense,

56 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XX

until 1837, when Riippell latinized it, but H. commersonii Cuvier
has priority." With this view I am unable to agree. Forskal was
not consistent in the form in which he listed the specific names he
published. However, he seems to have had a certain general scheme
that was followed with considerable regularity. Generic names
were printed in large capitals, specific names were in small capitals,
variety names had the initial letter a large capital, followed by
small capitals. The generic name might or might not be repeated
before the variety name. For example, his No. 42 is Perca summana,
with varieties fusco-guttata and areolata. Of these two, fusco-guttata
is treated in exactly the same manner as far in Esox (No. 98), while
areolata, which follows it, is treated in the same manner as belone,
under Esox. On page XI, in the list of "Pisces descripti," we find
Perca with species h, summana, variety /3, fusco-guttata, and variety
5, areolata. On page XIII of the same list we find Esox with species
marginatus, variety /3, far, and variety 7, belone. It seems clear
that he considered far and belone varieties of Esox marginatus, just
as he considered fusco-guttata and areolata varieties of Perca sum-
mana. I cannot see any reason for considering that far was used in
a vernacular sense by Forskal. If it was so used, then fusco-guttata,
under No. 42 of the fishes, was also used as a vernacular name.

Ardeapiscis Whitley (1931).

Type Hemirhamphus welsbyi Ogilby 1908.

(1) Nasal barbel concealed in the nasal fossa. (2) Anal fin not
modified in the male (?). (3) Caudal fin very deeply forked, the
lower lobe not described, probably larger than upper. (4) Beak a
little longer than rest of head. (5) Airbladder not described. (6)
Ventral insertion much nearer hypural joint than gill opening, the
fin nearly as long as the pectoral. (7) Body elongate, compressed,
back not described, probably rounded. (8) Pectoral moderate,
probably inserted high on body. (9) Dorsal fin inserted in front of
anal, the dorsal base longer than anal base, last ray produced,
distal margin not described. (10) Teeth not on beak, form not
described. (11) Premaxillary plate wider than long. (12) Scales large.

Hemirhamphus Cuvier (1817).

Type Esox brasiliensis Linnaeus 1758.

(1) Nasal barbel concealed in nasal fossa, the reflexed tip of
expanded distal portion covering a portion of nasal fossa. (2)
Anal fin not modified in the male, more or less falcate. (3) Caudal

1933 FAMILY HEMIRHAMPHIDAE — WEED 57

fin deeply forked, the lower lobe wider and longer than the upper.
(4) Beak usually not shorter than rest of head. (5) Airbladder
cellular. (6) Ventral insertion nearer to hypural joint than to gill
opening. (7) Body rather robust, deeper than wide, sides nearly
vertical and parallel, back rounded, more or less. (8) Pectorals
moderate, inserted high on side. (9) Dorsal fin inserted about over
front of anal, the bases of the two about equal in length, the last
ray somewhat produced, distal margin concave, more or less falcate.
(10) Teeth simple, in bands on premaxillary plate and on the rim
of mouth on lower jaw, none on beak. (11) Premaxillary plate
about as long as broad. (12) Scales moderate.

It is likely that valid specific differences may be found in the
form of the nasal fossa and the nasal barbel in this genus as well
as in Hyporhamphus. In some forms examined in this connection
the nasal fossa is hardly deeper or more pronounced than it is in
Zenarchopterus. In that case, the barbel seems to be usually quite
small. In other forms the nasal fossa may be very deep and con-
tinued back under the face bones to a considerable distance. In
such cases the nasal barbel seems to be larger, with its expanded
tip more definitely reflexed and roofing over the rear portion of the
fossa. It seems evident that the reflexed portion of the barbel can
be straightened up so that it stands well above the top of the head.

Rhynchorhamphus Fowler (1928).

Type Hemirhamphus georgii Cuvier and Valenciennes 1846.

(1) Nasal barbel concealed in nasal fossa. (2) Anal fin not
modified in the male. (3) Caudal not deeply forked, lower lobe
somewhat longer than upper. (4) Beak very long, much longer
than rest of head. (5) Airbladder not described. (6) Ventral
fins inserted nearer to hypural joint than to gill opening, about half
as long as pectorals. (7) Body long, slender, somewhat compressed,
back not described, probably rounded and fairly broad. (8) Pectoral
moderate, inserted high on side. (9) Dorsal origin slightly in advance
of that of anal, last ray not produced, distal margin convex anteriorly,
straight posteriorly. (10) Teeth simple, in bands in jaws, none on
beak. (11) Premaxillary plate longer than broad. (12) Scales large.

Described as a subgenus of Hemirhamphus and apparently con-
sidered of equal rank with the subgenera Hyporhamphus and
Hemirhamphus, although nearer to the latter. Fowler based the
separation of this form from the subgenus Hemirhamphus on the
great length of the beak and of the premaxillary plate. It seems

58 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XX

that the form of the dorsal and caudal fins should also be considered
generic characters in this case.

Hyporhamphus Gill (1859).

Type Hyporhamphus tricuspidatus Gill 1859.

(1) Nasal barbel concealed in nasal fossa (see p. 57). (2) Anal
fin not modified in the male. (3) Caudal forked, the lower lobe
longer than the upper. (4) Beak usually not longer than the rest
of the head. (5) Airbladder simple. (6) Ventrals inserted about
midway between gill opening and base of caudal, small. (7) Body
elongate, compressed, sides not notably flattened, back fairly
broad, rounded. (8) Pectorals moderate, inserted high on side.

(9) Dorsal fin inserted about over anal and not differing greatly
from it in length, last ray not produced, the distal margin concave.

(10) Teeth tricuspid, in bands, none on beak. (11) Premaxillary
plate not notably elongate, varying from a little longer than broad
to a little broader than long. (12) Scales moderate to large.

Eulepidorhamphus Fowler (1919).

Type Hemirhamphus sajori Temminck and Schlegel 1846.

(1) Nasal barbel concealed in nasal fossa. (2) Anal fin probably
not modified in male. (3) Caudal fin little forked, lower lobe little,
if any, longer or wider than upper. (4) Beak shorter than rest of head.
(5) Airbladder not described. (6) Ventral fins about midway between
gill opening and hypural joint. (7) Body elongate, compressed,
sides probably flattened, back probably rounded. (8) Pectoral
rather small, inserted high on side. (9) Dorsal fin inserted directly
over anal, about as long as anal, last ray not produced, distal margin
straight. (10) Teeth in jaws small and tricuspid, in villiform bands,
none on beak. (11) Premaxillary plate not described. (12) Scales
very small.

Described by Fowler as a subgenus of Hyporhamphus, based on
the small scales. Jordan and Hubbs (1925) redescribed Hypor-
hamphus kurumeus Jordan and Starks (1903), giving the number
of scales as about 90 instead of 70 as in the original description.
This brings the species very close to Eulepidorhamphus, as is also
indicated by the original figure of the species. The tail of kurumeus
is indicated as being somewhat more deeply forked than that of
sajori but with the lobes approximately equal, the lower being
one-third the diameter of the eye longer than the upper. A more
complete study might well indicate that sajori and kurumeus are

1933 FAMILY HEMIRHAMPHIDAE — WEED 59

congeneric and that they form as well differentiated a group as is
found in Hemirhamphus or Hyporhamphus. The outline of generic
characters above is taken from the excellent figure by Kumataro
I to and the accompanying text (1931). Some of the species included
by Whitley (1931) in his proposed genus Reporhamphus seem, except
for the larger scales, quite close to Eulepidorhamphus.

Reporhamphus Whitley (1931).

Type Hemirhamphus australis Steindachner 1866.

(1) Nasal barbel concealed in nasal fossa. (2) Anal fin not
modified in male. (3) Caudal fin moderately forked in R. australis,
little forked in the other three; the lower lobe distinctly larger in
R. australis, the lobes nearly equal in the other three. (4) Beak
about as long as rest of head. (5) Airbladder simple in R. melanochir
(Cuvier and Valenciennes 1846), not described in the other three.
(6) Ventral fins nearer hypural joint than gill opening in R. australis
and R. melanochir, nearer gill opening than hypural joint in R.
ardelio Whitley 1931 and R. regularis Giinther 1866. (7) Moderately
elongate, compressed, very flat-sided, almost rectangular in cross
section. (8) Pectoral fin moderate, inserted high on side. (9) Dorsal
fin inserted about over anal and about the same length; last ray
not produced; distal margin deeply concave, falcate in R. australis,
moderately concave, not falcate in R. ardelio, nearly straight in R.
melanochir, not described in R. regularis. (10) Teeth tricuspid in
R. australis and R. melanochir, not described in the other two.
(11) Premaxillary plate wider than long. (12) Scales moderate.

It is difficult, from the literature and without the possibility of
examining specimens, to determine the basis on which this genus is
founded (see p. 48). Apparently the author started with the premise
that tropical Atlantic and southern Pacific halfbeaks cannot be
congeneric. The genotype seems to be separated from the other
three included species by the deeply forked tail. There are no
characters given in the published descriptions that will separate it
clearly from Hyporhamphus Gill or Hemirhamphus Cuvier and
Valenciennes. The other three species have the tail only slightly
forked, with the lobes approximately equal. In this character they
do not seem to be different from Eulepidorhamphus Fowler, which
has, however, very much smaller scales. There is room for a differ-
ence of opinion as to whether Whitley's (1931, p. 314, line 6) state-
ment is a "definite bibliographic reference to a summary of
characters" or whether McCulloch's (1922, p. 40) key is a "summary

60 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XX

of characters . . . which differentiate or distinguish the genus . . .
from other genera" in a sense to satisfy the requirements of Article
25, section c, parts 1 and 2 of the International Rules of Zoological
Nomenclature as amended September 4-9, 1927.

Oxyporhamphus Gill (1863).
Hemiexocoetus Fowler (1901).

These two generic names are placed in the family Hemirham-
phidae in Jordan's (1923) list, but it is extremely probable that both
belong in the Exocoetidae (see p. 43). Therefore, a summary of
the characters assigned to them would add nothing to our under-
standing of the present family.

The characters mentioned in the above summaries are, of course,
not all that might be used in separating groups of halfbeaks. They
were selected because they happen to have been used in most of the
papers dealing with this group, or in the most important papers. The
revision of the genus Zenarchopterus by Mohr (1926) is the best
by far that has come to my attention and it is to be hoped that the
entire family may be worked over as carefully in the near future.

Among other characters that should be mentioned in generic
and specific diagnoses in this group are the number, form, and
arrangement of the gill rakers and the form of the interorbital area
and its relation to the position of the eye. The arrangement of the
lateral bands of teeth on the premaxillary plate seems to offer
possibilities in generic description. In some cases these bands are
broad at the forward end, with a triangular patch of different teeth
between them. In other cases this patch of teeth is absent.
Sometimes these bands are pointed at front, with a toothless
area between them. This character seems to offer much promise
as an aid in generic diagnosis in the family Hemirhamphidae.

Somewhere on the back, in all Hemirhamphidae, there is a change
in the imbrication of the scales. One scale has a free edge all around.
The scales in front of it have the free edge forward instead of to the
rear, while those around it are variously arranged to make up for
the change. The position of this scale and the arrangement
of the ones around it may be of considerable value in the separation
of various groups.

Hyporhamphus hildebrandi Jordan and Evermann (1927).

In "The Marine Fishes of Panama," page 239, Meek and Hilde-
brand (1923), under the name Hyporhamphus roberti (Cuvier and

1933 FAMILY HEMIRHAMPHIDAE — WEED 61

Valenciennes) (1846), described a halfbeak that Jordan and Evermann
(1927, p. 503) on the basis of the previous description consider a new
species, giving it the name Hyporhamphus hildebrandi, "Type
No. ?, Field Museum Nat. Hist., a specimen collected by Meek
and Hildebrand at Toro Point, Fox Bay, Colon." All the specimens
(seven) of this species in Field Museum are labeled "Fox Bay, Colon,
Panama," without any indication as to the precise location in the
bay. Jordan and Evermann plainly state, "This species is based
upon specimens obtained by Meek and Hildebrand at Toro Point,
Fox Bay, Colon, and identified by them with Hemirhamphus roberti
Cuv. & Val., 1846, . . . ." It seems necessary, therefore, to consider
the entire lot in Field Museum, or the entire lot studied by Meek
and Hildebrand, as cotypes, with no type specimen designated.
I, therefore, designate as lectotype the largest specimen of this
species in Field Museum and submit the following description of
this specimen:

FIG. 2. Hyporhamphus hildebrandi. Dorsal view of head.

Type from Fox Bay, Colon, Panama. No. 8260 Field Museum
of Natural History. Length, total 240 mm. (from tip of premaxillary
plate to hypural joint, 105 mm.). Collected March 25, 1911, by
S. E. Meek and S. F. Hildebrand.

D. 15; A. 17; P. 10; V. 6; scales in lateral series about 70 (60
to 70 according to method of counting), rather deciduous and about
half of those in lateral series lost. Gill rakers about 30 on the lower
arm and about 5 on the upper arm of gill arch. There is a fairly clear
division between the upper and lower arms of the gill arch.
Branchiostegals 12. Beak .33 (.22 of total length), 1.65 times rest
of head. Head from tip of premaxillary plate .22. Entire head .53
(.36 of total length). Eye .06 (3.5 in head). Length of premaxillary
plate .048 (4.6 in head), its width .057 (3.9 in head). Interorbital
width .067 (3.3 in head), flat, sharply curved down to the top of the
orbit on each side, the general level notably above the orbits. Depth
.10 (2.2 in head). Width of body .076 (2.9 in head, 1.3 in depth).
Length of pectoral .11 (2 in head). Length of ventral .086 (2.5 in
head, 1.3 in pectoral). Distance from origin of ventral to hypural
joint .38, to gill opening .40. Dorsal origin directly over that of anal .

62 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XX

Anal slightly and dorsal apparently not scaled at base. Distal
margins of both fins concave forward, about straight for more than
half the length at rear, the front rays somewhat produced, not
falcate, the last rays not or slightly produced. Caudal moderately
forked, the lower lobe longer and wider than the upper. The back
broadly rounded. The lateral line does not, perhaps, form quite
so well defined a keel as in some other species but does rather sharply
indicate the edges of the narrow, flat belly. The sides below
the lateral stripe are rather flat and converge towards the belly. The
premaxillary plate, measured from the top of the head, with the
mouth open, is considerably shorter than the width across the wings.
The tip is slightly truncate. The teeth are in a broad band on each
side of the premaxillary plate. Those on the wings do not seem to
extend downward quite so far as in Odontorhamphus and are con-
cealed by fleshy lips. So far as can be observed without careful dis-
section, the lower end of the band is above the level of the top of
the mandible. The lateral bands of teeth on the premaxillary plate
are separated by a rather broad space at the front, with a very small,
triangular patch uniting them at the extreme front. This patch
does not extend as far to the sides as the width of the truncated
portion of the front of the plate. The teeth in both jaws are sharply
tricuspid. The teeth of the lower jaw are in a moderately broad
band which is mainly on top of the mandible around the mouth
depression. The band extends down into the mouth on the sides of
the depression but is sufficiently well developed on top of the man-
dible to be readily seen from the side. At the rear, the band of teeth
tapers to a sharp point that is well behind the corner of the mouth.
The premaxillary plate is scaled nearly to its tip, only the edges of
the upper surface all around being naked.

The nasal barbel is somewhat flattened, between spatulate and
club-shaped. The membrane on which it stands is not sunk so
deeply into the nasal fossa as in most species of this group of Hemi-
rhamphidae. It seems likely that the barbel stands well above the
surface of the head in life.

The silvery color has been entirely destroyed in the preservative,
leaving only a black band along the side. This band is very narrow
forward but is about as wide as the pupil between the dorsal and anal
fins. It is plainly visible from the upper angle of the gill opening to
the base of the caudal, where it widens slightly to form a roundish
spot a little larger than the pupil.

1933 FAMILY HEMIRHAMPHIDAE — WEED 63

REFERENCES
BLEEKER, P.

1849. Bijdrage tot de kennis der ichthyologische fauna van het eiland Madura,
met beschrijving van eenige nieuwe species. Verb. Bat. Genootsch., 22,
pp. 1-16.

1850. Bijdrage tot de kennis der ichthyologische fauna van Midden- en Oost-
Java, met beschrijving van eenige nieuwe species. Verh. Bat. Genootsch.,
23, pp. 1-23.

1852. Bijdrage tot de kennis der ichthyologische fauna van Blitong (Biliton),
met beschrijving van eenige nieuwe soorten van zoetwatervisschen. Nat.
Tidschr. Ned. Ind., 3, pp. 87-100.

1852. Bijdrage tot de kennis der Snoekachtige visschen van den Soenda-
Molukschen Archipel. Verh. Bat. Genootsch., 24, pp. 1-28.

1858. Vierde bijdrage tot de kennis der vischfauna van Biliton. Nat. Tidschr.
Ned. Ind., 15, pp. 219-240.

1866. Revision des Hemirhamphes de 1'Inde archipelagique. Nederl. Tijdschr.

Dierk., 3, pp. 136-170.
1866-72. Atlas ichthyologique des Indes Orientales Neerlandaises, 6, pp. 51-67,

plates.

CUVIER, G.

1817. Le regne animal, 2.

CUVIER, G. and VALENCIENNES, A.

1846. Histoire naturelle des poissons, 19.

FORSKAL, P.

1775. Descriptiones animalium.

FOWLER, H. W.

1901. Description of a New Hemirhamphid. Proc. Acad. Nat. Sci. Phila.,

p. 293.

1905. Some Fishes from Borneo. Proc. Acad. Nat. Sci. Phila., pp. 455-523.
1919. Notes on Synentognathous Fishes. Proc. Acad. Nat. Sci. Phila., 71,

pp. 2-15.

1928. The Fishes of Oceania. Mem. Bernice P. Bishop Mus., 10, pp. 74-79.

1929. Note on Fodiator Jordan and Meek. Copeia, No. 170, p. 3.

FOWLER, H. W. and BEAN, B. A.

1922. Fishes from Formosa and the Philippine Islands. Proc. U. S. Nat.
Mus., 62, art. 2, pp. 1-73.

GILBERT, C. H. and STARKS, E. C.

1904. The Fishes of Panama Bay. Mem. Calif. Acad. Sci., 4, pp. 1-304,
pis. 1-32.

GILL, T.

1859. Description of Hyporhamphus, a New Genus of Fishes Allied to Hemi-
rhamphus Cuvier. Proc. Acad. Nat. Sci. Phila., p. 131.

1859. Description of a Third Genus of Hemirhamphinae. Proc. Acad. Nat.

Sci. Phila., p. 156.
1863. Note on the Genera of Hemirhamphinae. Proc. Acad. Nat. Sci. Phila.,

p. 156.

GOODE, G. B. and BEAN, T. H.

1882. Descriptions of Twenty-five New Species of Fish from the Southern
United States, and Three New Genera, Letharchus, loglossus and Chriodorus.
Proc. U. S. Nat. Mus., 5, pp. 412-437.

64 FIELD MUSEUM OF NATURAL HISTORY— ZOOLOGY, VOL. XX

GtJNTHER, A.

1866. Catalogue of Fishes in the British Museum, 6, pp. 259-277.

HASSELT, J. C. VAN

1823. Uittreksel uit een' brief van den Heer J. C. van Hasselt, aan den Heer
C. J. Temminck, geschreven uit Tjecande, Residentie Bantam, den 29sten
December 1822. Algemeene Konst- en Letter-bode, 2, pp. 130-133
(Dermogenys named and described on p. 131).

1824. Extrait d'une seconde lettre sur les poissons de Java, ecrite par M. van
Hasselt a M. C. J. Temminck. Bull. Sci. Nat., 2, p. 374. Apparently this
is a reprint of the letter published in Algemeene Konst- en Letter-bode,
August, 1823.

HlLDEBRAND, S. F. and SCHROEDER, W. C.

1928. Fishes of Chesapeake Bay. Bull. U. S. Bureau of Fisheries, 43, pp. 1-366.

ITO, KUMATARO

1931. Illustrations of Japanese Aquatic Plants and Animals. Fisheries Society
of Japan. Hyporhamphus sajori (Temminck and Schlegel) is figured, 1,
pi. 21, fig. 4. The plates in this work are by Kumataro ltd and the text seems
to be a compilation by a committee of scientists.

JORDAN, D. S.

1923. A Classification of Fishes, Including Families and Genera as Far as
Known. Stanford Univ. Pub., Univ. Ser., Biol. Sci., 3, No. 2, pp. 79-243, 1-X.

JORDAN, D. S. and EVERMANN, B. W.

1896. The Fishes of North and Middle America. Bull. U. S. Nat. Mus., 47,

part 1, pp. 718-724.
1927. New Genera and Species of North American Fishes. Proc. Calif. Acad.

Sci., (4), 16, pp. 501-507.

JORDAN, D. S. and HUBBS, C. L.

1925. Record of Fishes Obtained by David Starr Jordan in Japan, 1922. Mem.
Carnegie Mus., 10, No. 2, pp. 93-346.

JORDAN, D. S. and SNYDER, J. O.

1901. A Preliminary Check-list of the Fishes of Japan. Annotationes Zool.
Japon., 3, pp. 31-159.

JORDAN, D. S. and STARKS, E. C.

1903. A Review of the Synentognathous Fishes of Japan. Proc. U. S. Nat.
Mus., 26, pp. 525-544.

LINNAEUS, C.

1758. Systema naturae, 1, ed. 10.

McCoY, F.
1887. Prodromus of the Zoology of Victoria, dec. XIV, pp. 133-134, pi. 135.

McCULLOCH, A. R.

1920, 1922. Check List of the Fishes and Fishlike Animals of New South
Wales. Australian Zool., 1 and 2 (part 1 was published in 1, pp. 217-227,
pis. 16 to 18 [1920]: part 2 in 2, pp. 24-68, pis. 4-24 [1921]; part 3 in 2, pp.
86-130, pis. 25-43 '[1922]).

MOHR, E.

1926. Die Gattung Zenar chapter us Gill. Zool. Jahrb. abt. syst., 52, p. 231.

MEEK, S. E. and HILDEBRAND, S. F.

1923. Marine Fishes of Panama. Field Mus. Nat. Hist., Zool. Ser., 15, part 1.

1933 FAMILY HEMIRHAMPHIDAE — WEED 65

NICHOLS, J. T. and BREDER, C.

1928. An Annotated List of the Synentognathi. Zoologica, 8, No. 7, pp. 423-
448, pis.

OGILBY, J. D.

1908. Descriptions of New Queensland Fishes. Proc. Roy. Soc. Queensland,
21, pp. 87-98.

ROUGHLEY, T. C.

1916. Fishes of Australia and Their Technology. Technical Education Ser.,
Technological Museum, Sydney, New South Wales, No. 21.

STEAD, D. G.

1908. Edible Fishes of New South Wales.

STEINDACHNER, F.

1866. Zur Fischfauna von Port Jackson in Australien. Sitzber. Akad. Wiss.
Wien, 53, pp. 424-481.

TEMMINCK, C. J. and SCHLEGEL, H.

1846. In SIEBOLD, P. F. VON: Fauna Japonica, poissons.

WEBER, M. and DE BEAUFORT, L. F.

1922. The Fishes of the Indo-Australian Archipelago, 4, pp. 135-172.

WHITLEY, G. P.

1930. Five New Names for Australian Fishes. Australian Zool., 6, p. 250.

1931. New Names for Australian Fishes. Australian Zool., 6, pp. 310-334.

i;;^ ifli^Y r? T:;^
F^B 261::-;

UNIVFRSITY Of ILLINOIS

UNIVERSITY OF ILLINOIS-URBANA