University of Kansas Publications

Museum of Natural History

Volume 13, No. 1, pp. 1-18

June 1, 1960

Five Natural Hybrid Combinations

in Minnows (Cyprinidae)

BY

FRANK B. CROSS AND W. L. MINCKLEY

University of Kansas

Lawrence

1960

University of Kansas Publications,
Museum of Natural History

Editors: E. Raymond Hall, Chairman, Henry S. Fitch,

Robert W. Wilson

Volume 13, No. 1, pp. 1-18

Published June 1, 1960

University of Kansas

Lawrence, Kansas

PRINTED IN

THE STATE PRINTING PLANT

TOPEKA, KANSAS

1960

28-3424

Five Natural Hybrid Combinations
in Minnows (Cyprinidae)

BY
FRANK B. CROSS AND W. L. MINCKLEY

The hybrid fishes described herein are Chrosomus erythrogaster
(Rafinesque) × Notropis cornutus frontalis (Agassiz), C. erythrogaster
× Semotilus atromaculatus (Mitchill), Campostoma
anomalum plumbeum (Girard) × S. atromaculatus, Gila nigrescens
(Girard) × Rhinichthys cataractae (Valenciennes), and Notropis
venustus venustus (Girard) × Notropis whipplei (Girard). Two
of the combinations have been reported, without descriptions, in
literature (citations below), and Hubbs (1955: Fig. 3) graphically
indicated hybridization between the same genera with which this
paper is concerned, but did not designate the species involved.

All specimens of C. erythrogaster × N. c. frontalis, C. erythrogaster
× S. atromaculatus, C. a. plumbeum × S. atromaculatus,
and N. v. venustus × N. whipplei were taken in a period of
severe drought in Kansas and Arkansas. All were from small,
spring-fed streams that support large populations of fishes. That
the drought of 1953-1956 had pronounced effects on stream habitats
in Kansas has been documented by Minckley and Cross (1959).
Satisfactory sites for spawning may have been few, but an abundance
of adult fishes persisted from earlier, wet years. Unusual
crowding of spawning fishes would increase the opportunity for
fertilization of the eggs of one species by sperm from another
species. We think that the hybrids reported here (excepting G.
nigrescens × R. cataractae) are explainable on the basis of crowding;
we have no information about stream-conditions where the
last-named hybrid was found. Generally, hybridization of fishes
seems most common in areas that have been subject to radical
climatic change in the past 20,000 or fewer years (Hubbs, 1955:18-19),
and in streams that have been altered recently by the
activities of man (Hubbs and Strawn, 1956:342, and others).
Streams from which we report hybrids probably were affected
by overgrazing of their watersheds; overgrazing was unusually
severe in the drought.

Most of the hybrids were recognized as unusual at the time
of capture, and were saved as part of numerically selective samples
[Pg 4]
from the streams (rather than being discovered in the laboratory,
in random samples).

Our measurements were made by methods defined by Hubbs
and Lagler (1958); values are expressed as thousandths of the
larger dimension.

Chrosomus erythrogaster × Notropis cornutus frontalis: KU 3872
(26.7 mm. in standard length) and KU 4170 (46.6 mm.) from
Deep Creek, Riley Co., Kansas, Sec. 23, T. 11S, R. 7E, Dec. 14,
1957, and Apr. 26, 1958, respectively; and KU 4185 (39.3 mm.)
from Bluff Creek, Pottawatomie Co., Kansas, Sec. 15, T. 6S, R.
8E, June 29, 1958. Compared in Table 1 with five specimens of
C. erythrogaster, KU 3914 (39.3 to 47.3 mm., mean 43.0 mm.)
from the same locality and of the same date as KU 3872 (above);
and with five specimens of N. c. frontalis, KU 4184 (41.0 to 46.5
mm., mean 42.5 mm.) from the same locality and of the same date
as KU 4185 (above). This cross has previously been recorded
by Trautman (1957:326, 355) and by Minckley (1959:431).

The head-lengths of the hybrids are greater than in specimens
of like size of C. erythrogaster or N. c. frontalis (Table 1). Hubbs
and Miller (1943:373-374) reported that hybrids of Gila orcutti
× Siphateles mohavensis have larger, more robust heads than either
of the parental species, perhaps because of heterosis. The enlarged
heads in hybrids of C. erythrogaster and N. c. frontalis result primarily
from elongation of the postorbital region, with lesser elongation
of the snout and orbit. The enlarged head affects measurements
obtained for other structures that are parts of the head
(and expressed as proportions of standard length or head-length),
causing a tendency toward N. c. frontalis when the head-part is
divided by standard length, and greater intermediacy or a tendency
toward C. erythrogaster when the head-part is divided by head-length.
In characters in which the parental species differ most
(size of eye, length of upper jaw, and width of gape), the hybrids
are intermediate between the parental species, regardless of
whether the measurements are expressed as proportions of head-length
or standard length; however, tendencies toward one or
the other of the parental species (dependent on the divisor) can
also been seen in these characters. Some experimentally propagated
hybrids show highly variable, and sometimes extreme
characters, rather than intermediacy of meristic and proportional
characters (Hubbs, 1956).

[Pg 5]
Table 1. Comparisons of Three Specimens of Chrosomus erythrogaster
× Notropis cornutus frontalis with Specimens of the Parental Species
(means are above, ranges in parentheses below)

[Pg 6]
In pigmentation, all three of the hybrids are intermediate between
the parental species. The mid-lateral band (which is dark
and discrete in C. erythrogaster, but faint, broad, and diffuse in
N. c. frontalis) is broader and fainter in the hybrids than in Chrosomus,
but is better developed than in N. c. frontalis. The dorsolateral
dark band of C. erythrogaster is present in the hybrids, but
is less distinct than in that species, and less distinct than the mid-lateral
band of the hybrids themselves. The dorsolateral band
is not present in N. c. frontalis. The color of the peritoneum in
the hybrids is the glossy, jet-black of C. erythrogaster in two specimens,
and the dusky-black of N. c. frontalis in one.

Chrosomus and Notropis differ greatly in the length and convolution
of the intestine. Chrosomus has a long, coiled gut, which
is crossed by the mid-ventral line eight or nine times; in N. c.
frontalis, the intestine forms a flat, S-shaped loop that does not
cross the mid-ventral line. In the two largest hybrids (KU 4170
and 4185), the gut is intermediate, crossing the mid-ventral line
four times. In the smaller hybrid (KU 3872) the gut crosses the
mid-ventral line twice but the configuration of the anterior loops
[Pg 7]
indicates that the same intestinal convolutions that were found
in the larger specimens would have developed in KU 3872 as the
gut elongated with increase in size of the fish.

Both Deep and Bluff creeks are clear, gravel-bottomed streams
draining parts of the Flint Hills Area of Kansas. A description of
Flint Hills streams and lists of fishes occurring in them have been
published by Minckley (1956 and 1959), and by Minckley and
Cross (1959).

Chrosomus erythrogaster × Semotilus atromaculatus: KU 2947
(28.0 mm. in standard length) from Mill Creek, Wabaunsee Co.,
Kansas, Sec. 30, T. 12S, R. 9E, Mar. 22, 1953. Compared in Table
2 with five specimens of C. erythrogaster, KU 2836 (27.2 to 31.0
mm., mean 28.5) from the same locality and of the same date as
KU 2947 (above); and with five specimens of S. atromaculatus,
KU 1954, 2499, 2703, and 2838 (25.5 to 31.1 mm., mean 28.9 mm.)
from streams in the same area.

This hybrid is intermediate between the two species in number
of scales and pharyngeal teeth, and has a composite of the pigmentation
found in the parental fishes (Table 2). For diagnostic
purposes, greater importance is attached to the characters mentioned
above than to proportional measurements, which are subject
to considerable error because of the small size of the specimens.
The few measurements that were taken indicate that this
hybrid, like C. erythrogaster × N. c. frontalis, has a larger head
than do specimens of like size of either parental species. The enlarged
head affects measurements obtained for other structures
that are parts of the head; only the length of the upper jaw,
which is greatly different in the parental species, is actually
intermediate in KU 2947.

Mill Creek is a clear stream, similar to Deep and Bluff creeks
but somewhat larger. Mill Creek had an exceptionally large
population of fishes at the time the hybrid was found, but Chrosomus
and Semotilus were neither unusually common nor rare.

Two other crosses, both of which have been described in the
literature, also have been found in Mill Creek. These are N. c.
frontalis × S. atromaculatus, and N. c. frontalis × Notropis rubellus
(Agassiz).

[Pg 8]
Table 2. Comparison of One Specimen of Chrosomus erythrogaster ×
Semotilus atromaculatus with Specimens of the Parental Species
(means are above, ranges in parentheses below)

Campostoma anomalum plumbeum × Semotilus atromaculatus:
KU 4013 (three males, 86.0 to 96.0 mm. in standard length, mean
89.5 mm.) from Timber Creek, Scott Co., Kansas, Sec. 2, T. 16S,
[Pg 9]
R. 33W, June 19, 1958. Compared in Table 3 with five specimens
of C. a. plumbeum, KU 4034 (85.7 to 93.1 mm., mean 90.2 mm.)
from the Smoky Hill River, Wallace Co., Kansas, Sec. 26, T. 13S,
R. 39W, June 20, 1958; and with five specimens of S. atromaculatus,
[Pg 10]
KU 4012 and 4047 (85.0 to 97.5 mm., mean 91.7 mm.) from the
same locality and of the same date as KU 4013 (above), and
Sappa Creek, Decatur Co., Kansas, Sec. 29, T. 2S, R. 28W, June
23, 1958, respectively. This hybrid combination has previously
been recorded by Johnson (1945).

Table 3. Comparisons of Three Specimens of Campostoma anomalum
plumbeum × Semotilus atromaculatus with Specimens of the Parental
Species (means are above, ranges in parentheses below)

The hybrids seem uniformly intermediate between the parental
species. Application of the hybrid index to the characters listed
in Table 3 results in a value of 55.7 when C. a. plumbeum is assigned
the value 0.

The pharyngeal arches of the hybrids are peculiarly deformed.
Expressed in terms of the one-or two-rowed arrangement common
to all North American cyprinids, tooth-counts of 0,5-4,1; 1,3(?)-4,0;
and 2,5-4,1 best fit the three fish. However, one arch bears only
three teeth, all deformed and badly aligned, plus a pit that presumably
represents a lost fourth tooth. At the other extreme,
one arch bears eight teeth, some of which are attached to the
arch between and behind others that are countable as part of the
basic main row. Supernumerary teeth and other deformities may
have resulted from abnormalities in the replacement process. In
some cases, replacement teeth probably failed to develop; in
others, replacement teeth seemingly developed, but attached to
the arch in abnormal positions, with or without loss of previous
teeth, causing irregularity in alignment. Hubbs (1951) described
an irregular (seemingly three-rowed) alignment in a fish that
Hay (1888:249) reported from western Kansas as Squalius elongatus.
However, Hubbs considered the specimen to be an aberrant
example of S. atromaculatus, and the characteristics that he lists
for it do not correspond closely with those of the hybrid specimens
that we have. Evans and Deubler (1955:32) found three
rows of teeth in two of 150 specimens of Semotilus, and attributed
the abnormality to failure of old teeth to fall out after formation
of new teeth. The teeth of Campostoma usually number 0,4-4,0,
and those of Semotilus 2,5-4,2. The pharyngeal arches are much
smaller in Campostoma than in Semotilus.

The peritoneum is mottled dark and silvery in the hybrids; it
has a composite of the coloration in the parental species rather
than a blended shade. The intestine has two diagonal loops crossing
the ventral part of the body cavity, and the hindgut lies high
in the cavity, along the left side of the air bladder. In Campostoma,
the long gut is transversely coiled around the air bladder, whereas
in Semotilus the gut forms a longitudinal, flattened, S-shaped loop,
ventral to the air bladder.

[Pg 11]In the hybrids, the mouth is slightly oblique and nearly terminal.
The lower lip is thick and fleshy, but has only a suggestion of the
projecting mandibular shelf that is unique in Campostoma. The
upper lip is uniform in width, not medially expanded as in S.
atromaculatus. One of the hybrids lacks barbels, one has a
Semotilus-like barbel on the right side only, and one has a vestigial
barbel on the right side and an anomalous barbel that is nearly
terminal on the left upper lip.

In coloration, the hybrids lack the spot in the anterior base of the
dorsal fin that is characteristic of Semotilus, but each has a poorly-developed
dark lateral band, and a weak basicaudal spot. This
band and spot are usually prominently developed in S. atromaculatus
and usually are absent in adults of C. a. plumbeum.

In the position and obliquity of the mouth, basic color pattern
(diffuse lateral band and basicaudal spot), and the presence in one
specimen of a nearly terminal, barbel-like structure, the hybrids
somewhat resemble Hybopsis biguttata (Kirtland), which occurs
rarely in the Kansas River Basin. These partial similarities are coincidental,
because other characters of the hybrids make relationship
with H. biguttata implausible. The high number of gill rakers
(Table 3) and the length and position of the gut indicate strongly
that the three specimens are hybrids with C. anomalum as one
parent; the pharyngeal arches, though deformed, indicate that the
other parental species has two rows of teeth, with five teeth in the
main row. Only S. atromaculatus, among species in the Kansas
River Basin, usually has such a dental formula, and other characters
of our three specimens fit expectations in a hybrid between that
species and C. a. plumbeum.

Timber Creek, where the three hybrids were collected, is a small,
spring-fed, sandy-bottomed tributary to Scott County State Lake
in the extreme southwestern part of the Kansas River Basin. The
stream was less than 10 feet wide and six inches deep, except in
three pools near road crossings. The hybrids were found in two of
these pools, along with numerous S. atromaculatus and one adult
C. a. plumbeum.

Another specimen of C. a. plumbeum × S. atromaculatus (KU
4841, 39.3 mm. in standard length) was taken in the North Platte
River at Lisco, Garden County, Nebraska, on September 11, 1959.
That specimen has 7 anal rays and 52 scales in the lateral line; otherwise,
it is similar to the three hybrids described above.

Gila nigrescens × Rhinichthys cataractae: KU 4253 (a male,
[Pg 12]
60.6 mm. in standard length), from New Mexico, Bernalillo
County, Rio Grande 12 mi. S Bernalillo on U. S. Highway 85
(Corraleo Bridge). Compared in Table 4 with six specimens of
G. nigrescens: KU 4251, 4254, and 4262 (63.1-72.4 mm. in standard
length, mean 66.4 mm.); and with five specimens of R. cataractae:
KU 4248, 4258, and 4264 (55.6-65.0 mm. standard length, mean 59.5
mm.). Comparative material was taken at the same locality as
KU 4253 and at nearby localities in the Rio Grande.

The hybrid is intermediate in almost all of the features in which
the parental species differ from each other. For six of the characters
included in Table 4, the hybrid index is 49.7 per cent, when
Gila is assigned the value 0 (height of dorsal fin and numbers of fin
rays and teeth excluded). There is no enlargement of the head in
KU 4253, such as was found in Gila orcutti × Siphateles mohavensis
(Hubbs and Miller, 1943:373), Chrosomus erythrogaster × Notropis
cornutus frontalis, and C. erythrogaster × Semotilus atromaculatus.
The height of the dorsal fin, which Hubbs and Miller
(loc. cit.) found to be extreme in G. orcutti × S. mohavensis, exceeds
the average for the parental species in G. nigrescens × R.
cataractae also; but, dorsal fins as high as that of the hybrid were
found in some individuals of both parental species. In R. cataractae,
all fins are more rounded and more expansive than in G. nigrescens,
and fins other than the dorsal have an intermediate size in the hybrid.
This intermediacy has doubtful significance, because fin-size
in Rhinichthys varies greatly with body-size, sex, and probably with
the state of sexual development. Rhinichthys matures at smaller
size than Gila, and never becomes so large as that species.

Gila nigrescens and R. cataractae differ strikingly in features
involving the snout and mouth, and these differences provide the
most conclusive evidence that KU 4253 is a hybrid of these species.
The projecting, fleshy snout of R. cataractae is bridged to the ventral
mouth by a frenum that is approximately 3 mm. wide in specimens
60 mm. in standard length. In Gila, the snout does not
project beyond the mouth, which is oblique, lacks a frenum, and
is larger than in Rhinichthys. The snout of the hybrid projects
less than in R. cataractae and is bridged to the upper lip by a
frenum 1.7 mm. wide. The mouth of the hybrid is intermediate
in size, obliquity, and thickness of the lips. Rhinichthys has
barbels, Gila lacks them, and the hybrid has one vestigial barbel,
on the right side. The lower surface of the head of Rhinichthys is
broad and flattened, with pronounced rugosity on the gular area
[Pg 13]
and isthmus. In Gila the underside of the head is convex, with
comparatively smooth membranes; the hybrid is intermediate,
but tends toward Gila.

Table 4. Comparisons of One Specimen of Gila nigrescens × Rhinichthys
cataractae with Specimens of the Parental Species (means are
above, ranges in parentheses below)

[Pg 14]
Table 5. Comparisons of One Specimen of Notropis v. venustus × Notropis
whipplei with Specimens of the Parental Species, and with N. lutrensis
× N. v. venustus. Measurements (lengths and depths) Are
Expressed as Thousandths of Standard Length (means above, ranges
in parentheses below)

[Pg 15]
The air bladder of KU 4253 is nearly as large as in Gila, and much
larger than the degenerate air bladder of R. cataractae. Although
the hybrid appears to be male, the gonads (especially the right
one) are poorly developed. The hybrid is intermediate in curvature
of the lateral line, which is nearly straight in Rhinichthys and
strongly decurved in Gila.

Specimen No. 4253 is mostly pallid, resembling Gila much more
than Rhinichthys in pigmentation. A mid-dorsal dark streak is
conspicuous in the hybrid, especially anteriorly, but is less intense
than in Gila. Rhinichthys lacks a well-developed dorsal stripe.
The preorbital and suborbital areas are more heavily pigmented
in the hybrid than in Gila, but not nearly so dark as in Rhinichthys.
The hybrid has a faint dark basicaudal spot that is variably developed
in Rhinichthys but absent in Gila.

Notropis venustus venustus × Notropis whipplei: KU 3516 (a
male, 47.8 mm. in standard length), from Arkansas, Sevier Co.,
Winters Creek where it is crossed by U. S. Highway 71, 5 mi. N
of Little River Bridge, March 8, 1956. Compared in Table 5
with four specimens of N. whipplei, KU 3517 (45.0-52.6 mm. in
standard length, mean 50.6 mm.), same locality and date as KU
3516; four specimens of N. v. venustus, KU 3510 (44.5-49.6 mm.
in standard length, mean 47.3 mm.), Louisiana, Winn Parish,
Little Naches Bayou on U. S. Highway 71, 8.8 mi. NW Montgomery,
March 4, 1956; three specimens of Notropis lutrensis (Baird
and Girard) × N. v. venustus, KU 3510 (43.3-47.3 mm. in
stand[Pg 16]ard
length, mean 44.7 mm.), same locality and date as N. v. venustus
above; and with tabulated data on N. v. venustus from Gibbs
(1957a:185-186). All specimens are from the lower Red River
Drainage; other series of N. whipplei, N. venustus, and N. lutrensis
× N. venustus, from the Red River Drainage and elsewhere, were
examined but are not tabulated because of differences in size,
and because of geographic variability that has been discussed by
Gibbs (1957a).

The Subgenus Cyprinella of Notropis, to which N. venustus and
N. whipplei belong, has been studied intensively by Gibbs (1957a
and b). Notropis venustus differs conspicuously from N. whipplei
in having a large dark basicaudal spot; also, N. venustus usually
has 8 (rather than 9) anal rays, and 15 (rather than 13) scales
above the lateral line immediately anterior to the dorsal fin. Specimens
of N. v. venustus from the Red River Drainage, where the
most robust representatives of the species are found, differ from
N. whipplei in depth of head, body, and caudal peduncle (Table
5).

KU 3516 has a composite of the 9-rayed anal fin of N. whipplei
and the caudal spot (albeit diffuse) of N. venustus; and, the hybrid
is intermediate in body-proportions that distinguish the two
species, especially depth of head, body, and caudal peduncle.
In other features KU 3516 has values within the overlapping
ranges of variation of whipplei and venustus except that the ratio
of postdorsal length to standard length is extremely long in the
hybrid, and the ratio of prepelvic length to standard length is
extremely short (Table 5). Both extreme values for the hybrid
seem to result from the cumulative influence of characters in
which the parental species differ slightly in mean value (especially
head-length, in which the hybrid is like whipplei, and caudal
peduncle-length, in which the hybrid approaches venustus, despite
the 9-rayed anal fin of the hybrid). The basicaudal spot of the
hybrid is like that of N. v. venustus except for being less intense.

Notropis venustus hybridizes extensively with N. lutrensis (Hubbs,
Kuehne, and Ball, 1953:226-230; Hubbs and Strawn, 1956), and that
combination occurs in streams near the locality where KU 3516 was
taken. KU 3516 resembles N. lutrensis × N. v. venustus in many
ways, but is more slender than the latter hybrid. The depth of
head, body, and caudal peduncle are greater in N. lutrensis than in
N. venustus (much greater than in N. whipplei); therefore,
speci[Pg 17]mens
of N. lutrensis × N. venustus are usually deeper than N.
venustus, whereas KU 3516 is less deep. KU 3516 has a rather
sharp snout and thin, straight lips that are strongly suggestive of
N. whipplei, rather than N. lutrensis, in which the snout is rounded
and the lips are more obliquely decurved. There is less pigment
underlying the anterior lateral-line scales in KU 3516 than in N.
lutrensis × N. venustus, and melanophores on the scale-pockets of
KU 3516 are arranged in narrower, more distinct submarginal bars
than in N. lutrensis × N. venustus. Because of the difference in
pigmentation, the lateral scales of N. whipplei (and of KU 3516)
appear more narrowly diamond-shaped than the lateral scales of
N. lutrensis or N. lutrensis × N. venustus. The lengths and heights
of the scales are approximately the same in all three species.

Winters Creek, where KU 3516 was taken, flowed approximately
five cubic feet per second at the time our collection was made; a
landowner on the stream stated that it had been dry, except for
pools, in the previous two summers. The water was somewhat
gray, but nearly clear. The habitat consisted mainly of short riffles,
with average depth of four inches, and pools to depths of two feet.
Twelve species of fish, including N. whipplei but not N. lutrensis
or N. venustus, were found; other minnows were Semotilus atromaculatus,
N. chalybaeus, N. cornutus, N. umbratilis, and Campostoma
anomalum.

LITERATURE CITED

Evans, H. E., and Deubler, Jr., E. E.

Gibbs, Jr., R. H.

Hay, O. P.

Hubbs, C. L.

Hubbs, C. L., and Lagler, K. F.

Hubbs, C. L., and Miller, R. R.

Hubbs, C.

Hubbs, C., Kuehne, R. A., and Ball, J. C.

Hubbs, C., and Strawn, K.

Johnson, R.

Minckley, W. L.

Minckley, W. L., and Cross, F. B.

Trautman, M. B.

Transmitted March 2, 1960.

◻

28-3424