Mate selection in two species of sunfish (Lepomis gibbosus and L . megalotis peltastes)

ROBERT G. STEELE AND MILES H. A. KEENLEYSIDE Department of Zoology, University of Western Ontario, London, Ontario

Received June 15, 1971

STEELE, R. G., and M. H. A. KEENLEYSIDE. 1971. Mate selection in two species of sunfish (Lepomis gibbosus and L. megalotispeltastes). Can. J. Zool. 49: 1541-1548.

The reactions of 20 male and 20 female pumpkinseed (Lepomis gibbosus) and longear sunfish (L. megalotispeltastes) to conspecifics and heterospecifics of the opposite sex were studied. Female pumpkin- seeds showed a preference for conspecific males in the absence of visual cues; female longears showed a preference for conspecific males when both types of male could be seen. Males of both species did not discriminate between the two kinds of females in either the presence or absence of visual cues. Female choice of a conspecific spawning partner may be an important ethological isolating mechanism between these two species. A mechanism for hybridization between them is proposed.

Les reactions de 20 miles et 20 femelles de crapets-soleil (Lepomisgibbosus) et de crapets-aux-longues- oreilles (L. megalotispeltastes) aux individus du sexe oppose de la mdme espke ou de I'autre espke ont kt6 Ctudiks. Les crapets-soleil femelles montrent une preference pour les mlles de la mdme espke en l'absence d'indications visuelles; les crapets-aux-longues-oreilles femelles montraient une prkfkrence pour les miles de la meme e s p k quand les deux types de mlles peuvent dtre vus. Les males de I'une ou de I'autre e s p h ne font pas de discrimination entre les deux e s p k s de femelles, que ce soit en la presence ou en I'absence de repbres visuels. Le choix des femelles pour un partenaire de fraye de la rndme espbe peut dtre un important mhnisme ethologique d'isolement entre ces deux e s p h s . Un mkanisme d'hybridation entre eux est propose.

Introduction

In the Middle Branch of the Thames River near Thamesford, Oxford County, Ontario, two species of s d s h occur: the pumpkinseed, Lepomis gibbosus (L.) and the northern longear L. megalotis peltastes Cope. In addition, hybrids between these two species are found which clearly indicates that reproductive isolation between the two species is less than perfect. Nevertheless, the presence of morphologically identifiable adults of the two species indicates some reproductive isolation must occur.

As postmating isolating mechanisms (Mayr 1970) are currently under investigation, they will not be discussed in this paper. Considering premating isolating mechanisms, seasonal isola- tion is probably unimportant as the breeding seasons of the two species broadly overlap (J. Cresswell, unpublished data). Male longears usually nest colonially close to shore in the relatively weedfree, slowly moving water of the main river, but have occasionally been seen nesting within I m of occupied pumpkinseed nests which are always found singly in quiet, weedy backwaters. The habitat preference exhibited by the males is less apparent in the

females as ripe female pumpkinseeds have fre- quently been seined from the edges of spawning longear colonies, and ripe longear females have often been seined from the backwaters. Thus, habitat isolation is at best an imperfect isolating mechanism. Mechanical isolation cannot operate as fertilization in sunfish is external.

Ethological isolation is the remaining premat- ing barrier to hybridization. In aquarium tests, Keenleyside (1967) found that pumpkinseed males, and to a much lesser extent longear males, visually distinguished between the two species of females by directing more aggressive and court- ship activities toward conspecsc females. Re- sponses of females were not studied. Observations of spawning in the study area suggested that mate selection by females might be more important to intraspecific breeding than mate selection by males for reasons discussed later.

This paper tests the hypothesis that both sexes of both species can visually discriminate each other. Where such discrimination occurs, it could be the basis for conspecific mate selection and hence the maintenance of species integrity; where such discrimination is not pronounced, it could be the basis for indiscriminant mate selec- tion and hence for hybridization.

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sant

a C

ruz

(UC

SC)

on 1

1/13

/14

For

pers

onal

use

onl

y.

1542 CANADIAN JOURNAL OF ZOOLOGY. VOL. 49, 1971

Materials and Methods S a e c t s

The sunfish used in these experiments were collected in the study area by seining during the summer of 1967, and all experiments were conducted in October and Novem- ber, 1967. Before use the fish were held in four 850-liter cement tanks in a greenhouse where they were separated by species and sex. Water temperatures in these tanks ranged between 15 and 20°C during the study. Photo- period in the greenhouse was natural (43" N lat.), so the fish received approximately 9) h of light per day.

The ikh were fed a diet of Strike Fish Food #4 Crumbles (Country Best, Agway Inc., Syracuse, N.Y.), occasionally supplemented with live Daphnia, small guppies, Tenebrio and mosquito larvae, and frozen brine shrimp. Pump- kinseed sunfish were larger than longear sunfish and long- ear males were consistently larger than longear females. No attempt was made to match experimental fish by size either within or between species, other than choosing fish within the adult range for their species (pumpkinseeds 7-20 cm, longears 5-10 cm, Trautman 1957).

Apparatus A cement and plate glass tank (335 X 152 X 66 cm deep)

was divided lengthwise into five equal compartments (Fig. 1). Each compartment was in turn divided into five sections by inserting a double diamond glass partition 91 cm from each end and a one-way glass partition 30 cm inward from the glass partitions towards the center of the compartment. The central sections (F-J. Fig. 1) were separated from each other by black plastic, while all other sections were separated by fine-mesh fiber glass screening. The tank bottom was covered by a layer of fine gravel to a depth of about 7 cm. Water depth above the gravel was about 24 cm.

Qch end of each of the five compartments was illu- minited by a 100-W bulb. The central sections were each illuminated by a 100-W bulb surrounded with an alumi- num foil reflector. All bulbs were about 30 cm from the water surface. The central bulbs were connected to a powerstat variable autotransformer (type 3PN116B, Superior Electric Co., Bristol, Conn.). All lights were connected to a time switch which provided a 16 h : 8 h, 1ight:dark cycle, beginning at 7 a.m., to approximate the local June photoperiod when sunfish spawning normally occurs.

The entire tank was surrounded by a black plas- tic curtain about 250 cm high and 30 cm from the edge of the tank. Tap water slowly trickled into each compartment at one end, and water level was maintained by a standpipe drain at the other end of the tank. Water temperature was 21.5 + 2OC at the inflow end and 24 + 3°C at the outjlow end of the experimental tank for 90% of the experimental time. Two inflow adjustments to cool the water in the tank when it rose above 27°C caused brief declines of 7 and 8".

Reactions of males and females of both species to conspecifics and to heterospecifics of the opposite sex were recorded in two experiments. A single fish to be tested was introduced into the central section of each compartment (test sections F-J, Fig. I), given a day to settle down, and observed the following day. Conspecific or heterospecific fish of the opposite sex were introduced into the end sections (stimulus sections A-E and K-0,

Fig. 1) in groups of three, as longear females do not en- counter singly nesting male longears. It was felt that the addition of more than three fish might unduly crowd the stimulus sections. These groups remained in the stimulus sections for the duration of each experiment. In each longitudinal compartment, then, there was a single fish under observation in the test section, three conspecific fish of the opposite sex in one stimulus section, and three heterospecific fish of the opposite sex in the other stimulus section.

Observations of test fish position and behavior were made through a slit in the black plastic curtain while reflections of the fish were watched in a mirror suspended over the test sections. These observations were made under two lighting conditions: control (Q with the central lights on, allowing each test fish to see only its own mirror image in the one-way glass partition at each end of the test section, and experimental (E) with the central lights off, allowing the test fish to see through the one-way glass partitions into the stimulus sections at each end of the test sections. The lights over the stimulus sections re- mained on during both C and E conditions. Under the C condition, stimulus fish could see into the central section, while under the E condition, stimulus fish could see their mirror images in the one-way glass partition.

Under lighting condition C, a test fish was observed for a metronome-timed 2-s interval during which its position and behavior were recorded (see below). Then the test fish in the next compartment was observed for a 2-s interval and its position and behavior recorded. This continued until all five test fish had been observed, at which time the first test fish was again observed for a 2-s interval. This procedure was repeated until a total of 200 observations had been recorded for each of the five test fish. The intensity of the lights over the test sections was then reduced to zero during a 30-9 period (producing lighting condition E) and the entire observation and recording procedure was repeated until a further 200 observations per test fish had been recorded. The five test fish were then discarded; no fish was used as a test fish more than once during the experiments. Fish behav- ior and position were recorded on five Adams nine unit laboratory counters (Clay-Adams Inc., New York, N.Y.). All data were collected between 11 a.m. and 12:30 p.m.

Position and behavior of test fish were recorded as one of the following categories.

1. Pointing conspecifically (PC) was recorded if the test fish was within one fish length of, and oriented toward, the one-way glass partition beyond which were the three conspecifics of the opposite sex (the conspecific partition).

2. Pointing heterospecifically (PH) was recorded if the test fish behaved as in 1, but oriented towards the hetero- specific partition.

3. Nosing conspecifically (NC) was recorded if the test fish touched the conspecific partition with its snout, and was either stationary or swimming slowly along the partition.

4. Nosing heterospecifically (NH) was recorded if the test fish behaved as in 3 but contacted the heterospecific partition.

5. Fluttering conspecifically (FC) was recorded if the test fish in addition to touching the conspecific partition with its snout, also engaged in rapid horizontal beats of its tail, as if trying to swim through the partition.

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sant

a C

ruz

(UC

SC)

on 1

1/13

/14

For

pers

onal

use

onl

y.

I

Bla

ck P

lost

lc

Cur

tain

100

Wat

t B

ul

b0

C

emen

t E

dge

of T

onk

n

Cle

ar

Glo

ss

Po

rt~t

lon

-

100

Wat

t B

ulb

w~

th Ref

lect

or 0 Plate

Gla

ss E

dg

e of

Tan

k -

One

-way

G

lass

Par

tlt~

on

-

Sta

nd

p~p

e Dra

in 0

Bla

ck P

lasl

~c P

art

~t~

on

-

Fibr

eglo

ss S

cree

n P

art~

tlo

n --

0

1 2

3

11

SC

AL

E b

t 30

80

9

0 c

m

FIG. 1.

Top

vie

w o

f ex

peri

men

tal t

ank

show

ing s

uwiv

isiw

s. S

timul

us se

ctio

ns: A

-E,

K-0

. T

est s

ectio

ns: F

-J.

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sant

a C

ruz

(UC

SC)

on 1

1/13

/14

For

pers

onal

use

onl

y.

1544 CANADIAN JOURNAL OF ZOOLOGY. VOL. 49, 1971

6. Fluttering heterospeciiically (FH) was recorded if the test fish behaved as in 5 against the heterospecific partition.

7. Neutral (N) was recorded if the test fish was not recorded in one of the above six categories. Such iish usually hovered close to the bottom gravel in the central part of the test section.

8. Crossover (XO) was recorded each time the test fish swam from one end of the test section to the other.

The XO data were collected at the same time as the data for the first seven behavior acts; however, only the latter were included in the 200 observations per test fish recorded under each of lighting conditions C and E.

If the test fish were discriminating between the two species of stimulus fish solely on the basis of visual cues, there should be no differences between comparable behav- ior acts (FC and FH, NC and NH, PC and PH) in the C phase, but there should be a relatively greater number of conspecific acts in the E phase.

Experiment I . Females as Test Fish Stimulus fish were alternated with each other by

introducing three male pumpkinseed sunfish into each of stimulus sections A, C, E, L, and N, and three longear sunfish into each of stimulus sections B, D, K, M, and 0 (Fig. 1). All 10 groups of males remained in these sections for the duration of the experiment, and were fed dried fish food each afternoon.

For the first day of observations, single female pump- kinseeds had been placed in test sections F, H, and J, while female longears had been placed singly into test sections G and I (Fig. 1). For the second day the alloca- tion of test sections was reversed; longears being placed in sections F, H, and J, and pumpkinseeds into sections G and I. Naive female test fish of the two species were alternated in this manner until the completion of experi- ment 1 at the end of eight observation days (during which

20 females of each species were observed). These test fish were not fed while in the experimental tank. The apparatus was then dismantled, cleaned, set up again as before, and the water aged for one week before experiment 2 began.

Experiment 2. Males as Test Fish The testing procedure for male sun6sh responses to

conspecific and heterospecific females was similar to that described for females. Male pumpkinseeds were originally introduced into test sections F, H, and J, and alternated with test sections G and I on successive observation days. Male longears were originally introduced into sections G and I, and alternated with sections F, H, and J on succes- sive observation days. Three female pumpkinseeds were placed in stimulus sections A, C, E, L, and N, while three female longears were placed in stimulus sec- tions B, D, K, M, and 0. The females remained in those sections for the duration of the experiment, which again consisted of 8 observation days. Twenty males of each species were observed.

Results

Table 1 presents the sums of behavior acts for males and females of each species under lighting conditions C and E. These values were greater in the E phase than in the C phase of the experi- ments for every category except N, indicating an increase in test fish activity under the E lighting conditions. In every case PH > NH > FH, and in every case but one (male pumpkinseeds, E phase) PC > NC > FC, indicating that these fish hointed more than they nosed, and nosed more than they fluttered under C and E conditions.

TABLE 1 Totals of behavior acts in experiments 1 and 2

Female

Pumpkinseeds (n = 20) C 3 32 828 2640 464 31 2 76 E 105 156 999 1533 849 202 156 254

Longears (n = 20) C 8 27 381 3050 500 27 7 46 E 85 389 1248 1588 612 61 17 184

Male

FC NC PC N PH NH FH XO

Pumpkinseeds (n = 20) C 15 32 271 3167 454 49 12 105 E 153 106 797 1731 95 1 157 105 248

Longears (n = 20) C 5 18 574 2840 539 15 9 178 E 39 146 872 1772 958 141 72 285

No re: FC. flutterconspecifically; NC, noseconspecifically; PC, point conspeclfically; N, ncutral; PH. point hetero- specifically; NH, nose heterospecifically; FH, flutter heterospecifically; XO, crossover; C, control; E. experimental.

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sant

a C

ruz

(UC

SC)

on 1

1/13

/14

For

pers

onal

use

onl

y.

STEELE AND KEENLEYSIDE: MATE SELECTION M SUNFISH 1545

A comparison of comparable behavior acts (FC and FH, NC and NH, PC and PH) in the C phase, using an analysis of variance1 revealed no sign5cant differences between those acts except between PC and PH for female pumpkinseeds (P < 0.025). This difference indicated that female pumpkinseeds discriminated between the two ends of the test section by pointing significantly more towards the conspecific end, in spite of the fact that under the C conditions they could not see the stimulus males behind the one-way glass partition.

1Populations, acts, and experimental conditions were treated as fixed effects, and fish as random effects for the analysis of comparable behavior acts.

Female Pumpkinseeds

A comparison of comparable behavior acts in the E phase of the experiments revealed no significant differences between those acts except between PC and PH, and NC and NH in female longears (P < 0.001). This indicated that of the four kinds of fish tested only female longears clearly distinguished between the two ends of the test section in the E phase by orienting towards and contacting the conspedic one-way glass partition more than they did the heterospecfic partition. Although FC was not s i d c a n t l y larger than FH for female longears in the E phase the trend to conspecific choice persisted. Certainly when all the conspecific acts (FC, NC, and PC) were compared with all the hetero- specific acts (PH, NH, and FH) under both

-m- Female Longears

Mole Pumpkinsssd$ Male Longears

(n=20) (n=20) -3m

FIG. 2. Percentage increases (vertical scale) in behavior acts between control and experimental conditions.

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sant

a C

ruz

(UC

SC)

on 1

1/13

/14

For

pers

onal

use

onl

y.

1546 CANADlAN JOURNAL OF ZOOLOGY. VOL. 49. 1971

conditions C and E, only female longears per- formed significantly more conspecific than heterospecific acts (P < 0.001).

Figure 2 was derived from the data of Table 1 and illustrates the percentage increase in be- havior acts between the C and E phases of the two experiments. Males scored higher XO values than females in the C phase (Table I), but the percentage increase in XO from phase C to phase E was higher for females than for males (Fig. 2). A comparison of XO values using an analysis of variance2 revealed no significant differences between females of the two species, or between males of the two species, or between females and males, when differences between the E and C conditions were considered. When mean XO values were considered, no significant differences were found between females of the two species, or between males of the two species, but the mean XO value for females was sig- nificantly lower (P < 0.025) than it was for males. This difference resulted from females having lower XO values than males during the C phase, and from male longears having higher XO values than female longears in the E phase (Table 1).

Figure 2 also shows that female longears ex- hibited both the largest percentage increase in a single behavior act (PC), and the greatest dif- ferences between comparable behavior acts (PC and PH, NC and NH), when they were able to perceive fish of the opposite sex under lighting condition E.

Discussion Sunfish are inactive at night in aquaria and in

the field (personal observation), and a decrease in light intensity could therefore be expected to promote a decrease in fish activity. In these experiments, however, a decrease in light in- tensity above the test fish (condition E) was followed by an increase in test fish activity (Table 1, Fig. 2). This increase was undoubtedly a result of the perception by the test fish of the stimulus sections behind the one-way glass partitions. Further, the clear positive response of female longears to the conspecific end of the

2Populations and experimental conditions were treated as fixed effects, and fish as random effects for analysis of the XO data which was transformed to the square root of the observed value plus 0.5.

test sections indicated the test fish could see the stimulus fish. At least part of their increased activity in the E phase may therefore have been caused by their responses to the stimulus fish.

Male suniish in this study failed to show dis- crimination between the two types of females (Table 1, Fig. 2), in contradiction of Keenley- side's (1967) results which showed that male pumpkinseeds, and to a lesser extent male longears, preferentially courted and were aggres- sive towards conspecific females. In the present study, conditions may not have permitted males to get close enough to females to discriminate between them. Alternately, male discrimination of females may be related to sexual maturity of the test males. Keenleyside (1967) tested males during the breeding season after they had dug nests in the aquarium gravel. The present experiments were done in the fall, 3 to 4 months after the end of the breeding season, and males did not dig nests in the test tank. Nest-digging by male sunfish is dependent on gonadal de- velopment, which is in turn dependent on photoperiod (Smith 1969, 1970). As these fish were held under natural photoperiod before testing, their gonads were probably not mature. If male discrimination of females depends on an advanced stage of sexual maturity, the males tested would not distinguish between the females.

Females, however, did discriminate between the two types of male sunfish. Female pumpkin- seeds in the C phase of the experiment pref- erentially oriented towards the conspeci6c end of the test section. Their inability to see the male fish in the C phase suggests they may have oriented towards non-visual stimuli emitted by the male pumpkinseeds. In the study area, male pumpkinseeds often nest in dense aquatic vegetation which must render a female's visual perception of these nests difficult. Hence the males would enhance their chances of mating with conspecific females if they emitted a non- visual stimulus to which female pumpkinseeds responded.

Two non-visual recognition cues which might be used are olfactory and auditory. While ol- factory cues have been associated with homing by longear sunfish (Gunning 1959) and in long- ear nest colony formation (Gerald 1970), its role in pumpkinseed reproduction is unknown. Olfactory cues would provide poor localization

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sant

a C

ruz

(UC

SC)

on 1

1/13

/14

For

pers

onal

use

onl

y.

STEELE AND KEENLEYSIDE: MATE SELECTION IN SUNFISH 1547

except in moving water, and water flow in the pumpkinseed nesting area (weedy backwaters) is usually negligible. Auditory cues on the other hand provide good localization. Gerald (1970) demonstrated grunt-like sound production in six of seven species of courting male sunfish which he suggests are species-specific and function to attract conspecifics. Sound production by stimu- lus male pumpkinseeds may well have been the non-visual stimulus which evoked the con- specitic orientation of the test female pumpkin- seeds in the present experiments.

Female longears did not differentiate between ends in the C phase of the experiment, but clearly directed more behavior acts towards the conspecific end during the E phase (Table 1, Fig. 2). This indicates that the female longears responded to the males visually by directing more behavior acts towards conspecific males.

Male longears nest in closely spaced colonies of up to several hundred nests. These groups of brightly colored, exceedingly active, territorial males could provide a powerful visual stimulus to ripe female sunfish seeking spawning partners. These male longears, however, particularly at the height of spawning in a crowded colony, have little or no time for active discrimination between females of the two species as they are too busy excluding intruders from their ter- ritories (Keenleyside 1971, 1972). In fact, under these conditions, females often dart into nests without any preliminary interaction with the guarding male and spawning begins immediately (Gerald 1970; personal observation). Ripe females hovering around the periphery of longear colonies would have more time to select conspedic spawning partners than male long- ears.

Male pumpkinseeds nest singly in the study area and are not as continually occupied with aggressive territorial defense as male longears. Consequently they could discriminate betweenp approaching females before they reached the nest.

Keenleyside (1967) showed that nesting male pumpkinseeds preferentially courted conspecific females, and that longear males less clearly courted longear females in preference to pump- kinseeds. The present experiments showed that female longears consistently oriented visually towards male longears in preference to male pumpkinseeds, and less clearly that female

pumpkinseeds oriented non-visually towards conspedic males. These results are summarized in Fig. 3.

The presence of hybrids in the study area (J. Cresswell, unpublished data) indicates crossing between the two species. As both female longears and male pumpkinseeds choose conspecifics under experimental conditions, it is doubtful that hybridization occurs between them. More likely hybridization occurs between female pumpkinseeds and male longears (Fig. 3). This hypothesis is supported by the fact that small ripe female pumpkinseeds comparable in size to large female longears have frequently been seined from spawning longear colonies. Ripe female longears have been seined from the back- waters near pumpkinseed nests only when male longears were also nesting nearby. Also on a few occasions interspecitic spawning has been ob- served in male longear nests; it has never been seen in pumpkinseed nests.

Male test fish of both species performed more crossovers than female test fish during the C phase of the experiments (Table 1). This may have been caused by the stronger response of males to their mirror images in the end walls of the test section. Male sudish are generally aggressive towards conspecifics, both in aquaria and in the field (Keenleyside 1967, 1971), and during the C phase male test fish occasionally moved across the center of the test section as though responding to their mirror image. That this response was not more pronounced may have been a result of their having habituated to the experimental condition under lighting con- dition C for a day before testing.

Stimulus fish, both males and females, did not respond to either the sight of the test fish

FIG. 3. Postulated model of hybridition between longear (L) and pumpkinseed (P) sunfish. K: Keenley- side's results (1967). S: Steele and Keenleyside's results. The wide vector indicates a marked preference for con- speci6cs; the narrow vector indicates a weaker preference for conspecifics.

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sant

a C

ruz

(UC

SC)

on 1

1/13

/14

For

pers

onal

use

onl

y.

1548 CANADIAN JOURNAL OF ZOOLOGY. VOL. 49. 1971

(lighting condition C) or their mirror image in the one-way glass (lighting condition E). This lack of response was probably a result of three factors: (1) stimulus fish had a longer time to habituate to the experimental conditions, (2) the clear glass partition prevented the stimulus fish from approaching the one-way glass par- tition closer than 30 cm, (3) stimulus fish, be- cause they were introduced in groups of three, could interact among themselves.

During the breeding season adult males of both species confine their movements to a relatively small area surrounding their nests. Females swim about on the periphery of longear colonies or pumpkinseed nests, and hence cover greater distances. This may explain why females in this study showed a relatively greater increase in XO behavior than males when exposed to the E condition (Table 1, Fig. 2).

Conspecific mate selection by male and female s d s h of both species is undoubtedly an im- portant factor in the maintenance of species integrity in the Thames River study area. When reproductive isolation occasionally breaks down between the two species, this breakdown likely occurs between male longears and female pumpkinseeds.

Acknowledgments The authors thank Dr. R. K. Misra for assist-

ance with analysis of the data and Dr. D. M. Scott for suggestions offered during manuscript prep- aration. This research was supported by a Province of Ontario Graduate Scholarship to R. G. Steele, and a National Research Council of Canada grant to M. H. A. Keenleyside.

GERALD, J. W. 1970. Species isolating mechanisms in the genus Lepomis. Ph.D. Thesis, Univ. Texas, Austin.

GUNNING, G. E. 1959. The sensory basis for homing in the longear sunfish, Lepomis megalotis megalotis (Rahesaue). Invest. Indiana Lakes Streams, 5: 103- - ?

130. KEENLEYSIDE, M. H. A. 1967. Behavior of male sun-

fishes (genus Lepomis) towards females of three species. Evolution, 21 : 688-695.

1971. Aggressive behavior of male longear sun- 6sh (Lepomis megolotis). Z. Tierpsychol. 28: 227-240. - 1972. Intraspecific intrusions into nests of spawning longear sunfish (Pisces: Centrarchidae). Copcis. In press.

MAYR, E. 1970. Populations, species and evolution. Harvard Univ. P m , Cambridge, Mass.

S m . R. J. F. 1969. Control of prespawning behaviour of su&h (Lepornis gibbom -and- L. megalotis). I. Gonadat androgen. Anim. Behav. 17: 279-285,

1970. Conrrol of prespawning behaviour of sun- fish (Lepomk g'bbosrn and Lepomis rnefulotis). Environmental factors. Anim. Behav. 18: 575-587.

W-, M. B. 1957. The fishes of Ohio. Ohio State Umv. Press, Columbus.

Can

. J. Z

ool.

Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

Sant

a C

ruz

(UC

SC)

on 1

1/13

/14

For

pers

onal

use

onl

y.