Journal of Fish Biology (1996) 48, 726–737
Sexual behaviour of immature male eastern mosquitofish:
a way to measure intensity of intra-sexual selection?
A. B*, A. P†‡, R. P†  G. M†
*Dipartimento di Psicologia Generale, Università di Padova, Piazza Capitaniato 3,
35139 Padova and †Dipartimento di Biologia, Università di Padova, via Trieste 75,
35121 Padova, Italy
(Received 22 November 1994, Accepted 30 June 1995)
Immature males of eastern mosquitofish Gambusia holbrooki start to be sexually active well
before their copulatory organ (gonopodium) has completely developed and before they become
able to transfer sperm. Sexual activity of males, consisting of copulatory attempts tending to
bypass female acceptance, is intense (one attempt per minute) and is likely to be energetically
very costly. The sexual behaviour of immature males relative to their maturation stage is
described and tested against two possible adaptive explanations. Sexual activity was present in
males from the beginning of the development of their gonopodium and increased during the
following stages of maturation. Two to three weeks before gonopodium development was
completed, sexual activity of immatures was as high as that of adults. Adult males showed
aggressive behaviour against a male attempting a copulation, irrespective of the maturity of the
latter. Since previous studies have shown that the reproductive success in this species is
negatively correlated with male size when male–male competition is low (i.e. when the sex ratio
is female biased), but decreases with male size when competition is high, the hypothesis was
tested that sexual activity of immature males functions as a way to predict their future
reproductive success if they mature at a given size. A second hypothesis tested was that
precocious sexual experience improves the efficiency of copulatory attempts. Results were more
in agreement with the first hypothesis, since size at maturity of males was influenced by the sex
ratio experienced during maturation and precocious experience gave very little advantage.
? 1996 The Fisheries Society of the British Isles
Key words: Gambusia holbrooki; sexual behaviour; immature males; aggressive behaviour;
sex ratio; experience.
INTRODUCTION
An increasing body of evidence indicates that, as assumed by life history theories,
individuals generally pay costs associated with reproduction (Clutton-Brock, 1991;
Marconato, 1991). Selection should therefore favour strategies that reduce costs to
a minimum, and individuals are expected to reproduce only when advantages
exceed costs in terms of reduced future fitness. In many bird species, for example,
secondary sexual traits develop only during the breeding season (Andersson, 1983),
or, as in many fish species, breeding coloration is shown or intensified only at the
appearance of females (e.g. Kodric-Brown, 1983; Gronell, 1989).
An obvious corollary to the above statements is that mating behaviour and
manifestations connected with reproduction (territoriality, breeding coloration,
courtship behaviour, etc.) should not appear until individuals reach functional
sexual maturity. The rare exceptions known so far have an adaptive explanation. In several primates, including humans, young infertile individuals
‡Author to whom correspondence should be addressed. Tel.: +39–49–827610; fax: +39–49–8276209;
email: [email protected]
726
0022–1112/96/040726+12 $18.00/0
? 1996 The Fisheries Society of the British Isles
     
727
include sexual acts in their play and in some cases a complete coitus is achieved
(Ford & Beach, 1951; Mason, 1965). Among chimpanzee (Pan troglodytes
Blumenbach), young with no previous sexual experience are not immediately able
to copulate successfully, which suggests a possible role for sexual play. In
aquarium studies, immature males of the eastern mosquitofish (Gambusia holbrooki Girard) exhibit precocious copulatory activity, and their rate of copulatory
attempts was comparable to that of adults (Bisazza, unpublished data).
Gambusia holbrooki is a small poeciliid fish commonly found in brackish and
fresh waters of Italy, where it was introduced from North America at the
beginning of this century. Fertilization is internal and the male uses its modified
anal fin (gonopodium) to transfer sperm to the female. As in most poeciliids
(Constantz, 1975; Farr, 1984), male sexual activity mainly consists in approaching the female from behind, moving the gonopodium forward and trying to
introduce forcibly the gonopodial tip into her genital pore (gonopodial thrusting). In this species, females rarely if ever show receptivity. Males make
copulatory attempts at a very high rate, but only a small proportion of them is
successful (Martin, 1975; Bisazza & Marin, 1995). Male mosquitofish, typical of
most poeciliid species, stop growing when they become sexually mature (Schultz,
1961; Borowsky, 1973a; Zulian, 1990) and body size attained at maturation
changes in response to the social environment (Borowsky, 1973b; Sohn, 1977).
To explain the behaviour of immature male mosquitofish, two hypotheses were
proposed. The first one postulates that, as observed in apes and other vertebrates,
experience gained when immature increases the efficiency of later sexual behaviour
as adults. The second hypothesis, which will be discussed in more detail later, is
derived from the peculiar reproductive biology of this species. The success of
mating attempts in G. holbrooki is in fact negatively correlated with the male’s size
(Bisazza, 1993; Bisazza & Marin, 1995), a relationship observed in other poeciliid
species and possibly a general phenomenon in this family (Farr et al., 1986).
Conversely, both intrasexual selection and female preference tend to favour the
largest in a group of males (Bisazza & Marin, 1991; 1995). Computer simulations
indicate that when males are rare and the encounter rate between them is low,
males should mature at a small size, but when population density is high and the
sex ratio is male biased, the influence of intra- and intersexual selection increases
and males should delay maturation to attain a large size (Bisazza & Marin, 1995).
It is suggested that by attempting to copulate with females, immature males gain
information about their probable reproductive success at their present body size,
and accordingly ‘ decide ’ whether to delay maturity or not.
In the first part of this paper information is given about the ontogeny of male
sexual behaviour and the interactions between adults and immature males
described in relation to their developmental stage. In the second part, two
experiments aimed at testing the two explanations for the presence of intense
sexual activity in immature males are presented.
MATERIALS AND METHODS
EXPERIMENTAL ANIMALS
The individuals used for this work were captured during the breeding season in three
different bodies of water of Northern Italy. Mostly handnets were used, with a 2·5 mm
728
.    .
mesh that retained specimens of every size. Before starting the experiments, individuals
were acclimatized for some days in 130-l. aquaria (stock aquaria), maintained at a
constant photoperiod (illuminated from 06.00 to 20.00 hours) and temperature
(26&1) C). Fish were fed twice a day with commercial fish food (Tetra Min). For testing
the influence of precocious experience on male sexual behaviour at maturity (Experiment
3), offspring born in captivity to some of the wild females were used. All individuals to
be used in the experiments were anaesthetized with MS222, sexed and measured to the
nearest 0·5 mm total length (..).
MATURATION STAGES OF MALES
Maturation stage of males was determined by examining the development of the
gonopodium under a dissecting microscope. Ten stages were identified (Dulzetto, 1931).
Briefly, the development of the gonopodium starts with the elongation of the 3rd, 4th and
5th rays of the anal fin (stage 1 to 4); at stage 5 the length of the gonopodium is similar
in length to that of mature males, and rays 3 to 5 begin to develop distally bifurcated
articules that, at stages 8 and 9, form a series of hooks. At stage 8 the gonopodium can
be turned forward. At stage 9 the gonopodium, previously opaque, becomes transparent.
At stage 10 the distal articles of the gonopodium become fused. The maturation of the
gonopodium is paralleled by the progressive maturation of the testes, which is completed
at stage 9, but males are able to inseminate females only at stage 10, when the
gonopodium is completely developed (Dulzetto, 1933).
FREQUENCY OF COPULATORY ATTEMPTS IN IMMATURE MALES
One hundred and forty-six males at different stages of sexual maturation were tested.
The day before each test, one male was introduced in one of six 60#36#36 cm aquaria,
each containing two acclimated adult females. Only females that had not delivered young
recently were used, because the frequency of attempts by males is significantly higher with
post-partum females (Bisazza et al., 1989). Each male was observed for two periods of 10
consecutive minutes, at least 30 min apart, and copulatory attempts were counted. At the
end of each test, the males were anaesthetized, measured, and the stage of maturation of
the gonopodium was determined.
FREQUENCY OF AGGRESSIVE BEHAVIOUR OF ADULT RESIDENT MALES
TOWARD IMMATURE MALES
One hundred and thirty-one males in different stages of sexual maturation were tested.
Experimental conditions were similar to those of above, the only difference being that one
adult male (resident) was also present in each experimental aquarium. Thus, each
aquarium contained three adults (one male and two females). The test male (from stage
1 to 10 of maturation) was introduced 1 day before being tested. The resident male was
always larger than the test male. During the two periods of observation recordings were
made of the number of attacks directed by the resident male to the test male, the number
of attacks from the test male to the resident, and the number of copulatory attempts by
the test male. At the end of each test, total length and maturation stage were determined
for both males.
EFFECT OF PRECOCIOUS SEXUAL EXPERIENCE ON THE EFFICIENCY OF
MALE COPULATORY BEHAVIOUR
Thirty males were used. Fifteen of them were grown from birth to sexual maturity in
130-l. aquaria containing adult females (experienced males). The other 15 males were
grown in similar aquaria, but adult females were not present (naïve males).
Upon reaching sexual maturity, each male was tested in aquaria (70#70#36 cm)
containing four adult females. Observations started 1 min after the male was introduced,
and continued for 30 min. Since the success of the copulatory attempts depends strongly
on the ratio of male to female length (Bisazza & Marin, 1991), the four females in the
experimental aquaria were similar in size, and the male to female length ratio was kept as
     
729
constant as possible. Copulatory attempts and successful gonopodial thrusts (observed
contact between gonopodium and female genital pore) were recorded for each male of the
two groups. As a measure of the efficiency of the copulation behaviour of a male, the
ratio between successful gonopodial thrusts and total number of copulatory attempts was
calculated.
EFFECT OF THE SEX-RATIO EXPERIENCED BY IMMATURE MALES ON
THEIR SIZE AT MATURITY
The effect of the sex ratio on male size at maturation was tested using 36 immature
males (stage 1–2), 72 adult males and 90 adult females. The experimental aquaria were
70#70#36 cm in dimension. The natural habitat of G. holbrooki is characterized by the
presence of dense vegetation. Aquaria were therefore subdivided into six sections, using
vertical partitions, in order to simulate the complexity of a natural situation. Partitions
allowed fish to move freely from one compartment to any other. Food was provided
twice a day ad libitum. One immature male (stage 1 or 2) and two adult males, whose size
exceeded by 4&0·5 and 7&0·5 mm that of the immature male, were introduced to each
aquarium. To simulate two situations with different sex ratios, one adult female was
introduced in some of the aquaria (adult male : female ratio=2 : 1, 19 replicates), and
four adult females in the rest of the aquaria (sex ratio 1 : 2, 17 replicates). Total length
of males was measured before the beginning of the experiment by using a microscope
with a micrometric ocular. Specimens were observed under the microscope in a Petri dish
filled with water, thus avoiding anaesthetizing males before the experiment. 15 days after
the immature male matured, it was removed from the aquarium, its total length was
measured and its maturation stage assessed as previously described.
The aquaria used for these experiments did not allow the continuous observation of the
males’ behaviour. To measure the frequency of aggressive and sexual interactions
between immature males and adult males, and between immature males and females,
similar aquaria were used in the middle of a dimly-lit room. Twenty-eight replicates were
run (14 with one female and 14 with four females), using immature males at stages 5 and
6 of gonopodium maturation. For each replicate, immatures were kept under observation for 30 min a day (three periods of 10 min, 10 to 15 min apart from each other), for
three consecutive days. Attacks by adult males towards immature, attacks by the
immature towards the adults, and gonopodial thrusts attempted by the immature were
scored, and the fraction of time that the immature spent in sectors of the aquarium where
one or two adult males were also present were measured.
STATISTICAL ANALYSIS
All variables were normally distributed and parametric tests were used. When variance
was not homogeneous between groups, data were logarithmically transformed. Proportional data were arcsin-root transformed. Statistics were computed with the SPSS/PC
statistical package.
RESULTS
FREQUENCY OF COPULATORY ATTEMPTS IN IMMATURE MALES (Table I)
The size of males was significantly different between groups (ANOVA,
F=17·7; d.f.=9, 145; P<0·001), and increased with maturation stage of the
gonopodium (trend analysis, F=146·2; d.f.=1, 145; P<0·001). However, differences between groups were not significant among males in stages 6 to 10 (a
posteriori analysis, Duncan procedure, P<0·05), and body size differences did
not affect the results of the experiment [ANCOVA, stage (factor): F=17·3;
d.f.=9, 145; P<0·001; total length (covariate): F=1·35; d.f.=1, 145; P=0·24].
.    .
730
T I. Frequency of copulatory attempts by males and total
body length (mean&..) at different stages of gonopodial
maturation. Maturation stages are defined according to Dulzetto
(1931)
Gonopodium
maturation
stage
1
2
3
4
5
6
7
8
9
10
No. of
individuals
Total length
(mm)
Copulatory
attempts
11
19
13
18
15
8
10
9
24
19
21·2&15·21
23·8&16·43
25·1&13·09
25·1&16·52
27·0&17·97
27·8&13·87
27·3&13·98
28·2&18·39
27·6&25·08
26·6&20·87
1·72& 3·19
2·10& 2·88
6·84& 7·55
8·44& 7·22
12·13& 6·79
18·00& 7·63
24·40&11·57
22·77& 8·91
26·45&15·28
33·36&15·27
T II. Frequency of aggression (mean&..) by adult males towards immature males
at different stages of gonopodial maturation and total body length (T.L.) (mean&..)
Gonopodium
maturation
stage
1
2
3
4
5
6
7
8
9
10
Number of
individuals
T.L. of adults
(mm)
T.L. of immatures
(mm)
Aggression
frequency
13
16
13
17
12
10
9
8
14
19
28·9&24·65
28·1&15·75
28·3&18·98
29·0&20·90
29·9&22·44
29·7&11·84
29·4&25·17
28·9&20·77
28·4&20·86
28·2&12·62
20·5&14·21
23·7&15·80
25·0&12·82
25·1&19·58
26·6&15·82
27·5&13·63
26·5&15·70
28·0&19·82
26·9&16·77
25·6&18·78
3·00& 7·14
11·18&24·00
13·84&19·46
25·70&29·14
21·08&22·62
20·60&14·44
52·88&27·96
48·37&24·23
44·28&35·85
49·15&38·41
The frequency of copulatory attempts differed significantly between stages
(ANOVA, F=34·3; d.f.=9, 145; P<0·001), and increased with the progress of
maturation from stage 1 to stage 6 (trend analysis, F=275·5; d.f.=1, 145;
P<0·001). After stage 6, differences between groups were not significant (a
posteriori analysis, Duncan procedure, P<0·05).
FREQUENCY OF AGGRESSIVE BEHAVIOUR OF ADULT MALES TOWARDS
IMMATURE MALES (Table II)
The total length of immature males increased with gonopodium maturation
from stage 1 to 6, but thereafter did not change significantly (ANOVA, F=20·8;
     
731
T III. Effect of precocious sexual experience on the efficiency of copulatory attempts
of adult males (mean number&..)
Naïve males
Experienced males
n
Ratio between
male T.L. and
females T.L.
(mm)
15
15
0·62&0·039
0·63&0·040
Copulatory
attempts
Successful
copulatory
attempts
Frequency of
successful
copulatory
attempts
(%)
136·46&26·53
75·73&33·95
7·26&3·71
4·06&1·33
5·49&2·8
6·01&2·5
d.f.=9, 130; P<0·001, a posteriori analysis, Duncan procedure, P<0·05). No
significant difference in length between resident males was found (ANOVA,
F=1·4, d.f.=9, 130; P=0·19).
Adult males attacked immatures at all stages of maturation of the
gonopodium. Variance between groups for attacks of adult males towards
immatures was not homogenous and so a logarithmic transformation was used.
Aggressive behaviour of adult males towards immatures increased with the
maturation of gonopodium (ANOVA, F=12·4; d.f.=9, 130; P<0·001; trend
analysis, F=105·6; d.f.=1, 130; P<0·001). After stage 7, the number of attacks
by adults towards immatures did not change significantly (a posteriori Duncan
procedure, P<0·05). Attack frequency by immatures towards adults was always
very low (0·2&0·58 min "1) and did not differ significantly between groups
(ANOVA, F=0·80; d.f.=9, 130; ..).
The number of copulatory attempts performed by the immature males
was negatively correlated with the number of attacks received from the adult
male (r= "0·36, n=131, P<0·001). In order to compare males at different stages
of maturation, the number of attempts was standardized as the ratio between
the number of attempts by each immature male at a given stage of maturation,
and the average number of attempts of immature males at the same stage.
Similar results were obtained when the absolute number of attempts by males at
similar stages of maturation were considered (stages 7 to 10, r= "0·49, n=50,
P<0·001).
EFFECT OF PRECOCIOUS SEXUAL EXPERIENCE ON THE EFFICIENCY OF
MALE COPULATORY BEHAVIOUR (Table III)
Experienced and naïve males did not differ in total length (t=0·75, n=30 ..;
Student’s t-test). Naïve males made more copulatory attempts than experienced
males (t=5·46, n=30, P<0·001), and more attempts were successful (t=3·14,
n=30, P=0·003). None the less, copulation efficiency (ratio between successful
attempts and total attempts) did not differ significantly between the two groups
(t=0·54, n=30, ..). If the observation period is divided into three parts of
10 min each (Fig. 1), in the first 10 min of observation experienced males did
better than naïve males (t=2·31, n=30, P=0·028), whereas this difference was
absent in the subsequent two periods (t=0·73, n=30, .., and t=0·48, n=30,
.., respectively).
.    .
732
Successful copulatory attempts (%)
10
P < 0.05
N.S.
N.S.
N.S.
7.5
5.0
2.5
To
ta
l
in
m
0
–3
20
10
0–
–2
10
0
m
m
in
in
0.0
F. 1. Frequency of successful gonopodial thrusts (mean with 1 ..) of males with precocious sexual
experience (/) compared with that of males grown in the absence of females (.).
T IV. Effect of sex ratio during male development on the body size (mean&..)
No. of
females
1
4
No. of
immatures
tested
Large male
Small male
Stage 1 to 2
Stage 10
Total length
increase
(%)
19
16
24·85&2·66
25·00&2·51
22·12&2·05
21·78&1·47
17·48&2·08
17·68&1·57
25·86&2·21
23·95&1·57
49·4&18·8
35·9&12·9
T.L. of adults (mm)
T.L. of immature (mm)
T V. Mean number (&..) of aggressive interactions with adult males, of copulatory attempts,
and of time spent in the same sector where at least an adult male was also present for immature males
in different sex ratio conditions
Adult
sex ratio
M:F
2:1
1:2
No. of
females
No. of
replicates
1
4
19
16
Aggressive interactions
Attacks by
adults
Attacks by
immature
41·7&26·4
37·9&23·3
6·2&11·5
8·7&13·9
Total
Time spent
with adult
male(s)
Copulatory
attempts
47·9&24·0
46·6&26·6
26·4&10·2
20·7&10·1
3·6& 5·2
13·1&11·7
Immature males used for this experiment were at stages 5 and 6 of their gonopodium maturation.
EFFECT OF THE SEX RATIO EXPERIENCED BY IMMATURE MALES ON
THEIR SIZE AT MATURITY (Tables IV and V)
Immature males that grew in aquaria with a low sex ratio increased on average
their body size at maturation by about 35% of their original size. Males that
grew in aquaria with a high sex ratio increased their size by about 50%; the
     
733
difference between the groups was highly significant (t=2·84, n=35, P<0·01,
Table IV). In 13 out of 19 cases, males grown in aquaria with a high sex ratio
matured at a size which was greater than that of both adult males present in the
same aquarium. Only six out of 17 immatures in the aquaria with a low sex ratio
matured at a size larger than that of the largest adult male. The difference was
only marginally significant (÷2 =2·74, P<0·10).
Aggressive interactions between immature (stage 5 and 6) and adult males did
not differ in aquaria with low and high sex ratios, in terms of attacks received
(t=0·4, n=30, ..), attacks performed (t=0·51, n=30, ..), and total number of
attacks (t=0·13, n=30, .., Table V). In addition, the time spent by immatures
in sectors of the aquaria where at least one adult male was present did not differ
in the two conditions (t=1·49, n=30, ..). However, immatures in aquaria with
a low sex ratio made a higher number of copulatory attempts than those in
aquaria where the sex ratio was high (t=2·75, n=30, P=0·01).
DISCUSSION
The results demonstrate that a male’s sexual behaviour (copulatory attempts)
starts well before it is able to fertilize females. On average, gonopodium
maturation from stage 1 to 10 takes about 20 to 30 days in the laboratory
(Zulian, 1990). Even during very early stages of maturation some sexual activity
was observed, and the frequency of attempts increased during the first five stages.
From stage 6, immatures were approximately as active as adults. The dynamics
of copulatory attempts were roughly constant during the entire maturation
period, the main difference being that the forward rotation of the gonopodium
occurred only after the 8th stage. As sexual activity of immature males
increased, they received an increasing number of attacks by adult males. From
stage 6 of gonopodial maturation onward, the frequency of attacks by adults did
not change significantly. It therefore seems that the frequency of attacks by
adult males depended on the sexual activity of the immatures and not on
gonopodial maturation stage, and that adult males do not discriminate between
immature males that are incapable of fertilizing the females and other adult
males. They apparently treat all males which are attempting to copulate with
females as possible competitors.
It is difficult to find an explanation for the high rate of copulatory attempts in
immature males. In theory, this behaviour could be non-adaptive, a consequence, perhaps, of increasing levels of blood androgen hormones. Precocious
copulatory activity can be induced by administration of testosterone in fish
(Liley & Seghers, 1975) and other vertebrates (Stone, 1940; Andrew, 1975).
Adult sexual behaviour can be elicited also in young of several birds and
mammals, but in most cases it appears to be a reflex response to peculiar
circumstances, and is observed very infrequently under natural conditions
(Andrew, 1966). A non-adaptive explanation is difficult to test, but in this case
it seems at least unlikely. During the breeding season adult male G. holbrooki
make on average one copulatory attempt per min, throughout the whole day
(Martin, 1975; Bisazza & Marin, 1991). It seems likely that such an intense
activity is energetically very expensive. Moreover, males engaging in sexual
activities are more visible, pay less attention to approaching predators (Endler,
734
.    .
1982), and probably suffer high mortality. Indirect evidence for these costs has
been found by Krumholz (1963) in populations of Gambusia manni (Hubbs). He
found that populations living in habitats with a high density of predators had a
sex ratio biased towards females, and a low mean size of adult males, compared
to those observed in populations living in habitats with low predator density.
Among the possible adaptive explanations is the ‘ training ’ hypothesis.
Copulatory experience during the stages preceding maturation may enhance the
copulatory efficiency of males. As far as we know, in fish there are no examples
of experience enhancing performance. However, it is known that young threespined sticklebacks (Gasterosteus aculeatus L.) fed in the laboratory with live
prey are more efficient foragers when they become adults than individuals fed
with frozen food (Ibrahim & Huntingford, 1992). Our results do not seem to
support an analogous explanation for copulatory activity. In fact, mosquitofish
males with precocious experience were more efficient in their copulatory attempts
than males grown in aquaria where females were not present, but only in the first
10 min after meeting a female. Later on, naïve males were as efficient as
experienced males. It seems therefore unlikely that immature males are sexually
so active for weeks in order to get trained, if after a few minutes of copulatory
attempts naïve males become capable of normal sexual behaviour.
The other possible explanation tested is that immature males, through an
intense sexual activity, gain information about the probable reproductive success
they would attain by maturing at their actual size. As in most poeciliid species,
male mosquitofish stop growing or grow little after reaching sexual maturity, but
considerable variation in this trait is generally found within populations.
Previous studies indicate that in G. holbrooki the success of copulatory attempts
is inversely related to male size when competition is absent (Bisazza, 1993;
Bisazza & Marin, 1995). On the contrary, when sex ratio is biased towards
males, larger males monopolize females and keep subordinate males far from
females (Bisazza & Marin, 1991). The reproductive success of males therefore
depends critically on the size at which they mature in relation to the social
environment: at low population densities and when the sex ratio is biased
towards females, males are expected to mature at a small size, and vice versa, as
shown by computer simulations and field study (Bisazza & Marin, 1995; Zulian
et al., 1995). In natural populations, the proportion of males and population
density usually increase from the beginning of the breeding season onwards.
None the less, there are frequent large variations both between and within
populations in different years (Zulian, 1990). A simple rule of thumb, e.g.
mature small at the beginning of the breeding season and at a larger size later on,
would not be very efficient.
Immature males that are able to test their potential reproductive success and
consequently ‘ decide ’ whether to mature or to keep growing, would therefore be
favoured. The present results seem to support this hypothesis: the size attained
at maturity by males is influenced by the frequency of sexual interactions with
females. Social control of body size at maturity is a well known phenomenon in
poeciliids. It has been shown that the frequency of aggressive interactions with
adult males influences the size of sexual maturation of males in other poeciliid
species (Borowsky, 1973b; Sohn, 1977; Farr, 1980), as well as in G. holbrooki
(Zulian et al., 1993). In the present experiments, the frequency of aggression by
     
735
adult males did not change signficantly between treatments (with one or four
adult females). None the less, the body size attained at maturation by males was
negatively correlated with the frequency of copulatory attempts, that, in turn,
depended on the sex ratio in the aquarium where they grew. These results
therefore suggest that social control mechanisms, at least in G. holbrooki, may be
more complicated, and perhaps more precise, than previously thought. In fact,
it could be argued that, for a maturing male, aggressive behaviour by adult males
may not be a precise measure of his possible reproductive success. During the
breeding season, adult males spend most of their time around females, attempting to copulate or trying to keep other males away from them (McPeek, 1992;
Bisazza & Marin, 1995). For this reason, the best way for an immature male to
test his potential reproductive success is by trying to copulate with females, and
by gaining information about copulations attempted, frequency of successful
copulations and of attacks received. Preliminary studies (Bisazza, unpublished
data) suggest that immature males also show sexual behaviour in other
poeciliids. In three species, Belonesox belizanus (Kner), Girardinus falcatus
(Eigenmann) and Phalloceros caudimaculatus (Hensel), sexual behaviour was
observed at an early stage of gonopodium development. Conversely, in Poecilia
reticulata (Peters) and in Heterandria formosa (Agassiz) males started to be
sexually active from 5 to 10 days after the gonopodium was completely
developed. Histological examination of the gonads showed that mature sperms
were present in P. reticulata and in H. formosa, but not in G. falcatus and
P. caudimaculatus. Interestingly, body size is the major factor determining the
outcome of male contests in G. holbrooki and in the other four species that we
found to have precocious sexual behaviour (Bisazza & Marin, 1995), but not in
the two species lacking it (Kodric-Brown, 1983; Zannini, 1994).
We would like to thank A. Kodric-Brown for helpful suggestions to a previous version
of the manuscript. This research was partially supported by a MURST 60% grant to
A. Bisazza.
References
Andersson, M. (1983). On the functions of conspicuous seasonal plumages in birds.
Animal Behaviour 31, 1262–1264.
Andrew, R. J. (1966). Precocious adult behaviour in the young chick. Animal Behaviour
14, 485–500.
Andrew, R. J. (1975). Effects of testosterone on the behaviour of the domestic chick. II.
Effects present in males but not in females. Animal Behaviour 23, 139–155.
Bisazza, A. (1993). Male competition, female mate choice and sexual size dimorphism in
poeciliid fishes. Marine Behaviour and Physiology 23, 257–286.
Bisazza, A., Marconato, A. & Marin, G. (1989). Male mate preferences in the
mosquitofish Gambusia holbrooki. Ethology 83, 335–343.
Bisazza, A. & Marin, G. (1991). Male size and female mate choice in the eastern
mosquitofish (Gambusia holbrooki: Poeciliidae). Copeia 1991, 730–735.
Bisazza, A. & Marin, G. (1995). Sexual selection and sexual size dimorphism in the
eastern mosquitofish Gambusia holbrooki (Pisces: Poeciliidae). Ethology, Ecology
and Evolution 7, 169–183.
Borowsky, R. L. (1973a). Relative size and the development of fin coloration in
Xiphophorus variatus. Physiological Zoology 46, 22–28.
Borowsky, R. L. (1973b). Social control of adult size in males Xiphophorus variatus.
Nature 245, 332–335.
736
.    .
Clutton-Brock, T. H. (1991). The Evolution of Parental Care. Princeton, NJ: Princeton
University Press.
Constantz, G. D. (1975). Behavioural ecology of mating in the male Gila topminnow,
Poeciliopsis occidentalis (Cyprinodontiformes: Poeciliidae). Ecology 56, 966–973.
Dulzetto, F. (1931). Sviluppo e struttura del gonopodio di Gambusia holbrooki. Pubblicazioni della Stazione Zoologica di Napoli 11, 1–85.
Dulzetto, F. (1933). La struttura del testicolo di Gambusia holbrooki (GRD) e la sua
evoluzione in rapporto con lo sviluppo del gonopodio. Archivio Zoologico Italiano
19, 405–431.
Endler, J. A. (1982). Convergent and divergent effects of natural selection on colour
patterns in two fish faunas. Evolution 36, 178–188.
Farr, J. A. (1980). The effect of juvenile social interaction on growth rate, size and age
at maturity, and adult social behaviour in Girardinus metallicus Poey (Pisces:
Poeciliidae). Zeitschrift für Tierpsychologie 52, 247–268.
Farr, J. A. (1984). Premating behaviour in the subgenus Limia (Pisces: Poeciliidae):
sexual selection and the evolution of courtship. Zeitschrift für Tierpsychologie 65,
152–165.
Farr, J. A., Travis, J. & Trexler, J. C. (1986). Behavioural allometry and interdemic
variation in sexual behaviour of the sailfin molly, Poecilia latipinna (Pisces:
Poeciliidae). Animal Behaviour 34, 497–509.
Ford, C. S. & Beach, F. A. (1951). Patterns of Sexual Behaviour. New York: Harper &
Brothers.
Gronell, A. M. (1989). Visiting behaviour by females of the sexually dichromatic
damselfish, Chrysiptera cyanea (Teleostei: Pomacentridae): a probable method of
assessing male quality. Ethology 81, 89–122.
Ibrahim, A. A. & Huntingford, F. A. (1992). Experience of natural prey and feeding
efficiency in three-spined sticklebacks (Gasterosteus aculeatus L.). Journal of Fish
Biology 41, 619–l625.
Kodric-Brown, A. (1983). Determinants of male reproductive success in pupfish
(Cyprinodon peconensis). Animal Behaviour 31, 128–137.
Krumholz, L. A. (1963). Relationships between fertility, sex ratio, and exposure to
predation in populations of the mosquitofish Gambusia manni Hubbs at Bimini,
Bahamas. Internationale Revue der gesamten Hydrobiologie 48, 201–256.
Liley, N. R. & Seghers, B. H. (1975). Factors affecting the morphology and behaviour
of guppies in Trinidad. In Function and Evolution in Behaviour (Baerends,
G. B., Beer, C. & Manning, A., eds), pp. 92–118. Oxford: Oxford University
Press.
Marconato, A. (1991). Life history tactics and population structure in freshwater fish
species. In Form and Function and Zoology (Lanzavecchia, G. & Valvassori,
R., eds). Selected Symposia and Monographs U.Z.I., pp. 379–393. Modena:
Mucchi.
Martin, R. G. (1975). Sexual and aggressive behavior, density and social structure in a
natural population of mosquitofish, Gambusia affinis holbrooki. Copeia 1975,
445–454.
Mason, W. A. (1965). The social development of monkeys and apes. In Primate
Behavior, Field Studies of Monkeys and Apes (DeVore, I., ed.), pp. 514–543. New
York: Holt, Rinehart & Winston.
McPeek, M. A. (1992). Mechanisms of sexual selection operating on body size in the
mosquitofish (Gambusia holbrooki). Behavioural Ecology 3, 1–12.
Schultz, R. J. (1961). Reproductive mechanisms of unisexual and bisexual strains of the
viviparous fish Poeciliopsis. Evolution 15, 302–325.
Sohn, J. J. (1977). Socially induced inhibition of genetically determined maturation in the
platyfish, Xiphophorus maculatus. Science 195, 199–201.
Stone, C. P. (1940). Precocious copulatory activity induced in male rats by subcutaneous
injections of testosterone propionate. Endocrinology 26, 511–515.
Zannini, L. (1994). Selezione sessuale e dimorfismo sessuale nei pecilidi. Doctoral
Thesis, University of Padova.
     
737
Zulian, E. (1990). Selezione e dimensioni nei maschi di Gambusia holbrooki (Pisces:
Poeciliidae). Analisi dei meccanismi che determinano la lunghezza alla maturita.
PhD Thesis, University of Padova.
Zulian, E., Bisazza, A. & Marin, G. (1993). Determinants of size in male eastern
mosquitofish (Gambusia holbrooki): Inheritance and plasticity of a sexual selected
character. Bollettino di Zoologia 60, 317–322.
Zulian, E., Bisazza, A. & Marin, G. (1995). Variations of male body size in natural
populations of Gambusia holbrooki. Ethology, Ecology and Evolution 7, 1–10.