Courtship as an honest indicator of
male parental quality in the
bicolor damselfish, Stegastes partitus
M
ales of many animal species have elaborate secondary sexual characters that
appear to be detrimental to their survival (see
Darwin, 1871). Increasing evidence suggests
that these ornaments often arise as a result of
sexual selection via female choice (see Bateson, 1983; Bradbury and Andersson, 1987 for
recent reviews). However, the mechanisms by
which female choice results in the evolution
of elaborate male characters are still a matter
of considerable debate (Bradbury and Andersson, 1987; Partridge and Harvey, 1986).
Sexual selection theory suggests two very
different mechanisms for the evolution of
elaborate ornaments in males. "Arbitrary
trait" models are based on the ideas of Fisher
(1930) and propose that the male characters
chosen by females are arbitrary and that the
coevolution of female preference and male
trait causes a runaway exaggeration of the trait.
The theoretical plausibility of such a runaway
"Fisher process" has since been confirmed
(Kirkpatrick, 1982; Lande, 1981; O'Donald,
1980). Empirical support for such a process,
however, is virtually nonexistent and may be
extremely difficult to obtain.
Alternatively, "adaptive" or "signaling"
models of sexual selection propose that female choice is nonarbitrary. These models
propose that females choose mates on the basis of characters that signal male quality, and
matings with high-quality males result in direct fitness benefits to females. These benefits
may result from increased offspring viability
(Andersson, 1986; Hamilton and Zuk, 1982;
Pomiankowski, 1988; Zahavi, 1975), female
fecundity (Grafen, 1990b), or offspring survival (Maynard Smith, 1985). Male quality may
be a function of a male's genotype (Andersson, 1986; Hamilton and Zuk, 1982; Pomiankowski, 1988; Zahavi, 1975) or phenotype
(Grafen, 1990b) and is advertised by the degree of ornament elaboration.
A critical aspect of these adaptive models is
that male advertisement must be honest. Although it remains debatable whether signals
can be immune from cheating (Dawkins and
Krebs, 1978; Krebs and Dawkins, 1984; Zahavi, 1975), recent population genetic (Grafen, 1990b) and evolutionarily stable strategy
(ESS) models (Grafen, 1990a) suggest that
honest signals are theoretically possible. In
these models, low-quality males do not "cheat"
by increasing their level of advertisement to
match that of high-quality males because a
given level of advertisement is assumed to be
more costly (e.g., in terms of survivorship) for
males of low quality than for males of high
quality. Therefore, although low-quality males
are capable of increasing their level of advertisement, they do not do so because this increase would not result in higher lifetime reproductive success.
Although several studies of mate choice have
suggested that signals serve as honest indicators of male quality (see below), few (e.g.,
Knapp and Kovach • Courtship and male parental quality
Roland A. Knapp
Department of Biological
Sciences and Marine
Science Institute,
University of California,
Santa Barbara, CA 93106,
USA
John T. Kovach
Neushul Mariculture,
475 Kellogg Way,
Goleta, CA 93117,
USA
Received 13 March 1991
Revised 27 May 1991
Accepted 28 May 1991
1045-2249/91/$4.00
© 1991 International Society
for Behavioral Ecology
295
Downloaded from http://beheco.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 10, 2016
Previous work on the bicolor damselfish, a species with exclusive male parental care of eggs,
suggested that female mate choice was based on male characteristics. The aims of this study
were to determine whether females discriminate among potential mates on the basis of courtship
and, if so, to determine whether courtship serves as an indicator of male parental quality.
Observations made over two reproductive cycles showed that courtship rates and mating success
of individual males are positively correlated and that males begin courting several days before
females begin laying eggs. Experimental manipulations showed that a male's courtship rate is
indicative of the subsequent egg survival from his nest. We suggest that observed differences
between males in their courtship rates and parental ability may be a result of differences in
their energy reserves. These results demonstrate the operation of honest advertising and lend
support to adaptive models of sexual selection. [Behav Ecol 1991;2:295-300]
296
Behavioral Ecology Vol. 2 No. 4
Only male courtship rate explained a significant amount of the variance in male mating success, and courtship rate and mating
success were positively correlated (Knapp and
Warner, 1991). This is in agreement with the
results of an earlier study of female mate choice
in the bicolor damselfish (Schmale, 1981). It
remains unclear from these results, however,
whether females receive any fitness benefits
by mating with high-courting males.
The objective of the present study, therefore, was to determine whether courtship rate
serves as an indicator of male parental quality.
We sought first to confirm the positive correlation between male courtship rate and mating success and to determine whether males
begin courting before the start of a reproductive cycle (i.e., before females begin laying
eggs). If male courtship precedes the start of
a reproductive cycle instead of beginning concurrently or afterward, females may be able
to use courtship to distinguish among potential mates before starting to lay eggs.
Having confirmed that courtship rate and
male mating success are positively correlated
and that males begin courting before females
begin laying eggs, we used an experiment to
test the prediction that courtship displays performed by males before the start of a reproductive cycle accurately reflect male parental
quality, i.e., that male courtship rate and subsequent egg survival are positively correlated.
Downloaded from http://beheco.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 10, 2016
Meller, 1990) provide the direct evidence necessary to conclude that offspring benefit from
the female preference for elaborate male
characters. In guppies and house finches, females choose males with the greatest amount
of orange carotenoid pigmentation (Endler,
1980; Hill, 1990; Kodric-Brown, 1989). Because carotenoids cannot be synthesized and
must be acquired through food, these authors
suggested that the degree of a male's orange
pigmentation may be an indicator of his feeding ability. However, whether or not male
feeding ability can have consequences for offspring fitness remains to be shown. In addition, several researchers investigating mate
choice in birds have suggested that some secondary sexual characters (courtship, conspicuous plumage) may advertise male parental
ability (Eckert and Weatherhead, 1987; Norris, 1990b). However, because these studies
provided only correlative evidence (e.g., males
with high courtship rates fed their offspring
at higher rates; Eckert and Weatherhead,
1987), it is difficult to eliminate the possibility
that intrinsic differences in brood quality gave
rise to the observed differences in male parental care quality. For example, if females
vary in the quality of offspring they produce
(e.g., size of offspring at hatching), male provisioning rates may be a function of offspring
quality and not intrinsic male quality. In the
present study we remove this potentially confounding factor by randomizing eggs among
nests and show that differences in egg survival
are due to intrinsic differences in male parental quality.
In the bicolor damselfish, Stegastes partitus,
males provide exclusive parental care of eggs.
Previous research on this species showed that
male size and the clutch area in nests are unimportant in female mate choice and egg survival (Knapp and Warner, 1991). Other researchers have also found no correlation
between male size and mating success (Itzkowitz and Makie, 1986; Sikkel, 1988) or between male size and egg survival (Petersen,
1990). These results are in marked contrast
to numerous other studies of female mate
choice in fishes with paternal care (e.g., Bisazza and Marconato, 1988; Cote and Hunte,
1989; Downhower and Brown, 1980; Gronell,
1989; Knapp and Sargent, 1989; Sargent,
1988). Male size may not influence egg survival in the bicolor damselfish because nest
takeovers by larger conspecific males, a frequent source of egg mortality in other fishes
with paternal care (e.g., Bisazza and Marconato, 1988; Gronell, 1989), are extremely uncommon. Egg survival may not be affected by
the clutch area in nests because most egg losses
occur at night when males are not defending
eggs.
METHODS
Study species
Bicolor damselfish are a common, sexually
monomorphic Caribbean reef fish (Randall,
1968) in which both males and females defend
permanent territories (Myrberg, 1972;
Schmale, 1981). Reproduction follows a lunar
cycle with spawning beginning a few days before the full moon and ending a few days after
the new moon (Knapp RA, unpublished data;
Robertson et al., 1988). During a reproductive cycle, females spawn every 2 days and
show high mate fidelity (Knapp and Warner,
1991). Spawning occurs during a 1-h period
starting at dawn during which females travel
to and lay eggs in male territories. Males have
never been seen to use alternative mating tactics (e.g., sneaking) during more than 250 h
of spawning observations (Knapp RA, personal observation). Eggs are demersal and are laid
in a discrete monolayer. Nests often contain
clutches in various stages of development because males are sequentially polygynous. Males
guard eggs until the eggs hatch as planktonic
larvae, 3.5 days after laying (Knapp RA, personal observations; Robertson et al., 1988).
In addition to defending eggs, males produce
vigorous courtship displays throughout the
day. During courtship, a male leaves his territory and swims toward a female while performing a series of rapid vertical dives termed
"dips" (Myrberg, 1972).
Study methods
We conducted the study over four reproductive cycles between May and August, 1990, on
the Teague Bay forereef, St. Croix, U.S. Virgin Islands (17°45' N, 64°35' W). The 680-m2
study site was delineated by the territories of
approximately 25 males and 30 females. Before the first reproductive cycle, we replaced
natural nests of all males with small clay flowerpots. The replacement of natural nests with
artificial nests does not appear to influence
female mate choice in this species (see Knapp
and Warner, 1991). All fish were individually
marked with subcutaneous injections of acrylic paint (Thresher and Gronell, 1978). To determine whether courtship rate and mating
success were correlated, we monitored the
courtship rates and mating success of 24 males
during the first and second reproductive cycles. We measured courtship rates daily for
males chosen at random by counting the number of dips performed per 10-min period. We
measured mating success daily by tracing the
perimeter of any new egg clutches onto transparent acetate sheets and later digitizing the
clutch areas (see Knapp and Warner, 1991,
for details). We calculated the mating success
of individual males by summing the areas of
all clutches received during a reproductive cycle. The courtship rate of individual males was
calculated by averaging the number of dips
per 10 min over all focal samples per reproductive cycle. Within a reproductive cycle, the
number of focal samples per male ranged from
5 to 19 (x ± SE= 11.5 ± 0.67).
To determine whether females could assess
males on the basis of courtship rate before
the start of a reproductive cycle, we compared
the temporal sequence of courtship with that
of egg laying over the first and second reproductive cycles. If males begin courting several
days before females begin laying eggs, females
would have the opportunity to distinguish
among potential mates on the basis of courtship.
To determine whether courtship rates of
males were indicative of subsequent egg survival in their nests, we performed an experiment in which males with differing courtship
rates were given standard presentations of
eggs to defend. This experiment consisted of
two manipulations, one at the beginning of
the third and one at the beginning of the fourth
reproductive cycle. For each manipulation, we
first measured the courtship rates of 16 males
by counting the number of dips performed
per 10-min period on each of 2 days before
the start of each reproductive cycle (Nlolal =
32). Nine of these 16 males were then chosen
for each manipulation to represent the range
of available courtship rates (Ntot3i = 18).
Measuring courtship rates before the onset
of a reproductive cycle allowed us not only to
determine whether females might be able to
assess male parental quality before laying eggs,
but also eliminated the possibility that courtship rates were influenced by the presence or
absence of eggs in nests, as occurs in the threespined stickleback, Gasterosteus aculeatus (Jamieson and Colgan, 1989).
On the second day of each cycle and just
after spawning ceased (0730 h), we obtained
nine nests from nonexperimental males, each
containing one newly laid clutch of eggs, which
we then substituted for those of the nine experimental males. Because eggs adhere tightly
to the nest surface immediately after being
laid, we could pick up and move nests among
males without disturbing the eggs. In addition, preliminary manipulations showed that
males readily defended unrelated eggs. Because egg survival in this species is not influenced by clutch size when clutches are > 10
cm2 (Knapp and Warner, 1991), males were
given clutches measuring >10 cm2. Because
egg losses from foster clutches occurred only
from around the clutch perimeters, we measured egg losses by tracing clutch perimeters
daily until the eggs hatched. Any new clutches
laid into the 18 nests during the experiment
were removed immediately after the spawning
period. If courtship serves as an indicator of
a male's parental ability, courtship rate and
egg survival should be positively correlated.
Of the 18 males used in the experiment, we
used two males in both manipulations. Both
males were observed courting during one reproductive cycle and not during the other.
This provided an opportunity to determine
the degree to which courtship served as an
indicator of the parental quality of the same
males over time. One male was never observed
to court during cycle 3 but courted during
cycle 4, and the second male courted during
cycle 3 but failed to court during cycle 4.
By randomizing clutches among males, we
ensured that any observed pattern of egg loss
did not result from differences in brood quality giving rise to differences in the quality of
male parental care. However, at the time of
the nest replacements, 12 of 13 courting males
were defending eggs, whereas 0 of 5 noncourting males were defending eggs. This
could potentially bias the results if egg survival
were influenced by differences in parental motivation between males that initially had and
did not have eggs. This did not appear to be
Knapp and Kovach • Courtship and male parental quality
297
Figure 1
Mating success (cm2 of eggs
received) versus courtship rate
(mean number of dips/10
min) for 24 male Stegastes
partitus during two
reproductive cycles. There are
six points at the origin.
Mating success and courtship
rate are significantly positively
correlated.
8
10
12
Courtship rate (mean no. of dips/10 mm)
RESULTS
Male courtship rates and mating success were
significantly positively correlated during both
reproductive cycles (Figure 1; cycle 1: N= 24,
F 1 2 2 = 81.5, R2 = .78, p < .0001; cycle 2: N
= 24, F, 22 = 19.9, R2 = .45, p < .0002; both
cycles combined: N = 48, F M 6 = 128.2, R2 =
.73, p < .0001; data for cycle 1 and for combined cycles were square-root transformed).
This relationship remained highly significant
when noncourting males were excluded from
the analysis (cycle 1: N = 20, F, 18 = 40.1, R2
= .67, p < .0001; cycle 2: N = 22, F1-20 =
10.9, R2 = .32, p < .004; both cycles combined: N = 42, F M 0 = 59.8, R2 = .59, p <
.0001; data for cycle 1 and for combined cycles were square-root transformed). These results suggest that females mate preferentially
with males with high courtship rates.
During both cycles, males began courting
Figure 2
Temporal relationship
between courtship rate (dips/
10 min) and the clutch area
laid (cm2) for S. partitus. Each
filled circle represents the
average courtship rate of all
males who were observed on a
particular day. Each filled
square represents the total
clutch area laid by all females
on the study site on a
particular day.
DISCUSSION
Downloaded from http://beheco.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 10, 2016
initially did not have eggs successfully hatched
his foster clutch. Furthermore, preliminary
nest replacement experiments involved three
courting males with empty nests, and these
males also hatched their foster clutches. Any
remaining bias was eliminated by analyzing the
data with and without the five noncourting
males.
4 to 6 days before females began spawning
(Figure 2). Courtship rates measured for males
before the start of a reproductive cycle were
significantly positively correlated with male
courtship rates measured over the first 5 days
of that reproductive cycle (JV = 18, r = .82, p
< .0001). Therefore, females could use male
courtship rate to discriminate among potential mates before the onset of a reproductive
cycle (i.e., before females begin laying eggs).
In the nest replacement experiments, all
males began defending their foster clutches
immediately after the manipulation. Male
courtship rate and egg survival were positively
related (Figure 3). This relationship was highly
significant regardless of whether the noncourting males were included in the analysis
(one-tailed Spearman rank test; all males: N
= 18, rs=.92,p < .0005; courting males only:
N = 13, r, = .78, p < .005). Five of 7 males
courting at rates above 4 dips/10 min hatched
100% of their foster eggs, but none of the
males courting at lower rates succeeded in
hatching more than 85%. The two males used
in both manipulations hatched similar proportions of their clutches to other males
courting at comparable rates (Figure 3). The
size of clutches given to males did not influence egg survival (cycle 3: N = 9, r = —.52,
p > .l;cycle4:N=9, r = -.01,p > .l;both
cycles combined: N = 18, r = — .15, p > .1).
These results strongly suggest that by mating
with males with higher courtship rates, females could increase the survivorship of their
offspring.
The results of this study show that females
mate preferentially with high-courting males
and that high-courting males hatch a greater
proportion of their eggs than males courting
at low rates. Therefore, a male's courtship rate
appears to serve as an honest advertisement
of his parental ability. This is in accordance
with the predictions of the "strategic handicap" model (Grafen, 1990b) in which males
use ornaments to voluntarily convey information to females, and advertising level is a
continuous and increasing function of male
quality.
Several other empirical studies of mate
choice lend support to the idea that advertisement may signal male quality. Female great
tits (Parus major) mate preferentially with males
with more conspicuous plumage (Norris,
1990a). Males with conspicuous plumage were
more likely to defend their broods, and matings with these males resulted in heavier fledglings, apparently because conspicuous males
delivered food at higher rates or of higher
quality to nestlings (Norris, 1990b). A similar
situation may exist in red-winged blackbirds
i
80 -
•J
V
60 -
•
072)
40 20-
/
02
4
6
8
10
Courtship rate (mean no. of dips/10 min)
were cannibalized by the guarding male within
6 h of the manipulation (the other three
clutches were eaten by neighboring S. diencaeus, against whom the much smaller 5. partitus were unable to defend their nests). Therefore, females may discriminate against
noncourting and low-courting males because
these males have low energy reserves and augment their food intake by eating eggs.
The possibility that male parental quality is
a function of energy reserves is further supported by preliminary results obtained from
fat extraction of males with known courtship
rates. Despite a small sample size, male fat
content and courtship rate were strongly and
positively correlated (iV=5,r=.97,/><.01;
fat extractions were carried out by R. C. Sargent at the University of Kentucky who had
no prior knowledge of the courtship rates of
each fish). Future experiments in which male
fat content is increased by supplemental feeding will aid in elucidating the role of energy
reserves in determining male parental quality.
These experiments may also provide insights
into what prevents low-courting males from
cheating and increasing their rates of advertisement.
From our study, we conclude that male bicolor damselfish accurately signal their parental quality by the vigor of their courtship displays. Females mate preferentially with
vigorously courting males, and such males have
higher offspring survival from their nests. In
addition, the observed variation among males
in parental quality may be a result of differences in their energy reserves. Our results
demonstrate the operation of honest advertisement and lend support to adaptive models
of sexual selection.
This work was supported by grants from the Lerner-Gray
Fund for Marine Research (American Museum of Natural
History) and Sigma Xi to R.A.K. and NSF grant BSR 8704351 to R. Warner. We thank R. C. Sargent for carrying
out the fat extractions and R. Warner, J. Endler, S. Rothstein, E. Schultz, C. St Mary, J. Caselle, A. Basolo, G.
Rosenqvist, and J. Reynolds for helpful comments on
earlier versions of the manuscript.
Knapp and Kovach • Courtship and male parental quality
l
g
was fitted using the NLIN
procedure of SAS software
(1988). Arrows indicate the
courtship rates and
percentage of egg survival of
two males used during cycle 3
and during cycle 4.
Downloaded from http://beheco.oxfordjournals.org/ at Penn State University (Paterno Lib) on May 10, 2016
courtship rates provision nestlings at higher
rates (Eckert and Weatherhead, 1987).
However, with the exception of the present
study, the only study of honest advertisement
we have found that eliminated any potential
female contribution to brood quality was that
by Moller (1990) in which he conducted partial cross-fostering experiments in barn swallows (Hirundo rustica). By exchanging a randomly chosen half of the offspring in one nest
with a randomly chosen half of the offspring
in another nest, Moller was able to determine
the male and female contribution to offspring
fitness. His experiments showed that male tail
length is negatively correlated with the intensity of infection by hematophagous mites, that
resistance to mites is at least partially heritable, and that matings with long-tailed males
produce offspring with higher mite resistance
(Moller, 1990). Therefore, tail length appears
to be an honest indicator of a male's genetic
quality, and females use tail length to discriminate among potential mates (Meller, 1988).
Although the results of the present study
show that male bicolor damselfish use courtship to advertise their quality, it is important
to determine the proximate basis for the observed differences in parental ability between
low- and high-courting males. Male parental
care in teleost fishes is often energetically expensive (DeMartini, 1987; Pedersen, 1979,
cited in Dominey and Blumer, 1984; Unger,
1983), and males with low energy reserves may
be unable to provide high-quality parental care
to eggs (Unger and Sargent, 1988). The possibility that the parental quality of male bicolor damselfish is a function of energy reserves is suggested by the patterns of clutch
loss in the nest replacement experiment. Foster clutches that suffered partial egg losses
during the experiment always lost eggs only
from around the clutch perimeter (clutches
in natural nests are also much more likely to
lose eggs from the perimeter than from within
the clutch; Knapp RA, unpublished data). Although these egg losses may be attributable
to the guarding male, small heterospecific
fishes also occasionally eat S. partitas eggs
(Knapp and Warner, 1991). However, in contrast to the observed pattern of clutch loss
from experimental nests, predation by heterospecifics generally results in the haphazard
removal of several clumps of eggs from inside
the clutch periphery (Knapp RA, personal observation). This difference suggests that partial egg losses from experimental nests are the
result of guarding males selectively eating portions of their clutches. In addition, two of the
five foster clutches given to noncourting males
m
Eggs urvivai (%)
(Agelaiicsphoeniceus), in which males with high
Figure 3
Percentage of egg survival
versus courtship rate (mean
number of dips/10 min) in 5.
partilus. There are five points
at the origin. Y= 100 (1 -
-*-
299
Krebs JR, Dawkins R, 1984. Animal signals: mind reading
and manipulation. In: Behavioural ecology: an evolutionary approach (Krebs JR, Davies NB, eds). Oxford:
Andersson M, 1986. Evolution of condition-dependent
Blackwell; 380-402.
sex ornaments and mating preferences: sexual selection
based on viability differences. Evolution 40:804-816.
Lande R, 1981. Models of speciation by sexual selection
on polygenic traits. Proc Natl Acad Sci USA 78:3721Bateson P, 1983. Mate choice. Cambridge: Cambridge
3725.
University Press.
Bisazza A, Marconato A, 1988. Female mate choice, maleMaynard Smith J, 1985. Sexual selection, handicaps and
male competition and parental care in the river bulltrue fitness. J Theor Biol 115:1-8.
head, Coitus gobio L. (Pisces, Cottidae). Anim Behav 36: Moller AP, 1988. Female choice selects for male sexual
1352-1360.
tail ornaments in the monogamous swallow. Nature
332:640-642.
Bradbury JW, Andersson MB, 1987. Sexual selection:
testing the alternatives—Dahlem Konferenzen. LonMoller AP, 1990. Effects of a haematophagous mite on
don: John Wiley.
the barn swallow (Hirundo rustica): a test of the Hamilton and Zuk hypothesis. Evolution 44:771-784.
Cote IM, Hunte W, 1989. Male and female mate choice
in the redlip blenny: why bigger is better. Anim Behav
MyrbergAAJr, 1972. Ethology of the bicolor damselfish,
38:78-88.
Eupomacentrus partitus (Pisces: Pomacentridae): a comparative analysis of laboratory andfieldbehaviour. Anim
Darwin C, 1871. The descent of man and selection in
Behav Monogr 5:197-283.
relation to sex. London: John Murray.
Norris KJ, 1990a. Female choice and the evolution of the
Dawkins R, Krebs JR, 1978. Animal signals: information
conspicuous plumage colouration of monogamous male
or manipulation. In: Behavioural ecology: an evolugreat tits. Behav Ecol Sociobiol 26:129-138.
tionary approach (Krebs JR, Davies NB, eds). Oxford:
Blackwell; 282-309.
Norris KJ, 1990b. Female choice and the quality of parental care in the great tit Parus major. Behav Ecol
DeMartini EE, 1987. Paternal defence, cannibalism and
Sociobiol 27:275-281.
polygamy: factors influencing the reproductive success
of painted greenling (Pisces, Hexagrammidae). Anim
O'Donald P, 1980. Genetic models of sexual selection.
Behav 35:1145-1158.
Cambridge: Cambridge University Press.
Partridge L, Harvey PH, 1986. Contentious issues in sexDominey WJ, Blumer LS, 1984. Cannibalism of early life
ual selection. Nature 323:580-581.
stages in fishes. In: Infanticide: comparative and evoPedersen RD, 1979. The seasonal cycle of body compolutionary perspectives (Hausfater G, Hrdy SB, eds).
sition in the mottled sculpin, Coitus bairdi (MS thesis).
New York: Aldine Press; 43-64.
Columbus: Ohio State University.
Downhower JF, Brown L, 1980. Mate preferences of female mottled sculpins, Cottus bairdi. Anim Behav 28: Petersen CW, 1990. The occurrence and dynamics of
clutch loss andfilialcannibalism in two Caribbean dam728-734.
selfish. J Exp Mar Biol Ecol 135:117-133.
Eckert CG, Weatherhead PJ, 1987. Male characteristics,
Pomiankowski A, 1988. The evolution of female mate
parental quality and the study of mate choice in the
preferences for male genetic quality. Oxf Surv Evol
red-winged blackbird (Agelaius phoeniceus). Behav Ecol
Biol 5:136-184.
Sociobiol 20:35-42.
Randall JR, 1968. Caribbean reef fishes. Neptune City,
Endler JA, 1980. Natural and sexual selection on color
New Jersey: TFH Publications.
patterns in Poecilia reticulata. Evolution 34:76—91.
Robertson DR, Green DG, Victor BC, 1988. Temporal
Fisher R, 1930. The genetical theory of natural selection.
coupling of production and recruitment of larvae of a
Oxford: Clarendon Press.
Caribbean reef fish. Ecology 69:370-381.
Grafen A, 1990a. Biological signals as handicaps. J Theor
Biol 144:517-546.
Sargent RC, 1988. Paternal care and egg survival both
increase with clutch size in the fathead minnow, PiGrafen A, 1990b. Sexual selection unhandicapped by the
mephales promelas. Behav Ecol Sociobiol 23:33-37.
Fisher process. J Theor Biol 144:473-516.
SAS Institute Inc, 1988. Cary, North Carolina: SAS InGronell AM, 1989. Visiting behaviour by females of the
stitute Inc.
sexually dichromatic damselfish, Chrysiptera cyanea (Teleostei: Pomacentridae): a probable method of assessing
Schmale MC, 1981. Sexual selection and reproductive
male quality. Ethology 81:89-122.
success in males of the bicolor damselfish, Eupomacentrus partitus (Pisces: Pomacentridae). Anim Behav 29:
Hamilton WD, Zuk M, 1982. Heritable true fitness and
1172-1184.
bright birds: a role for parasites? Science 218:384-387.
Sikkel PC, 1988. Factors influencing spawning site choice
Hill GE, 1990. Female house finches prefer colourful
by female garibaldi, Hypsypops rubicundus (Pisces: Pomales: sexual selection for a condition-dependent trait.
macentridae). Copeia 1988:710-718.
Anim Behav 40:563-572.
Thresher RE, Gronell AM, 1978. Subcutaneous tagging
Itzkowitz M, Makie D, 1986. Habitat structure and reof small reef fishes. Copeia 1978:352-353.
productive success in the beaugregory damselfish. J
Unger LM, 1983. Nest defense by deceit in the fathead
Exp Mar Biol Ecol 97:305-312.
minnow, Pimephales promelas. Behav Ecol Sociobiol 13:
Jamieson IG, Colgan PW, 1989. Eggs in the nests of males
125-130.
and their effect on mate choice in the three-spined
Unger LM, Sargent RC, 1988. Allopaternal care in the
stickleback. Anim Behav 38:859-865.
fathead minnow, Pimephales promelas: females prefer
Kirkpatrick M, 1982. Sexual selection and the evolution
males with eggs. Behav Ecol Sociobiol 23:27-32.
of female choice. Evolution 36:1—12.
Zahavi A, 1975. Mate selection—a selection for a handKnapp RA, Sargent RC, 1989. Egg-mimicry as a mating
icap. J Theor Biol 53:205-214.
strategy in the fantail darter, Etheostomaflabellare:females prefer males with eggs. Behav Ecol Sociobiol 25:
321-326.
Knapp RA, Warner RR, 1991. Male parental care and
female choice in the bicolor damselfish, Slegastes partitus: bigger is not always better. Anim Behav 41:747756.
Kodric-Brown A, 1989. Dietary carotenoids and male
mating success in the guppy: an environmental component to female choice. Behav Ecol Sociobiol 25:393401.
REFERENCES
300
Behavioral Ecology Vol. 2 No. 4