males ELISA BET FORSGR EN { Department of Zoology, Uppsala University,VillavÌgen 9, S-752 36 Uppsala, Sweden
SU M M A RY
Female preference for males successful in male ^ male competition is generally assumed to result in mating with high quality males. Here I report results from an experiment disentangling the e¡ects of intra- and intersexual selection in the sand goby, Pomatoschistus minutus, a marine ¢sh that exhibits paternal care. I show that large males are successful in male ^ male competition, but contrary to what one would expect, dominants are not preferred by females and are not better at taking care of the eggs. Female preference, however, correlated with the subsequent hatching success of the eggs. Thus, female choice selects for good parenting. Hence, direct bene¢ts in the form of superior paternal care can explain female choice in this species, supporting a good parent process of sexual selection. However, choosing on the outcome of male ^ male competition does not enable females to mate with the `best' males.
1. I N T RO DUC T ION
Female preference for dominant males has been found in various animal taxa; for example, mammals (Cox & Le Boeuf 1977; Horne & YlÎnen 1996), insects (Breed et al. 1980), ¢sh (Bisazza & Marin 1991) and birds (Alatalo et al. 1991). It is generally thought that by inciting male contests and/or mating with winners of male ^ male competition, females ensure mating with high quality males, thereby gaining indirect bene¢ts (Cox & Le Boeuf 1977; Callahan 1981; Poole 1989; Montgomerie & Thornhill 1989; Alatalo et al. 1991) or direct bene¢ts, such as reduced risk of parasite transmission (Freeland 1981) or better nest defence (Bisazza et al. 1989). Despite considerable evidence for discriminatory mating among animals (Andersson 1994), the predicted bene¢ts from carrying out mate choice are, however, not as well supported empirically (Pomiankowski 1987; Kirkpatrick & Ryan 1991). Intra- and intersexual selection are often assumed to reinforce each other, and characters selected by male contests are expected to be important cues also in female choice because they may signal quality honestly (Berglund et al. 1996). However, because of the confounding e¡ect of male ^ male competition on female mate choice, it is generally di¤cult to tease apart the e¡ects of intra- and intersexual selection on mating patterns. Here I show results from experiments carried out on a marine ¢sh which exhibits paternal care. The aim of the study was to test whether males successful in intrasexual competition are preferred by females, and whether they are better fathers (Hoelzer 1989), i.e. whether they are better at defending the nest and { Present address: Department of Marine Ecology, GÎteborg University, Kristineberg Marine Research Station, S-450 34 FiskebÌckskil, Sweden. (
[email protected])
Proc. R. Soc. Lond. B (1997) 264, 1283^1286 Printed in Great Britain
taking care of the eggs. The experimental design allowed me to separate intra- and intersexual selection and to evaluate the ¢tness consequences of mate choice. 2 . M ET HOD S The sand goby, Pomatoschistus minutus (Pallas) (Pisces, Gobiidae), is a marine ¢sh that is common along the coasts of Europe (Miller 1986). Sand gobies live for 1^2 years and usually reproduce during one season only (Healy 1971). Both males and females can reproduce repeatedly during the breeding season, which lasts from May through June in my study area. Males build nests under empty mussel shells by covering them with sand and excavating underneath. They attract females by a courtship display showing their colourful ¢ns. During spawning, the female attaches her eggs to the nest in a monolayer. There is little variation in egg size among females, and egg size is not associated with female size (C. Kvarnemo, unpublished data). The sand goby exhibits exclusive paternal care. The males care for the eggs by guarding against egg predators, fanning and cleaning them until they hatch, but males may also cannibalize their own eggs. Thus, the hatching success of the eggs is largely dependent on the performance of the male parent (Forsgren et al. 1996a). After hatching, there is no further parental care, and the fry have to manage on their own. Males are the more competitive sex (Kvarnemo et al. 1995; Forsgren et al. 1996b) and mate indiscriminately (C. Kvarnemo and E. Forsgren, unpublished data), while females are selective in mate choice (Forsgren 1992; Forsgren et al. 1996a; Forsgren 1997) The study was conducted in May ^ June 1994, at Klubban Biological Station, Sweden (588 15' N, 118 28' E). Sand gobies used in the experiment were caught with a hand trawl in a shallow sandy bay. The sexes were kept in separate tanks (100^150 l) with continuously renewed seawater. They were fed daily ad libitum with fresh mussel (Mytilus edulis) meat before the experiment. The experimental procedure consisted of four consecutive steps:
1283
& 1997 The Royal Society
and allowed to compete for it. When a nest was built by one of the males (the `winner'), the aquarium was divided by a Plexiglass screen and a new nest site was provided in the empty side with the non-nest-holding male (the `loser') to enable him to build a nest. After this, the experiment was continued (step 2) using the same males. (2) Female choice (¢gure 1b): the aquarium was divided into three, and a ripe female was put alone in the middle section. After acclimatization, 15 observations of female position (left or right) were made during each day (Forsgren 1992). The male with which the female was found close to in the majority of the observations was given a preference score of 1, whereas the other male was given the score 0. (3) Spawning (¢gure 1c): the female of step 2 was put into one of the males' compartments (chosen randomly), and another ripe female was added to the other male. Time before spawning was recorded. After spawning the ¢sh were caught. Total body length was measured to the nearest millimetre, and male breeding coloration was scored on a scale from 1 (dull) to 3 (bright). The males were weighed to the nearest milligram on a digital balance. Replicates where spawning had not taken place after 48 h were terminated and excluded from the data set. Preferred males were assumed to start spawning more quickly compared to unpreferred males. Accordingly, males which started spawning within less than 30 min were given a preference score of 1, males which started
given a score of 0. A total female preference score was then calculated by taking into account her behaviour both when she was kept apart from the males (step 2), and when she was allowed to interact and assess the males at a closer distance (step 3): the two scores of each male were summed to yield the total preference score with possible values of 0, 0.5, 1.0, 1.5 and 2.0. (4) Paternal care (¢gure 1d): I removed the nests, outlined the egg clutches with a pencil ( egg area 1) and then put the nest sites back into the aquarium, together with the males. When both males had rebuilt their nests (or after a maximum of 48 h), I removed the partition and added a small crab Carcinus maenas (width 10^15 mm) as a potential egg predator. The males were then allowed to care for their eggs. Thus, during the parental phase the eggs could be preyed upon by the crab, the other male, or by the parent male. Males were fed mussel meat every third day. Daily checks were made to monitor the development of the eggs. Just before hatching, after 6^10 days (depending on the water temperature), I removed the nest and outlined the egg clutch ( egg area 2). The egg area was traced onto paper and hatching success was calculated for each male as the proportion of the original clutch that reached hatching (egg area 2/egg area 1). I used hatching success as a measure of a male's parenting ability. After the experiment all ¢sh and fry were released.
3. R E SU LT S
(a)
(b)
(c)
(d)
Figure 1. The experimental aquaria (40 l). (a) Male^male competition. (b) Female choice. (c) Spawning. (d) Paternal care. Proc. R. Soc. Lond. B (1997)
Larger males enjoyed an advantage in male ^ male competition, as they managed to occupy the nest site provided more often (mean length s.d. of winners 58.6 4.9, losers 55.2 6.3 mm; paired t 2.91, n 35, p50.01). However, females did not prefer the winners of male ^ male competition, as the preference scores for winners (mean s.d.; 1.04 0.64) and losers (1.10 0.61) were practically identical (paired t ÿ 0.3, n 24, p 0.77). Winners were not preferred more often in step 2 ( 2 0.00008, d.f. 1, p 0.99), nor did they spawn quicker in step 3 ( 2 3.30, d.f. 2, p 0.19). Furthermore, female preference score was not associated with male body length (rs ÿ 0.09, n 58, p 0.51), colour intensity (rs ÿ 0.02, p 0.88) or condition (rs 0.03, p 0.85). Thus, it is unknown which cues were used in mate choice. The number of eggs spawned by females did not vary in relation to the attractiveness of the male, as the area of the egg clutch did not correlate with the female preference score of the male (rs ÿ 0.02, n 58, p 0.87).This is in accordance with earlier observations that females lay their whole clutch with a single male (Forsgren 1992). There was a positive correlation between the area of the initial clutch and the hatching success of the eggs (rs 0.39, n 57, p50.01). Thus, larger egg clutches had proportionally higher success as compared to smaller ones.The clutch size of winners and losers in male ^ male competition did not di¡er signi¢cantly (mean area s.d. of winners 10.71 2.51cm3, losers = 10.25 þ 2.25 cm3, paired t 0.78, n 24, p 0.44). There was no signi¢cant di¡erence in hatching success (median values) between winners (94%) and losers (98%) (Wilcoxon signed-rank test, z 0.23, n 23, p 0.82). Thus, the females gained no direct bene¢ts by mating
hatching success than other males (¢gure 2). Also, controlling for clutch size, hatching success was positively correlated with the female preference score (table 1). Thus, males which exhibited better parenting were preferred by the females. None of the measured male characters (body length, coloration, and condition) were associated with hatching success when clutch size was controlled for (table 1). 4 . DI S C U S S ION
Male size was important in male ^ male competition, because larger males more often won over smaller ones in competition over nest sites. This con¢rms earlier ¢ndings on the same species (LindstrÎm 1988). However, females did not prefer dominant males (i.e. those successful in male ^ male competition). Observations of mate sampling behaviour of females that had been released in the ¢eld also suggest that females pay little attention to male ^ male competition, as they seldom spawned with males they had observed in ¢ghts or interactions with other males (Forsgren 1997). Furthermore, the potential for female choice is high in this population as there are many nest-holding males to choose from (Forsgren et al. 1996b). The winners of male ^ male competition were not better at taking care of their eggs as they did not achieve a higher hatching success than the losers. Thus, females gained no direct bene¢ts when mating with males successful in male ^ male competition. Hatching success was, however, positively correlated with female preference score. Strongly preferred males had a higher hatching success than other males. Thus, males of higher parental quality were preferred by the females, resulting in substantial direct ¢tness bene¢ts.This supports a good parent process of sexual selection (Hoelzer1989). Female ¢shes may gain direct bene¢ts through increased egg survival by their choice of mate, for example, by choosing males already guarding eggs (reviews in Jamieson 1995; Petersen 1995). However, very few previous studies have
Figure 2. Female preference versus subsequent hatching success of the eggs (mean s.e.) (Kruskal^Wallis ANOVA, H 7.12, n 57, p 0.028). Weak preference includes preference scores 0 and 0.5 (n 13), intermediate 1.0 and 1.5 (n 33) and strong 2.0 (n 11). Proc. R. Soc. Lond. B (1997)
initial clutch size was held constant (n 57). (Male condition was de¢ned as the residual from a regression of male body length on mass (mass 0.07 body length ÿ 2.47, r 0.96, n 58, p 0.0001.)) variable male body length male coloration male condition female preference score
correlation coe¤cient
p
ÿ 0.06 0.01 0.11 0.22
40.20 40.25 40.10 50.01
conclusively shown female choice for males with high parental quality resulting in actual ¢tness gains (Knapp & Kovach 1991). One problem with many ¢sh studies is that the received egg area is an important confounding variable. It is common in ¢shes that larger broods have a higher hatching success because the parental investment increases with the reproductive value of the brood (Coleman et al.1985; Petersen & Marchetti1989; Forsgren et al. 1996a). This was also found in the present study. Consequently, preferred males, who often have a high mating success and receive more eggs, may have a high hatching success as a consequence of guarding large broods and not because of superior parenting quality. Another problem of many studies, not only on ¢sh, is to control for the e¡ect of the female. O¡spring survival and male provisioning rates may be in£uenced by the quality of the female. To reduce this problem, females could be randomly given to di¡erent males, or broods could be swapped between males, as in the study by Knapp & Kovach (1991). It is unclear why some males were better fathers than others, since hatching success was not correlated with any of the measured male traits that could potentially have acted as cues in mate choice. In the ¢eld, courtship intensity was found to be the most important cue in mate choice (Forsgren 1997). Thus, females may select good fathers by assessing male courtship. Whether males courting at a high level honestly signal their parental quality, for example, by indicating large fat reserves, as in damsel¢sh (Knapp & Kovach 1991), is unknown. In another sand goby population, large males fanned the eggs more, and larger males were preferred by females (LindstrÎm & HellstrÎm 1993). Whether this resulted in a higher hatching success for large males is, however, unknown. In the present study, as well as in damsel¢sh (Knapp & Warner 1991), large males did not provide better parental care as measured by their hatching success, and were not preferred by females. The large variation found in the quality of paternal care in sand gobies suggests strong direct selection on female preferences. The recent debate on sexual selection has been dominated by a focus on indirect bene¢ts; whether mate preferences are explained by arbitrary trait models, or by good genes models (reviews in Bradbury & Andersson 1987; Andersson 1994). As an explanation of the evolution of mate preferences, direct bene¢ts seems to have been overlooked. Nevertheless, mate choice may often be explicable through non-genetic bene¢ts, which directly a¡ect female reproductive success or survival
In the ¢eld, egg predation may be an important factor, and to mimic this I added inter- as well as intraspeci¢c potential egg predators in my experiment. There may, of course, exist quantitative di¡erences between the hatching success in my experiment and in the ¢eld, although there is no reason to assume qualitatively di¡erent results. A factor, which I did not include in my experiment, that might be important in the ¢eld is variation in nest quality. If, for example, dominant males occupy high-quality nest sites and, these are preferred by females, dominant males would end up with a higher mating success than other males, as in damsel¢sh (Hoelzer 1990). This does not, however, seem to be a problem because characteristics of the oviposition site (shell size and species) had no e¡ect on acceptance or rejection of mates (Forsgren 1997). Furthermore, mating success in the ¢eld does not seem to be a result of male dominance because nest sites are available in excess in the study population, thereby reducing the importance of male ^ male competition (Forsgren et al. 1996b). However, in another population (in the Baltic), male ^ male competition is much more important in determining the mating success of males (Forsgren et al. 1996b; LindstrÎm 1988). Thus, the dynamics seem very di¡erent in di¡erent populations. In conclusion, male sand gobies that are successful in intrasexual competition are not preferred by females. Instead, females choose good fathers, which manage to hatch a higher proportion of their eggs. Females would not have gained these direct bene¢ts if mating with the winners of male ^ male competition. Thus, direct bene¢ts in the form of competent paternal care can explain female choice in this species, and good genes models of sexual selection need not be invoked. My ¢ndings show that, in contrast to the common view, choosing on the outcome of male ^ male competition does not necessarily enable females to mate with the `best' males. I thank P. Karlsson and A. NordlÎf for help in the ¢eld, and T. Amundsen, A. Berglund, J. HÎglund, B. Sheldon and S. Ulfstrand for valuable comments on the manuscript. The Royal Swedish Academy of Sciences and Uppsala University (the Inez Johansson Foundation) provided ¢nancial support.
R E F E R E NC E S Alatalo, R.V., HÎglund, J. & Lundberg, A. 1991Lekking in the black grouseöa test of male viability. Nature 352,155^156. Andersson, M.1994 Sexual selection. Princeton University Press. Berglund, A., Bisazza, A. & Pilastro, A. 1996 Armaments and ornaments: an evolutionary explanation of traits of dual utility. Biol. J. Linn. Soc. 58, 385^399. Bisazza, A. & Marin, G. 1991 Male size and female mate choice in the eastern mosquito¢sh (Gambusia holbrooki: Poeciliidae). Copeia 3, 730^735. Bisazza, A., Marconato, A. & Marin, G. 1989 Male competition and female choice in Padogobius martensi (Pisces, Gobiidae). Anim. Behav. 38, 406^413. Bradbury, J. W. & Andersson, M. B. 1987 Sexual selection: testing the alternatives. Chichester, UK: Wiley. Breed, M. D., Smith, S. K. & Gall, B. G. 1980 Systems of mate selection in a cockroach species with male dominance hierarchies. Anim. Behav. 28, 130^134. Proc. R. Soc. Lond. B (1997)
Coleman, R. M., Gross, M. R. & Sargent, R. C. 1985 Parental investment decision rules: a test in bluegill sun¢sh. Behav. Ecol. Sociobiol. 18, 59^66. Cox, C. R. & Le Boeuf, B. J. 1977 Female incitation of male competition: a mechanism in sexual selection. Am. Nat. 111, 317^335. Forsgren, E. 1992 Predation risk a¡ects mate choice in a gobiid ¢sh. Am. Nat. 140, 1041^1049. Forsgren, E. 1997 Mate sampling in a population of sand gobies. Anim. Behav. 53, 267^276. Forsgren, E., Karlsson, A. & Kvarnemo, C. 1996a Female sand gobies gain direct bene¢ts by choosing males with eggs in their nests. Behav. Ecol. Sociobiol. 39, 91^96. Forsgren, E., Kvarnemo, C. & LindstrÎm, K. 1996b Mode of sexual selection determined by resource abundance in two sand goby populations. Evolution 50, 646^654. Freeland, W. J. 1981 Parasitism and behavioral dominance among male mice. Science 213, 461^462. Healy, M. C.1971Gonad development and fecundity of the sand goby, Gobius minutus Pallas.Trans. Am. Fish. Soc. 100, 520^526. Hoelzer, G. A. 1989 The good parent process of sexual selection. Anim. Behav. 38, 1067^1078. Hoelzer, G. A. 1990 Male ^ male competition and female choice in the Cortez damsel¢sh, Stegastes rectifraenum. Anim. Behav. 40, 339^349. Horne,T. J. & YlÎnen, H.1996 Female bank voles (Clethrionomys glareolus) prefer dominant males; but what if there is no choice ? Behav. Ecol. Sociobiol. 38, 401^405. Jamieson, I. 1995 Do female ¢sh prefer to spawn in nests with eggs for reasons of mate choice copying or egg survival ? Am. Nat. 145, 824^832. Kirkpatrick, M. & Ryan, M. J. 1991 The evolution of mating preferences and the paradox of the lek. Nature 350, 33^38. Knapp, R. A. & Kovach, J. T. 1991 Courtship as an honest indicator of male parental quality in the bicolor damsel¢sh, Stegastes partitus. Behav. Ecol. 2, 295^300. Knapp, R. A. & Warner, R. R. 1991 Male parental care and female choice in the bicolor damsel¢sh, Stegastes partitus: big is not always better. Anim. Behav. 41, 747^756. Kvarnemo, C., Forsgren, E. & Magnhagen, C. 1995 E¡ects of sex ratio on intra- and intersexual behaviour in sand gobies. Anim. Behav. 50, 1455^1461. LindstrÎm, K. 1988 Male ^ male competition for larger nest sites in the sand goby, Pomatoschistus minutus. Oikos 53, 67^73. LindstrÎm, K. & HellstrÎm, M. 1993 Male size and parental care in the sand goby, Pomatoschistus minutus. Ethol. Ecol. Evol. 5, 97^106. Miller, P. J. 1986 Gobiidae. In Fishes of the northeastern Atlantic and the Mediterranean, vol. 3 (ed. P. J. P. Whitehead), p.1074. Paris: UNESCO. Montgomerie, R. & Thornhill, R. 1989 Fertility advertisement in birds: a means of inciting male ^ male competition? Ethology 81, 209^220. Petersen, C. W. 1995 Male mating success and female choice in permanently territorial damsel¢shes. Bull. Mar. Sci. 57, 690^704. Petersen, C. W. & Marchetti, K. 1989 Filial cannibalism in the Cortez damsel¢sh Stegastes rectifraenum. Evolution 43, 158^168. Pomiankowski, A. 1987 The cost of choice in sexual selection. J.Theor. Biol. 128, 195^218. Poole, J. H. 1989 Mate guarding, reproductive success and female choice in African elephants. Anim. Behav. 37, 842^849. Reynolds, J. D. & Gross, M. R.1990 Costs andbene¢ts of female mate choice: is there a lek paradox ? Am. Nat. 136, 230^243. Received 28 April 1997; accepted 15 May 1997
