Offspring quality and female choice in the guppy, Poecilia reticulata

Transcription

1 Anim. Behav., 1995, 49, Offspring quality and female choice in the guppy, Poecilia reticulata PAUL F. NICOLETTO* Department of Biology, University of New Mexico, Albuquerque, NM 87131, U.S.A. (Received 7 September 1993; initial acceptance 21 October 1993; final acceptance 24 January 1994; MS. number: A6624R) Abstract. This study evaluated the offspring viability prediction of the condition-dependent and Fisherian models of female choice in the guppy. Families of full-sibling females were bred with the male they preferred or did not prefer in a choice experiment. The physical condition, sexual behaviour and coloration of the offspring were evaluated. There were no significant differences between offspring attributable to the type of sire. However, there were significant family and sire-type by family interactions for physical condition, male mating behaviour and coloration. These significant effects indicate that consistency within families may be due to genetic effects. Genetic analyses indicate that genetic variation probably exists for prolonged swimming performance, physical condition and display rate. The results of this study and other studies on other fish have shown that these measures of constitution are correlated with components of viability. These results are interpreted in the framework of the condition-dependent and Fisherian models of female choice. *Present address: Department of Biology, Shippensburg University, Shippensburg, PA 17257, U.S.A. The Fisherian and condition-dependent models of female choice are often cited to explain the evolution of elaborate male ornamentation in species without male parental care. The Fisherian model contends that male ornaments evolve because females have a genetically based mating preference for certain male traits. Male traits are arbitrary. The only advantage a female receives from her preference is the production of ornamented sons which will have greater mating success in the next generation (Lande 1981; Kirkpatrick 1982, 1987; Heisler et al. 1987). The conditiondependent model contends that females use the development of male ornaments as indicators of genetic quality (Zahavi 1975; Andersson 1982, 1986; Hamilton & Zuk 1982; Kodric-Brown & Brown 1984; Nur & Hasson 1984; Pomiankowski 1987a, b, 1988). Male advertising traits provide females information about the male s ability to exploit the current environment. Females indirectly enhance their fitness by producing high-quality female and male offspring. Learning which of these models best explains the evolution of male ornamentation has been problematic, because the two models make many similar predictions (Heisler et al. 1987; Balmford & Read 1991; Kirkpatrick & Ryan 1991). However, they do differ with respect to the relationship between a female s choice of a mate and the viability or fitness of her offspring (Heisler et al. 1987; Balmford & Read 1991; but see Nicoletto 1993). The Fisherian hypothesis predicts that there is no relationship between female choice and offspring fitness. The condition-dependent model predicts that there is a positive relationship between female choice and offspring fitness (Zahavi 1975; Kodric-Brown & Brown 1984; Heisler et al. 1987). Viability is used in the genetic models of Fisherian and condition-dependent female choice as a measure of offspring quality (Anderson 1982; Arnold 1985; Maynard Smith 1985; Kirkpatrick 1987). Because of logistical contraints on the researcher, however, and the limitations imposed by an organism s life history most empirical studies of female choice have been unable to measure offspring viability directly (but see Partridge 1980; Boake 1989). Most researchers use variables that are related to an animal s phenotypic condition or constitution, for example swimming performance (Nicoletto 1991, 1993), condition factor (Andersson 1989; Milinski & Bakker 1990; Nicoletto 1993) and parasite resistance (Kennedy et al. 1987; Ligon et al. 1990; Zuk et al. 1990; Houde & Torio 1992) because they are components or correlates of overall viability. Constitution is defined as the physical makeup of the /95/ $08.00/ The Association for the Study of Animal Behaviour 377

2 378 Animal Behaviour, 49, 2 individual comprising inherited qualities modified by environment (Webster s New Collegiate Dictionary 1981). I will use constitution as a general term for components or correlates of viability such as those mentioned above. The objective of this study was to evaluate the relationship between female choice and offspring viability in the guppy. I tested the prediction that there is a significant difference in constitution between offspring of females mated to preferred males and offspring of females mated to nonpreferred males. Guppies are particularly appropriate to study because females use a variety of male characteristics during mate choice and some of these characteristics are condition-dependent. Females respond to males on display rate (Farr 1980; Houde 1988), colour pattern (Endler 1980, 1983; Kodric-Brown 1985, 1989; Houde 1987, 1988; Long & Houde 1989; Houde & Endler 1990), dorsal and caudal fin size and shape (Bischoff et al. 1985), and colour intensity (Kodric-Brown 1989; Houde & Torio 1992). Male display rate, area and intensity of orange coloration, and overall ornament complexity are positively correlated to prolonged swimming performance in males (Nicoletto 1991, 1993). Display rate is also correlated with male physical condition (Nicoletto 1993) and negatively correlated with parasite load (Kennedy et al. 1987; McMinn 1990). MATERIALS AND METHODS Female Choice Experiment I conducted a female choice experiment to choose sires for a breeding experiment. Each laboratory-reared female was presented with four males, simultaneously, which were size-matched (within 2 mm), had individually recognizable colour patterns, and had dorsal and caudal fins of similar size and shape. Female choice trials were performed in an aquarium measuring cm with a central compartment ( cm) for the female and four peripheral compartments for the males. The four male compartments were made by placing opaque partitions from the corners of the choice tank to the corners of the female compartment at a 45 angle. This aquarium had an undergravel filter, and a dirty white coral substrate. Each male compartment contained a submersible heater. The water temperature in each compartment ranged between 27 and 29 C. A mirror (60 60 cm) was placed 0 66 m above the tank at a slight angle to enable simultaneous observation of all compartments. A VHS camcorder placed on the floor in front of the tank was focused on the mirror and used to record the experiment. Four males and a single female were allowed to acclimatize to the choice aquarium for 24 h before the female choice trial. Opaque partitions, similar to those separating males, were placed around the central female compartment before the choice experiment and prevented the female from seeing the males. These partitions were removed at the beginning of the experiment. Each female s choice of a male was determined by filming the female for a 25-min observation period. Data were not recorded for the first 5 min to reduce the effects of disturbance caused by removing the partitions. I ranked males based on how much time the female spent within 2 cm of the male s partition during the 20-min observation period. The male with which a female spent the most time was recorded as a positive choice (preferred male). The male with which a female spent the least time was recorded as a negative choice (non-preferred male). The female was then mated to the least preferred or most preferred male. Each guppy was only used once in this experiment. Male display rate was not recorded because the displays could not reliably be detected on the television screen. All choice experiments took place between 0800 and 1000 hours in a greenhouse under natural light. I collected the males used in the female choice experiment from a feral population in McCauley Hot Spring in the Jemez Mountains of New Mexico (see Nicoletto 1993). I captured males at approximately monthly intervals and allowed them to acclimatize to laboratory conditions for at least 2 weeks before testing. The females I used in the choice and breeding experiment were from a third-generation laboratory colony derived from the Jemez Mountain population. All of the females used in the female choice and breeding experiments were virgins reared without males present in their aquarium. I raised 16 families of females and arbitrarily chose six females (full-siblings) from each family, for a total of 96 females. Each of these females was used as a dam in the breeding experiment. A damfamily consisted of all six females from one

3 Nicoletto: Offspring quality and female choice in guppies 379 family, thus there were 16 dam-families each containing six full-siblings. The use of sets of laboratory-reared sisters minimized phenotypic and genotypic variation within a family of females because they were full-siblings and reared in constant conditions. Breeding Experiment I designed a breeding experiment to determine whether there were any significant differences between the offspring of females mated with their preferred male versus the offspring of females mated to their non-preferred male. Three sisters in each family were mated with the males they preferred during their female choice experiment. The other three sisters in each family were mated to the males they did not prefer during their female choice experiment. I used 42-litre aquaria that were divided in half as breeding and rearing tanks. Each tank contained a single preferred pair and a single nonpreferred pair from the same dam-family. Each of these tanks contained an undergravel filter that allowed water to circulate between the two compartments. A plastic or glass aquarium lid was placed over the aquarium to prevent fish from jumping between compartments. A submersible heater was placed next to the partition dividing the aquarium. The water temperature was kept at 28 C and differed by less than 2 C on either side of the partition. A floating plastic plant in each compartment provided a refuge for offspring. Sires remained with the dams for at least 20 days to ensure insemination. The female remained with the brood and often had multiple litters. These litters were not separated or counted. Within 2 weeks after the first male offspring reached sexual maturity, as determined by male gonopodial and colour development, I arbitrarily chose two males and two females from each litter. These fish were housed individually in one of four compartments of a 42-litre aquarium and were visually isolated. All fish were fed Tetramin tropical fish food at least twice daily and were fed brine shrimp approximately every other day. I measured the swimming performance and calculated a condition factor for each male and female offspring (see below). I also photographed male offspring and used them in a female sexual response experiment as described below. Prolonged Swimming Performance I determined the prolonged swimming performance of the parents and offspring by measuring critical swimming speed. Critical swimming speed is the maximum speed that a fish can sustain for a set period in a laboratory flow chamber (Brett 1964). Critical swimming speed is a common measure of prolonged swimming performance and reportedly correlates well with health, active metabolism and endurance in other fish species (Brett 1964; Smit 1965; Brett & Glass 1973; Jones et al. 1974; Beamish 1978). I measured critical swimming speed by increasing the water velocity in the flow chamber every 3 min until the fish became exhausted. A description of the flow chamber and the methodology for measuring swimming speed are given in Nicoletto (1991, 1993). Condition Factor The physical condition of parents and offspring was estimated by calculating a condition factor. Condition factors are used to estimate the relative physical condition or stoutness of fish (Bolger & Connolly 1989; Milinski & Bakker 1990). These factors assume that, at a given standard length, heavier fish are in better physical condition. I calculated the equation for the condition factor separately for the field-caught male parents (Fig. 1) and laboratory-reared female parents and all offspring (Fig. 2). Standard lengths of all fish were measured with dial calipers to the nearest 0 01 mm. Mass was measured to the nearest g with an electronic balance. The equation for the condition factor was calculated as Condition=Mass (g) 100/(Length b (cm) ), where b is equal to the slope. The slope was calculated by fitting a multiplicative regression model to the relationship between mass and standard length (Figs 1, 2). Colour Quantification I measured the colour patterns of male parents and offspring by projecting a colour slide of each fish on a computer digitizing tablet. Details of the photography and measurement procedures are given in Nicoletto (1993). The proportion of the fish s body that was covered by each colour was used in the analysis to correct for differences in body size. I used ornament complexity (the area

4 380 Animal Behaviour, 49, 2 two separate sets of rankings, one set for males from the female choice experiment and one set for the male offspring. The number of structural spots was not included in this analysis because these spots are usually continuously distributed on a male s body. Figure 1. The relationship between standard length and mass for field-caught sires used in the breeding experiment. Figure 2. The relationship between standard length and mass for laboratory-reared dams and their mature male and female offspring that were used in the breeding experiment. and number of orange and black colour spots and the area of structural colours) to quantify ornamentation. Ornament complexity was calculated by ranking each colour variable separately and then summing the ranks for each colour for each male. Therefore, ornament complexity is a variable that ranks the complexity of a male s total ornamentation relative to others in the sample from which it came (Nicoletto 1993). I performed Female Sexual Response to Male Offspring This experiment used a no-choice design to evaluate the attractiveness of male offspring and all sires to females. In a no-choice design, females are presented a single male and are not given a choice between different males. The details of this experiment are given in Nicoletto (1993). This experiment consisted of presenting each male with a virgin female and recording the male s display behaviour and female s sexual response (Liley 1966; Houde 1987, 1988; Reynolds & Gross 1992). Virgin females were obtained from a stock aquarium and were not the daughters or siblings of the males. The variables recorded during each 3-min observation period were: number of male displays, copulatory attempts, and the number of female sexual responses. A female sexual response was counted when the female oriented towards the male and either glided or slowly swam towards him (Houde 1987, 1988). These female behaviour patterns indicate that the female is receptive to the male s advances and are related to male mating success (Liley 1966; Houde 1987, 1988). This experiment took place in aquaria under full spectrum Vita lights placed 6 25 cm above the aquarium. I evaluated the female sexual response for the 107 male offspring and 96 sires from the breeding experiment. Statistical Analysis The characteristics of preferred and nonpreferred males (N=192) of individual females from the female choice experiment were analysed using Wilcoxon matched-pairs signed-ranks tests. The characteristics of preferred and non-preferred sires (N=96) used in the breeding experiment were compared using Wilcoxon rank-sum tests. These analyses test the hypothesis that the constitution and ornamentation of preferred males is greater than that of non-preferred males, thus the P-values resulting from these analyses are onesided. The one-sided tests are justified because both the Fisherian and condition-dependent models of female choice predict that females will

5 Nicoletto: Offspring quality and female choice in guppies 381 choose males with the best developed ornamentation, and the condition-dependent model predicts that these males will also be in the best physical condition (Heisler et al. 1987). All data collected on the offspring from the breeding experiment were analysed using a 2 2 factorial analysis of variance (ANOVA). The effects or treatments in the model were the sire type (i.e. either a preferred or non-preferred), dam-family and the interaction between sire type and dam-family. Male and female offspring were combined for the critical swimming speed and condition-factor analyses. The critical swimming speed and condition factors of male and female offspring were not significantly different (Wilcoxon rank-sums test, P=0 402 and 0 978, respectively). Therefore, I did not use sex as a main effect in these analyses. Critical swimming speed and the condition factor were normally distributed. The ornament data and the female response data were rank transformed (Conover & Iman 1981) and the ranks were analysed in the ANOVAs on these variables. The P-values for the overall F-test for each ANOVA were sequentially Bonferroni adjusted to protect against type 1 error. The P-values within each ANOVA were not adjusted. There were different sample sizes for the various ANOVAs, because 14 mated pairs failed to produce offspring. Therefore the data set was unbalanced and analysed with PROC GLM (SAS 1988). Eight of these pairs were preferred matings and six were non-preferred. Some additional pairs failed to produce the required four offspring. In addition, the female response experiment was only performed on 15 families of offspring. The ANOVAs revealed significant dam-family effects that were probably due to genetic variation and not environmental variation. Genetic variation is important because the condition-dependent model of female choice contends that ornament development is correlated with components of male constitution that can be passed on to offspring (Zahavi 1975; Kodric-Brown & Brown 1984; Heisler et al. 1987). Thus, there must be heritable genetic variation in the components of constitution that are correlated with the development of a male s ornamentation. Therefore, I performed quantitative genetic analyses to determine whether they would also support the genetic interpretation of the ANOVAs. Because this study was not designed to measure heritability, the heritability values reported here are either broadsense heritabilities or biased estimates of narrowsense heritabilities. Broad-sense heritability is an estimate of the total phenotypic variation in a character that is due to all genetic factors, that is, additive, dominance, epistatic and maternal variance components (Becker 1984). Narrow-sense heritability is an estimate of the total phenotypic variation in a character that is due to additive variance components (Becker 1984; Falconer 1989). The broad-sense heritability of critical swimming speed and the condition factor were estimated by using the methodology described by Becker (1984, pp ) for single pair matings. This analysis is appropriate when pairs of individuals are chosen randomly from a population and mated together and each mated pair produces several offspring. This is how the 16 dam-families consisting of six full-sisters (dams in the sire-type breeding experiment) were obtained. The narrow-sense heritability analyses for the constitution, ornaments and female response variables were done with mean parent offspring regressions (Becker 1984; Falconer 1989). The narrow-sense heritability estimates for the sire son and sire daughter regressions should be interpreted with caution because the sires were fieldcaught and the offspring were laboratory reared. Dam son and dam daughter heritability estimates may contain additive and maternal components because guppies are live bearing, and maternal effects have been demonstrated for offspring growth rate and mass (Rocchetta et al. 1985). RESULTS Female Choice Experiment In the female choice experiment, differences between preferred and non-preferred males in critical swimming speed, ornament complexity and black area tended towards significance. The ornament complexity and orange number of the sires also tended towards significance. However, there were no significant differences in any of these variables after a sequential Bonferroni adjustment (Tables I and II). Breeding Experiment The breeding experiment tested the prediction that there was a significant difference in

6 382 Animal Behaviour, 49, 2 Table I. Wilcoxon matched-pairs signed-rank test on the results of the mate choice experiment Average rank Variable Preferred Non-preferred z-statistic P-value Critical swimming speed Condition factor Ornament complexity Orange area Black area Structural area Orange number Black number Note that no P-value was statistically significant after a sequential Bonferroni adjustment. Table II. The results of Wilcoxon rank-sums tests on the males used as sires in the breeding experiment Average rank Variable Preferred Non-preferred z-statistic P-value Critical swimming speed Condition factor Ornament complexity Orange area Black area Structural area Orange number Black number Note that no P-value was statistically significant after a sequential Bonferroni adjustment. constitution, ornamentation and sexual behaviour between the offspring of preferred males and non-preferred males. The ANOVAs on all of the variables failed to support this prediction, as there were no significant sire-type effects (Table III). However, there was a significant dam-family effect for critical swimming speed, condition factor, ornament complexity, black area, structural area, orange number and copulation attempts. The dam-family effect for display number was marginally significant. The significant dam-family effect suggests that members of a family tended to be more similar to each other than to another family. There were also significant sire-type by damfamily interactions for the condition factor, ornament complexity, black area, orange number and copulation attempts. The significant interaction indicates that for some variables the differences observed in the offspring were attributable to differences between families and differences between sire type. This means that in some cases there are sire-type effects in the next generation but that this effect is not consistent. The overall F-tests for the orange area, black number and female response ANOVAs were not significant (Table III). The quantitative genetic analyses tend to support the genetic interpretation of the ANOVAs (Table IV). The broad-sense estimates for both critical swimming speed and the condition factor were greater than zero. The narrow-sense heritability analyses for critical swimming speed, display number and the female sexual response all yielded narrow-sense heritability estimates that were greater than zero (Table V). The narrow-sense heritability analyses for the condition factor,

7 Nicoletto: Offspring quality and female choice in guppies 383 Table III. The results of the two-way ANOVAs on offspring from the breeding experiment Overall Variable F-value P-value Source F-value P-value Critical swimming speed * Sire type Dam-family Interaction Condition factor * Sire type Dam-family Interaction Ornament complexity * Sire type Dam-family Interaction Structural area * Sire type Dam-family Interaction Orange number * Sire type Dam-family Interaction Black number Sire type Dam-family Interaction Display rate Sire type Dam-family Interaction Copulation attempts * Sire type Dam-family Interaction *P<0 05 with a sequential Bonferroni adjustment. Table IV. The results of one-way ANOVAs to determine the genetic variance of guppies from the dam-families Source Variable N family N dams V G SE Dam-family Critical swimming speed Condition factor V G : Heritability in the broad sense that contains variance components due to additive, dominance, epistatic and maternal factors. copulation attempts and the ornament variables were not significant. Thus, genetic variation apparently exists in both critical swimming speed and some sexual behaviour, but how much of this variation may be additive is uncertain. DISCUSSION The objective of this study was to determine whether offspring sired by preferred males differed from offspring of non-preferred males in their constitution, ornamentation and sexual behaviour. There was no significant difference between offspring of preferred and non-preferred sires for any constitution variable that I measured in this study. This result conforms best to the prediction of the Fisherian models of female choice. The Fisherian hypothesis predicts that there is no relationship between a female s choice of a mate and the constitution of her offspring. The significant interaction between sire type and dam-family

8 384 Animal Behaviour, 49, 2 Table V. The results of parent offspring regressions Variable Potential variance components Regression h 2 SE Critical swimming speed V A +V E Sire*son * V A +V E Sire*daughter V A +V M Dam*son V A +V M Dam*daughter * V A +V M +V E Mean * Condition factor V A +V E Sire*son V A +V E Sire*daughter V A +V M Dam*son V A +V M Dam*daughter V A +V M +V E Mean Display number V A +V E Sire*son * Copulation attempts V A +V E Sire*son Female response V A +V E Sire*son Orange area V A +V E Sire*son Black area V A +V E Sire*son Structural area V A +V E Sire*son V A : Variance due to additive genetic effects; V M : variance due to maternal genetic effects; V E : variance due to environmental effects. *Indicates heritabilities that were significantly greater (P<0 05) than zero. is consistent with this prediction. In some cases, matings with preferred males resulted in offspring with better constitutions and more ornamentation, and in other cases, matings with nonpreferred males yielded the same result. In other words, the lack of a consistent relationship (positive or negative) at the individual level between female choice and constitution yields no relationship at the population level. However, the interaction shows that sire type and therefore female preferences can have an effect in the next generation. The results of the analyses of the female choice experiment revealed no differences in the constitution and ornamentation of preferred and nonpreferred males or sires. Thus, females may have been choosing males randomly and therefore the lack of differences between the offspring would be expected. However, we cannot eliminate the possibility that females were choosing males because several variables tended towards statistical significance and females could have been choosing males based upon some unmeasured characteristic. In the Jemez Mountain population, display rate is probably the proximal cue that females use to choose males (Nicoletto 1993). I was unable to measure display rate during the female choice experiment so there is a possibility that females were using display behaviour to choose males. The lack of differences between offspring of preferred and non-preferred sires in this breeding experiment are consistent with two unpublished studies of guppies by A. Houde & A. Gong (J. A. Endler & A. Gong, personal communication). These studies found similar results with different experimental designs. However, Reynolds & Gross (1992) found that female preferences based on body size in guppies led to larger offspring size, faster growth rates and higher female fecundity. The lack of differences between offspring of different sire types does not mean that the condition-dependent model can be dismissed in this study. Research on guppies has shown that females use display rate (Farr 1980; Nicoletto 1993), orange ornamentation (Endler 1980, 1983; Kodric-Brown 1985, 1989; Houde 1988; Long & Houde 1989) and ornament complexity (Endler 1980, 1983; Nicoletto 1993) during mate choice. Many of these ornaments are conditiondependent as measured by prolonged swimming performance, the condition factor (Nicoletto 1993) and parasite load (Kennedy et al. 1987; McMinn 1990; Houde & Torio 1992). In addition female guppies respond sexually significantly more often to males in better physical condition (Nicoletto 1993) and this sexual response is correlated with mating success (Liley 1966; Houde

9 Nicoletto: Offspring quality and female choice in guppies , 1988). I have shown that genetic variation probably exists for prolonged swimming performance and physical condition. However, how much of this variation is additive remains a question. Studies on other fish species have shown that prolonged swimming ability and the condition factor are both correlated with survival components of fitness (Webb 1975; Beamish 1978; Wootton et al. 1978; Booth & Keast 1986; Bolger & Connolly 1989). Thus the potential seems to exist for the condition-dependent model to have an effect in the next generation. It seems that all the requirements exist for condition-dependent female choice to select for elaborate ornamentation. If this is the case, than what explains the lack of differences between the offspring of the different sire types? There are at least three possible reasons for the lack of differences between offspring. First, there were no differences between offspring of different sire types because the sires did not differ. Second, this experiment took place over one generation. The Reynolds & Gross (1992) study took place over three generations. Thus, the relationships between ornamentation and constitution reported elsewhere (Kennedy et al. 1987; Kodric-Brown 1989; McMinn 1990; Nicoletto 1991, 1993; Houde & Torio 1992) may not have been large enough to yield significant differences between offspring given the sample sizes of this study. Third, condition-dependent ornaments are thought to reflect a male s ability to survive and procure resources in the habitat and conditions under which it was born or has lived for a time (Zahavi 1977; Kodric-Brown & Brown 1984; Andersson 1986; Zeh & Zeh 1988). Guppy ornamentation, particularly ornament complexity, orange colour and display rate, are probably condition-dependent (Endler 1980, 1983; Kodric- Brown 1989; Nicoletto 1991, 1993; Houde & Torio 1992). For simplicity I am calling male display rate an ornament, because it probably functions to attract the female s attention. Females choosing males based on conditiondependent ornaments are potentially evaluating information about the environment in which those males live. Offspring reared in the laboratory encountered different environmental conditions from their field-caught sires. A laboratory environment, with its abundant food and constant conditions, may have permitted the expression of ornaments in male offspring that were present, but not expressed, in field-caught sires. It is also possible that the laboratory environment in this study may have been good enough that all the offspring of all the sires had robust constitutions. In either case, the potential information contained in the sire s ornaments did not reflect laboratory conditions that the offspring were going to encounter. If the offspring had been reared in the same environment as the sires, differences may have been observed. In other words, in the laboratory environment we may find an uncoupling of ornamentation and constitution, so differences in the ornamentation of offspring of different sire types were not observed. The uncoupling argument may also explain why the offspring of preferred males were not more ornamented than the offspring of non-preferred males as both the Fisherian and adaptive models predict. There are two lines of evidence that support the idea of uncoupling of ornamentation, orange colour in particular, and constitution. The relative area of orange spots in guppies has Y-linked inheritance and high heritability (Houde 1992) but the phenotypic expression of orange is affected by diet (Endler 1980, 1983; Kodric-Brown 1989) and by parasite load (Kennedy et al. 1987; McMinn 1990; Houde & Torio 1992). In this study, the field-caught sires had significantly fewer orange spots with less relative area than their laboratoryreared sons (Wilcoxon matched-pairs signedranks test, comparing the value of the sires with the average value of the two male offspring, N=81, orange number: P=0 0008; orange area: P=0 0007) and the heritability estimate for orange was not significantly different from zero, indicating little or no heritability. This suggests that the sires may have been unable to express their orange ornamentation because of adverse environmental conditions. The environment of the Jemez Mountain sires is highly oligotrophic and the fish feed primarily on algae. I examined the gut contents of 27 guppies in 1987 from the Jemez population and found all to contain algae, and only one large female contained a single unidentified insect. Although nothing is known about the development of orange ornamentation in guppies reared exclusively on a diet of algae, they are known to have slower growth rates than guppies fed Daphnia or Tetramin (Dussault & Kramer 1981). The controversy between the Fisherian and adaptive schools of female choice has stimulated

10 386 Animal Behaviour, 49, 2 much theoretical and empirical research in the past 10 years. There is now both theoretical (Andersson 1982, 1986; Hamilton & Zuk 1982; Pomiankowski 1987a, b, 1988; Grafen 1990) and empirical (Andersson 1989; Kodric-Brown 1989; McLennan & McPhail 1989; Hill 1990, 1991; Ligon et al. 1990; Milinski & Bakker 1990; Møller 1990; Zuk et al. 1990) support for adaptive female choice, but one major problem still remains. That problem is demonstrating paternal effects on offspring quality. Paternal effects have been shown on the growth rates of offspring in toads (Mitchell 1990), junglefowl (Zuk et al. 1990) and guppies (Reynolds & Gross 1992); however, as yet no study has demonstrated a positive relationship between ornament expression and offspring viability or fitness. The breeding experiment reported here showed that there were maternal effects on offspring quality, but I was unable to detect paternal effects. ACKNOWLEDGMENTS I thank A. Kodric-Brown, J. Endler, L. Hawkins, M. Molles and two anonymous referees for making suggestions on the manuscript. This research was supported by the Biology Department, Student Research Allocation Committee and the Biology Graduate Research Allocation Committee of the University of New Mexico. REFERENCES Andersson, M Sexual selection, natural selection and quality advertisement. Biol. J. Linn. Soc., 17, Andersson, M Evolution of condition-dependent sex ornaments and mating preferences: sexual selection based on viability differences. Evolution, 40, Andersson, S Sexual selection and cues for female choice in leks of Jackson s widowbird, Euplectes jacksoni. Behav. Ecol. Sociobiol., 25, Arnold, S. J Quantitative genetic models of sexual selection. Experientia, 41, Balmford, A. & Read, A. F Testing alternative models of sexual selection through female choice. Trends Ecol. Evol., 6, Beamish, F. W. H Swimming capacity. In: Fish Physiology, Vol. 7 (Ed. by W. S. Hoar & D. J. Randall), pp New York: Academic Press. Becker, W. A Manual of Quantitative Genetics. Pullman, Washington: Academic Enterprises. Bischoff, R. J., Gould, J. L. & Rubenstein, D. I Tail size and female choice in the guppy (Poecilia reticulata). Behav. Ecol. Sociobiol., 17, Boake, C. R. B Repeatability: its role in evolutionary studies of mating behavior. Evol. Ecol., 3, Bolger, T. & Connelly, P. L The selection of suitable indices for the measurement and analysis of fish condition. J. Fish Biol., 34, Booth, D. J. & Keast, J. A Growth energy partitioning by juvenile bluegill sunfish, Lepomis macrochirus Rafinesque. J. Fish Biol., 28, Brett, J. R The respiratory metabolism and swimming performance of young sockeye salmon. J. Fish. Res. Bd Can., 21, Brett, J. R. & Glass, N. R Metabolic rates and critical swimming speeds of sockeye salmon (Oncorhynchus nerka) in relation to size and temperature. J. Fish. Res. Bd Can., 30, Conover, W. J. & Iman, R. L Rank transformations as a bridge between parametric and nonparametric statistics. Am. Stat., 35, Dussault, G. V. & Kramer, D. L Food and feeding behavior of the guppy, Poecilia reticulata (Pisces: Poeciliidae). Can. J. Zool., 59, Endler, J. A Natural selection on color patterns in Poecilia reticulata. Evolution, 34, Endler, J. A Natural and sexual selection on color patterns in poeciliid fishes. Environ. Biol. Fish., 9, Falconer, D. S Introduction to Quantitative Genetics. New York: John Wiley. Farr, J. A Social behavior patterns as determinants of reproductive success in the guppy, Poecilia reticulata Peters (Pisces: Poeciliidae): an experimental study of the effects of intermale competition, female choice, and sexual selection. Behaviour, 74, Grafen, A Sexual selection unhandicapped by the Fisher process. J. theor. Biol., 144, Hamilton, W. D. & Zuk, M Heritable true fitness and bright birds: a role for parasites? Science, 218, Heisler, I. L., Andersson, M. B., Arnold, S. L., Boake, C. R., Borgia, G., Hausfater, G., Kirkpatrick, M., Lande, R., Maynard Smith, J., O Donald, P., Thornhill, R. & Weissing, F. J The evolution of mating preferences and sexually selected traits. In: Sexual Selection: Testing the Alternatives (Ed. by J. W. Bradbury & M. B. Andersson), pp New York: John Wiley. Hill, G. E Female house finches prefer colourful males: sexual selection for a condition-dependent trait. Anim. Behav., 40, Hill, G. E Plumage coloration is a sexually selected indicator of male quality. Nature, Lond., 350, Houde, A. E Mate choice based upon naturally occurring color-pattern variation in a guppy population. Evolution, 41, Houde, A. E The effects of female choice and male male competition on the mating success of male guppies. Anim. Behav., 36,

11 Nicoletto: Offspring quality and female choice in guppies 387 Houde, A. E Sex-linked heritability of sexually selected character in a natural population of Poecilia reticulata (Pisces: Poeciliidae) (guppies). Heredity, 69, Houde, A. E. & Endler, J. A Correlated evolution of female mating preferences and male color patterns in the guppy Poecilia reticulata. Science, 248, Houde, A. E. & Torio, A. J Effect of parasitic infection on male color pattern and female choice in guppies. Behav. Ecol., 3, Jones, D. R., Kiceniuk, J. W. & Bamford, O. S Evaluation of swimming performance of several fish species from the Mackenzie river. J. Fish. Res. Bd Can., 31, Kennedy, C. E. J., Endler, J. A., Poynton, S. & McMinn, H Parasite load predicts mate choice in guppies. Behav. Ecol. Sociobiol., 21, Kirkpatrick, M Sexual selection and the evolution of female choice. Evolution, 36, Kirkpatrick, M Sexual selection by female choice in polygynous animals. A. Rev. Ecol. Syst., 18, Kirkpatrick, M. & Ryan, M. J The evolution of mating preferences and the paradox of the lek. Nature, Lond., 350, Kodric-Brown, A Female preference and sexual selection for male coloration in the guppy (Poecilia reticulata). Behav. Ecol. Sociobiol. 17, Kodric-Brown, A Dietary carotenoids and male mating success: an environmental component of female choice. Behav. Biol. Sociobiol., 25, Kodric-Brown, A. & Brown, J. H Truth in advertising: the kinds of traits favored by sexual selection. Am. Nat., 124, Lande, R Models of speciation by sexual selection on polygenic traits. Proc. natn. Acad. Sci. U.S.A., 78, Ligon, L. D., Thornhill, R. & Zuk, M Male male competition, ornamentation and the role of testosterone in sexual selection in red jungle fowl. Anim. Behav., 40, Liley, N. R Ethological isolating mechanisms in four sympatric species of poeciliid fishes. Behav. Suppl., 13, Long, K. D. & Houde, A. E Orange spots as a visual cue for female mate choice in the guppy Poecilia reticulata. Ethology, 82, McLennan, D. A. & McPhail, J. D Experimental investigations of the evolutionary significance of sexually dimorphic nuptial coloration in Gasterosteus aculeatus (L.): the relationship between male color and male behavior. Can. J. Zool., 67, McMinn, H Effects of the nematode Camallanus cotti on sexual and non-sexual behavior in the guppy (Poecilia reticulata). Am. Zool., 30, Maynard Smith, J Sexual selection, handicaps and true fitness. J. theor. Biol., 115, 1 8. Milinski, M. & Bakker, C. M Female sticklebacks use male coloration in mate choice and hence avoid parasitized males. Nature, Lond., 344, Mitchell, S. L The mating system genetically affects offspring performance in Woodhouse s toad (Bufo woodhousei). Evolution, 44, Møller, A. P Effects of a haematophagous mite on the barn swallow (Hirundo rustica): a test of the Hamilton and Zuk hypothesis. Evolution, 44, Nicoletto, P. F The relationship between male ornamentation and swimming performance in the guppy, Poecilia reticulata. Behav. Ecol. Sociobiol., 28, Nicoletto, P. F Female sexual response to condition-dependent ornaments in the guppy, Poecilia reticulata. Anim. Behav., 46, Nur, N. & Hasson, O Phenotypic plasticity and the handicap principle. J. theor. Biol., 110, Partridge, L Mate choice increases a component of offspring fitness in fruit flies. Nature, Lond., 283, Pomiankowski, A. N. 1987a. The costs of choice in sexual selection. J. theor. Biol., 128, Pomiankowski, A. N. 1987b. Sexual selection: the handicap principle does work, sometimes. Proc. R. Soc. Lond. Ser. B, 231, Pomiankowski, A. N The evolution of female mate preferences for male genetic quality. Oxf. Surv. evol. Biol., 5, Reynolds, J. D. & Gross, M. R Female mate preference enhances offspring growth and reproduction in a fish, Poecilia reticulata. Proc. R. Soc. Lond. Ser. B, 250, Rocchetta, G., Vanelli, M. L. & Pancaldi, C Body weight and growth rate in laboratory lines of Poecilia reticulata reared on two different diets. Theor. appl. Genet., 69, SAS SAS/STAT User s Guide. Release 6.03 edn. Cary, North Carolina: SAS Institute. Smit, H Some experiments on the oxygen consumption of goldfish (Carassius auratus L.) in relation to swimming speed. Can. J. Zool., 43, Webb, P. W Hydrodynamics and energetics of fish propulsion. Bull. Fish. Res. Bd Can., 190, Webster s New Collegiate Dictionary Springfield, Missouri: G. & C. Merriam Company. Wootton, R. J., Evans, G. W. & Mills, L Annual cycle in female three-spined sticklebacks (Gasterosteus aculeatus L.) from an upland and lowland population. J. Fish Biol., 12, Zahavi, A Mate selection: a selection for a handicap. J. theor. Biol., 53, Zahavi, A The cost of honesty. J. theor. Biol., 67, Zeh, D. W. & Zeh, J. A Condition-dependent sex ornaments and field test of sexual-selection theory. Am. Nat., 132, Zuk, M., Thornhill, R. & Ligon, J. D Parasites and mate choice in red junglefowl. Am. Zool., 30,