Abstract

The phenotype-linked fertility hypothesis suggests that a female can benefit directly from mate choice when the cues she uses indicate the quantity and/or quality of his spermatozoa. We tested the link between sperm quality and male body size and coloration in the resource-free mating system of the guppy, a tropical fish characterized by strong female choice. Larger males possessed larger testes and are therefore predicted to produce larger numbers of spermatozoa than smaller males. Larger male guppies also produced longer spermatozoa than smaller males. Degree of carotenoid coloration did not predict either the quantity or the quality of a male's spermatozoa. These results are consistent with a previous study that showed that female guppies in the study population prefer larger males to brightly colored males. The male-size directed increase in spermatozoon size may be the result of interplay between sperm competition and the coevolution of spermatozoon traits with the female reproductive tract.

Much work on sexual selection has been concerned with understanding the evolution and maintenance of precopulatory (female) mate choice and in particular the possible benefits that females are likely to gain from such behavior (Andersson 1994). Female choice has been said to evolve because it provides females with direct (material) and/or indirect (genetic) benefits (Kirkpatrick and Ryan 1991; Kokko et al. 2003). Direct benefits may include increased access to resources or a reduction in mating costs, whereas indirect benefits are said to accrue through the increased attractiveness or viability of a female's offspring when mated with a more attractive/fitter male (Kirkpatrick and Ryan 1991).

The guppy, Poecilia reticulata, is a live-bearing poeciliid fish native to Trinidad and Tobago and northern parts of South America and is widely used as a model organism in sexual selection studies due to the conspicuous color patterns and courtship displays that characterize males of this species (Houde 1997). The carotenoid colors are the male phenotypic component most consistently used by females during mate choice (reviewed in Houde 1997), but male size has also been demonstrated to be an important trait for female choice in some wild and feral populations (Reynolds and Gross 1992; Houde 1997; Watt et al. 2001).

The resource-free nature of the guppy mating system has led many to believe that precopulatory female choice of male guppies evolved primarily through the acquisition of indirect genetic benefits to a female's offspring (Houde 1997). However, although a number of studies have presented data that are consistent with the acquisition of genetic benefits, the potential role of direct benefits (e.g., reduced risks of infection and predation), and in particular direct fertility benefits, has not been fully evaluated (Houde 1997).

In birds, it has been suggested that females can gain direct fertility benefits from extrapair copulations if male functional fertility covaries with aspects of the male phenotype used by females in mate choice (Sheldon 1994). However, although little evidence among bird species supports this phenotype-linked fertility hypothesis (Sheldon 1994; Peters et al. 2004), there is growing evidence among fish species that a male's phenotype reflects his underlying fertility (Engen and Folstad 1999; Kortet et al. 2004; Måsvær et al. 2004). In the guppy, male phenotypic traits consistently used by female guppies during mate choice, such as the degree of carotenoid coloration, have been shown to correlate positively with both stripped (Pitcher and Evans 2001) and natural (Pilastro et al. 2002) ejaculate size, as well as with the share of paternity in sperm competition experiments (Evans et al. 2003). However, at present no study has tested empirically whether spermatozoon quality in the guppy is reflected in the male phenotype.

The aim of this paper was to assess whether the phenotypic traits used by female guppies during precopulatory mate choice reflect the underlying quantity and, more importantly, quality of a male's spermatozoa. Spermatozoon quality, defined as the size and velocity of spermatozoa, was examined along with the males' standard length and relative area of carotenoid coloration. Spermatozoon size and motility have been shown to have important roles not only in the fertilization rates of single pair matings (Levitan 2000; Kime et al. 2001; Vladić et al. 2002) but also in deciding the outcome of sperm competition during multimale matings across a range of taxa (Radwan 1996; Donoghue et al. 1999; Oppliger et al. 2003; Gage et al. 2004).

METHODS

Experimental animals

Males (n = 65) in this study were descendants of a population of feral fish kept at the University of Leicester Botanical Gardens that originated from a group of 10 males and 10 females introduced to the gardens from Trinidad back in the late 1970s. All males were sexually mature, around 6 months old, and maintained in male-biased (approximately 3:1) mixed-sex aquaria, at 26 ± 1 °C with a 12:12 h light:dark lighting regime. Food consisting of freshly hatched brine shrimp and tropical flake food was provided ad libitum twice a day. Males were isolated from females 3 days prior to ejaculate stripping in order to ensure that spermatozoa reserves were replenished (Pilastro and Bisazza 1999).

Ejaculate stripping

The ejaculate stripping procedure was as described in Matthews et al. (1997) with 2 exceptions. First, males were anesthetized with a 0.09% solution of 2-phenoxyethanol. Second, the ejaculate was stripped into a drop of guppy saline solution (207 mM NaCl, 5.4 mM KCl, 1.3 mM CaCl2, 0.49 mM MgCl2, 0.41 mM MgSO4, 10 mM Tris (Cl), pH 7.5; Gardiner 1978) placed near the base of the gonopodium. This drop of saline helped draw the spermatozoa away from the body of the fish and also reduced their adhesion to the receptacle on which the fish was being stripped. Male standard length (distance between the tip of the snout and the end of the tail root or caudal peduncle) was also measured during this procedure using manual calipers accurate to 0.02 mm. Spermatozoa were obtained from 57 of the 65 males used in this study.

Assessment of spermatozoon motility

The spermatozoa of the guppy are produced in the form of unencapsulated sperm aggregations known as spermatozeugmata (Constantz 1984). On insemination into the female reproductive tract, the spermatozeugmata show a rapid (approximately 60 s) dissociation into free-swimming spermatozoa, and within 15 min, these spermatozoa migrate to the ovarian cavity (Hamlett WC, Greven H, and Schindler J, unpublished data). The fertilization dynamics of the guppy are poorly understood. The eggs appear to be fertilized over a 2- to 8-day period following parturition of the previous brood (Turner 1937; Thibault and Schultz 1978), but the fate of inseminated spermatozoa during this fertilization period is not known. It is possible that they are temporarily stored until eggs are ready to be fertilized, either within an anterodorsally located region known as the “seminal receptacle” or in narrow invaginations of the ovarian epithelia, called “delles,” which penetrate down toward each egg (Jalabert and Billard 1969; Kobayashi and Iwamatsu 2002). It is not clear, however, how the spermatozoa then penetrate through the delle epithelia layers to meet the egg surface (Kobayashi and Iwamatsu 2002).

The speed at which a male's spermatozoa initially dissociate may be important in deciding which spermatozoa reach the delles first and therefore have the best chance of fertilizing an egg. Similarly, the rate at which a male's spermatozoa move following arrival in the delles may further influence their chances of maintaining their position in the delles prior to penetrating the epithelia and reaching the egg surface. Spermatozoon motility measurements were therefore carried out for 15 min from the point of initial activation to mimic the time period during which naturally inseminated spermatozoa are likely to reach and take up position in the delles. Spermatozoon motility was quantified during the first 60 s of activity (the period of initial dissociation described above, hereafter called ID motility) and again 15 min later (the period of ovarian migration described above, hereafter called OM motility).

Spermatozoa were activated using a solution of 150 mM KCl containing dried skimmed milk powder (Marvel, Birmingham, UK; 150 mg per 10 ml of 150 mM KCl). To prevent adhesion of the spermatozoa, glass slides and coverslips were wiped with a water repellent solution (Rain-X, Pennzoil Quaker State Ltd., Bishop's Stortford, UK). The addition of 4 wax “pillars” approximately 0.5 mm thick to the slide prevented a coverslip squashing the spermatozeugma prior to activation.

Spermatozoon motility was recorded on a standard VHS video recorder connected to a Hitachi CCD black and white video camera (Model KP-M1E/K) and an Olympus BX41 negative phase contrast microscope with a X20 negative phase objective. All recordings of motility were carried out at 26 ± 1 °C in order to mimic the body temperature of females in which normal spermatozoa motility would occur and to limit any temperature-mediated changes in spermatozoon motility.

To record ID motility, 10 μl of the activation solution was added to a pretreated glass slide along with an individual spermatozeugma isolated from the guppy saline solution using a pulled and fire-polished glass micropipette (Brand microhaematocrit tubes, Na-hep, 75 mm long) and covered with a coverslip. Filming started once spermatozoa were observed to leave the spermatozeugma and continued for 90 s. Where possible, 3 spermatozeugmata were filmed for each male.

OM motility was recorded by adding 4 spermatozeugmata to a 0.5-ml microfuge tube containing 40 μl of the activation solution. After 15 min, the tube was gently agitated to ensure a homogenous mix of the spermatozoa in the solution without damaging them. A 10-μl sample of the agitated mixture was then removed and added to a pretreated glass slide and filming was started immediately and motility recorded over a 30-s period. Again, where possible 3 replicates were filmed for each male.

Quantifying spermatozoon motility

Spermatozoon motility was quantified using the Hobson Sperm Tracker package (Hobson Tracking Systems Ltd, Sheffield, UK). Kime et al. (2001) recommend that the 3 most useful parameters for assessing spermatozoon motility in fish are 1) the curvilinear velocity (VCL − average velocity [μms−1] measured over the whole track), 2) the straight-line velocity (VSL − average velocity [μms−1] measured over a straight line from the start to the end of the track), and 3) the linearity (LIN − average value of the ratio VSL/VCL [%], which estimates the proximity of the cell's track to a straight line). We found strong correlations among these parameters at each time point, and therefore, only the VSL was used in the subsequent analyses.

ID motility analysis was quantified for 60 s out of the 90-s recording sequences to capture a representative sample of the variation during the dissociation period. OM motility analysis was restricted to the first 15 s of the 30-s recorded period because spermatozeugmata had fully dissociated at this time point.

Assessment of spermatozoon size

Free spermatozoa were obtained by dissociating 10–20 spermatozeugmata from each male in 100–200 μl of activation solution and fixing them in a 5% formalin solution after 15 min. A 5-μl sample of the fixed spermatozoa was placed onto a standard glass slide with a coverslip and left to dry for 10 min. The sample was viewed at 400× magnification using a Leitz Laborlux S microscope, and only spermatozoa that appeared intact without broken flagella or distorted heads/midpieces were selected for analysis. Images of these spermatozoa were captured using an attached Spot CCD camera (Insight QE Model 4.2, Diagnostic Instruments Inc., Sterling Heights, MI) connected to the Leica IM50 image manager software package (Version 1.2, Leica Microsystems AG, Glattbrugg, Switzerland) and measured using the Scion image analysis software (www.scioncorp.com).

We measured the length (μm) of the head, midpiece, and flagellum of the spermatozoa, as well as the areas (μm2) of the head and midpiece region. Twenty spermatozoa were measured for each male (n = 54), except in the case of 3 males where only 19, 19, and 13 spermatozoa were measured, respectively (total of 1131 spermatozoa measured).

Assessment of male phenotype

Approximately 1 month following the ejaculate-stripping procedure, males were euthanased via overanesthetization and body mass (MB) and testes mass (MT) were measured (wet masses accurate to 0.1 mg). Testes mass was used as a surrogate of the size of a male's spermatozoa reserves because the need to use large numbers of spermatozeugmata for other purposes meant that total stripped ejaculate could not be quantified. Still images of each male's right lateral view were also captured using a Vista CCD camera (Model NCL5132DSP) attached to a standard VHS video recorder. Due to fish losses, and the use of some males in another study, body mass and testes mass measurements were obtained from only 37 of the original 57 males.

Still images of each male's lateral view were digitized through an imagenation frame grabber card (Model PXC200A) using the PXCBASIC software provided. Images were analyzed, and the extent of the carotenoid-colored regions of the total lateral body area (excluding fins) were quantified using the Scion image analysis software described previously. Calibration of measurements (in mm2) was carried out using the standard length measurements recorded earlier. To control for any effects of body size, the absolute extent of the carotenoid-colored regions of the body was converted to relative extents by dividing the absolute values by the total lateral body area (excluding fins). The image obtained from one male was of insufficient quality to be able to accurately quantify the degree of carotenoid coloration, and so measurements were made from the remaining 36 males only.

Statistical analysis

All data were checked for normality (Kolmogorov–Smirnov test). As a result, relative body area measures for the carotenoid-colored regions of each male were arcsine square root transformed to increase their approximation to a normal distribution. Comparisons between males for both the size of spermatozoon components and the speed of spermatozoon motility were conducted using nonparametric Kruskal–Wallis tests due to heterogeneity of variance. In addition, the significance of correlation analyses was adjusted for the effects of multiple comparisons using the sequential Bonferroni correction (Rice 1989). All statistical analyses were conducted using SPSS version 11.0.1. Descriptive values where reported are mean ± standard deviation.

RESULTS

Spermatozoon motility parameters

Spermatozoon motility estimates for ID and OM motilities were available for 35 males and 26 males, respectively. This was due to low numbers of spermatozeugmata and the lack of intact spermatozeugmata in the stripped ejaculates of some males.

VSL averaged 94.47 ± 22.84 and 38.54 ± 12.67 μms−1 for the ID and OM motility, respectively. More variation occurred between males than within males for this motility parameter at both time points (ID: H = 53.25, degrees of freedom [df] = 32, P = 0.011; OM: H = 55.01, df = 24, P < 0.001), and so the mean estimates for each male at each time point were used in the subsequent analyses.

Spermatozoon size parameters

Males differed significantly in the mean size of all spermatozoon measures (Table 1). Males with a longer flagellum had significantly shorter midpieces (Table 2; r = −0.48, n = 57, P < 0.001), but midpiece area was significantly larger in males with larger VSL for OM motility (r = 0.58, n = 26, P = 0.002).

Table 1

The mean (±standard deviation [SD]) dimensions of the different components measured on the spermatozoa of the guppy across all males sampled (n = 57)

Spermatozoon component Mean ± SD Kruskal–Wallis 
  H df P 
Head length (μm) 4.01 ± 0.14 497.73 56 <0.001 
Midpiece length (μm) 4.79 ± 0.70 528.64 56 <0.001 
Flagellum length (μm) 45.76 ± 1.56 596.01 56 <0.001 
Area of head region (μm24.87 ± 0.55 548.62 56 <0.001 
Area of midpiece region (μm24.59 ± 0.63 420.16 56 <0.001 
Total spermatozoon length (μm) 54.56 ± 1.39 624.58 56 <0.001 
Spermatozoon component Mean ± SD Kruskal–Wallis 
  H df P 
Head length (μm) 4.01 ± 0.14 497.73 56 <0.001 
Midpiece length (μm) 4.79 ± 0.70 528.64 56 <0.001 
Flagellum length (μm) 45.76 ± 1.56 596.01 56 <0.001 
Area of head region (μm24.87 ± 0.55 548.62 56 <0.001 
Area of midpiece region (μm24.59 ± 0.63 420.16 56 <0.001 
Total spermatozoon length (μm) 54.56 ± 1.39 624.58 56 <0.001 

Kruskal–Wallis results refer to tests of the differences among males in the mean size of each spermatozoon component.

Table 2

Correlation matrix for mean spermatozoon size components (head length, midpiece length, flagellum length, head area, midpiece area, and total length)

Spermatozoon component Head length Midpiece length Flagellum length Total length Head area 
Midpiece area 0.195(0.15) 0.608(0.00) −0.184(0.17) 0.119(0.38) 0.261(0.05) 
Head area 0.701(0.00) −0.098(0.47) 0.293(0.027) 0.350(0.008)  
Total length 0.206(0.12) −0.031(0.82) 0.888(0.00)   
Flagellum length 0.080(0.56) 0.479(0.00)    
Midpiece length 0.028(0.83)     
Spermatozoon component Head length Midpiece length Flagellum length Total length Head area 
Midpiece area 0.195(0.15) 0.608(0.00) −0.184(0.17) 0.119(0.38) 0.261(0.05) 
Head area 0.701(0.00) −0.098(0.47) 0.293(0.027) 0.350(0.008)  
Total length 0.206(0.12) −0.031(0.82) 0.888(0.00)   
Flagellum length 0.080(0.56) 0.479(0.00)    
Midpiece length 0.028(0.83)     

Numbers outside brackets represent the Pearson correlation value, whereas numbers inside brackets represent the associated P values (n = 57 in all cases). Probabilities that remained significant following Bonferroni correction for multiple comparisons are highlighted in bold.

Male phenotype

Male mean length was 17.39 ± 1.67 mm (range 13.38–20.31 mm; n = 57), and mean MB was 82.01 ± 20.82 mg (n = 37). Mean MT of the 37 males measured was 3.51 ± 1.76 mg with a mean gonadosomatic index [GSI: (MT/MB) × 100] of 4.28 ± 1.89. Larger males had larger absolute testes mass than small males (Figure 1). There was, however, no significant difference in GSI between different sized males (r = 0.060, n = 37, P = 0.72).

Figure 1

Relationship between testes mass and body mass (Regression: y = 0.044x − 0.13, r = 0.53, n = 37, P = 0.001).

Figure 1

Relationship between testes mass and body mass (Regression: y = 0.044x − 0.13, r = 0.53, n = 37, P = 0.001).

Carotenoid-colored patches covered a mean 14.41 ± 4.80% of a male's total body area, but the relative extent of these patches was unrelated to male standard length (r = 0.27, n = 36, P = 0.11) or MT (r = 0.020, n = 36, P = 0.91). VSL was unrelated to male standard length (ID: r = 0.048, n = 35, P = 0.79; OM: r = 0.040, n = 26, P = 0.85) and relative carotenoid area (ID: r = −0.18, n = 28, P = 0.35; OM: r = 0.004, n = 21, P = 0.99).

Correlation analyses were used to investigate whether male phenotypic traits (standard length, relative carotenoid area) were good indicators of the size of each spermatozoon component. There were no significant correlations between the relative size of the carotenoid regions and any of the spermatozoon components measured. However, male standard length correlated positively with both the length of the flagellum (Table 3; r = 0.37, n = 57, P = 0.005) and the total length of the spermatozoon (Figure 2).

Figure 2

Correlation between mean spermatozoon total length and male standard length (r = 0.41, n = 57, P = 0.001).

Figure 2

Correlation between mean spermatozoon total length and male standard length (r = 0.41, n = 57, P = 0.001).

Table 3

Correlations between male phenotypic traits (standard length and relative carotenoid area) and mean spermatozoon size components (head length, midpiece length, flagellum length, head area, midpiece area, and total length)

 Standard length Relative carotenoid area 
Head length 0.25(0.058) −0.16(0.36) 
Midpiece length −0.046(0.74) −0.18(0.29) 
Flagellum length 0.37(0.005) 0.19(0.27) 
Head area 0.32(0.014) −0.11(0.52) 
Midpiece area 0.22(0.100) −0.16(0.34) 
Total length 0.41(0.001) 0.082(0.63) 
 Standard length Relative carotenoid area 
Head length 0.25(0.058) −0.16(0.36) 
Midpiece length −0.046(0.74) −0.18(0.29) 
Flagellum length 0.37(0.005) 0.19(0.27) 
Head area 0.32(0.014) −0.11(0.52) 
Midpiece area 0.22(0.100) −0.16(0.34) 
Total length 0.41(0.001) 0.082(0.63) 

Numbers outside brackets represent the Pearson correlation value, whereas numbers inside brackets represent the associated P values (n = 57 and n = 36 for the standard length and relative carotenoid area analyses, respectively). Probabilities that remained significant following Bonferroni correction for multiple comparisons are highlighted in bold.

DISCUSSION

Male size was found to be a good predictor of both the potential size of a male's spermatozoa reserves and also the length of the spermatozoa contained within those reserves, with larger males producing larger numbers and longer spermatozoa than smaller males. There was no relationship between either the potential size or the quality of the spermatozoa reserves and the relative size of a male's carotenoid-colored area. Male size therefore better predicted a male's potential fertility than carotenoid coloration, consistent with the evidence that females in this population show preferences for large males rather than brightly colored males (Watt et al. 2001).

The number of spermatozoa in the stripped ejaculate of the guppy has been shown previously to correlate with a number of aspects of male phenotype, including body size (Kuckuck and Greven 1997; Pilastro and Bisazza 1999; Pitcher and Evans 2001). Mature fish testes consist mainly of spermatozoon cells (Billard 1986), and so it is generally assumed that testes mass will be a good indicator of the number of spermatozoa produced, as has been shown in a number of fish species (Marconato and Shapiro 1996; Zbinden et al. 2001) including the guppy (Haubruge et al. 2000). Relative testes mass (measured here as GSI) is predicted to increase with sperm competition risk and/or intensity (Parker 1998), and comparative studies have demonstrated this across species of fish (Stockley et al. 1997), birds (Birkhead 1998), and mammals (Gomendio et al. 1998). In the present study, however, the lack of a correlation between male standard length and GSI implies that sperm competition intensity does not differ across the range of male sizes studied and is therefore unlikely to explain the male-size directed increases in either spermatozoa number or size.

Spermatozoon size is predicted to evolve around narrow optima under the influence of natural selection and/or sexual selection and can be best explained by the marginal value theorem (Parker 1993). For example, in the moth Plodia interpunctella, when nutritional resources are limited, males maintain spermatozoon size but alter spermatozoa numbers, suggesting the former spermatozoon trait is more critical to fitness (Gage and Cook 1994). Similarly, sperm competition theory predicts that under increasing levels of sperm competition risk/intensity, the size of spermatozoa will remain the same but the number of spermatozoa will increase (Parker 1993). However, spermatozoon size can either increase or decrease under increasing levels of sperm competition depending on the mode of fertilization and how spermatozoon size affects spermatozoon longevity (Parker 1998). Although comparative studies across species have demonstrated both positive (birds, Briskie et al. 1997; frogs, Byrne et al. 2003; butterflies, Gage 1994; mammals, Gomendio and Roldan 1991; moths, Morrow and Gage 2000) and negative (fish, Stockley et al. 1997) effects of increasing sperm competition on spermatozoon size, within species there has generally been less evidence of any sperm competition effect (Simmons et al. 1999; LaMunyon and Ward 2002).

The female reproductive tract may also have important influences on the size of a male's spermatozoa, particularly under sperm competition (Briskie and Montgomerie 1992; Morrow and Gage 2000; Miller and Pitnick 2002; Pitnick et al. 2003). For instance, the dimensions of a female's reproductive tract are likely to influence whether large spermatozoa are competitively advantaged (through increased swimming speed; Gomendio and Roldan 1991) or disadvantaged (through reduced longevity; Gomendio and Roldan 1993), relative to smaller spermatozoa. Furthermore, increasing the size of the spermatozoon storage vessel (seminal receptacle) in female fruit flies, Drosophila melanogaster, can lead to an increase in the size of a male's spermatozoa through postcopulatory female choice favoring large spermatozoa (Miller and Pitnick 2002). It is possible that a similar mechanism maintains the male-size directed increase in spermatozoon size observed in this study.

Large male guppies produce offspring with faster growth rates in at least one guppy population (Reynolds and Gross 1992). If large male guppies also possess larger spermatozoa, as suggested here, and these are competitively advantaged, then female preference for large males is likely to also select for an increase in spermatozoon size. Larger males will produce larger females, which will in turn select for males with larger spermatozoa. However, for this mechanism to operate, 2 assumptions must be met. First, larger males must produce larger spermatozoa. Second, larger spermatozoa must be at a competitive advantage when competing against smaller spermatozoa.

Theory suggests that longer spermatozoa will produce larger propulsive forces and swim faster (Katz and Drobnis 1990), though supporting empirical data are lacking (but see Gomendio and Roldan 1991; Casselman and Montgomerie 2004). Evidence from a range of taxa suggests, however, that both the size (Radwan 1996; Oppliger et al. 2003) and motility (Donoghue et al. 1999; Gage et al. 2004) of spermatozoa can influence their relative success during sperm competition, with larger/faster swimming spermatozoa outcompeting their smaller/slower competitors.

The assumption that larger males produce longer spermatozoa is harder to justify. The available evidence suggests that spermatozoon size is not condition dependent (Gage and Cook 1994; Hellriegel and Blanckenhorn 2002), and so it is unlikely that larger male guppies are simply in better condition. Because guppies exhibit XY sex determination with males as the heterogametic sex (Volff and Schartl 2001), spermatozoon size could be sex linked, as suggested for other species (Ward 2000; Morrow and Gage 2001; Simmons and Kotiaho 2002). Furthermore, body size in guppies shows significant father–son heritability (Reynolds and Gross 1992), and heritability estimates are sufficiently high to suggest Y-linked inheritance (Nakajima and Taniguchi 2002). Perhaps, genes conferring males with large spermatozoa are closely associated with or are the same genes on the Y chromosome that control male body size. Female preference for large male size could therefore indirectly select for males with large spermatozoon size, particularly if such spermatozoa are competitively advantaged.

We found no correlations between spermatozoon quality or quantity and the relative area of carotenoid coloration, implying that this coloration is a poor indicator of a male's potential fertility. This result is somewhat surprising given the consistent preference shown by female guppies for carotenoid coloration (Houde 1997). Guppy color patterns have evolved under the conflicting forces of natural (predation) selection and sexual selection, with males from high-predation environments evolving fewer and smaller color spots (Houde 1997). In addition, carotenoid availability in the environment is thought to further limit the expression of carotenoid-based color patches (Grether et al. 1999). Athough carotenoids in the diet do not affect the area of carotenoid coloration in guppies, they do affect the color saturation, or chroma, of these colors (Grether 2000). Furthermore, the chroma of carotenoid-colored regions is also an important component used during mate choice in the guppy (Kodric-Brown 1989). It is possible, therefore, that we might have found different results had we measured the chroma of carotenoid patches.

Previous work on the study population has, however, found no female preference for either the relative area of carotenoid-colored patches or their chroma (Watt et al. 2001). Grether (2000) suggested that the indicator value of carotenoid ornaments should be higher in environments with limited carotenoid availability because increases in carotenoid uptake can produce considerable increases in the chroma of carotenoid-based ornaments. In contrast, in environments with high carotenoid availability, an increase in carotenoid uptake leads to little change in the chroma of carotenoid-based ornaments (Grether 2000). It is therefore possible that the population used in this study has sufficiently high carotenoid availability to render the carotenoid ornament of little value in conveying information to females. No data are available on carotenoid availability in this population to test this idea.

The results discussed here suggest a potential direct fertility benefit to female choice of male size in the present study population, a fact hitherto unknown in this resource-free mating system. Females in the study population prefer larger males to brightly colored males (Watt et al. 2001), and here we show that male size was a better predictor of both the potential size of spermatozoa reserves and the total length of spermatozoa contained within those reserves, than the degree of carotenoid coloration. The idea that a male's phenotype may reflect the underlying quality of his spermatozoa has important implications for the mating system of the guppy because it provides a potential mechanism by which females can ensure their offspring are sired by the male of their choice (i.e., if males favored during precopulatory mate choice are also favored during postcopulatory sperm competition). Given that sexual selection theory suggests that direct selection on female mating preferences is a relatively greater force than indirect selection (Kirkpatrick and Barton 1997; Kokko et al. 2003), we believe that direct fertility benefits in the guppy deserve further attention before we can fully understand the evolution of this model mating system.

We thank Terry Burke and Klaus Reinhardt for helpful and constructive comments on an earlier draft of this manuscript. We would also like to thank Tim Birkhead for the loan of equipment and valuable advice. All work was conducted under Home Office License PPL40/2520 and supported by a University of Sheffield studentship awarded to A.M.J.S.

References

Andersson
M
Sexual selection
 , 
1994
Princeton (NJ)
Princeton University Press
Billard
R
Spermatogenesis and spermatology of some teleost fish species
Reprod Nutr Dev
 , 
1986
, vol. 
26
 (pg. 
877
-
920
)
Birkhead
TR
Birkhead
TR
Møller
AP
Sperm competition in birds: mechanisms and function
Sperm competition and sexual selection
 , 
1998
London
Academic Press
(pg. 
579
-
622
)
Briskie
JV
Montgomerie
R
Sperm size and sperm competition in birds
Proc R Soc Lond B Biol Sci
 , 
1992
, vol. 
247
 (pg. 
89
-
95
)
Briskie
JV
Montgomerie
R
Birkhead
TR
The evolution of sperm size in birds
Evolution
 , 
1997
, vol. 
51
 (pg. 
937
-
945
)
Byrne
PG
Simmons
LW
Roberts
JD
Sperm competition and the evolution of gamete morphology in frogs
Proc R Soc Lond B Biol Sci
 , 
2003
, vol. 
270
 (pg. 
2079
-
2086
)
Casselman
SJ
Montgomerie
R
Sperm traits in relation to male quality in colonial spawning bluegill
J Fish Biol
 , 
2004
, vol. 
64
 (pg. 
781
-
789
)
Constantz
GD
Smith
RL
Sperm competition in poeciliid fishes
Sperm competition and the evolution of animal mating systems
 , 
1984
Orlando (FL)
Academic Press
(pg. 
465
-
485
)
Donoghue
AM
Sonstegard
TS
King
LM
Smith
EJ
Burt
DW
Turkey sperm mobility influences paternity in the context of competitive fertilization
Biol Reprod
 , 
1999
, vol. 
61
 (pg. 
422
-
427
)
Engen
F
Folstad
I
Cod courtship song: a song at the expense of the dance?
Can J Zool
 , 
1999
, vol. 
77
 (pg. 
542
-
550
)
Evans
JP
Zane
L
Francescato
S
Pilastro
A
Directional postcopulatory sexual selection revealed by artificial insemination
Nature
 , 
2003
, vol. 
421
 (pg. 
360
-
363
)
Gage
MJG
Associations between body size, mating pattern, testis size and sperm lengths across butterflies
Proc R Soc Lond B Biol Sci
 , 
1994
, vol. 
258
 (pg. 
247
-
254
)
Gage
MJG
Cook
PA
Sperm size or numbers? Effects of nutritional stress upon eupyrene and apyrene sperm production strategies in the moth Plodia interpunctella (Lepidoptera: Pyralidae)
Funct Ecol
 , 
1994
, vol. 
8
 (pg. 
594
-
599
)
Gage
MJG
Macfarlane
CP
Yeates
S
Ward
RG
Searle
JB
Parker
GA
Spermatozoal traits and sperm competition in Atlantic salmon: relative sperm velocity is the primary determinant of fertilization success
Curr Biol
 , 
2004
, vol. 
14
 (pg. 
44
-
47
)
Gardiner
D
Utilization of extracellular glucose by spermatozoa of two viviparous fishes
Comp Biochem Physiol
 , 
1978
, vol. 
59A
 (pg. 
165
-
168
)
Gomendio
M
Harcourt
AH
Roldan
ERS
Birkhead
TR
Møller
AP
Sperm competition in mammals
Sperm competition and sexual selection
 , 
1998
London
Academic Press
(pg. 
667
-
756
)
Gomendio
M
Roldan
ERS
Sperm competition influences sperm size in mammals
Proc R Soc Lond B Biol Sci
 , 
1991
, vol. 
243
 (pg. 
181
-
186
)
Gomendio
M
Roldan
ERS
Coevolution between male ejaculates and female reproductive biology in eutherian mammals
Proc R Soc Lond B Biol Sci
 , 
1993
, vol. 
252
 (pg. 
7
-
12
)
Grether
GF
Carotenoid limitation and mate preference evolution: a test of the indicator hypothesis in guppies (Poecilia reticulata)
Evolution
 , 
2000
, vol. 
54
 (pg. 
1712
-
1724
)
Grether
GF
Hudon
J
Millie
DF
Carotenoid limitation of sexual coloration along an environmental gradient in guppies
Proc R Soc Lond B Biol Sci
 , 
1999
, vol. 
266
 (pg. 
1317
-
1322
)
Haubruge
E
Petit
F
Gage
MJG
Reduced sperm counts in guppies (Poecilia reticulata) following exposure to low levels of tributyltin and bisphenol A
Proc R Soc Lond B Biol Sci
 , 
2000
, vol. 
267
 (pg. 
2333
-
2337
)
Hellriegel
B
Blanckenhorn
WU
Environmental influences on the gametic investment of yellow dung fly males
Evol Ecol
 , 
2002
, vol. 
16
 (pg. 
505
-
522
)
Houde
AE
Sex, color and mate choice in guppies
 , 
1997
Princeton (NJ)
Princeton University Press
Jalabert
B
Billard
R
Etude ultrastucturale du site de conservation des spermatozoides dans l'ovaire de Poecilia reticulata (Poisson Teleosteen)
Ann biol anim biochim biophys
 , 
1969
, vol. 
9
 (pg. 
273
-
280
)
Katz
DF
Drobnis
EZ
Bavister
BD
Cummins
J
Roldan
ERS
Analysis and interpretation of the forces generated by spermatozoa
Fertilization in mammals
 , 
1990
Norwell (MA)
Sereno Symposia
(pg. 
125
-
137
)
Kime
DE
Van Look
KJW
McAllister
BG
Huyskens
G
Rurangwa
E
Ollevier
F
Computer-assisted sperm analysis (CASA) as a tool for monitoring sperm quality in fish
Comp Biochem Physiol
 , 
2001
, vol. 
130C
 (pg. 
425
-
433
)
Kirkpatrick
M
Barton
NH
The strength of indirect selection on female mating preferences
Proc Natl Acad Sci USA
 , 
1997
, vol. 
94
 (pg. 
1282
-
1286
)
Kirkpatrick
M
Ryan
MJ
The evolution of mating preferences and the paradox of the lek
Nature
 , 
1991
, vol. 
350
 (pg. 
33
-
38
)
Kobayashi
H
Iwamatsu
T
Fine structure of the storage micropocket of spermatozoa in the ovary of the guppy Poecilia reticulata
Zoolog Sci
 , 
2002
, vol. 
19
 (pg. 
545
-
555
)
Kodric-Brown
A
Dietary carotenoids and male mating success in the guppy: an environmental component to female choice
Behav Ecol Sociobiol
 , 
1989
, vol. 
25
 (pg. 
393
-
401
)
Kokko
H
Brooks
R
Jennions
MD
Morley
J
The evolution of mate choice and mating biases
Proc R Soc Lond B Biol Sci
 , 
2003
, vol. 
270
 (pg. 
653
-
664
)
Kortet
R
Vainikka
A
Rantala
MJ
Taskinen
J
Sperm quality, secondary sexual characters and parasitism in roach (Rutilus rutilus L.)
Biol J Linn Soc
 , 
2004
, vol. 
81
 (pg. 
111
-
117
)
Kuckuck
C
Greven
H
Notes on the mechanically stimulated discharge of spermiozeugmata in the guppy, Poecilia reticulata: a quantitative approach
Z Fischkunde
 , 
1997
, vol. 
4
 (pg. 
73
-
88
)
LaMunyon
CW
Ward
S
Evolution of larger sperm in response to experimentally increased sperm competition in Caenorhabditis elegans
Proc R Soc Lond B Biol Sci
 , 
2002
, vol. 
269
 (pg. 
1125
-
1128
)
Levitan
DR
Sperm velocity and longevity trade off each other and influence fertilization in the sea urchin Lytechinus variegatus
Proc R Soc Lond B Biol Sci
 , 
2000
, vol. 
267
 (pg. 
531
-
534
)
Marconato
A
Shapiro
DY
Sperm allocation, sperm production and fertilization rates in the bucktooth parrotfish
Anim Behav
 , 
1996
, vol. 
52
 (pg. 
971
-
980
)
Måsvær
M
Liljedal
S
Folstad
I
Are secondary sex traits, parasites and immunity related to variation in primary sex traits in the arctic charr?
Proc R Soc Lond B Biol Sci
 , 
2004
, vol. 
271
 (pg. 
S40
-
S42
)
Matthews
IM
Evans
JP
Magurran
AE
Male display rate reveals ejaculate characteristics in the Trinidadian guppy Poecilia reticulata
Proc R Soc Lond B Biol Sci
 , 
1997
, vol. 
264
 (pg. 
695
-
700
)
Miller
GT
Pitnick
S
Sperm-female coevolution in Drosophila
Science
 , 
2002
, vol. 
298
 (pg. 
1230
-
1233
)
Morrow
EH
Gage
MJG
The evolution of sperm length in moths
Proc R Soc Lond B Biol Sci
 , 
2000
, vol. 
267
 (pg. 
307
-
313
)
Morrow
EH
Gage
MJG
Artificial selection and heritability of sperm length in Gryllus bimaculatus
Heredity
 , 
2001
, vol. 
87
 (pg. 
356
-
362
)
Nakajima
M
Taniguchi
N
Genetic control of growth in the guppy (Poecilia reticulata)
Aquaculture
 , 
2002
, vol. 
204
 (pg. 
393
-
405
)
Oppliger
A
Naciri-Graven
Y
Ribi
G
Hosken
DJ
Sperm length influences fertilization success during sperm competition in the snail Viviparus ater
Mol Ecol
 , 
2003
, vol. 
12
 (pg. 
485
-
492
)
Parker
GA
Sperm competition games: sperm size and sperm number under adult control
Proc R Soc Lond B Biol Sci
 , 
1993
, vol. 
253
 (pg. 
245
-
254
)
Parker
GA
Birkhead
TR
Møller
AP
Sperm competition and the evolution of ejaculates: towards a theory base
Sperm competition and sexual selection
 , 
1998
London
Academic Press
(pg. 
1
-
54
)
Peters
A
Denk
AG
Delhey
K
Kempenaers
B
Carotenoid-based bill colour as an indicator of immunocompetence and sperm performance in male mallards
J Evol Biol
 , 
2004
, vol. 
17
 (pg. 
1111
-
1120
)
Pilastro
A
Bisazza
A
Insemination efficiency of two alternative male mating tactics in the guppy (Poecilia reticulata)
Proc R Soc Lond B Biol Sci
 , 
1999
, vol. 
266
 (pg. 
1887
-
1891
)
Pilastro
A
Evans
JP
Sartorelli
S
Bisazza
A
Male phenotype predicts insemination success in guppies
Proc R Soc Lond B Biol Sci
 , 
2002
, vol. 
269
 (pg. 
1325
-
1330
)
Pitcher
TE
Evans
JP
Male phenotype and sperm number in the guppy (Poecilia reticulata)
Can J Zool
 , 
2001
, vol. 
79
 (pg. 
1891
-
1896
)
Pitnick
S
Miller
GT
Schneider
K
Markow
TA
Ejaculate-female coevolution in Drosophila mojavensis
Proc R Soc Lond B Biol Sci
 , 
2003
, vol. 
270
 (pg. 
1507
-
1512
)
Radwan
J
Intraspecific variation in sperm competition success in the bulb mite: a role for sperm size
Proc R Soc Lond B Biol Sci
 , 
1996
, vol. 
263
 (pg. 
855
-
859
)
Reynolds
JD
Gross
MR
Female mate preference enhances offspring growth and reproduction in a fish, Poecilia reticulata
Proc R Soc Lond B Biol Sci
 , 
1992
, vol. 
250
 (pg. 
57
-
62
)
Rice
WR
Analyzing tables of statistical tests
Evolution
 , 
1989
, vol. 
43
 (pg. 
223
-
225
)
Sheldon
BC
Male phenotype, fertility, and the pursuit of extra-pair copulations by female birds
Proc R Soc Lond B Biol Sci
 , 
1994
, vol. 
257
 (pg. 
25
-
30
)
Simmons
LW
Kotiaho
JS
Evolution of ejaculates: patterns of phenotypic and genotypic variation and condition dependence in sperm competition traits
Evolution
 , 
2002
, vol. 
56
 (pg. 
1622
-
1631
)
Simmons
LW
Tomkins
JL
Hunt
J
Sperm competition games played by dimorphic male beetles
Proc R Soc Lond B Biol Sci
 , 
1999
, vol. 
266
 (pg. 
145
-
150
)
Stockley
P
Gage
MJG
Parker
GA
Møller
AP
Sperm competition in fishes: the evolution of testis size and ejaculate characteristics
Am Nat
 , 
1997
, vol. 
149
 (pg. 
933
-
954
)
Thibault
RE
Schultz
RJ
Reproductive adaptations among viviparous fishes (Cyprinodontiformes: Poeciliidae)
Evolution
 , 
1978
, vol. 
32
 (pg. 
320
-
333
)
Turner
CL
Reproductive cycles and superfetation in poeciliid fishes
Biol Bull
 , 
1937
, vol. 
72
 (pg. 
145
-
164
)
Vladić
TV
Afzelius
BA
Bronnikov
GE
Sperm quality as reflected through morphology in salmon alternative life histories
Biol Reprod
 , 
2002
, vol. 
66
 (pg. 
98
-
105
)
Volff
JN
Schartl
M
Variability of genetic sex determination in poeciliid fishes
Genetica
 , 
2001
, vol. 
111
 (pg. 
101
-
110
)
Ward
PI
Sperm length is heritable and sex-linked in the yellow dung fly (Scathophaga stercoraria)
J Zool
 , 
2000
, vol. 
251
 (pg. 
349
-
353
)
Watt
PJ
Shohet
AJ
Renshaw
K
Female choice for good genes and sex-biased broods in guppies
J Fish Biol
 , 
2001
, vol. 
59
 (pg. 
843
-
850
)
Zbinden
M
Largiadèr
CR
Bakker
TCM
Sperm allocation in the three-spined stickleback
J Fish Biol
 , 
2001
, vol. 
59
 (pg. 
1287
-
1297
)