SEARCH

SEARCH BY CITATION

Keywords:

  • heterozygosity;
  • microsatellite;
  • multiple paternity;
  • Poecilliidae;
  • superfetation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References

Heterandria formosa Agassiz, exhibits internal fertilization, internal brooding of embryos, sperm storage and an extreme level of superfetation. In this study we used microsatellite markers to examine variation among seven populations that exhibited significant variance in their histories of population density. We found that the populations were genetically distinct and that the heterozygosity increased as population density increased. We also examined paternity in three of those populations and found that the number of sires per female and the number of sires per brood increased with population density. Overall, the rates of multiple paternity are quite low relative to other species. The correlations with population density suggest that contact rates play a critical role in the breeding system in this species but the low rates of multiple paternity suggest that females may exert control over fertilization of their ova.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References

Polyandrous females and multiply sired broods of offspring characterize the overwhelming majority of animal breeding systems (e.g. Birkhead, 2000). However, the reasons for the prevalence of multiple mating and concurrent multiple paternity (CMP), and the causes of the notable exceptions to the pattern are not fully understood.

One of the traditional hypotheses to explain high rates of CMP is that this phenomenon reflects intense scramble competition among males (or their sperm) for fertilizations (Parker, 1979; Ball & Parker, 1996). In this view, low rates of CMP are attributed to social behaviors that allow a single male to dominate other males and limit access to a female (Hutchings & Myers, 1988; Brockmann et al., 1994; Jaeger et al., 2002). Other theories propose that females can play a large role in determining rates of CMP and that their optimal mating levels may be different than those of males. On the one hand, females can derive a variety of benefits from multiple inseminations (Murie, 1995; Evans & Magurran, 2000; Jennions & Petrie, 2000) and they can exert considerable effort to attract additional males (Berteaux et al., 1999). On the other hand, increased mating activity can be costly to females for a variety of reasons (Daley, 1978; Constantz, 1989; Laumann et al., 1994; Rowe, 1994; Sikstrom et al., 1996; Stouthamer et al., 1999; Wallander et al., 2001) and there are striking examples of females evolving attributes that restrict mating levels (Arnqvist & Rowe, 1995).

Of course, these behavioral interactions occur in a variety of ecological contexts and it is clear that breeding systems and levels of CMP may covary with those contexts (Keller & Reeve, 1994; Moller & Birkhead, 1994; Kelly et al., 1999; Crim et al., 2002; Garner et al., 2002). One of the most important of these ecological contexts is variation in adult density. Consistent variation in adult density is expected to govern a host of evolutionary processes including standing levels of heterozygosity (e.g. Smith et al., 1975), life histories (e.g. Roff, 2002), intensity of sexual selection (e.g. Conner, 1989) and level of sperm competition (e.g. Levitan, 1998; McCarthy & Lindenmayer, 1998). Despite the prominent role that density variation is expected to play, it is unclear how density variation among natural populations affects the level of sexual conflict over mating frequency (sensuHolland & Rice, 1998) thereby ultimately altering levels of CMP.

To understand the role of density variation on breeding systems in natural populations one must find a system in which there is enough variation in each of those parameters to permit rigorous study. Some of the species of livebearing topminnows (Poeciliidae) offer striking opportunities for meeting this challenge. In poeciliids, fertilization is internal, and fertilized ova are retained within the ovary until the embryos are large enough to be self-sufficient and free swimming (Rosen & Bailey, 1963). While males devote enormous efforts in mating activity in all species, there is substantial interspecific variation in the extent to which male activity is divided between attempts at forced insemination and attempts to elicit female cooperation via courtship (Farr, 1989; Ptacek & Travis, 1998); in some species there is substantial intraspecific variation as well (Ptacek & Travis, 1996). High levels of mating activity can be costly to females, especially when forced insemination is common (Constantz, 1989), and females can vary markedly in their preferences for and responses to males with different body sizes and behavioral phenotypes (e.g. Houde & Endler, 1990; Bisazza et al., 2001; Horth, 2003). Sperm storage, which is considered universal among poeciliid females, and superfetation (successive brood overlap in the ovary of a single female), which is a feature of the life history of several poeciliid species, creates substantial opportunities for sperm competition. Within a single female with superfetation there may be several broods, sired by the same male or by different males; it is even possible that offspring within broods are each sired by different males. Finally, several species of poeciliids exhibit substantial, consistent, intraspecific variation in adult density and temporal patterns of density variation (Chick et al., 1992; Leips & Travis, 1999; Trexler et al., 2002) and there is some indication that density variation can exert a considerable influence on sexual and social interactions (e.g. Martin, 1975).

Prior studies of CMP in poeciliids present intriguing patterns that further bolster their potential value for as model systems for understanding the interplay of density and CMP. While the general pattern is for populations to exhibit very high rates of CMP (50–100% of females; Table 1), this rate is known to vary substantially among conspecific populations of one species (Trexler et al., 1997). In some cases, there are significant phenotypic differences between females producing singly sired broods and those producing multiply sired broods (Borowsky & Kallman, 1976; Borowsky & Khouri, 1976; Travis et al., 1990; Greene & Brown, 1991; Trexler et al., 1997).

Table 1.  Estimates of multiple paternity based on genetic-based colour polymorphisms, allozymes, or microsatellites.
SpeciesDegree of multiple paternityMethod of detectionReference
Gambusia affinis (Baird and Girard)56% of females, but probably near 100%AllozymesChesser et al. (1984)
62.5% of broods, 80% of females, 2.13 males per female, 1.75 males per brood Robbins et al. (1987)
49–81% of femalesAllozymesGreene & Brown (1991)
Gambusia holbrooki (Girard )90% of females 2.2 males per femaleMicrosatellitesZane et al. (1999)
Xiphophorus maculatus (Gunther)66% of females 2.18 males per femaleColour patternsBorowsky & Kallman (1976)
Xiphophorus variatus (Meek)42% of femalesColour patternsBorowsky & Khouri (1976)
Poeciliopsis monacha (Miller)23% of femalesAllozymesLeslie & Vrijenhoek (1977)
Poecilia reticulata (Peters)Average of 24–64% femalesMicrosatellitesKelly et al. (1999)
0–100% of femalesMicrosatellitesEvans & Magurran (2001)
Poecilia latipinna (Lesueur)9–85% of femalesAllozymesTrexler et al. (1997)

Heterandria formosa Agassiz, the least killifish, is potentially a very important species for understanding the distribution of paternity rates within the poeciliid fishes and for dissecting the role of density variation in the evolution of breeding systems. It is a small (12–30 cm) live-bearing fish found throughout the coastal plain of the southeastern US in a variety of habitats (Martin, 1980) and at a wide range of densities (Leips & Travis, 1999). Females of this species display the most advanced levels of superfetation of all the poeciliids (D. Reznick, personal communication). As many as six broods of embryos may be found simultaneously in a single female, although brood number typically ranges from two to four (Travis et al., 1987) and varies among populations (Leips & Travis, 1999). Males of this species rely mainly on forced insemination attempts rather than courtship for mating (Farr, 1989). The suggestion that male–male competition and sneak copulations are more important to male mating success than is female preference (Bisazza & Pilastro, 1997; but see Aspbury & Basolo, 2002) leads to the expectation that H. formosa will exhibit high rates of multiple paternity.

In this paper we report novel results on the covariation of adult density regimes, levels of genetic heterozygosity, and patterns of CMP among seven populations of H. formosa. The populations vary in historical regimes of population density, average female size and average brood size. We demonstrate that genetic heterozygosity of these populations increases as population density increases. For three focal populations, which reflect the range of demographic parameters found in the seven populations as a whole, we report rates of multiple paternity, the number of males contributing to the offspring of each female, and the number of males contributing to the offspring of each brood for those females with multiple broods. We show that all three measures increase as historical population density increases.

Samples

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References

We collected individual H. formosa from seven locations in northern Florida (Fig. 1). These were: Trout Pond (TP), a large permanent pond in the Apalachicola National Forest; Little Lake Jackson (LLJ), a small lake at the northern end of Tallahassee in Leon County; the St Mark's River (SMR) in Wakulla County; the headwaters of the Wacissa River (WR) in Jefferson County; MacBride's Slough (MBS), a spring-fed creek near the Wakulla River, and two seasonally fluctuating wetlands in the St Mark's National Wildlife Refuge: Tram Road (TR) wetland, and Gambo Bayou (GB).

image

Figure 1. Location of the seven sampling sites in northern Florida, USA.

Download figure to PowerPoint

The collections were made in June 2001 at all seven locations, and a total of 15 individuals (both males and females) were collected at each site. In addition, the sampling scheme involved more intensive sampling at three focal populations: WR (high population density), TP (low population density) and GB (intermediate population density). From these three populations, we collected at least 20 females and 10 males; approximately 10 females and five males at each site were collected in June 2001, and the remainder was collected in April 2002. The fish were killed in MS222, sexed and measured, and all brooded embryos were dissected from the gravid females. All individuals were immediately frozen at −80 °C.

For each population, we genotyped all the adult males and females. For the three focal populations, we also genotyped offspring, but only those from females which had two or more broods and at least one of those broods with three or more offspring.

DNA extraction and microsatellite amplification

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References

The tail tissue of adults (approximately 20 mg) and all the tissue of eyed-stage embryos (stages 3–5 in Travis et al., 1987) was digested in cetyltrimethylammonium bromide extraction buffer with proteinase K. Total genomic DNA was extracted twice with phenol–chloroform–isoamyl alcohol (25 : 24 : 1) and once with chloroform–isoamyl alcohol (24 : 1). The DNA was precipitated using cold 95% ethanol and then washed with 70% ethanol. For some samples, DNA was extracted using either the DNAEasy Tissue Kit (Qiagen, Inc., Venlo, The Netherlands) (50 samples) or the G-Nome DNA Kit (Qbiogene Inc., Montreal, Canada) (50 samples). Additionally, the DNA from embryos at early stages of development (stage 2 in Travis et al., 1987) was extracted by placing an entire embryo in a solution of Perkin-Elmer PCR Buffer (Perkin Elmer Inc., Wellesley, MA, USA), Tween-80, and proteinase K, and digested at 65 °C for 60 min and 95 °C for 15 min (procedure modified from Simpson et al., 1999). All pellets were resuspended in 10 mm Tris–HCl and stored at −20 °C. DNA was extracted from a total of 516 individuals.

Twenty-six pairs of microsatellite primers from previously published studies on poeciliid fish were tested on the DNA from 30 of the adult H. formosa. We then chose three pairs of primers (Table 2) that amplified the most variable microsatellite regions (greatest number of alleles) in fish from the same and different localities. These three microsatellite regions were then amplified in all individuals.

Table 2.  The primers used to amplify microsatellite regions of Heterandria formosa.
Primer sequenceAnnealing temperature (°C)Reference
TSS005 F: CTT TAA TAC CCA ATC AGT GG53Seckinger et al. (2002)
TSS006 R: CAA CTG GAA GAG GAG TTG TC
TSS013 F: TCA TCT GGA GCA GGC ACA TG58Nakamura (2001)
TSS014 R: GCG TTT GGT TTC CTA CTG AC
TSS051 F: CGC CGC TTA CCA GAA CTT AAT58Nakamura (2001)
TSS052 R: TCA GGC TCT CTG TTT GTC CA

Each forward primer was 5′-end-labelled with HEX (TSS005), 6-FAM (TSS013), or NED (TSS051). Polymerase chain reactions (PCR) were performed following the Gene Scan Reference Guide (Chapter 8, Perkin-Elmer). After incubation at 95 °C for 12 min, reactions were cycled 35 times through 94 °C for 30 s, 53 °C (TSS005) or 58 °C (TSS013 and TSS051) for 30 s, 72 °C for 1 min. The samples were then incubated at 72 °C for 30 min. Each of the resulting PCR products was then analysed on an ABI Prism 310 sequencer (Applied Biosystems Inc., Foster City, CA, USA). The Genescan 400 HD ROX was used as internal marker and fragment sizes were analysed using the Genescan 3.7 software (Perkin-Elmer). The data were imported into the Genotyper 3.6 (Perkin-Elmer) software for automatic genotyping of alleles, and then manually checked for accuracy.

Population density

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References

We estimated the density of adults in each population with methods described in detail in Leips & Travis (1999). We visited each population in May and September of each year (2000, 2001, 2002); on each visit, we threw a 0.50 m2 trap in suitable H. formosa habitat (in water <1 m deep, usually with substantial vegetation cover) three times at each site and counted all fish caught within each trap. At two locations, TP and WR, we sampled more intensively, throwing the trap three times at each of three sites. For this report, we use the average density at each site to make replication levels equal for all populations in all years so that we can examine these data in a balanced analysis. Analyses of otolith increments in these populations of H. formosa indicate that adults mature at 50–60 days of age with a maximum longevity of about 120 days (J. Travis and R. Allman, unpublished data). These data indicate that the adults counted at each visit represent separate generations and are independent estimates of the density regime in each population.

Data analysis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References

For all the seven populations included in this study, estimates of allele frequencies and heterozygosities were performed using the program Cervus 2.0 (Marshall et al., 1998). Estimates for agreement of allele frequencies with Hardy–Weinberg equilibrium and analysis of population differentiation were obtained from the Genepop on the Web program (Raymond & Rousset, 1995; Hendrie et al., 1998).

For the three focal populations we identified paternal alleles by comparing offspring genotypes against those of a known mother. Estimates of minimum sire number for each brood were obtained from the number of segregating paternal alleles; for example, ztwo males were inferred to have sired a brood displaying three or four paternal alleles. We looked for evidence of siring by an additional male for a brood by inspecting the allele combinations at multiple loci; this would be the case, for example, if a unique paternal allele at one locus was paired in a particular offspring with three paternal alleles at another locus. Additionally, for some of the broods in which only two paternal alleles were observed, we looked for evidence of paternity by multiple males due to significant departures from Mendellian expectations for a singly fathered brood. We also calculated, D, the probability of not detecting multiple paternity when it was present, by using the allele frequencies of adults at each site to determine the probability that two or more males shared the same multilocus genotype (as in Westneat et al., 1987; Hanotte et al., 1991).

For all the three focal populations, we calculated the frequency of multiple paternity and the number of males per female based on females that had more than three offspring, and we calculated the number of males per brood based on broods composed of more than three individuals.

We analysed the population density data with a random effects model analysis of variance following Hicks (1973). We considered population to be a random effect with seven levels, year a random effect with three levels, and sampling period within year a random effect with two levels per year. The expected mean squares for this design indicate that the population–year interaction is the appropriate error term for the main effect of population. This interaction in its turn may be tested over the interaction of population and sampling period within year. The effect of year is tested over the effect of sampling period within year, which in its turn is testable over the population–sampling period interaction. The full table of expected mean squares is available upon request.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References

Average adult densities varied by an order of magnitude among populations (Table 3), with WR exhibiting the highest densities and SMR the lowest. The variation among populations (F6, 12 = 7.94, P < 0.01) accounted for 35% of the total variance in the data. The population–sampling period interaction was the only other effect that was significant (F18, 84 = 2.55, P < 0.01) and accounted for 21% of the total variance in the data. These results indicate that despite local heterogeneity in dynamic patterns (the population–sampling period interaction), the variation among populations in adult density was sufficiently consistent that we can consider these populations to have distinct density regimes.

Table 3.  Density of adult fish, female standard length and observed heterozygosity (HO) for the seven populations in this study. Note that the standard error for population density is calculated before removing effects of year and sampling period and overestimates the genuine error.
SiteDensity (fish/0.5 m2)Standard length (mm)Average HOHOTSS005HOTSS013HOTSS051
  1. TP, Trout Pond; GB, Gambo Bayou; WR, Wacissa River; MBS, MacBride's Slough; LLJ, Little Lake Jackson; SMR, St Mark's River; TR, Tram Road.

TP3.5 ± 0.720.35 ± 3.650.5960.5260.5790.684
GB4.9 ± 2.117.14 ± 3.610.7790.6820.7870.868
WR36.6 ± 10.714.59 ± 1.900.8210.7710.9440.750
MBS3.1 ± 1.917.53 ± 3.250.7290.4310.8700.886
LLJ1.7 ± 0.818.00 ± 4.670.6940.6680.7320.684
SMR1.1 ± 0.620.59 ± 2.410.5570.5590.4360.677
TR3.6 ± 0.914.55 ± 2.720.7540.7820.6830.796

Of the three loci used in the analysis, locus TSS051 proved to be most informative, with an average of eight alleles per population. Loci TSS005 and TSS013 had an average of 5.4 and 5.7 alleles per locus, respectively. The heterozygosities (both observed and expected) for the individual loci at each site ranged from 43.1 to 94.4% and are summarized in Table 3 (as determined by the program Cervus 2.0; Marshall et al., 1998).

Allelic and genotypic frequencies in all of the populations conformed to Hardy–Weinberg equilibrium expectations (all P-values greater than 0.10). All seven of the populations were genetically distinct from one another when all three loci were considered (Table 4), although some pairs of populations were genetically indistinguishable at individual loci (as determined by Genepop on the Web program; Raymond & Rousset, 1995; Hendrie et al., 1998).

Table 4.  Probability that the observed allelic distributions are the same between all pairs of sites (using Genepop on the Web; Raymond & Rousset, 1995; Hendrie et al., 1998). The test employs a Markov chain method using the following parameters: dememorization, 1000; batches, 100; iterations per batch, 1000). All standard errors associated with the P-values are <0.01.
 LLJGBMBSTPTRWR
  1. LLJ, Little Lake Jackson; GB, Gambo Bayou; MBS, MacBride's Slough; TP, Trout Pond; TR, Tram Road; WR, Wacissa River; SMR, St Mark's River.

Locus TSS005
 GS0.872     
 MBS<0.0010.002    
 TP<0.001<0.001<0.001   
 TR0.1900.191<0.001<0.001  
 WR0.0290.0380.040<0.0010.090 
 SMR0.3130.2310.009<0.0010.0510.019
Locus TSS013
 GS0.003     
 MBS<0.001<0.001    
 TP<0.001<0.001<0.001   
 TR<0.0010.196<0.001<0.001  
 WR0.0570.0050.433<0.001<0.001 
 SMR<0.001<0.001<0.001<0.001<0.0010.025
Locus TSS051
 GS<0.001     
 MBS0.020<0.001    
 TP<0.001<0.001<0.001   
 TR0.002<0.001<0.001<0.001  
 WR<0.0010.030<0.001<0.001<0.001 
 SMR<0.001<0.001<0.001<0.0010.002<0.001

For these microsatellite loci, the level of genetic variation in a population was directly related to the historical population density regime. The average heterozygosity in each population was positively correlated with the adult density in the population, averaged across sampling periods and years (Fig. 2, Spearman rank correlation = 0.89, P < 0.05, d.f. = 5). It might be argued that the geometric mean adult density is a more genetically relevant measure of the population density regime. We calculated geometric means in two ways, by backtransforming the arithmetic averages of log-transformed data (after adding 0.1 to all values to adjust for replicate trap throws samples in which no adults were captured) and by calculating a geometric mean of the average adult density in each sampling period. Both methods gave the same rank order of populations for the geometric mean density, which differed slightly from the rank order of the arithmetic means provided in Table 3 (TP and TR reversed ranks from the arithmetic means and LLJ and MBS reversed ranks). The average heterozygosity remained positively correlated with the geometric mean adult density (Spearman rank correlation = 0.75, d.f. = 5, NS) but the correlation could be considered significant only with a 0.10 type I error rate.

image

Figure 2. The average observed heterozygosity (HO) of the seven populations in this study as a function of the historical density of each population. Note that the standard error for population density is calculated before removing effects of year and sampling period and overestimates the genuine error.

Download figure to PowerPoint

In the three focal populations, the rate of multiple paternity, the average number of sires per female, and the average number of sires per brood increased as average population density increased (Table 5). The TP population had the lowest rate of multiple paternity (15%), while the WR population had the highest (66%) and GB was intermediate (54%). The number of males siring offspring per female and the number of sires per brood are significantly different (α=0.05) among all pairs of sites (T’ method of multiple comparisons among pairs of mean values with unequal sample sizes; Sokal & Rohlf, 1995). The average numbers of sires per female and per brood are low in comparison with what other studies of poeciliids (Table 1) have suggested.

Table 5.  Estimates of the number of males who contributed sperm to the offspring of females (within females and within broods) and the level of multiple paternity in each focal population. The number of females and the number of broods used to calculate the statistics are in parentheses.
SiteDensity (fish/0.5 m2)Number of sires per female (N)Number of sires per brood (N)Rate of multiple paternity (%)
  1. TP, Trout Pond; GB, Gambo Bayou; WR, Wacissa River.

TP3.5 ± 0.71.15 ± 0.38 (11)1.09 ± 0.29 (20)15
GB4.9 ± 2.11.64 ± 0.67 (13)1.39 ± 0.58 (23)54
WR36.6 ± 10.71.75 ± 0.62 (12)1.66 ± 0.55 (15)66

It is unlikely that the levels of multiple paternity or numbers of sires per female or per brood are substantially higher than these data indicate. The probability that additional males in each population had identical genotypes to identified sires at all three loci (and were therefore undetected in the paternity analysis) is small in each population (0.005 at WR, 0.011 at GB, and 0.033 at TP). It is unlikely the pattern of paternity rates with density is an artefact of small brood sizes; brood sizes are inversely related to density (see also Leips & Travis, 1999) and if our ability to identify multiple paternity were purely a function of brood size variation among populations, we should have found the highest levels at TP and the lowest levels at WR rather than the reverse pattern.

There is unequal representation of alleles from individual sires in each brood of offspring, across all three populations. In females carrying offspring that are sired by multiple males, the primary male accounted for 70.3 ± 11.6% of all the offspring, while the second male accounted for 27.4 ± 10.3% of the offspring.

There is a weak positive relationship between the size of a female and whether her offspring were sired singly or by multiple males (TP: n1 = 4, n2 = 16, 0.05 < P < 0.10; WRA: n1 = 8, n2 = 13, 0.05 < P < 0.10; TP: n1 = 10, n2 = 11, P = 0.10; Kolmogorov–Smirnov two-sample test; Sokal & Rohlf, 1995, p. 434).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References

Our microsatellite markers provided much greater resolution of heterozygosity within populations of H. formosa than that found in previous studies that employed allozymes (Baer, 1998). However, our finding that these seven populations may be quite genetically distinct form one another, even when contained within the same drainage (e.g. TR and GB), bolsters the conclusions of those earlier studies. It also suggests, in view of these substantial distinctions, that it would be worthwhile to revisit the phylogeography of H. formosa.

The positive correlation among populations between average density and average microsatellite heterozygosity is noteworthy, even if it is not surprising (cf. Young et al., 1996). While historical bottlenecks in density have been postulated to explain low current levels of genetic variation in many case studies, attempts to find correlations between the level of genetic variation and the pattern of numerical fluctuations among populations or species have had mixed success (reviewed in Travis, 1990). The failure of many studies to find the predicted correlations is attributable to two factors, the failure to account for how levels of gene flow among populations change as population densities change and the failure of a snapshot of current density variation to reflect longer term variation in numerical dynamics. The available genetic data for H. formosa is consistent with negligible rates of current gene flow beyond very small scales (Baer, 1998). The sampling data presented here reveal density differences that are consistent across at least six generations, and the comparison of these data to older data on some of these populations (Leips & Travis, 1999) suggests that the density differences detected here are longstanding ones. It is possible but, in our view, less likely, that these distinctions were created by founder effects in the process of re-colonizing these coastal plain habitats (Baer, 1998) rather than by the genetic drift that follows long-term density differences.

Across species, genetic diversity explains much of the variance in fitness, and loss of heterozygosity has a deleterious effect on population fitness (Reed & Frankham, 2003). If heterozygosity is correlated with historical population density, as it is in H. formosa, then even rare events of population decline can affect the average fitness of the population. Low rates of multiple paternity may further reduce the average genetic diversity in a population, and could have severe consequences on the evolvability of a population. This study is one of the first to demonstrate these combined effects of population density.

Microsatellite analysis was also useful for determining paternity in H. formosa. This study is the first to report levels of multiple paternity in this species. Even with the better resolving power of microsatellites over other methods for detecting paternity, the level of multiple paternity in H. formosa is quite low relative to other poeciliids (Table 1), indeed the lowest levels reported for this entire group. Given that H. formosa has the highest level of superfetation among poeciliids, it is surprising that the ability to brood multiple clutches simultaneously does not correlate with a higher rate of multiple paternity.

The incidence of multiple paternity in this study is low in populations with low density, and higher in populations with high density. This is consistent with a model in which the frequency of multiple mating is proportional to the incidence of encounters between males and females. In low-density populations, a female may have little opportunity to be selective among potential mates and there may be little male–male competition. In high-density populations, females may have more selective ability, and there may be more competition among males for females and more opportunities for sperm competition. It is possible that higher multiple paternity rates at higher densities reflect merely a higher cumulative diversity of sires from stored sperm. Two factors make this hypothesis difficult to assess; adult longevity decreases dramatically with increases in density (Leips et al., 2000) and paternity was divided very unequally between sires. Whether female responses to males differ among these populations because of their historical density differences and any type of density-dependent selective effects on females remains to be determined.

It is important to recognize that the rates of multiple paternity in populations with low heterozygosity can be underestimates because of our inability to detect additional males with identical genotypes. However, the probabilities of nondetection of multiple males are quite low in all populations, and do not affect the results qualitatively. Moreover, the larger broods in lower density populations should have facilitated our ability to detect multiple paternity in those females.

Within populations of H. formosa there was a tendency for larger females to be the ones carrying multiply sired broods. This relationship would be consistent with results of other studies when multiple paternity rates do not approach 100% (see Trexler et al., 1997). This suggests that males copulate preferentially with larger females and that larger females are more likely to be the foci of male–male interactions.

Despite indications that most matings in H. formosa are the result of forced male copulations rather than female choice (Bisazza & Pilastro, 1997), females probably do exhibit some preference for males. This is indicated by the relatively low level of multiple paternity within females and even within broods when compared with other poeciliid species. There also appeared to be an unequal representation of male genotypes within each female. This could result from sperm precedence (sperm from the last copulation is used primarily for fertilization of ova; see Constantz, 1984), or it may be that females have some internal control over fertilization of embryos by stored sperm.

Although this study provides the first estimates of multiple paternity in H. formosa it is still unclear whether females are able to use sperm differentially from preferred vs. less preferred males. Future research should address how copulation attempts and successes relate to the actual paternity of offspring in H. formosa. This should provide a clear picture of whether females are able to control sperm usage. This study also highlights the utility of microsatellite markers for paternity assessment in poeciliids.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References

This work was supported by the National Science Foundation through award DEB 9903925 to J. Travis. The authors would like to thank Joel Trexler and Kyoko Nakamura for sharing their primer sequences with us. Brian Storz helped with the DNA extractions, and Margaret Gunzburger, Matt Aresco, and Brian Storz helped to collect the fish. We are indebted to Dr Laura Keller for allowing us to use her laboratory space and equipment for the DNA analysis, and to Dr Don Levitan for use of the analysis software. Thanks to Charlie Baer for guidance and advice with fragment analysis.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Samples
  6. DNA extraction and microsatellite amplification
  7. Population density
  8. Data analysis
  9. Results
  10. Discussion
  11. Acknowledgments
  12. References
  • Arnqvist, G. & Rowe, L. 1995. Sexual conflict arms races between the sexes – a morphological adaptation for control of mating in a female insect. Proc. R. Soc. Lond. B Biol. Sci. 261: 123127.
  • Aspbury, A.S. & Basolo, A.L. 2002. Repeatable female preferences, mating order and mating success in the poeciliid fish, Heterandria formosa. Behav. Ecol. Sociobiol. 51: 238244.
  • Baer, C.F. 1998. Population structure in a south-eastern US freshwater fish, Heterandria formosa. II. Gene flow and biogeography within the St. Johns River drainage. Heredity 81: 404411.
  • Ball, M.A. & Parker, G.A. 1996. Sperm competition games: external fertilization and ‘adaptive’ infertility. J. Theor. Biol. 180: 141150.
  • Berteaux, D., Bety, J., Rengifo, E. & Bergeron, J.M. 1999. Multiple paternity in meadow voles (Microtus pennsylvanicus): investigating the role of the female. Behav. Ecol. Sociobiol. 45: 283291.
  • Birkhead, T. 2000. Promiscuity: an Evolutionary History of Sperm Competition. Harvard University Press, Cambridge, MA.
  • Bisazza, A. & Pilastro, A. 1997. Small male mating advantage and reversed size dimorphism in poeciliid fishes. J. Fish Biol. 50: 397406.
  • Bisazza, A., Vaccari, G. & Pilastro, A. 2001. Female mate choice in a mating system dominated by male sexual coercion. Behav. Ecol. 12: 5964.
  • Borowsky, R. & Kallman, K.D. 1976. Patterns of mating in natural populations of Xiphophorus from Belize and Mexico. Evolution 30: 693706.
  • Borowsky, R. & Khouri, J. 1976. Patterns of mating in natural populations of Xiphophorus II. X. variatus from Tamaulipas, Mexico. Copeia 1976: 727734.
  • Brockmann, H.J., Colson, T. & Potts, W. 1994. Sperm competition in horseshoe crabs (Limulus polyphemus). Behav. Ecol. Sociobiol. 35: 153160.
  • Chesser, R.K., Smith, M.W. & Smith, M.H. 1984. Biochemical genetics of mosquitofish. III. Incidence and significance of multiple insemination. Genetica 64: 7781.
  • Chick, J.H., Jordan, F., Smith, J.P. & McIvor, C.C. 1992. A comparison of four enclosure traps and methods used to sample fishes in aquatic macrophytes. J. Freshw. Ecol. 7: 353361.
  • Conner, J. 1989. Density dependent sexual selection in the fungus beetle Bolitotherus cornutus. Evolution 43: 13781386.
  • Constantz, G.D. 1984. Sperm competition in poeciliid fishes. In: Sperm Competition and the Evolution of Animal Mating Systems (R. L.Smith, ed.), pp. 465485. Academic Press, New York.
  • Constantz, G.D. 1989. Reproductive biology of poeciliid fishes. In: Ecology and Evolution of Livebearing Fishes (Poeciliidae) (G. K.Meffe & F. F.SnelsonJr, eds), pp. 3350. Prentice-Hall, Englewood Cliffs, NJ.
  • Crim, J.L., Spotila, L.D., Spotila, J.R., O'Connor, M., Reina, R., Williams, C.J. & Paladino, F.V. 2002. The leatherback turtle, Dermochelys coriacea, exhibits both polyandry and polygyny. Mol. Ecol. 11: 20972106.
  • Daley, M. 1978. The cost of mating. Am. Nat. 112: 171174.
  • Evans, J.P. & Magurran, A.E. 2000. Multiple benefits of multiple mating in guppies. Proc. Natl. Acad. Sci. USA 97: 1007410076.
  • Evans, J.P. & Magurran, A.E. 2001. Patterns of sperm precedence and predictors of paternity in the Trinidadian guppy. Proc. R. Soc. Lond. B Biol. Sci. 268: 719724.
  • Farr, J.A. 1989. Sexual selection and secondary sexual differentiation in poeciliids: determinants of male mating success and the evolution of female choice. In: Ecology and Evolution of Livebearing Fishes (Poeciliidae) (G. K.Meffe & F. F.Snelson Jr, eds), pp. 91123. Prentice-Hall, Englewood Cliffs, NJ.
  • Garner, T.W.J., Gregory, P.T., McCracken, G.F., Burghardt, G.M., Koop, B.F., McLain, S.E. & Nelson, R.J. 2002. Geographic variation of multiply paternity in the common garter snake (Thamnophis sirtalis). Copeia 2002: 1523.
  • Greene, J.M. & Brown, K.L. 1991. Demographic and genetic characteristics of multiply inseminated female mosquitofish (Gambusia affinis). Copeia 1991: 434444.
  • Hanotte, O.T., Burke, T., Armour, J.A.L. & Jeffreys, A.J. 1991. Hypervariable minisatellite DNA sequences in the Indian pea fowl Pavo cristatus. Genomics 9: 587597.
  • Hendrie, R., Abercrombie-Higgins, J., Chang, T., Maddocks, R., Truong, B.T., Morgan, E. & Wetherall, J. 1998. Genepop on the Web. http://www.wbiomed.curtin.edu.au/genepop/
  • Hicks, C.R. 1973. Fundamental Concepts in the Design of Experiments. Holt, Rinehart, and Winston, New York.
  • Holland, R. & Rice, W.R. 1998. Chase-away sexual selection: antagonistic seduction versus resistance. Evolution 52: 17.
  • Horth, L.A. 2003. Melanic body colour and aggressive mating behaviour are correlated traits in male mosquitofish (Gambusia holbrooki). Proc. R. Soc. Lond. B Biol. Sci. 270: 10331040.
  • Houde, A.E. & Endler, J.A. 1990. Correlated evolution of female mating preferences and male color patterns in the guppy Poecilia reticulata. Science 248: 14051408.
  • Hutchings, J.A. & Myers, R.A. 1988. Mating success of alternative maturation phenotypes in male Atlantic salmon, Salmo salar. Oecologia 75: 169174.
  • Jaeger, R.G., Gillette, J.R. & Cooper, R.C. 2002. Sexual coercion in a territorial salamander: males punish socially polyandrous female partners. Anim. Behav. 63: 871877.
  • Jennions M.D. & Petrie, M. 2000. Why do females mate multiply? A review of the genetic benefits. Biol. Rev. 75: 2164.
  • Keller L. & Reeve, H.K. 1994. Partitioning of reproduction in animal societies. Trends Ecol. Evol. 9: 98102.
  • Kelly, C.D., Godin, J.-G.J. & Wright, J.M. 1999. Geographical variation in multiple paternity within natural populations of the guppy (Poecilia reticulata). Proc. R. Soc. Lond. B Biol. Sci. 266: 24032408.
  • Laumann, E., Gagnon, J. Michael, R. & Michael, S. 1994. The Social Organization of Sexuality. The University of Chicago Press, Chicago, IL.
  • Leips, J. & Travis, J. 1999. The comparative expression of life-history traits and its relationship to the numerical dynamics of four populations of the least killifish. J. Anim. Ecol. 68: 595616.
  • Leips, J., Travis, J. & Rodd, F.H. 2000. Genetic influences on experimental population dynamics of the least killifish. Ecol. Monogr. 70: 289309.
  • Leslie, J.F. & Vrijenhoek, R.C. 1977. Genetic analysis of natural populations of Poeciliopsis monacha. J. Hered. 68: 301306.
  • Levitan, D.R. 1998. Sperm limitation, sperm competition and sexual selection in external fertilizers. In: Sperm Competition and Sexual Selection (T.Birkhead & A.Moller, eds), pp. 173215. Academic Press, San Diego.
  • McCarthy, M.A. & Lindenmayer, D.B. 1998. Population density and movement data for predicting mating systems of arboreal marsupials. Ecol. Modeling 109: 193202.
  • Marshall, T., Slate, J., Kruuk, L. & Pemberton, J.M. 1998. Statistical confidence for likelihood-based paternity inference in natural populations. Mol. Ecol. 7: 639655.
  • Martin, R.G. 1975. Sexual and aggressive behaviour, density and social structure in a natural population of mosquitofish, Gambusia affinis holbrooki. Copeia 1975: 445454.
  • Martin, F.D. 1980. Heterandria formosa Agassiz, least killifish. In: Atlas of North American Fishes (D. S.Lee, C. R.Gilbert, C. H.Hocutt, R. E.Jenkins, D. E.McAllister & J. R.StaufferJr, eds), pp. 547. North Carolina State Museum of Natural History, Raleigh, NC.
  • Moller, A.P. & Birkhead, T.R. 1994. The evolution of plumage brightness in birds is related to extrapair paternity. Evolution 48: 10891100.
  • Murie, J.O. 1995. Mating behavior of Columbian ground squirrels. 1. Multiple mating by females and multiple paternity. Can. J. Zool. 73: 18191826.
  • Nakamura, K. 2001. Isolation and Characterization of Nuclear Microsatellite Markers in the Sailfin Molly, Poecilia latipinna. Department of Biology, Florida International University, Miami, FL. 58 p.
  • Parker, G.A. 1979. Sexual selection and sexual conflict. In: Sexual Selection and Reproductive Competition in Insects (M. S.Blum & N. B.Blum, eds), pp. 123166. Academic Press, New York.
  • Ptacek, M.B. & Travis, J. 1996. Interpopulation variation in male mating behaviours in the sailfin molly, Poecilia latipinna. Anim. Behav. 52: 5971.
  • Ptacek, M.B. & Travis, J. 1998. Hierarchical patterns of covariance between morphological and behavioural traits. Anim. Behav. 56: 10441048.
  • Raymond, M. & Rousset, F. 1995. GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. J. Hered. 86: 248249.
  • Reed, D.H. & Frankham, R. 2003. Correlation between fitness and genetic diversity. Conserv. Biol. 17: 230237.
  • Robbins, L.W., Hartman, G.D. & Smith, M.H. 1987. Dispersal, reproductive strategies, and the maintenace of genetic variability in mosquitofish (Gambusia affinis). Copeia 1987: 156164.
  • Roff, D. 2002. Life History Evolution. Sinauer Associates, Sunderland, MA.
  • Rosen, D.E. & Bailey, R.M. 1963. The poeciliid fishes (Cyprinodontiformes), their structure, zoogeography, and systematics. Bull. Am. Mus. Nat. Hist. 126: 1176.
  • Rowe, L. 1994. The costs of mating and mate choice in water striders. Anim. Behav. 48: 10491056.
  • Seckinger, J., Brinkmann, H. & Meyer, A. 2002. Microsatellites in the genus Xiphophorus, developed in Xiphophorus montezumae. Mol. Ecol. Notes 2: 46.
  • Sikstrom, B., Hellberg, D., Nilsson, S., Brihmer, C. & Mardh, P.A. 1996. Sexual risk behavior in women with cervical human papillomavirus infection. Arch. Sex. Behav. 25: 361372.
  • Simpson, R., Wilding, C. & Grahame, J. 1999. Simple protocol for extracting nuclear DNA from single embryos of a marine snail. Biotechniques 26: 10501052.
  • Smith, M.H., Garten, C.T. & Ramsey, P.R. 1975. Genic heterozygosity and population dynamics in small mammals. In: Isozymes IV. Genetics and Evolution (C. L.Markert, ed.), pp. 85102. Academic Press, New York.
  • Sokal, R.R. & Rohlf, F.J. 1995. Biometry, 3rd edn. W.H. Freeman and Co., New York.
  • Stouthamer, R., Breeuwer, J.A.J. & Hurst, G.D.D. 1999. Wolbachia pipientis: microbial manipulator of arthropod reproduction. Ann. Rev. Microbiol. 53: 71102.
  • Travis, J. 1990. The interplay of population dynamics and the evolutionary process. Philos. Trans. R. Soc. Lond. B Biol. Sci. 330: 253259.
  • Travis, J., Farr, J.A., Henrich, S. & Cheong, R.T. 1987. Testing theories of clutch overlap with the reproductive ecology of Heterandria formosa. Ecology 68: 611623.
  • Travis, J., Trexler, J.C. & Mulvey, M. 1990. Multiple paternity and its correlates in female Poecilia latipinna (Poeciliidae). Copeia 1990: 722729.
  • Trexler, J.C., Travis, J. & Dinep, A. 1997. Variation among populations of the sailfin molly in the rate of concurrent multiple paternity and its implications for mating-system evolution. Behav. Ecol. Sociobiol. 40: 297305.
  • Trexler, J.C., McElroy, T.C., Loftus, W.F., Jordan, F., Chick, J.H., Kandl, K.L. & Bass, O.L. Jr 2002. Ecological scale and its implications for freshwater fish in the Everglades. In: The Everglades, Florida Bay, and Coral Reefs of the Florida Keys: an Ecosystem Sourcebook (J. W.Porter & K. G.Porter, eds), pp. 153181. CRC Press, Boca Raton, FL.
  • Wallander, J., Blomqvist, D. & Lifjeld, J.T. 2001. Genetic and social monogamy: does it occur without mate guarding in the ringed plover? Ethology 107: 561572.
  • Westneat, D.F., Fredrick, P.C. & Haven, W.R. 1987. The use of genetic markers to estimate the frequency of successful alternative reproductive tactics. Behav. Ecol. Sociobiol. 21: 3545.
  • Young, A., Boyle, T. & Brown, T. 1996. The population genetic consequences of habitat fragmentation for plants. Trends Ecol. Evol. 11: 413418.
  • Zane, L., Nelson, W.S., Jones, A.G. & Avise, J.C. 1999. Microsatellite assessment of multiple paternity in natural populations of a live bearing fish, Gambusia holbrooki. J. Evol. Biol. 12: 6169.