Life history of Gambusia vittata (Pisces: Poeciliidae).
Viviparous fishes, such as those from the family Poeciliidae, are excellent models for understanding the evolution of life cycles. Nearly 220 species of poeciliid fish are known (Pires et al., 2011). This abundance of species makes it easy for researchers to conduct comparative studies and to find links between life-history evolution and causes of natural selection (Johnson and Bagley, 2011). Life-history traits of poeciliid fishes vary not only across species, but also across populations of a single species. The study of life histories of different populations in the same species can help us to recognize the driving forces behind the observed differences among populations in life-history traits (Johnson and Bagley, 2011). This is because within the few species that have been thoroughly studied (e.g., Poecilia reticulata, Gambusia affinis, Brachyrhaphis rhabdophora), the species and populations evolved the same traits independently and as a response to the same environmental factors such as predation, stream velocity, or population density (Reznick and Endler, 1982; Johnson and Belk, 2001; Reznick et al., 2002; ZUniga-Vega et al., 2007).
To gain a complete comprehension of the diversity of life cycles of viviparous fishes and of the evolutionary causes of this diversity, we must expand our knowledge to multiple species within the same family. Researchers have examined only a few species of poeciliid fishes in terms of spatial and temporal variation in their life histories (Johnson and Bagley, 2011). Of those examined, several have life-history descriptions based on a single site or for collections made at a single point in time (e.g., Johnson, 2002; Patimar et al., 2011; Gkenas et al., 2012). However, for a complete understanding of the evolution of life histories within the family Poeciliidae we need thorough descriptions of variation among multiple populations through both space and time, with a particular emphasis on species for which no life-history data are available.
In this paper we describe spatial and temporal variation in life-history traits of the viviparous fish Gambusia vittata. We collected data from three sites over 3 months and tested for differences in life-history traits among sites. Two of our study sites, although separated by approximately 120 km (linear distance), were located within the same river drainage. Therefore, we predicted that the two sites located in the same river drainage would be more similar to each other, in terms of life-history traits, compared to our third site, which is located in a smaller river that has no inland contact with the other. We also tested for temporal variation in the life history of G. vittata. We had data for January, March, and April, which correspond to the dry season. However, we still predicted differences among these months. Finally, we compared the life history of G. vittata to other poeciliid fishes that are thoroughly studied in an attempt to understand the causes that are driving temporal and spatial variation in the life history of this species.
MATERIALS AND METHODS--Study Species--Gambusia vittata is a viviparous fish of the family Poeciliidae. In males the anal fin is modified to form a structure that is used to fertilize females (gonopodium). Gambusia vittata is endemic to a small region in eastern Mexico. It occurs mainly in two river drainages: the Tamesi River and the Panuco River. Gambusia vittata inhabits streams, canals, rivers, and ditches that are normally clear. Occasionally the currents in these water systems can be strong, but typically they are light to moderate (Miller et al., 2005).
Gambusia vittata was originally described by Hubbs (1926), based only on the morphological examination of preserved females. Later, he was able to collect males and, based on characteristics of the male gonopodium, assign this species to a different monotypic genus. Hubbs named the species Flexipenis vittatus (in Rivas, 1963:334). According to this author, the gonopodial structures of F. vittatus differ substantially from those observed in other members of the genus Gambusia and, in contrast, resemble those of males of the genera Belonesox and Heterophallus. Rauchenberger (1989) reinstated this species within Gambusia, because her detailed morphological examination revealed that G. vittata exhibits the synapomorphies that she described for the genus Gambusia. Rauchenberger (1989) interpreted the distinct gonopodium of G. vittata as retention of a primitive character. Phylogenetic analyses based on genetic data confirmed the current taxonomic status of G. vittata (Lydeard et al., 1995; Hrbek et al., 2007).
We know little of the biology of G. vittata. Darnell (1962) examined the stomach contents of several specimens of G. vittata and found that 50% of the material encountered was filamentous algae. Only 20% of the stomach contents consisted of arthropods. The remaining 30% consisted of detritus and undetermined organic matter. These feeding habits are unusual compared to other congeneric species, which feed predominantly on arthropods (Darnell, 1962). Apparently, females of G. vittata undergo a period of fasting when embryos approach full term. After the young are born, females feed heavily on algae, presumably to avoid eating their own offspring (Darnell, 1962). A description of the external female genitalia of G. vittata was provided by Peden (1972). According to this author, females of G. vittata exhibit the most distinct and differentiated genitalia of all Gambusia species (similar to the male gonopodium of G. vittata). Males mate with females by means of gonopodial thrusts, with no apparent courtship behavior (Bisazza, 1993). The karyotype of this species consists of 48 (2n) nonmetacentric chromosomes (Campos and Hubbs, 1971). A description of helminth parasites present in G. vittata has been provided by Salgado-Maldonado et al. (2004). The exotic and highly pathogenic tapeworm Bothriocephalus acheilognathi has been collected from the intestine of G. vittata specimens (Salgado-Maldonado and Pineda-Lopez, 2003).
Study Sites and Field Methods--We collected individuals of G. vittata from three different sites in the eastern region of Mexico. Table 1 shows a detailed description of the three study sites, including a list of other fish species present. We used seines (1.3 m deep x 5 m long, 8-mm mesh) to capture individuals. Following capture, we euthanized each fish in 3amenobenzoic acid ethyl ester (also known as MS-222). Rather than preserving the fish in formalin, we fixed specimens in the field in 95% ethanol and stored them in the laboratory in 70% ethanol. This allowed us to preserve samples so that they could subsequently be used for DNA sequencing.
We collected G. vittata individuals from three sites across 3 months during the dry season. Our first collections were in January 2012 from all three sites. Though we collected females in January from site 3, only two females contained embryos. Therefore, we removed these two females from all analyses. We collected again from all three sites in April 2012. In March 2013 we were able to collect again from site 2 only. The chronological order in which we collected specimens is January 2012, April 2012, and March 2013. Site 2 is the only site with data collected for the month of March (refer to Table 1 for detailed descriptions).
Quantifying Life-History Traits--We quantified five life-history traits for each population each month: 1) number of embryos per reproductive female (brood size), 2) individual embryo mass (dry mass), 3) size at maturity for both males and females, 4) reproductive allotment (RA), and 5) the amount of maternal transfer of nutrients to developing embryos (i.e., degree of matrotrophy). We collected all data by dissecting preserved females, with one exception: we used preserved males to estimate male size at maturity. We determined brood size by counting the number of embryos each female contained. We identified the stage of development following Haynes (1995). We estimated individual embryo mass as the average dry weight of embryos from each female. To estimate female size at maturity, we classified all our females into 2-mm size classes, based on their standard length (SL). The minimum size at maturity corresponded to the size class in which at least 50% of the females contained embryos or vitellogenic follicles (Reznick and Endler, 1982; Johnson and Belk, 2001). Previous studies have estimated male size at maturity as the average SL of adult males, because apparently males cease growing upon maturity (Johnson and Belk, 2001; Jennions et al., 2006; Scott and Johnson, 2010). However, in G. vittata we have observed wide variation in the sizes of mature males (15-26 mm SL). Therefore, we estimated the minimum size at maturity for males following a similar procedure as females. We classified all males into 2-mm size classes. To estimate male size at maturity, we examined the development of the gonopodium. The minimum size at maturity corresponded to the size class in which at least 50% of the males exhibited completely developed gonopodia. We calculated the RA of each female as the total dry weight of the brood. We also measured the dry mass of reproductive females. We measured dry masses for all females and embryos after desiccation in an oven for 24 h at 55[degrees]C.
We estimated the amount of nutrient transfer to developing embryos using the matrotrophy index (MI; Reznick et al., 2002; Thompson et al., 2002). We calculated an MI for each population in each month. The MI was calculated by dividing the estimated dry mass of the embryos near time of birth (stage 11 according to Haynes, 1995) by the estimated dry mass of the embryos at the time of fertilization (stage 4). We estimated these dry masses using a regression between stage of development as the independent variable and embryo mass (log-transformed) as the dependent variable for each site and for each month.
Researchers have characterized maternal provisioning of embryos using three categories: matrotrophic, lecithotrophic, and incipient matrotrophic (Blackburn, 1992; Reznick et al., 2002; Thompson et al., 2002). Matrotrophic organisms provide considerable amounts of nutrients to their developing embryos. Lecithotrophic organisms provide all nutrients necessary for development before fertilization in the form of yolk. Incipient matrotrophic organisms transfer small quantities of nutrients to developing embryos to offset some of the metabolic costs. To be considered matrotrophic, the MI for a population must range between 1.0, meaning that a small amount of nutrients was transferred to the embryos after fertilization, and >100, which implies that a large amount of nutrients was given to the embryos postfertilization. Lecithotrophic populations will have an MI of around 0.6-0.7, values consistent with the mass lost by the embryos during development due to metabolic expenses and lack of female nutrient transfer. Finally, if a population has an MI between 0.71.0 then they are considered to be incipient matrotrophic. This means that they will lose less mass during development compared to lecithotrophic embryos because they have a small amount of nutrient transfer from their mothers during development (Reznick et al., 2002; Thompson et al., 2002).
Statistical Analyses--To compare brood size, individual embryo mass, and RA among populations and between months, we used general linear models. We conducted one linear model for each of these three traits. In all three analyses, we used female dry mass as a covariate, and both source population and sampling month (January 2012, April 2012, or March 2013) were used as categorical factors. The factor ''sampling month'' was nested in the factor "population." We used stage of development as an additional covariate when analyzing individual embryo mass. We log-transformed individual embryo mass, RA, and female mass to meet the assumptions of normality and variance homogeneity. We square-root transformed brood size. To facilitate interpretation and visualization of the general linear model results, we present figures and least-square means obtained from similar analyses conducted on untransformed data.
To compare the degree of maternal transfer of nutrients to developing embryos, we constructed 95% confidence intervals for our MI values by means of a resampling procedure (as per Zuniga-Vega et al., 2011). First, for each population and sampling month we bootstrapped our individuals to obtain a new sample. Second, based on this new sample, we fitted a regression between stage of development and embryo mass. Third, from this regression, we estimated embryo mass at fertilization (stage 4) and embryo mass at birth (stage 11) and, based on these estimates, we calculated a new MI. We repeated this procedure 1,000 times, generating 1,000 MI values for each month and population. To construct 95% confidence intervals, we used the 25th and 975th sorted values of the resulting distribution of MI values as the lower and upper limits, respectively.
RESULTS--Average size of mature females across all sites and months was 22.2 mm SL whereas the average size of mature males across all sites and months was 20.8 mm SL. The average dry mass of mature females (excluding digestive tract and embryos) across all sites and months was 0.049 g. Considering only females of adult size (i.e., larger than the minimum size at maturity), the proportion of reproductive females varied from around 0.77-1.0 across sites and months (Fig. 1). The largest difference among months occurred at site 1. In January 2012, 77.8% of adult-sized females had developing embryos. In contrast, in April 2012 a larger percentage of adult-sized females (98.3%) contained developing embryos (Fig. 1a). At site 2, the proportions of reproducing females were similar in January 2012 (0.83) and March 2013 (0.79), but at this site again in April 2012, we found a higher proportion of adult-sized females that were reproductive (0.96; Fig. 1b). At site 3 in April 2012, we found that all adult-sized females contained developing embryos (Fig. 1c). Although not included in our analyses, we noted that the two adult females collected from site 3 in January 2012 contained developing embryos (Fig. 1c).
We found significant differences among months for brood size ([F.sub.3,172] = 3.42, P = 0.019; Table 2). Within sites, brood size increased in March and April compared to January (Fig. 2a). Significant variation in brood size was present across sites as well ([F.sub.2,172] = 4.06, P = 0.019; Table 2). We observed the largest mean brood size at site 1 (adjusted least-square mean = 5.62 young), whereas we observed the smallest brood size at site 2 (adjusted least-square mean = 4.88 young). Average brood size for site 3 was 5.19. We also observed a significant effect of female mass on brood size ([F.sub.1,172] = 325.93, P < 0.0001; Table 2). Heavier (larger) females produced larger broods (Fig. 3a). The interaction between site and female mass significantly affected brood size ([F.sub.2,172] = 4.03, P = 0.02). In spite of this significant slope heterogeneity, the least-square means derived from this model were comparable among sites because the slopes crossed outside the biologically meaningful values of female mass.
Significant differences in embryo size occurred among months ([F.sub.3,169] = 3.67, P = 0.014; Table 2). For instance, embryo mass at site 2 decreased in April 2012 compared to January 2012 and March 2013 (Fig. 2b). In contrast, there were no significant differences in embryo mass across sites ([F.sub.2,169] = 0.48, P = 0.62; Table 2). Mean embryo mass (adjusted least-square means) varied across sites (0.0013, 0.0017, and 0.0018 for sites 1, 2, and 3, respectively). However, this intersite variation was not statistically significant. Female mass had a significant effect on individual embryo mass ([F.sub.1,169] = 24.02, P < 0.0001; Table 2). Larger females produced larger embryos (Fig. 3b). However, the effect was weaker on this trait than that observed on brood size and RA (Figs. 3a and 3c). Stage did not affect embryo size ([F.sub.1,169] = 0.61, P = 0.43; Table 2). This indicates that the embryos stay at a constant mass during development.
We found significant differences in RA among months within sites ([F.sub.3,172] = 6.17, P = 0.0005; Table 2). At site 1, RA during April 2012 was significantly higher than RA during January 2012. At site 2, RA increased in March 2013 compared to January 2012 and April 2012 (Fig. 2c). In contrast, we found no significant differences in RA across sites ([F.sub.2,172] = 0.099, P = 0.91). Mean size-adjusted RA was 0.0070 g in site 1, 0.0081 g in site 2, and 0.0099 g in site 3. Female mass had a significant effect on RA as well ([F.sub.1,172] = 234.96, P < 0.0001; Table 2). Larger females exhibited greater RA (Fig. 3c).
Female size at maturity varied across sites. At sites 1 and 3 females matured on average at a larger size (21 mm SL) compared to site 2 (17 mm SL). Our monthly variation data within sites was limited. Females collected in April 2012 from site 1 did not result in an estimate of female size at maturity because we did not find small-sized females. Similarly, females collected in January 2012 from site 3 only included two reproducing females (14 females were collected; only 2 of which were reproductive). Thus, no estimate of minimum size at maturity could be obtained for January at site 3. At site 2 our monthly estimates of female size at maturity were equal among months (17 mm SL in all 3 months; Fig. 2d).
Male size at maturity also varied across sites. Similar to our results for female size at maturity, at sites 1 and 3 males matured on average at larger sizes (20 and 21 mm SL, respectively) compared to site 2 (15.7 mm SL). We found slight variation among months within sites. At site 1 minimum size of mature males was 19 mm SL for January 2012 and 21 mm SL for April 2012 (Fig. 2e). At site 2, the smallest sizes at maturity were observed in both January 2012 and March 2013 (15 mm SL), whereas the largest for this site was observed in April 2012 (17 mm SL; Fig. 2e). Males collected in January 2012 from site 3 did not result in an estimate of male size at maturity because we only collected three males, and all were large-sized (>24 mm SL) and mature. Therefore, for site 3 we estimated male size at maturity only for April 2012 (21 mm SL).
We found minimal variation in MI values across sites: 1.06 for site 3, 1.17 for site 2, and 1.63 for site 1, with confidence intervals of 1.05-2.47 for site 1, 0.80-1.61 for site 2, and 0.77-1.18 for site 3. These MI values per site were not significantly different as indicated by their overlapping confidence intervals. This result was consistent with the nonsignificant interaction between site and stage of development affecting embryo mass ([F.sub.2,169] = 1.77, P = 0.17; Table 2), which indicated that the relationship between embryo mass and developmental stage (i.e., the rate of mass change of embryos during development) was similar among sites. Similarly, among months within sites, all confidence intervals for the MI values overlapped, indicating the lack of significant differences (Fig. 2f). These monthly MI values ranged between 0.91 in site 2 during January 2012 and 3.96 in site 1 during January 2012.
DISCUSSION--In this study we provide the first description of the life history traits of Gambusia vittata, a poeciliid fish endemic to a small region in eastern Mexico. We document variation among sites and among months within sites for reproductive features. Notice that our sampling scheme included monthly variation as well as yearly variation because our samples from January and April corresponded to 2012, whereas our samples from March corresponded to 2013. For this study we assumed that variation among months was greater than variation between years and, therefore, interpret our monthly differences as variation as the dry season progresses regardless of year. However, we recognize that our results for March 2013 might include a year effect and hence suggest a cautious interpretation.
The proportion of reproductive females was above 0.50 for all months in all sites, which indicates that G. vittata are reproductive at least from January to April. April 2012 had the largest proportion of reproductive females across all sites and months, which suggests an increase in the number of reproductive females from January to April. In most of the Mexican territory, April corresponds with the late phases of the dry season (Garcia, 2004). Therefore, our data suggest that before the beginning of the rainy season more G. vittata females produce offspring. studies have found that during the wet season there is a higher abundance of resources for the young and higher water volume, causing predator and prey distributions to be more regular (Winemiller, 1993; Machado et al., 2002). similar studies have concluded that females may possess ecological and physiological sensors allowing them to sense changes in the environment (such as the beginning of the rainy season). In response to these senses, females reproduce at higher rates to give their offspring a higher chance of survival (Kusano, 1982; Winemiller, 1993). Therefore, the increase in proportion of reproductive females may be due to the beginning of the rainy season and associated physiological drivers.
Brood size differed among sites and among months within sites. At sites 1 and 2, brood size increased from January to April. This increase may also contribute to higher rates of reproduction just before the rainy season. The mean brood size for G. vittata ranged from 4.88 to 5.62 across sites. Compared to other Gambusia species of similar sizes, the brood size of G. vittata is small. Most other Gambusia species have an average brood size of 10 or more offspring, which is twice as large as G. vittata brood sizes (e.g., Gambusia puncticulata puncticulata [range of brood size across sites] = 1-70, Abney and Rakocinski, 2004; G. affinis [average number of embryos per brood] = 34.48, Swenton and Kodric-Brown, 2012; Gambusia holbrooki, = 50.8, Gkenas et al., 2012; Gambusia sexradia = 17.34, Riesch et al., 2010). An exception is Gambusia hubbsi, a species that exhibits a similar brood size compared to G. vittata (variation across low and high predation environments = 3.43-7.26, Riesch et al., 2013). However, when compared to other poeciliid species of similar sizes, G. vittata has a relatively average brood size. Several other poeciliid fishes have brood sizes ranging from around two to nine offspring (Brachyrhaphis episcopi = 2.9-8.74, Jennions et al., 2006; Poeciliopsis prolifica = 4.2, Pires et al., 2007; Poeciliopsis baenschi = 2.7, Scott and Johnson, 2010; Heterophallus milleri = 8.49, Riesch et al., 2011). Statistically significant differences in brood size were found across sites. Females at sites 1 and 3 have larger broods (5.62 and 5.19 newborns), on average, than their counterparts at site 2 (4.88 newborns). This result is contrary to what was expected. Although sites 1 and 2 are in the same river drainage system, site 3 is more similar in brood size to site 1 than is site 2. This lower brood size at site 2 might represent an adaptation or a plastic response to local conditions. For instance, low resource (food) availability at site 2 might constrain the number of newborns that females can produce.
We emphasize that all the females that we dissected contained a single brood of developing embryos, indicating the lack of superfetation in this species. Superfetation is the ability of a female to simultaneously carry several broods at different developmental stages (Turner, 1937), and is present in several members of the family Poeciliidae (Reznick and Miles, 1989; Pires et al., 2011). The lack of superfetation in G. vittata is consistent with other Gambusia species (Reznick and Miles, 1989; Pires et al., 2011).
Mean individual embryo mass varied from 0.0013-0.0018 g across sites. However, these differences in individual embryo mass across sites were not statistically significant. This lack of variation in embryo mass across sites is consistent with other poeciliid species (e.g., Poeciliopsis baenschi, Scott and Johnson, 2010; Poecilia butleri, Zuniga-Vega et al., 2011). Compared to other Gambusia species (G. p. puncticulata = 0.00129-0.005 g, Abney and Rakocinski, 2004; G. sexradiata [mean embryo mass] = 0.0014 g, Riesch et al., 2010; G. hubbsi [across low- and high-predation environments] = 0.0013-0.0046 g, Riesch et al., 2013), the individual embryo mass for G. vittata is relatively similar. However, when compared to the individual embryo mass of a population of Gambusia eurystoma living in a sulfidic habitat (0.0064 g, Riesch et al., 2010) the embryo mass of G. vittata is significantly smaller. When compared to other poeciliid species of about the same size, G. vittata has a similar individual embryo mass (Heterophallus milleri [average embryo mass] = 0.0012 g, Riesch et al., 2011; Poecilia butleri [range of variation across populations] = 0.0012-0.0021, Zuniga-Vega et al., 2011).
Stage of development did not affect embryo mass. The mass of individual embryos remains constant throughout development. This indicates that G. vittata is slightly matrotrophic. Therefore, females transfer a small amount of nutrients to the developing embryos postfertilization to offset the loss of mass experienced during development due to metabolic costs (Reznick et al., 2002; Thompson et al., 2002). The MI values that we calculated also indicate moderate matrotrophy in G. vittata because all of these values were equal to or slightly larger than unity. This indicates a relatively similar embryo weight between fertilization and birth. The mean MI values varied little across sites (1.06-1.63) and only the value for site 1 (1.63) was significantly higher than unity, evidencing a slight increase in mass from fertilization to birth. With respect to monthly variation, MI values ranged between 0.91 and 3.96. However, all of these values were not statistically different from each other, and most were not significantly different than unity, indicating relatively constant embryo mass throughout development in all studied populations and across all studied months. The nonsignificant interaction between site and stage of development affecting individual embryo mass also indicates that the mass of developing embryos remains relatively constant throughout development in all three sites.
Gambusia taxa were not expected to exhibit matrotrophy (Reznick and Miles, 1989). However, more recent studies using tritiated leucine (Marsh-Matthews et al., 2005) found that members of the Gambusia clade do exhibit some amount of postfertilization nutrient transfer from mothers to embryos. Some of these species include G. affinis (DeMarais and Oldis, 2005; Marsh-Matthews et al., 2005), G. holbrooki (Edwards et al., 2006; Marsh-Matthews et al., 2010), Gambusia clarkhubbsi, Gambusia gaigei, Gambusia geiseri, Gambusia nobilis (Marsh-Matthews et al., 2010), and G. p. puncticulata (Abney and Rakocinski, 2004). We conclude that G. vittata is a matrotrophic species as well. We calculated two variables that support this conclusion. First, all calculated MI values for site and month were statistically equal to or greater than 1. Second, we found no significant effect of stage of development on embryo mass. Several other poeciliid species also have MI values that indicate small amounts of maternal transfer of nutrient to embryos just as we observed in G. vittata (Poeciliopsis latidens = 0.86, Poeciliopsis lucida = 1.34, Poeciliopsis occidentalis = 1.12, Reznick et al., 2002; Poecilia butleri = 1.07-5.84, Zuniga-Vega et al., 2011). Also, the lack of significant variation in MI values across sites that we observed in G. vittata is congruent with studies of other poeciliid species (e.g., Poecilia butleri, Zuuiga-Vega et al., 2011). This lack of variation suggests that in some poeciliid species the degree of matrotrophy is not a plastic trait and that within these species there is no genetic variation in the degree of maternal transfer of nutrients to embryos (but see Pires et al., 2007).
RA did not vary significantly across sites. However, it did vary significantly across months (Table 2). In both sites 1 and 2, there was an increase in RA from January to April. This trend supports our hypothesis for an increase in reproductive effort right before the rainy season, presumably for increased survival of offspring during the months with higher food availability and less density of predators and competitors (Winemiller, 1993; Machado et al., 2002). RA also increased with female size. This trend is common in other poeciliid fishes as well (Jennions et al., 2006; Johnson and Belk, 2001; Zuniga-Vega et al., 2011). The mean size-adjusted values per site for RA of G. vittata varied from 0.0070-0.0099 g. These values represent between 12.5% and 16.8% of the total dry mass of females. Other Gambusia species exhibit similar RA (G. eurystoma = 15.63%, Riesch et al., 2010; G. hubbsi [range of variation across populations] = 10.1915.16%, Riesch et al., 2013). However, G. sexradiata exhibits more extreme RA values depending on the particular environment it inhabits. In a relatively toxic environment, G. sexradiata exhibits low RA (10.46%), whereas in a nontoxic environment its RA value is higher compared to G. vittata and other Gambusia species (23.03%, Riesch et al., 2010). Other poeciliid species of similar sizes show wide variation in RA. Compared to G. vittata, some species appear to invest less effort into reproduction (Poeciliopsis baenschi [range of variation across populations] = 0.0027-0.0044 g; Scott and Johnson, 2010), whereas females from other species make larger investments in reproduction (Brachyrhaphis parismina = 0.015-0.018 g, Belk et al., 2011; Heterophallus milleri = 18.87% of the total female dry mass, Riesch et al., 2011).
Females matured between 17 and 21 mm SL across all sites. Males matured between 15.7 and 21 mm SL across all sites. Females of G. vittata matured at similar sizes compared to other Gambusia species (G. affinis [range of variation across sites] = 16-26 mm SL, Stockwell and Vinyard, 2000; G. holbrooki = 20.35 mm SL, Gkenas et al., 2012; G. sexradiata = 24.19 mm SL, Riesch et al., 2010). Males of G. vittata also matured at similar sizes to those of other Gambusia species (e.g., G. holbrooki = 16.44 mm SL, Gkenas et al., 2012). At site 2 both males and females matured at smaller sizes, and presumably earlier, than their counterparts at sites 1 and 3. This difference is worth noting because sites 1 and 2 are in the same river drainage system. However, sites 1 and 3 are more similar to each other than are sites 1 and 2 for size at maturity. This may represent an adaptation or plastic response to local environmental conditions. Earlier studies found that predation had a large effect on life-history traits. In environments with high predation and high adult mortality rates, decreased size at maturity is favored because rapidly maturing individuals will be more likely to reproduce prior to death (Reznick and Endler, 1982; Reznick et al., 1990; Roff, 1992; Johnson and Belk, 2001). Thus, a likely explanation for the differences that we observed in size at maturity between populations that inhabit the same river drainage (sites 1 and 2) is the occurrence of high predator-driven adult mortality in site 2. Future demographic and genetic studies on different populations of G. vittata would improve our understanding of the intraspecific patterns of life-history variation that we documented here for this species.
We thank the following people for field assistance: P. Frias-Alvarez, A. Hernandez-Rosas, A. Molina-Moctezuma, L. Vazquez-Vega, C. Olivera-Tlahuel, K. Villa-Meza, I. ZapataMoran, and P. Garcia-Aviles. Laboratory assistance was provided by B. Zuniga-Ruiz, M. Hernandez-Apolinar, and P. Mendoza-Hernandez. Fieldwork was conducted under permit DGO-PA.07010.210612.1749 issued by Comision Nacional de Acuacultura y Pesca, Secretaria de Agricultura, Ganaderia, Desarrollo Rural, Pesca y Alimentacion-Mexico. Funding for this study came from Consejo Nacional de Ciencia y Tecnologia and Secretaria de Educacion Publica-Mexico through project 129675 (SEP-CONACyT Ciencia Basica 2009). MLW thanks Brigham Young University, specifically the David M. Kennedy Center for International Studies, for funding her trip and stay in Mexico City and for taking care of the logistics that enabled her to conduct research on the Universidad Nacional Autonoma de Mexico campus.
ABNEY, M. A., AND C. RAKOCINSKI. 2004. Life history variation in Caribbean gambusia, Gambusia puncticulata puncticulata (Poeciliidae) from the Cayman Islands, British West Indies. Environmental Biology of Fishes 70:67-79.
BELK, M. C., E. C. NANCE, AND J. B. JOHNSON. 2011. Life history of Brachyrhaphis parismina: variation within and among populations. Copeia 3:372-378.
BISAZZA, A. 1993. Male competition, female mate choice and sexual size dimorphism in poeciliid fishes. Pages 257-286 in The behavioural ecology of fishes (F. Huntingford and P. Torricelli, editors). Harwood Academic Press, Chur, Switzerland.
BLACKBURN, D. G. 1992. Convergent evolution of viviparity, matrotrophy, and specializations for fetal nutrition in reptiles and other vertebrates. American Zoologist 32:313-321.
CAMPOS, H. H., AND C. HUBBS. 1971. Cytomorphology of six species of gambusiine fishes. Copeia 1971:566-569.
DARNELL, R. M. 1962. Fishes of the Rio Tamesi and related coastal lagoons in east-central Mexico. Publications of the Institute of Marine Science 8:299-365.
DEMARAIS, A., AND D. OLDIS. 2005. Matrotrophic transfer of fluorescent microspheres in poeciliid fishes. Copeia 3:632636.
EDWARDS, T. M., H. D. MILLER, AND L. J. GUILLETTE. 2006. Water quality influences reproduction in female mosquitofish (Gambusia holbrooki) from eight Florida springs. Environmental Health Perspectives 114:69-75.
GARCIA, E. 2004. Modificaciones al sistema de clasificacion climatica de Koppen. Fifth edition. Instituto de Geografia Universidad Nacional Autonoma de Mexico, Mexico City, Mexico.
GKENAS, C., A. OIKONOMOU, A. ECONOMOU, F. KIOSSE, AND I. LEONARDOS. 2012. Life history pattern and feeding habits of the invasive mosquitofish, Gambusia holbrooki, in Lake Pamvotis (NW Greece). Journal of Biological Research-Thessaloniki 17:121-136.
HAYNES, J. L. 1995. Standardized classification of poeciliid development for life-history studies. Copeia 1995:147-154.
HRBEK, T., J. SECKINGER, AND A. MEYER. 2007. A phylogenetic and biogeographic perspective on the evolution of poeciliid fishes. Molecular Phylogenetics and Evolution 43:986-998.
HUBBS, C. L. 1926. Studies of fishes of the order Cyprinodontes. VI. Material for a revision of the American genera and species. Miscellaneous Publications of the Museum of Zoology, University of Michigan 16:1-86.
JENNIONS, M. D., B. B. M. WONG, A. COWLING, AND C. DONNELLY. 2006. Life-history phenotypes in a live-bearing fish Brachyrhaphis episcopi living under different predator regimes: seasonal effects? Environmental Biology of Fishes 76:211-219.
JOHNSON, J. B. 2002. Divergent life histories among populations of the fish Brachyrhaphis rhabdophora: detecting putative agents of selection by candidate model analysis. oikos 96:82-91.
JOHNSON, J. B., andJ. C. Bagley. 2011. Ecological drivers of lifehistory divergence. Pages 38-49 in Ecology and evolution of poeciliid fishes (J. P. Evans, A. Pilastro, and I. Schlupp, editors). University of Chicago Press, Chicago, Illinois.
JOHNSON, J. B., AND M. C. BELK. 2001. Predation environment predicts divergent life-history phenotypes among populations of the livebearing fish Brachyryaphis rhabdophora. oecologia 126:142-149.
KUSANO, T. 1982. Postmetamorphic growth, survival, and age at first reproduction of the salamander, Hynobius nebulosus tokyoensis Tago in relation to a consideration on the optimal timing of first reproduction. Researches on Population Ecology 24:329-344.
LYDEARD, C., M. C. WOOTEN, AND A. MEYER. 1995. Cytochrome b sequence variation and a molecular phylogeny of the live-bearing fish genus Gambusia (Cyprinodontiformes: Poeciliidae). Canadian Journal of Zoology 73:213-227.
MACHADO, G., A. A. GIARETTA, AND K. G. FACURE. 2002. Reproductive cycle of a population of the guaru, Phallocerus caudimaculatus (Poeciliidae), in southeastern Brazil. Studies on Neotropical Fauna and Environment 37:15-18.
MARSH-MATTHEWS, E., M. BROOKS, R. DEATON, AND H. TAN. 2005. Effects of maternal and embryo characteristics on post-fertilization provisioning in fishes of the genus Gambusia. Oecologia 144:12-24.
MARSH-MATTHEWS, E., R. DEATON, AND M. BROOKS. 2010. Survey of matrotrophy in lecithotrophic poeciliids. Pages 255-257 in Viviparous fishes II (M. C. Uribe and H. J. Grier, editors). New Life Publications, Homestead, Florida.
MILLER, R. R., W. L. MINCKLEY, AND S. M. NORRIS. 2005. Freshwater fishes of Mexico. University of Chicago Press, Chicago, Illinois.
PATIMAR, R., M. GHORBANI, A. GOL-MOHAMMADI, AND H. AZIMI-GLUGAHI. 2011. Life history pattern of mosquitofish Gambusia holbrooki (Girard, 1859) in the Tajan River (southern Caspian Sea to Iran). Chinese Journal of Oceanology and Limnology 29:167-173.
PEDEN, A. E. 1972. Differences in the external genitalia of female gambusiin fishes. Southwestern Naturalist 17:265-272.
PIRES, M. N., A. I. BANET, B. J. A. POLLUX, AND D. N. REZNICK. 2011. Variation and evolution of reproductive strategies. Pages 2837 in Ecology and evolution of poeciliid fishes (J. P. Evans, A. Pilastro, and I. Schlupp, editors). University of Chicago Press, Chicago, Illinois.
PIRES, M. N., K. E. MCBRIDE, AND D. N. REZICK. 2007. Interpopulation variation in life-history traits of Poeciliopsis prolifica: implication for the study of placental evolution. Journal of Experimental Zoology 307A:113-125.
RAUCHENBERGER, M. 1989. Systematics and biogeography of the genus Gambusia (Cyprinodontiformes: Poeciliidae). American Museum Novitates 2951:1-74.
REZNICK, D. N., AND J. A. ENDLER. 1982. The impact of predation on life history evolution in Trinidadian guppies (Poecilia reticula). Evolution 36:160-177.
REZNICK, D. N., AND D. B. MILES. 1989. A review of life history patterns in poeciliid fishes. Pages 125-148 in Ecology and evolution of livebearing fishes (Poeciliidae) (G. K. Meffe and F. F. Snelson, Jr., editors). Prentice Hall, Englewood Cliffs, New Jersey.
REZNICK, D. N., H. BRYGA, AND J. A. ENDLER. 1990. Experimentally induced life-history evolution in a natural population. Nature 346:357-359.
REZNICK, D. N., M. MATEOS, AND M. S. Springer. 2002. Independent origins and rapid evolution of the placenta in the fish genus Poeciliopsis. Science 298:1018-1020.
RIESCH, R., T. J. COLSTON, B. L. JOACHIM, AND I. SCHLUPP. 2011. Natural history and life history of the Grijalva gambusia Heterophallus milleri Radda, 1987 (Teleostei: Poeciliidae). Aqua, International Journal of Ichthyology 17:95-102.
RIESCH, R., R. A. MARTIN, AND R. B. LANGERHANS. 2013. Predation's role in life-history evolution of a livebearing fish and a test of the Trexler-DeAngelis model of maternal provisioning. American Naturalist 181:78-93.
RIESCH, R., M. PLATH, F. GARCIA DE LEON, AND I. SCHLUPP. 2010. convergent life-history shifts: toxic environments result in big babies in two clades of poeciliids. Naturwissenschaften 97:133-141.
RIVAS, L. R. 1963. Subgenera and species groups in the poeciliid fish genus Gambusia Poey. Copeia 1963:331-347.
ROFF, D. A. 1992. The evolution of life histories: theory and analysis. Chapman and Hall, New York.
ROFF, D. A. 2002. Life history evolution. Sinauer Associates Inc., Sunderland, Massachusetts.
SALGADO-MALDONADO, G., AND R. F. PINEDA-LOPEZ. 2003. The Asian fish tapeworm Bothriocephalus acheilognathi: a potential threat to native freshwater fish species in Mexico. Biological Invasions 5:261-268.
SALGADO-MALDONADO, G., G. CABANAS-CARRANZA, E. SOTO-GALERA, R. F. PINEDA-LOPEZ, J. M. CASPETA-MANDUJANO, E. AGUILAR-CASTELLANOS, AND N. MERCADO-SILVA. 2004. Helminth parasites of freshwater fishes of the Panuco river basin, east central Mexico. Comparative Parasitology 71:190-202.
SCOTT, L. E., AND J. B. JOHNSON. 2010. Does sympatry predict life history and morphological diversification in the Mexican livebearing fish Poeciliopsis baenschi? Biological Journal of the Linnean Society 100:608-618.
STEARNS, S. C. 1992. The evolution of life histories. Oxford University Press, Oxford, United Kingdom.
STOCKWELL, C. A., AND G. L. VINYARD. 2000. Life history variation in recently established populations of western mosquitofish (Gambusia affinis). Western North American Naturalist 60:273-280.
SWENTON, D. M., AND A. KODRIC-BROWN. 2012. Habitat and life history difference between two species of Gambusia. Environmental Biology of Fishes 94:669-680.
THOMPSON, M.B., J.R. STEWART, B. K. SPEAKE, M. J. HOSIE, AND C. R. MURPHY. 2002. Evolution of viviparity: what can Australian lizards tell us? Comparative Biochemistry and Physiology B 131:631-643.
TURNER, C. L. 1937. Reproductive cycles and superfetation in poeciliid fishes. Biological Bulletin 72:145-164.
WINEMILLER, K. O. 1993. Seasonality of reproduction by live-bearing fishes in tropical rainforest streams. oecologia 95:266-276.
ZUNIGA-VEGA, J. J., D. N. REZNICK, ANDJ. B. JOHNSON. 2007. Habitat predicts reproductive superfetation and body shape in the livebearing fish Poeciliopsis turrubarensis. Oikos 116:995-1005.
ZUNIGA-VEGA, J.J., M. SUAREZ-RODRESUEZ, H. ESPINOSA-PEREZ, AND J. B. JOHNSON. 2011. Morphological and reproductive variation among populations of the Pacific molly Poecilia butleri. Journal of Fish Biology 79:1029-1046.
Submitted 26 September 2013.
Acceptance recommended by Associate Editor, Mark Pyron, 15 January 2014.
MEAGHAN L. WELDELE, J. JAIME ZUNIGA-VEGA, * AND JERALD B. JOHNSON
Evolutionary Ecology Laboratories, Department of Biology, Brigham Young University, Provo, UT 84602 (MLW, JBJ) Departamento de Ecologia y Recursos Naturales, Facultad de Ciencias, Universidad Nacional Autonoma de Mexico, Ciudad Universitaria 04510, Distrito Federal, Mexico JJZ-V)
* Correspondent: email@example.com
TABLE 1--Description of sampling locations for Gambusia vittata. Values for pH, temperature, and salinity correspond to average values across sampling months. Site numbers are used throughout the paper for identification purposes. Site Geographical Elevation State Collecting number coordinates (m) months: sample size (males; females) 1 21[degrees] 52.8 San Luis January 2012: 14'2.2"N, Potosi 63(16; 47) 98[degrees] April 2012: 70 52'35.1"W (9; 61) 2 21[degrees] 157.4 Veracruz January 2012: 16'41.8"N, 52(24; 28) 97[degrees] April 2012: 57'18.1"W 61(6; 55) March 2013: 74(42: 32) 3 21[degrees] 93.6 Veracruz January 2012: 19'45"N, 17(3; 14) 97[degrees] April 2012: 44'51.2"W 61(19; 42) Site pH Temperature Salinity number ([degrees]C) (ppt) 1 8.04 29.2 0.48 2 7.61 28.2 0.58 3 7.73 24.9 0.45 Site Other species Descriptive number present notes 1 Poecilia mexicana (a) River with clear water and fast Pseudoxiphophorus jonesii (a) current. Substrate Astyanax mexicanus (b) of rocks and gravel. Herichthys pantostictus (c) Abundant algae. Depths to 0.7 m. 2 Poecilia mexicana (c) River with turbid water. Currents are Poeciliopsis gracilis (a) slight to moderate. Muddy substrate. Pseudoxiphophorus jonesii (a) Abundant riparian Xiphophorus hirchmanni (a) vegetation and water Xiphophorus nezahualcoyotl (a) hyacinth (Eichhornia Xiphophorus variatus (a) crassipes). Depths Astyanax mexicanus (b) to 0.9 m. 3 Gambusia panuco (a) River with clear water and moderate Poecilia mexicana (a) current. Substrate Xiphophorus variatus (a) of mud and bedrock. Astyanax mexicanus (b) Abundant algae and Herichthys pantostictus (c) water hyacinth. Depths to 0.3 m. (a) Poeciliidae. (b) Characidae. (c) Cichlidae. TABLE 2--Results of analysis of covariance model examining brood size, individual embryo mass, and reproductive allotment of Gambusia vittata at three sites in eastern Mexico. Life-history Effect SS (a) df MS trait Brood size Female mass 40.26 1 40.26 Month (site) 1.27 3 0.42 Site 1.00 2 0.50 Site x female 1.0 2 0.50 mass Error 21.25 172 0.1235 Individual Female mass 2.91 1 2.91 embryo mass Stage of 0.07 1 0.07 embryos Month (site) 1.33 3 0.44 Site 0.12 2 0.06 Site x 0.43 2 0.21 stage of development Site x female 0.21 2 0.10 mass Error 20.48 169 0.12 Reproductive Female mass 47.40 1 47.40 allotment Month (site) 3.73 3 1.24 Site 0.04 2 0.02 Site x female 0.03 2 0.02 mass Error 34.70 172 0.20 Life-history Effect F P trait Brood size Female mass 325.93 <0.0001 Month (site) 3.42 0.019 Site 4.06 0.019 Site x female 4.03 0.020 mass Error -- -- Individual Female mass 24.02 <0.0001 embryo mass Stage of 0.61 0.43 embryos Month (site) 3.67 0.014 Site 0.48 0.62 Site x 1.77 0.17 stage of development Site x female 0.86 0.42 mass Error -- -- Reproductive Female mass 234.96 <0.0001 allotment Month (site) 6.17 0.0005 Site 0.099 0.91 Site x female 0.08 0.92 mass Error -- -- (a) SS indicates sum of squares; MS, mean square
|Printer friendly Cite/link Email Feedback|
|Author:||Weldele, Meaghan L.; Zuniga-Vega, J. Jaime; Johnson, Jerald B.|
|Date:||Dec 1, 2014|
|Previous Article:||Reproductive characteristics of two Gambusia congeners in West Texas.|
|Next Article:||Survival and growth rates of stocked razorback suckers (Xyrauchen texanus) in a wetland of the Green River, Utah.|