Molecular variation and evolutionarily significant units in the endangered Gila topminnow

View Description

View PDF & Text

PDF
Text

Text Search...

Final Technical Report: Microsatellite Variation in Topminnows (I95051)
Arizona Game and Fish Department
Heritage Grant IIPAM No. I95051
Molecular Variation and Evolutionarily Significant Units in the Endangered Gila Topminnow
KAREN M. PARKER, RUBY J. SHEFFER, AND PHILIP W. HEDRICK
Department of Biology, Arizona State University, Tempe, AZ 85287-1501
Running head: Variation in the Gila topminnow
Word count: 7,240
Key words: evolutionarily significant units, Gila topminnow, major histocompatibility complex, microsatellite loci
Address correspondence to Philip W. Hedrick, email hedrick@hedricklab.la.asu.edu
Address: as above
Tel: 602-965-0799
Fax: 602-965-2519 Abstract: Variation in microsatellite loci for the endangered Gila topminnow from the four watersheds in Arizona which they are still naturally extant was examined. Bylas Spring had quite low variation while the other populations all had significant variation. All of the populations had "diagnostic" alleles and the genetic divergence, based on several quantitative measures, between populations was substantial. The amounts and patterns of genetic variation are consistent with known historical and physical differences between sites. Further, the sites differ in a number of important factors in their physical habitat, biota, and the life-history of the topminnows. Based on these considerations, we recommend that the four watersheds all be considered as evolutionarily significant units and managed in a manner appropriate for separate evolutionarily significant units.
Introduction
Highly variable molecular genetic variants have recently been used to address a number of issues previously difficult to answer in conservation and evolutionary biology (Avise & Hamrick 1996; Smith & Wayne 1996). Determining the unit of conservation is particularly controversial, with some suggesting the species as the appropriate unit (e.g. Caughley & Gunn 1996) while others consider evolutionarily significant units (ESUs) that are on independent evolutionary trajectories, whether they be populations, stocks, subspecies, or species (e.g. Moritz 1994; Waples 1995), should be the units of conservation. Until recently, molecular markers most commonly used were allozymes or mitochondrial DNA (mtDNA) variants. In recent years, highly variable nuclear molecular markers are being widely applied. Although it is tempting to use the latest molecular findings exclusively to determine ESUs, other important historical, ecological, and distributional data need to be included to provide a comprehensive perspective (Waples 1995). Here we attempt to provide such an integrated perspective for the endangered Gila topminnow.
2
The Gila topminnow (Poeciliopsis o. occidentalis), once one of the most abundant fishes in the Gila river drainage, is federally and state-listed as endangered and exists in the United States in only four isolated Arizona watersheds (Figure 1) (e.g. Sheffer et al. 1997a). Major factors influencing their decline were the loss and fragmentation of adequate shallow-water habitat and the establishment of the non-native central mosquito fish, Gambusia affinis (Meffe et al. 1983). The present report is one of a series in which we have examined genetic variation and fitness-related traits in the Gila topminnow in an effort to understand some of the genetic factors important for its conservation (see Discussion).
Below we describe genetic variation in polymorphic microsatellite loci identified in the Gila topminnow and indicate what evolutionary factors appear to have resulted in the pattern of variation observed. Variation in microsatellite loci is compared to that found in previous studies for allozymes (Vrijenhoek et al. 1985; Meffe & Vrijenhoek 1988), mtDNA (Quattro et al. 1996), and a potentially adaptive major histocompatibility complex (MHC) locus (Hedrick & Parker 1998). In light of these findings and other data on distribution, habitats, and life history of extant Gila topminnow populations, we discuss appropriate ESUs and potential conservation for the Gila topminnow.
Materials and Methods
Microsatellite loci, which have short, tandem repeats of variable number that are considered different alleles (Ashley & Dow, 1994), have become an important nuclear molecular markers to determine population differentiation. Variation at microsatellite loci is particularly appropriate for this application because they generally evolve rapidly, are not within transcribed genes, and allelic variation at these loci is thought to be selectively neutral.
3
Using standard techniques (Parker et al. 1998), we identified 10 microsatellite loci in Gila topminnows, five of which are polymorphic in our present samples. Three of the five polymorphic loci (Pooc-6-10, Pooc-C15, and Pooc-LL53) are simple dinucleotide repeats while the other two (Pooc-G49 and Pooc-OO56) are composed of dinucleotide repeats interrupted by other sequences.
Natural populations of Gila topminnows exist in the United States in only four Arizona watersheds, represented here by samples from Bylas Spring, Cienega Creek, Monkey Spring, and Sharp Spring (see Discussion below and in Sheffer et al. 1997a). For the present study, we examined variation at microsatellite loci in 160 individuals, 40 individuals sampled from stocks derived from each population - the same fish previously examined for variation at a MHC locus (Hedrick and Parker 1998). We also compare variation in these individuals for microsatellite loci and the MHC locus.
Results
Variation within populations
Table 1 gives the observed allelic frequencies for the five polymorphic loci in the four samples (allele designations indicate the number of base pairs for that allele). First, notice that each population has several alleles (in boldface) that appear "diagnostic" for the specific population. For example for Bylas Spring, allele Pooc-G49-161 was absent in other samples and allele Pooc-C15-240 was fixed in Bylas Spring and in very low frequency in Cienega Creek and Sharp Spring. Overall, Bylas Spring has diagnostic alleles at two loci, Cienega Creek at three, Monkey Springs at four, and Sharp Spring at two.
The loci vary both in sharing alleles over different populations and the amount of variation. The highest sharing is for loci Pooc-G49 and Pooc-OO56 which both have an
4
allele common in all, or nearly all, of the samples. The other three loci are quite divergent, with Monkey Spring fixed for an allele at locus Pooc-6-10 (and different by five dinucleotide repeats) not found in the others and high divergence between nearly all pairs of populations for Pooc-C15 and Pooc-LL53, although the size range and total number of alleles is higher for Pooc-C15.
The amounts of within-population variation, as determined by observed and expected heterozygosity and the observed number of alleles, is in Table 2. In all locus-population combinations, observed heterozygosity was not significantly different than that expected from Hardy-Weinberg proportions. The lowest variation was in the Bylas Spring sample, polymorphic for only one locus. However, At this locus (Pooc-G49), however, the Bylas Spring sample had the highest heterozygosity of all four because of a second allele in substantial frequency that differed by a single dinucleotide repeat. The other three samples were each polymorphic for three loci, with Cienega Creek and Sharp Spring having the highest average heterozygosity. Sharp Spring had the most alleles averaged over loci (3.8), primarily because of 12 different alleles at Pooc-C15. The locus with the lowest within-population variation was Pooc-6-15, fixed for an allele in all populations, while the most variable was Pooc-C15, with 6.2 alleles per sample, even though Bylas Spring was fixed for this locus.
We examined the variation over individuals within a population in several different ways to determine if there was interlocus association. First, we calculated the correlation coefficient for individual homozygosity (scored as 0) versus heterozygosity (scored as 1) for each polymorphic pair of loci over all 40 individuals within a sample. The average of these nine correlation coefficients was near zero (0.049) and they ranged from -0.125 to 0.200 (all non-significant). Second, we calculated linkage (gametic) disequilibrium between pairs of polymorphic loci for which we had sufficient statistical power to detect it. For example, for Pooc-C15 the 40 individuals from Sharp Spring were represented by 20 different genotypes, making linkage disequilibrium estimation
5
infeasible for any locus pair that included Pooc-C15. In fact, estimation was possible only for the three locus pairs in Cienega Creek and there was no evidence of linkage disequilibrium for any of these using the approach of Slatkin & Excoffier (1996). Finally, we calculated the observed variance in individual heterozygosity over all five loci and compared it to that randomly generated from a sample of size 40 (Table 3). In all cases, observed variance was close to that expected and the probability of obtaining a variance greater than that observed by chance using a Monte Carlo approach ranged from 0.20 to 0.42, with a mean of 0.31, consistent with the hypothesis that there were no interlocus associations (e.g. Brown et al. 1980).
Variation between populations
We examined differences between populations using several different measures. First, the probability of a unique allele per locus was calculated, which indicates the frequency of allelic differences in kind for a given population compared to another population (Hedrick 1971). Table 4 gives probabilities of a unique allele for the five polymorphic loci for all sample pairs, the values range from 0.150 for the probability of a unique allele in Bylas Spring when compared to Cienega Creek to 0.597 for Monkey Spring when compared to Bylas Spring. Averages across the bottom vary from 0.350 for Bylas Spring with the least unique alleles compared to other samples, to 0.590 for Monkey Spring which has the highest proportion. The largest asymmetry in the averages across the bottom of the table and those on the right side is for Bylas Spring, indicating that Bylas Spring has fewer unique alleles when compared to other samples (0.350) than the other samples have when compared to Bylas Spring (0.486).
We also calculated genetic distance (Nei 1972) and FST (Nei 1987) between pairs of samples for the ten microsatellite loci. We did not calculate a size-based genetic-distance measure because of large differences between loci in the variation in size (see
6
discussion in Boyce et al. 1997; see also Takezaki & Nei 1996). Genetic distance and FST values varied from 0.098 and 0.223 for Cienega Creek - Sharp Spring comparison to 0.440 and 0.712 for the Bylas Spring - Monkey Spring comparison (Table 5). For all pairs of populations, the FST values were statistically significant.
Using the genetic distance values from Table 5 for the microsatellite loci, Figure 2 is a phylogenetic tree using the UPGMA approach (Kumar et al. 1993). Monkey Spring is the most divergent, consistent with that seen from calculating the probability of a unique allele in Table 4. Cienega Creek and Sharp Spring cluster together, with Bylas Spring also closely related. Using a neighbor-joining approach to obtain a phylogenetic tree (not shown), Monkey Spring is again quite divergent but Bylas Spring and Cienega Creek first cluster together, with Sharp Spring being closely related.
Comparison of MHC and Microsatellite Variation in the Same Individuals
We examined the association of microsatellite and MHC variation in the same individuals in several ways. First, we calculated the correlation coefficient of individual heterozygosity for each locus pair in each population in which they were variable (nine different combinations) and none was significant. The mean value of these correlations was -0.018 and ranged from -0.15 to 0.16. Second, we calculated the extent of linkage disequilibrium between the microsatellite loci and the MHC locus in the only three combinations for which the calculation could be done. All were in the Cienega Creek, and none showed a statistically significant association. Finally, we calculated the variance of individual heterozygosity including the MHC locus and the five polymorphic microsatellite loci. The probability that the variance was greater than that expected at random was similar to that for microsatellite loci alone, with an average probability of 0.39. Overall then, there appears to be no association between variation among individuals within a population for microsatellite loci and the MHC locus.
7
Discussion
Gila topminnows have become one of the most thoroughly examined endangered species genetically, with molecular genetic studies of allozymes, mtDNA, a MHC locus, and microsatellite loci and research on traits related to fitness by Quattro & Vrijenhoek (1989), Sheffer et al. (1997a), Sheffer et al. (1997b), and Cardwell et al. (1998). Along with the extensive information from earlier studies on the natural history and ecology of the species and the factors influencing its decline (see Meffe et al. 1983 and references therein), it appears possible to understand relationships between present-day populations and make well-educated recommendations for future conservation efforts.
Comparison of Data from Different Molecular Markers
New, within-population genetic data from the microsatellite loci (expected heterozygosity and observed number of alleles) are summarized in Table 6 along with the data from previous studies (allozymes from Meffe & Vrijenhoek, 1988; mtDNA from Quattro et al. 1996; and MHC from Hedrick & Parker, 1998). These data are generally consistent over types of loci, allowing for sampling error over loci, thus making several generalizations possible. However, there are several observations that merit discussion.
First, Bylas Spring has the lowest overall variation (the only polymorphic locus for this sample was Pooc-G49) and had the lowest variation for both microsatellite loci and the MHC locus. The other three populations all have substantial polymorphism, with Sharp Spring somewhat higher for both measures of genetic variation. Vrijenhoek et al. (1985), after their allozyme survey, suggested that Sharp Spring as the most genetically variable of the four sites, and the other three had low variation. Sharp Spring still appears
8
the most genetically variable of the four sites when examining all types of loci given in Table 6, but both Cienega Creek and Monkey Spring have substantial overall variation.
Second, genetic distance is substantial for all pairs of populations for microsatellite loci, with Bylas Spring-Monkey Spring the largest (Table 5). This comparison did not share alleles at the MHC locus, making the MHC genetic distance infinite. Vrijenhoek et al. (1985) found the different sites were similar for allozyme loci. In fact, Bylas and Monkey Springs shared alleles (and were monomorphic for them) at all 25 allozyme loci examined, making allozyme genetic distance zero between these populations. Sharing of alleles for the allozyme loci is surprising because Bylas and Monkey Springs now are the most physically isolated from each other. The lack of any differences may reflect common ancestry in the distant past and a low rate of evolution for these loci, along with low power to determine the level of differentiation because of generally low variation at allozyme loci.
Third, variation in mtDNA region examined by Quattro et al. (1996) shows far less variation within and between populations than either microsatellite loci or the MHC locus. It is thought generally that mtDNA evolves rapidly but this does not appear the case for a number of fishes (e.g. Moritz et al. 1987). In addition, Quattro et al. (1996) used six-cutter restriction endonucleases, so that their results were of relatively low resolution, probably not recognizing differentiation that occurred in the last several hundred thousand years (Wilson et al. 1985).
Finally, there is evidence that genes in the MHC are under balancing selection (Hedrick 1994), most likely from parasite of pathogen resistance (Hedrick & Kim 1997). If so, the pattern of variation within and between populations may reflect this phenomenon, given reasonably strong balancing selection. The pattern of variation within individuals for the MHC locus, however, is consistent with Hardy-Weinberg expectations for the samples from Cienega Creek and Monkey Spring (Bylas Spring is not variable) and shows a deficiency of heterozygotes for Sharp Spring, the opposite of the expectation
9
for balancing selection (possibly due to the presence of a null allele, Hedrick & Parker 1998).
Between-population differentiation for the MHC locus is significant and appears to be as large or larger than for the polymorphic microsatellite loci (see discussion of measures of differentiation in Hedrick & Parker 1998). While it is possible that the MHC differentiation may have resulted from divergent selection among sites, the amount of differentiation is not unlike the pattern observed for microsatellite loci Pooc-C15 and Pooc-LL53, suggesting the pattern of MHC differentiation may be primarily influenced by non-selective factors (see below). In fact, the pattern of variation over samples for microsatellite locus Pooc-LL53 is nearly identical to that observed for the MHC locus, i.e., both are monomorphic in Bylas Spring, share an allele in high frequency between Bylas Spring and Cienega Creek, are nearly fixed for a third allele in Monkey Spring, and Sharp Spring has several different alleles but shares one with Monkey Spring. Because the measures of association within individuals for Pooc-LL53 and the MHC locus were non-significant, it is likely that these nearly identical patterns of genetic variation for these two loci are similar reflections of population history and are not the result of linkage. When comparing the patterns of microsatellite and MHC variation over samples in bighorn sheep, Boyce et al. (1997) concluded that the greater similarity of MHC variation as compared to the microsatellite variants over samples provided support that MHC had a similar selective history over the different populations. Such a pattern is not obvious in the Gila topminnow data.
Explanation of the Observed Variation Within and Between Populations
There are number of non-selective factors that may play a role in determining variation within and between populations. Table 7 lists primary ones that may have been important in the four populations and gives an estimate of the relative effect for each factor.
10
Probably the most important factor determining lower variation observed in Bylas Spring is the low effective population size. This is both a function of the small habitat, the amount of water flowing from the spring (actually the middle spring or S-II) averages only a few liters per minute which can support at most a few hundred adults, and the recurring bottlenecks observed since the site was discovered (Marsh & Minckley 1990; Minckley et al. 1991). In fact, during one of these bottleneck periods in 1990, Gila topminnows became extinct at Bylas Spring (our stock was derived from a refugium population). In addition, because its location is elevated from the Gila River, it may be that there were only a few founders to originally colonize the site and that immigration has since been unlikely.
Monkey Spring appears to have been isolated from a tributary of Sonoita Creek through formation of a natural, 10-meter high travertine dam perhaps 10,000 years bp (Minckley et al. 1991). In other words, there may have been little or no immigration into Monkey Spring for more than 20,000 generations, assuming that there is two or more generations per year in the warm, constant-temperature habitat. Although the size of the site is somewhat reduced from its time of discovery by water development, population size is substantial and varies seasonally less than other sites because of its constant temperature and flow, and absence of floods. Long-term isolation with little or no immigration provides an explanation for the high probability of unique microsatellite alleles observed in Monkey Spring.
Both Sharp Spring and Cienega Creek have large amounts of variation shared in part with other sites. Because the population sizes have been large historically and there was substantial connectedness between these and other sites, such patterns of variation are not unexpected. Sharp Spring has several MHC alleles that differ substantially from those at other sites (Hedrick & Parker 1998) and has a large number of alleles at microsatellite locus Pooc-C15, suggesting Sharp Spring (and the upper Santa Cruz River that it now represents) may have had a very large population size in the recent past (also
11
indicated by museum collections) and may have been somewhat isolated from other populations.
Evolutionarily Significant Units in the Gila Topminnow
The molecular genetic studies of microsatellite and MHC variation generally support the importance of all four populations as significant ESUs. Monkey Spring, which has the most unique alleles and is quite genetically divergent, and Sharp Spring, which has the most genetic variation, are the most strongly supported as separate ESUs by these data. In addition, there are a number of other non-genetic attributes that support their significance (Table 8).
First, Bylas Spring is the only remaining population representing the mainstem Gila River stock (see Marsh & Minckley 1990; Minckley et al. 1991), isolated from others by at least 580 km by stream channel, much of which now is dry except during flooding. The other populations are also isolated, with 250 km separating Cienega Creek and Monkey Spring and 101 km separating Monkey Spring and Sharp Spring. There are extensive dry reaches between all sites except during floods. In other words, even though there may have been ancestral exchange between these sites though intermediate populations, it is unlikely that there has been exchange in recent decades and natural exchange is unlikely in the future.
Monkey Spring, besides appearing to be the most distinctive population based on the microsatellite data, also was occupied by a now-extinct species of pupfish, genus Cyprinodon (Minckley 1973; Minckley et al. 1991), and a now extinct, morphologically distinct form of Gila intermedia, the Gila chub (Rinne 1976; DeMaris 1986). Monkey Spring topminnows appear to have relatively low fecundity in nature (e.g. Constanz 1979) although this difference disappeared under laboratory conditions (Sheffer et al. 1997a). On the other hand, Cardwell et al. (1998) found that in a common laboratory
12
environment, male development time was approximately 50% longer in Monkey Spring topminnows than for males from any of the other three sites. Because Gila topminnows are members of a tropical genus, perhaps delayed development is a relict trait retained (or expressed) because of the warm, constant temperature at Monkey Spring. Perhaps, other populations evolved faster male development because of higher male-male competition for mates in seasonally variable environments.
Finally, neither Cienega Creek nor Monkey Spring has been infested with central mosquito fish while Bylas and Sharp Spring have been impacted by these non-native fish. Presence of mosquito fish not only introduces a competitor and predator but potentially, non-native parasites and pathogens. The suite of selection pressures experienced by the populations with or without mosquito fish may be substantial and itself of evolutionary significance.
Moritz (1994) suggested that genetic criteria for evolutionarily significant units is that they show phylogeographic differentiation for mtDNA variants and significant divergence of allele frequencies at nuclear loci. The four populations examined here show significant divergence both at the microsatellite loci and the MHC locus. Although no variation was found for mtDNA by Quattro et al. (1996) and there was little phylogeographic differentiation at the MHC locus, there is some indication of the size of the microsatellite alleles sorting out geographically. For example, at Pooc-LL53 the three largest alleles are found in the Sharp Spring sample and for Pooc-C15 in Monkey Spring there is cluster of seven alleles with sizes from 204 to 224 while Sharp Spring has three small alleles and nine alleles with sizes between 226 and 240.
Let us put these conclusions in perspective. After a laboratory study of fitness correlates by Vrijenhoek et al. (1985) indicated Sharp Spring fish had higher fitness than Monkey Spring fish, they recommended "the Sharp Spring stock currently offers the best choice for stocking the Gila River system." As a result, the U. S. Fish and Wildlife Service stopped using Monkey Spring fish for reintroductions and has since used only
13
Sharp Spring stock. Genetic results presented here, those of Sheffer et al. (1997a), Hedrick & Parker (1998), and Cardwell et al (1998) are, on the other hand, consistent with the recommendations by Simons et al. (1989) that "at least one representative lineage is preserved from each of the four geographic areas in Arizona." These genetic results support the suggestion in the draft recovery plan (USFWS 1993) that reintroductions be "from the hydrographically nearest natural population."
On a cautionary note, the new, highly variable markers used here are considered indicative of recent, non-selective factors. It is therefore possible that significant differences between populations may be observed that are not correlated with variation in adaptively important traits (Hedrick 1996; Hedrick & Savolainen 1996). However, the size of the genetic differences observed in these topminnow populations for both microsatellite loci and the MHC locus suggest that their isolation has been long enough term to reasonably suggest that the populations are on independent evolutionary trajectories, i.e., that they are significantly different evolutionarily significant units.
14
Acknowledgments
We appreciate the knowledge, inspiration, and comments on this manuscript by W. L. Minckley and the statistical assistance of Steven Kalinowski. This research was supported in part through funding from the Arizona Game and Fish Department Heritage Fund.
Disclaimer
The findings, opinions, and recommendations in this report are those of the investigators who have recieved partial or full funding from the Arizona Game and Fish Department Heritage Fund. The findings, opinions, and recommendations do not necessarily reflect those of the Arizona Game and Fish Commission or the Department, or necessarily represent official Department policy or management practice. For further information, please contact the Arizona Game and Fish Department.
15
Literature Cited
Ashley, M. V., and B. D. Dow. 1994. The use of microsatellite analysis in population biology: background, methods, and potential applications. Pages 185-201 in B. Sherwater, B. Streit, G. P. Wagner, and R. DeSalle (eds.). Molecular ecology and evolution: approaches and applications. Birkhauser, Basel, Switzerland.
Avise, J. C., and J. L Hamrick (editors). 1996. Conservation genetics: case histories from nature. Chapman and Hall, New York.
Boyce, W. M., P. W. Hedrick, N. E. Muggli-Cockett, S. Kalinowski, M. C. Penedo, and R. R. Ramey. 1997. Genetic variation of major histocompatibility complex and microsatellite loci: a comparison in bighorn sheep. Genetics 145:421-433.
Brown, A. H. D., M. F. Feldman, and E. Nevo. 1980. Multilocus structure of natural populations of Hordeum spontaneum. Genetics 96:523-536.
Cardwell, T. N., R. J. Sheffer, and P. W. Hedrick. 1998. Differences in male development time among populations of the endangered Gila topminnow. Journal of Heredity (submitted).
Caughley, G., and A. Gunn. 1996. Conservation biology in theory and practice. Blackwell Science. Cambridge, Massachusetts.
Constanz, G. D. 1979. Life history patterns of a livebearing fish in contrasting environments. Oecologia 40:189-201.
DeMarais, B. D. 1986. Morphological variation in Gila (Pisces: Cyprinidae) and geologic history: Lower Colorado river basin. M.S. Thesis. Arizona State University, Tempe, Arizona.
Hedrick, P. W. 1971. A new approach to measuring genetic similarity. Evolution 25:276-280.
Hedrick, P. W. 1994. Evolutionary genetics of the major histocompatibility complex. American Naturalist 143:945-964 .
16
Hedrick, P. W. 1996. Conservation genetics and molecular techniques: a perspective. Pages 459-477 in T. B. Smith and R. K. Wayne, editors. Molecular genetic approaches in conservation. Oxford University Press, New York.
Hedrick, P. W., and T. Kim. 1997. Genetics of complex polymorphisms: parasites and maintenance of MHC variation. in R. S. Singh and C. K. Krimbas, editors. Evolutionary genetics from molecules to morphology. Cambridge University Press, New York.
Hedrick, P. W., and K. M. Parker. 1998. MHC variation in the endangered Gila topminnow. Evolution (in press).
Hedrick, P. W., and O. Savolainen. 1996. Molecular and adaptive variation: a perspective for endangered plants. Pages 92-102. in J. Machinski, D. H. Hammond, and L. Holter, Editors. Southwestern rare and endangered plants: proceedings of the second conference. General technical report RM-GTR-283. U. S. Department of Agriculture, Forest Service, Fort Collins, Colorado.
Kumar, S., K. Tamura, and M. Nei. 1993. MEGA: molecular evolutionary genetics analysis, version 1.01. Pennsylvania State University, University Park, Pennsylvania.
Marsh, P. C., and W. L. Minckley. 1990. Management of endangered Sonoran topminnow at Bylas Springs, Arizona: description, critique, and recommendations. Great Basin Naturalist 50:265-272.
Meffe, G. K., D. A. Hendrickson, W. L. Minckley, and J. N. Rinne. 1983. Factors resulting in decline of the endangered Sonoran topminnow, Poeciliopsis occidentalis (Atheriniformes: Poeciliidae), in the United States. Biological Conservation 25:135-159.
Meffe, G. K., and R. C. Vrijenhoek. 1988. Conservation genetics in the management of desert fishes. Conservation Biology 2:157-169.
17
Minckley, W. L. 1973. Fishes of Arizona. Arizona Game and Fish Department. Phoenix Arizona.
Minckley, W. L., G. K. Meffe, and D. L. Soltz. 1991. Conservation and management of short-lived fishes: the cyprinodontoids. Pages 247-282. in W. L. Minckley and J. E. Deacon, editors. Battle against extinction: native fish management in the American west. University of Arizona Press, Tucson, AZ.
Moritz, C. 1994. Defining "evolutionary significant units" for conservation. Trends in Ecology and Evolution 9:373-375.
Moritz, C., T. E. Dowling, and W. M. Brown. 1987. Evolution of animal mitochondrial DNA: relevance for population biology and systematics. Annual Reviews of Ecology and Systematics 18:269-292.
Nei, M. 1972. Genetic distance between populations. American Naturalist 106:283-292 .
Nei, M. 1987. Molecular evolutionary genetics. Columbia University Press, New York.
Parker, K. M., K. Hughes, T. J. Kim, and P. W. Hedrick. 1998. Isolation and characterization of microsatellite loci from the Gila topminnow (Poeciliopsis o. occidentalis) and examination in guppies (Poecilia reticulata). Molecular Ecology (in press).
Quattro, J. M., P. L. Leberg, M. E. Douglas, and R. C. Vrijenhoek. 1996. Molecular evidence for a unique evolutionary lineage of endangered Sonoran desert fish (genus Poeciliopsis). Conservation Biology 10:128-135.
Quattro, J. M., and R. C. Vrijenhoek. 1989. Fitness differences among remnant populations of the endangered Sonoran topminnow. Science 245:976-978.
Rinne, J. N. 1976. Cyprinid fishes of the genus Gila from the lower Colorado River basin. Wassman Journal of Biology 34: 65-107.
Sheffer, R. J., P. W. Hedrick, W. L. Minckley, and A. L. Velasco. 1997a. Fitness in the endangered Gila topminnow. Conservation Biology 11:162-171.
18
Sheffer, R. J., P. W. Hedrick and C. Shirley. 1997b. No bilateral asymmetry in wild-caught, endangered Gila topminnows. Heredity (in press).
Simons, L. H., D. A. Hendrickson, and D. Papoulias. 1989. Recovery of the Gila topminnow: a success story? Conservation Biology 3:11-15.
Slatkin, M., and L. Excoffier. 1996. Testing for linkage disequilibrium in genotypic data using the Expectation-Maximinization algorithm. Heredity 76:377-383.
Smith, T. B., and R. K Wayne (editors). Molecular genetic approaches in conservation. Oxford University Press, New York.
Takezaki, N, and M. Nei. 1996. Genetic distances and reconstruction of phylogenetic trees from microsatellite DNA. Genetics 144:389-399.
U. S. Fish and Wildlife Service. 1993. Draft Gila topminnow recovery plan. U. S. Fish and Wildlife Service. Albuquerque, New Mexico.
Vrijenhoek, R. C., M. E. Douglas, and G. K. Meffe. 1985. Conservation genetics of endangered fish populations in Arizona. Science 229:400-402 .
Waples, R. S. 1995. Evolutionary significant units and the conservation of biological diversity under the Endangered Species Act. American Fisheries Society Symposium 17:8-27.
Wilson, A. C., R. L. Cann, S. Carr, M. George, U. B. Gyllensten, et al. 1985. Mitochondrial DNA and two perspectives on evolutionary genetics. Biological Journal of the Linnean Society 26:375-400.
19
Table 1. Allelic frequencies for the five polymorphic microsatellite loci in the four populations. Frequencies in boldface are for "diagnostic" alleles, those only in a single population (except those in low frequency) or those in substantially higher in a given sample than all others.
Locus
Allele
Bylas Spring
Cienega Creek
Monkey Spring
Sharp Spring
Average
Pooc-G49
149
---
---
---
0.038
0.009
159
0.250
1.000
1.000
0.962
0.803
161
0.750
---
---
---
0.188
Pooc-6-10
287
---
---
1.000
---
0.250
297
1.000
1.000
---
1.000
0.750
Pooc-C15
202
---
---
---
0.050
0.012
204
---
---
0.025
0.200
0.056
208
---
---
0.088
0.012
0.025
210
---
---
0.012
---
0.003
214
---
---
0.612
---
0.153
216
---
---
0.225
---
0.056
218
---
0.012
---
---
0.003
222
---
---
0.012
---
0.003
224
---
---
0.012
---
0.003
226
---
---
---
0.012
0.003
228
---
---
---
0.100
0.025
230
---
---
---
0.012
0.003
232
---
0.362
0.012
0.025
0.100
234
---
---
---
0.100
0.025
236
---
---
---
0.338
0.084
238
---
0.612
---
0.112
0.181
240
1.000
0.012
---
0.025
0.259
246
---
---
---
0.012
0.003
Pooc-OO56
143
---
0.200
---
---
0.050
145
1.000
0.800
0.762
1.000
0.890
149
---
---
0.238
---
0.060
Pooc-LL53
142
---
---
0.988
---
0.247
144
---
0.488
---
---
0.122
146
1.000
0.512
---
---
0.378
150
---
---
0.012
0.425
0.109
154
---
---
---
0.550
0.138
164
---
---
---
0.025
0.006 20
Table 2. The observed heterozygosity (HO), expected heterozygosity (HE), and observed number of alleles (n) for the five polymorphic microsatellite loci in the four populations.
Locus
Measure
Bylas Spring
Cienega Creek
Monkey Spring
Sharp Spring
Average
Pooc-G49
HO
0.350
0.000
0.000
0.075
0.106
HE
0.375
0.000
0.000
0.075
0.112
n
2
1
1
2
1.5
Pooc-6-10
HO
0.000
0.000
0.000
0.000
0.000
HE
0.000
0.000
0.000
0.000
0.000
n
1
1
1
1
1.0
Pooc-C15
HO
0.000
0.600
0.600
0.800
0.500
HE
0.000
0.493
0.625
0.812
0.476
n
1
4
8
12
6.2
Pooc-OO56
HO
0.000
0.300
0.325
0.000
0.171
HE
0.000
0.320
0.362
0.000
0.171
n
1
2
2
1
1.5
Pooc-LL53
HO
0.000
0.625
0.025
0.500
0.288
HE
0.000
0.500
0.025
0.516
0.260
n
1
2
2
3
2.0
Average
HO
0.070
0.305
0.190
0.275
0.210
HE
0.075
0.263
0.202
0.281
0.205
n
1.2
2.0
2.8
3.8
2.45
21
Table 3. The number of heterozygous loci observed per individual for the five microsatellite loci and the mean over loci (H) and the mean observed and expected variance of heterozygosity per individual, OsH(2
) and EsH(2
), respectively, and the probability (P) that the observed variance is greater than that expected at random.
Heterozygosity
Population
0
1
2
3
4 H
OsH()2
EsH()2
P
Bylas Spring
26
14
-
-
-
0.35
0.23
0.23
0.41
Cienega Creek
4
17
13
6
-
1.52
0.77
0.70
0.21
Monkey Spring
12
18
10
-
-
0.95
0.56
0.50
0.20
Sharp Spring
3
21
14
2
-
1.38
0.50
0.49
0.42
22
Table 4. Probability of a unique allele for a given population at the top of the table when compared to the population at the side, averaged over the five polymorphic microsatellite loci.
Bylas Spring
Cienega Creek
Monkey Spring
Sharp Spring
Average
Bylas Spring
-
0.335
0.600
0.522
0.486
Cienega Creek
0.150
-
0.597
0.370
0.372
Monkey Spring
0.550
0.567
-
0.470
0.529
Sharp Spring
0.350
0.242
0.572
-
0.388
Average
0.350
0.381
0.590
0.454
23
Table 5. The genetic distance (D) and FST between the population pairs for the four types of loci where the number of loci is given in parentheses.
Comparison
Microsatellites (10)
Allozymes (25)
mtDNA (1)
MHC (1)
D FST
D FST
D FST
D FST
Bylas-Cienega
0.188 0.479
0.080 1.000
0.0 0.0
0.35 0.33
Bylas-Monkey
0.440 0.712
0.000 0.000
0.0 0.0
∞ 0.87
Bylas-Sharp
0.174 0.506
0.016 0.300
0.0 0.0
∞ 0.46
Cienega-Monkey
0.294 0.494
0.080 1.000
0.0 0.0
∞ 0.51
Cienega-Sharp
0.098 0.223
0.030 0.431
0.0 0.0
∞ 0.24
Monkey-Sharp
0.271 0.466
0.020 0.300
0.0 0.0
2.35 0.37
Overall FST
- 0.591
- 0.750
- 0.0
- 0.55
24
Table 6. The expected heterozygosity (HE) and observed number of alleles (n) for the different types of loci and the four populations where the number of loci in given in parentheses.
Loci (number)
Bylas Spring
Cienega Creek
Monkey Spring
Sharp Spring
HE n
HE n
HE n
HE n
Microsatellite (10)
0.038 1.10
0.132 1.50
0.101 1.90
0.140 2.40
Allozyme (25)
0.000 1.00
0.000 1.00
0.000 1.00
0.037 1.08
mtDNA (1)
0.000 1.00
0.000 1.00
0.000 1.00
0.000 1.00
MHC (1)
0.000 1.00
0.500 2.00
0.141 3.00
0.729 5.00
25
Table 7. Characteristics that are potentially relevant to understanding the amount of genetic variation within and the differentiation between the populations.
Characteristic
Bylas Spring
Cienega Creek
Monkey Spring
Sharp Spring
Population size (number of adults)
Small (<100)
Large (>>10,000)
Intermediate (<10,000)
Large (>>10,000)
Age of population (years before present)
Recent (<100)
Old (>>10,000)
Old (~10,000)
Old (>>10,000)
Relative historic immigration from other populations
Low
High
Low
High
Number of founders
Very low
High
Low
High
26
Table 8. Characteristics of the four populations divided into those of the physical habitat, the biota, and the life history of Gila topminnows.
Characteristic
Bylas Spring
Cienega Creek
Monkey Spring
Sharp Spring
Physical
Location
Mainstem Gila
Tributary to Santa Cruz
Above natural dam, tributary to Santa Cruz
Headwaters of Santa Cruz
Temperature*
Variable
Variable
Constant, warm spring
Variable
Habitat size
Very small
Large
Small
Large
Flooding frequency
Very low
High
Low
Intermediate
Biota
Endemic vertebrates
None
None
Pupfish, distinct chub
None
Mosquito fish
Present
Absent
Absent
Present
Food supply
Seasonally variable
Seasonally variable
Constant
Seasonally variable
Life history
Male development**
Fast
Fast
Slow
Fast
Fecundity***
High
High
Low
High
Female reproduction
Seasonal
Seasonal
Year-round
Seasonal
* All these habitats are thermally ameliorated, either by low-volume spring inflow or groundwater exchange.
**Under laboratory conditions (Cardwell et al. 1998).
***In the field.
27
Figure Legends
Figure 1. Map showing the location of the four samples used.
Figure 2. The phylogenetic tree for the four populations using the UPGMA approach based on the allelic frequencies at the five polymorphic microsatellite loci.
28
29
Appendix Table 1.
Locus
Allele
Bylas Spring
Cienega Creek
Monkey Spring
Sharp Spring
Pgd
a
1.00
-
1.00
0.45
f
-
1.00
-
0.55
Est-4
a
1.00
-
1.00
0.70
b
-
1.00
-
0.30
MHC - DRB
Pooc-1
1.000
0.500
-
-
Pooc-2
-
-
-
0.325
Pooc-3
-
-
-
0.175
Pooc-4
-
-
0.025
-
Pooc-5
-
0.500
-
-
Pooc-6
-
-
0.925
0.050
Pooc-7
-
-
0.050
-
Pooc-8
-
-
-
0.350
Pooc-9
-
-
-
0.100

Description

Rating
TITLE	Molecular variation and evolutionarily significant units in the endangered Gila topminnow
CREATOR	Parker, Karen M.
SUBJECT	Killifishes--Arizona--Genetics; Rare fishes--Arizona--Genetics;
Browse Topic	Land and resources
DESCRIPTION	29 pages (PDF version). File size: 598 KB
Language	English
Contributor	Sheffer, Ruby J.; Hedrick Philip W.; Arizona State University Dept. of Biology; Arizona Game and Fish Dept. Heritage Fund
Publisher	Arizona Game and Fish Dept.
TYPE	Text
Material Collection	State Documents
Acquisition Note	Heritage grant IIPAM no. I95051
RIGHTS MANAGEMENT	Copyright to this resource is held by the creating agency and is provided here for educational purposes only. It may not be downloaded, reproduced or distributed in any format without written permission of the creating agency. Any attempt to circumvent the access controls placed on this file is a violation of United States and international copyright laws, and is subject to criminal prosecution.
DATE ORIGINAL	1998 ca.
Time Period	1990s (1990-1999)
ORIGINAL FORMAT	Born Digital
Source Identifier	GF 4.3:I 23
Location	o57394122
DIGITAL IDENTIFIER	I95051.pdf
DIGITAL FORMAT	PDF (Portable Document Format)
REPOSITORY	Arizona State Library, Archives and Public Records--Law and Research Library.
File Size	611668 Bytes
Full Text	Final Technical Report: Microsatellite Variation in Topminnows (I95051) Arizona Game and Fish Department Heritage Grant IIPAM No. I95051 Molecular Variation and Evolutionarily Significant Units in the Endangered Gila Topminnow KAREN M. PARKER, RUBY J. SHEFFER, AND PHILIP W. HEDRICK Department of Biology, Arizona State University, Tempe, AZ 85287-1501 Running head: Variation in the Gila topminnow Word count: 7,240 Key words: evolutionarily significant units, Gila topminnow, major histocompatibility complex, microsatellite loci Address correspondence to Philip W. Hedrick, email hedrick@hedricklab.la.asu.edu Address: as above Tel: 602-965-0799 Fax: 602-965-2519 Abstract: Variation in microsatellite loci for the endangered Gila topminnow from the four watersheds in Arizona which they are still naturally extant was examined. Bylas Spring had quite low variation while the other populations all had significant variation. All of the populations had "diagnostic" alleles and the genetic divergence, based on several quantitative measures, between populations was substantial. The amounts and patterns of genetic variation are consistent with known historical and physical differences between sites. Further, the sites differ in a number of important factors in their physical habitat, biota, and the life-history of the topminnows. Based on these considerations, we recommend that the four watersheds all be considered as evolutionarily significant units and managed in a manner appropriate for separate evolutionarily significant units. Introduction Highly variable molecular genetic variants have recently been used to address a number of issues previously difficult to answer in conservation and evolutionary biology (Avise & Hamrick 1996; Smith & Wayne 1996). Determining the unit of conservation is particularly controversial, with some suggesting the species as the appropriate unit (e.g. Caughley & Gunn 1996) while others consider evolutionarily significant units (ESUs) that are on independent evolutionary trajectories, whether they be populations, stocks, subspecies, or species (e.g. Moritz 1994; Waples 1995), should be the units of conservation. Until recently, molecular markers most commonly used were allozymes or mitochondrial DNA (mtDNA) variants. In recent years, highly variable nuclear molecular markers are being widely applied. Although it is tempting to use the latest molecular findings exclusively to determine ESUs, other important historical, ecological, and distributional data need to be included to provide a comprehensive perspective (Waples 1995). Here we attempt to provide such an integrated perspective for the endangered Gila topminnow. 2 The Gila topminnow (Poeciliopsis o. occidentalis), once one of the most abundant fishes in the Gila river drainage, is federally and state-listed as endangered and exists in the United States in only four isolated Arizona watersheds (Figure 1) (e.g. Sheffer et al. 1997a). Major factors influencing their decline were the loss and fragmentation of adequate shallow-water habitat and the establishment of the non-native central mosquito fish, Gambusia affinis (Meffe et al. 1983). The present report is one of a series in which we have examined genetic variation and fitness-related traits in the Gila topminnow in an effort to understand some of the genetic factors important for its conservation (see Discussion). Below we describe genetic variation in polymorphic microsatellite loci identified in the Gila topminnow and indicate what evolutionary factors appear to have resulted in the pattern of variation observed. Variation in microsatellite loci is compared to that found in previous studies for allozymes (Vrijenhoek et al. 1985; Meffe & Vrijenhoek 1988), mtDNA (Quattro et al. 1996), and a potentially adaptive major histocompatibility complex (MHC) locus (Hedrick & Parker 1998). In light of these findings and other data on distribution, habitats, and life history of extant Gila topminnow populations, we discuss appropriate ESUs and potential conservation for the Gila topminnow. Materials and Methods Microsatellite loci, which have short, tandem repeats of variable number that are considered different alleles (Ashley & Dow, 1994), have become an important nuclear molecular markers to determine population differentiation. Variation at microsatellite loci is particularly appropriate for this application because they generally evolve rapidly, are not within transcribed genes, and allelic variation at these loci is thought to be selectively neutral. 3 Using standard techniques (Parker et al. 1998), we identified 10 microsatellite loci in Gila topminnows, five of which are polymorphic in our present samples. Three of the five polymorphic loci (Pooc-6-10, Pooc-C15, and Pooc-LL53) are simple dinucleotide repeats while the other two (Pooc-G49 and Pooc-OO56) are composed of dinucleotide repeats interrupted by other sequences. Natural populations of Gila topminnows exist in the United States in only four Arizona watersheds, represented here by samples from Bylas Spring, Cienega Creek, Monkey Spring, and Sharp Spring (see Discussion below and in Sheffer et al. 1997a). For the present study, we examined variation at microsatellite loci in 160 individuals, 40 individuals sampled from stocks derived from each population - the same fish previously examined for variation at a MHC locus (Hedrick and Parker 1998). We also compare variation in these individuals for microsatellite loci and the MHC locus. Results Variation within populations Table 1 gives the observed allelic frequencies for the five polymorphic loci in the four samples (allele designations indicate the number of base pairs for that allele). First, notice that each population has several alleles (in boldface) that appear "diagnostic" for the specific population. For example for Bylas Spring, allele Pooc-G49-161 was absent in other samples and allele Pooc-C15-240 was fixed in Bylas Spring and in very low frequency in Cienega Creek and Sharp Spring. Overall, Bylas Spring has diagnostic alleles at two loci, Cienega Creek at three, Monkey Springs at four, and Sharp Spring at two. The loci vary both in sharing alleles over different populations and the amount of variation. The highest sharing is for loci Pooc-G49 and Pooc-OO56 which both have an 4 allele common in all, or nearly all, of the samples. The other three loci are quite divergent, with Monkey Spring fixed for an allele at locus Pooc-6-10 (and different by five dinucleotide repeats) not found in the others and high divergence between nearly all pairs of populations for Pooc-C15 and Pooc-LL53, although the size range and total number of alleles is higher for Pooc-C15. The amounts of within-population variation, as determined by observed and expected heterozygosity and the observed number of alleles, is in Table 2. In all locus-population combinations, observed heterozygosity was not significantly different than that expected from Hardy-Weinberg proportions. The lowest variation was in the Bylas Spring sample, polymorphic for only one locus. However, At this locus (Pooc-G49), however, the Bylas Spring sample had the highest heterozygosity of all four because of a second allele in substantial frequency that differed by a single dinucleotide repeat. The other three samples were each polymorphic for three loci, with Cienega Creek and Sharp Spring having the highest average heterozygosity. Sharp Spring had the most alleles averaged over loci (3.8), primarily because of 12 different alleles at Pooc-C15. The locus with the lowest within-population variation was Pooc-6-15, fixed for an allele in all populations, while the most variable was Pooc-C15, with 6.2 alleles per sample, even though Bylas Spring was fixed for this locus. We examined the variation over individuals within a population in several different ways to determine if there was interlocus association. First, we calculated the correlation coefficient for individual homozygosity (scored as 0) versus heterozygosity (scored as 1) for each polymorphic pair of loci over all 40 individuals within a sample. The average of these nine correlation coefficients was near zero (0.049) and they ranged from -0.125 to 0.200 (all non-significant). Second, we calculated linkage (gametic) disequilibrium between pairs of polymorphic loci for which we had sufficient statistical power to detect it. For example, for Pooc-C15 the 40 individuals from Sharp Spring were represented by 20 different genotypes, making linkage disequilibrium estimation 5 infeasible for any locus pair that included Pooc-C15. In fact, estimation was possible only for the three locus pairs in Cienega Creek and there was no evidence of linkage disequilibrium for any of these using the approach of Slatkin & Excoffier (1996). Finally, we calculated the observed variance in individual heterozygosity over all five loci and compared it to that randomly generated from a sample of size 40 (Table 3). In all cases, observed variance was close to that expected and the probability of obtaining a variance greater than that observed by chance using a Monte Carlo approach ranged from 0.20 to 0.42, with a mean of 0.31, consistent with the hypothesis that there were no interlocus associations (e.g. Brown et al. 1980). Variation between populations We examined differences between populations using several different measures. First, the probability of a unique allele per locus was calculated, which indicates the frequency of allelic differences in kind for a given population compared to another population (Hedrick 1971). Table 4 gives probabilities of a unique allele for the five polymorphic loci for all sample pairs, the values range from 0.150 for the probability of a unique allele in Bylas Spring when compared to Cienega Creek to 0.597 for Monkey Spring when compared to Bylas Spring. Averages across the bottom vary from 0.350 for Bylas Spring with the least unique alleles compared to other samples, to 0.590 for Monkey Spring which has the highest proportion. The largest asymmetry in the averages across the bottom of the table and those on the right side is for Bylas Spring, indicating that Bylas Spring has fewer unique alleles when compared to other samples (0.350) than the other samples have when compared to Bylas Spring (0.486). We also calculated genetic distance (Nei 1972) and FST (Nei 1987) between pairs of samples for the ten microsatellite loci. We did not calculate a size-based genetic-distance measure because of large differences between loci in the variation in size (see 6 discussion in Boyce et al. 1997; see also Takezaki & Nei 1996). Genetic distance and FST values varied from 0.098 and 0.223 for Cienega Creek - Sharp Spring comparison to 0.440 and 0.712 for the Bylas Spring - Monkey Spring comparison (Table 5). For all pairs of populations, the FST values were statistically significant. Using the genetic distance values from Table 5 for the microsatellite loci, Figure 2 is a phylogenetic tree using the UPGMA approach (Kumar et al. 1993). Monkey Spring is the most divergent, consistent with that seen from calculating the probability of a unique allele in Table 4. Cienega Creek and Sharp Spring cluster together, with Bylas Spring also closely related. Using a neighbor-joining approach to obtain a phylogenetic tree (not shown), Monkey Spring is again quite divergent but Bylas Spring and Cienega Creek first cluster together, with Sharp Spring being closely related. Comparison of MHC and Microsatellite Variation in the Same Individuals We examined the association of microsatellite and MHC variation in the same individuals in several ways. First, we calculated the correlation coefficient of individual heterozygosity for each locus pair in each population in which they were variable (nine different combinations) and none was significant. The mean value of these correlations was -0.018 and ranged from -0.15 to 0.16. Second, we calculated the extent of linkage disequilibrium between the microsatellite loci and the MHC locus in the only three combinations for which the calculation could be done. All were in the Cienega Creek, and none showed a statistically significant association. Finally, we calculated the variance of individual heterozygosity including the MHC locus and the five polymorphic microsatellite loci. The probability that the variance was greater than that expected at random was similar to that for microsatellite loci alone, with an average probability of 0.39. Overall then, there appears to be no association between variation among individuals within a population for microsatellite loci and the MHC locus. 7 Discussion Gila topminnows have become one of the most thoroughly examined endangered species genetically, with molecular genetic studies of allozymes, mtDNA, a MHC locus, and microsatellite loci and research on traits related to fitness by Quattro & Vrijenhoek (1989), Sheffer et al. (1997a), Sheffer et al. (1997b), and Cardwell et al. (1998). Along with the extensive information from earlier studies on the natural history and ecology of the species and the factors influencing its decline (see Meffe et al. 1983 and references therein), it appears possible to understand relationships between present-day populations and make well-educated recommendations for future conservation efforts. Comparison of Data from Different Molecular Markers New, within-population genetic data from the microsatellite loci (expected heterozygosity and observed number of alleles) are summarized in Table 6 along with the data from previous studies (allozymes from Meffe & Vrijenhoek, 1988; mtDNA from Quattro et al. 1996; and MHC from Hedrick & Parker, 1998). These data are generally consistent over types of loci, allowing for sampling error over loci, thus making several generalizations possible. However, there are several observations that merit discussion. First, Bylas Spring has the lowest overall variation (the only polymorphic locus for this sample was Pooc-G49) and had the lowest variation for both microsatellite loci and the MHC locus. The other three populations all have substantial polymorphism, with Sharp Spring somewhat higher for both measures of genetic variation. Vrijenhoek et al. (1985), after their allozyme survey, suggested that Sharp Spring as the most genetically variable of the four sites, and the other three had low variation. Sharp Spring still appears 8 the most genetically variable of the four sites when examining all types of loci given in Table 6, but both Cienega Creek and Monkey Spring have substantial overall variation. Second, genetic distance is substantial for all pairs of populations for microsatellite loci, with Bylas Spring-Monkey Spring the largest (Table 5). This comparison did not share alleles at the MHC locus, making the MHC genetic distance infinite. Vrijenhoek et al. (1985) found the different sites were similar for allozyme loci. In fact, Bylas and Monkey Springs shared alleles (and were monomorphic for them) at all 25 allozyme loci examined, making allozyme genetic distance zero between these populations. Sharing of alleles for the allozyme loci is surprising because Bylas and Monkey Springs now are the most physically isolated from each other. The lack of any differences may reflect common ancestry in the distant past and a low rate of evolution for these loci, along with low power to determine the level of differentiation because of generally low variation at allozyme loci. Third, variation in mtDNA region examined by Quattro et al. (1996) shows far less variation within and between populations than either microsatellite loci or the MHC locus. It is thought generally that mtDNA evolves rapidly but this does not appear the case for a number of fishes (e.g. Moritz et al. 1987). In addition, Quattro et al. (1996) used six-cutter restriction endonucleases, so that their results were of relatively low resolution, probably not recognizing differentiation that occurred in the last several hundred thousand years (Wilson et al. 1985). Finally, there is evidence that genes in the MHC are under balancing selection (Hedrick 1994), most likely from parasite of pathogen resistance (Hedrick & Kim 1997). If so, the pattern of variation within and between populations may reflect this phenomenon, given reasonably strong balancing selection. The pattern of variation within individuals for the MHC locus, however, is consistent with Hardy-Weinberg expectations for the samples from Cienega Creek and Monkey Spring (Bylas Spring is not variable) and shows a deficiency of heterozygotes for Sharp Spring, the opposite of the expectation 9 for balancing selection (possibly due to the presence of a null allele, Hedrick & Parker 1998). Between-population differentiation for the MHC locus is significant and appears to be as large or larger than for the polymorphic microsatellite loci (see discussion of measures of differentiation in Hedrick & Parker 1998). While it is possible that the MHC differentiation may have resulted from divergent selection among sites, the amount of differentiation is not unlike the pattern observed for microsatellite loci Pooc-C15 and Pooc-LL53, suggesting the pattern of MHC differentiation may be primarily influenced by non-selective factors (see below). In fact, the pattern of variation over samples for microsatellite locus Pooc-LL53 is nearly identical to that observed for the MHC locus, i.e., both are monomorphic in Bylas Spring, share an allele in high frequency between Bylas Spring and Cienega Creek, are nearly fixed for a third allele in Monkey Spring, and Sharp Spring has several different alleles but shares one with Monkey Spring. Because the measures of association within individuals for Pooc-LL53 and the MHC locus were non-significant, it is likely that these nearly identical patterns of genetic variation for these two loci are similar reflections of population history and are not the result of linkage. When comparing the patterns of microsatellite and MHC variation over samples in bighorn sheep, Boyce et al. (1997) concluded that the greater similarity of MHC variation as compared to the microsatellite variants over samples provided support that MHC had a similar selective history over the different populations. Such a pattern is not obvious in the Gila topminnow data. Explanation of the Observed Variation Within and Between Populations There are number of non-selective factors that may play a role in determining variation within and between populations. Table 7 lists primary ones that may have been important in the four populations and gives an estimate of the relative effect for each factor. 10 Probably the most important factor determining lower variation observed in Bylas Spring is the low effective population size. This is both a function of the small habitat, the amount of water flowing from the spring (actually the middle spring or S-II) averages only a few liters per minute which can support at most a few hundred adults, and the recurring bottlenecks observed since the site was discovered (Marsh & Minckley 1990; Minckley et al. 1991). In fact, during one of these bottleneck periods in 1990, Gila topminnows became extinct at Bylas Spring (our stock was derived from a refugium population). In addition, because its location is elevated from the Gila River, it may be that there were only a few founders to originally colonize the site and that immigration has since been unlikely. Monkey Spring appears to have been isolated from a tributary of Sonoita Creek through formation of a natural, 10-meter high travertine dam perhaps 10,000 years bp (Minckley et al. 1991). In other words, there may have been little or no immigration into Monkey Spring for more than 20,000 generations, assuming that there is two or more generations per year in the warm, constant-temperature habitat. Although the size of the site is somewhat reduced from its time of discovery by water development, population size is substantial and varies seasonally less than other sites because of its constant temperature and flow, and absence of floods. Long-term isolation with little or no immigration provides an explanation for the high probability of unique microsatellite alleles observed in Monkey Spring. Both Sharp Spring and Cienega Creek have large amounts of variation shared in part with other sites. Because the population sizes have been large historically and there was substantial connectedness between these and other sites, such patterns of variation are not unexpected. Sharp Spring has several MHC alleles that differ substantially from those at other sites (Hedrick & Parker 1998) and has a large number of alleles at microsatellite locus Pooc-C15, suggesting Sharp Spring (and the upper Santa Cruz River that it now represents) may have had a very large population size in the recent past (also 11 indicated by museum collections) and may have been somewhat isolated from other populations. Evolutionarily Significant Units in the Gila Topminnow The molecular genetic studies of microsatellite and MHC variation generally support the importance of all four populations as significant ESUs. Monkey Spring, which has the most unique alleles and is quite genetically divergent, and Sharp Spring, which has the most genetic variation, are the most strongly supported as separate ESUs by these data. In addition, there are a number of other non-genetic attributes that support their significance (Table 8). First, Bylas Spring is the only remaining population representing the mainstem Gila River stock (see Marsh & Minckley 1990; Minckley et al. 1991), isolated from others by at least 580 km by stream channel, much of which now is dry except during flooding. The other populations are also isolated, with 250 km separating Cienega Creek and Monkey Spring and 101 km separating Monkey Spring and Sharp Spring. There are extensive dry reaches between all sites except during floods. In other words, even though there may have been ancestral exchange between these sites though intermediate populations, it is unlikely that there has been exchange in recent decades and natural exchange is unlikely in the future. Monkey Spring, besides appearing to be the most distinctive population based on the microsatellite data, also was occupied by a now-extinct species of pupfish, genus Cyprinodon (Minckley 1973; Minckley et al. 1991), and a now extinct, morphologically distinct form of Gila intermedia, the Gila chub (Rinne 1976; DeMaris 1986). Monkey Spring topminnows appear to have relatively low fecundity in nature (e.g. Constanz 1979) although this difference disappeared under laboratory conditions (Sheffer et al. 1997a). On the other hand, Cardwell et al. (1998) found that in a common laboratory 12 environment, male development time was approximately 50% longer in Monkey Spring topminnows than for males from any of the other three sites. Because Gila topminnows are members of a tropical genus, perhaps delayed development is a relict trait retained (or expressed) because of the warm, constant temperature at Monkey Spring. Perhaps, other populations evolved faster male development because of higher male-male competition for mates in seasonally variable environments. Finally, neither Cienega Creek nor Monkey Spring has been infested with central mosquito fish while Bylas and Sharp Spring have been impacted by these non-native fish. Presence of mosquito fish not only introduces a competitor and predator but potentially, non-native parasites and pathogens. The suite of selection pressures experienced by the populations with or without mosquito fish may be substantial and itself of evolutionary significance. Moritz (1994) suggested that genetic criteria for evolutionarily significant units is that they show phylogeographic differentiation for mtDNA variants and significant divergence of allele frequencies at nuclear loci. The four populations examined here show significant divergence both at the microsatellite loci and the MHC locus. Although no variation was found for mtDNA by Quattro et al. (1996) and there was little phylogeographic differentiation at the MHC locus, there is some indication of the size of the microsatellite alleles sorting out geographically. For example, at Pooc-LL53 the three largest alleles are found in the Sharp Spring sample and for Pooc-C15 in Monkey Spring there is cluster of seven alleles with sizes from 204 to 224 while Sharp Spring has three small alleles and nine alleles with sizes between 226 and 240. Let us put these conclusions in perspective. After a laboratory study of fitness correlates by Vrijenhoek et al. (1985) indicated Sharp Spring fish had higher fitness than Monkey Spring fish, they recommended "the Sharp Spring stock currently offers the best choice for stocking the Gila River system." As a result, the U. S. Fish and Wildlife Service stopped using Monkey Spring fish for reintroductions and has since used only 13 Sharp Spring stock. Genetic results presented here, those of Sheffer et al. (1997a), Hedrick & Parker (1998), and Cardwell et al (1998) are, on the other hand, consistent with the recommendations by Simons et al. (1989) that "at least one representative lineage is preserved from each of the four geographic areas in Arizona." These genetic results support the suggestion in the draft recovery plan (USFWS 1993) that reintroductions be "from the hydrographically nearest natural population." On a cautionary note, the new, highly variable markers used here are considered indicative of recent, non-selective factors. It is therefore possible that significant differences between populations may be observed that are not correlated with variation in adaptively important traits (Hedrick 1996; Hedrick & Savolainen 1996). However, the size of the genetic differences observed in these topminnow populations for both microsatellite loci and the MHC locus suggest that their isolation has been long enough term to reasonably suggest that the populations are on independent evolutionary trajectories, i.e., that they are significantly different evolutionarily significant units. 14 Acknowledgments We appreciate the knowledge, inspiration, and comments on this manuscript by W. L. Minckley and the statistical assistance of Steven Kalinowski. This research was supported in part through funding from the Arizona Game and Fish Department Heritage Fund. Disclaimer The findings, opinions, and recommendations in this report are those of the investigators who have recieved partial or full funding from the Arizona Game and Fish Department Heritage Fund. The findings, opinions, and recommendations do not necessarily reflect those of the Arizona Game and Fish Commission or the Department, or necessarily represent official Department policy or management practice. For further information, please contact the Arizona Game and Fish Department. 15 Literature Cited Ashley, M. V., and B. D. Dow. 1994. The use of microsatellite analysis in population biology: background, methods, and potential applications. Pages 185-201 in B. Sherwater, B. Streit, G. P. Wagner, and R. DeSalle (eds.). Molecular ecology and evolution: approaches and applications. Birkhauser, Basel, Switzerland. Avise, J. C., and J. L Hamrick (editors). 1996. Conservation genetics: case histories from nature. Chapman and Hall, New York. Boyce, W. M., P. W. Hedrick, N. E. Muggli-Cockett, S. Kalinowski, M. C. Penedo, and R. R. Ramey. 1997. Genetic variation of major histocompatibility complex and microsatellite loci: a comparison in bighorn sheep. Genetics 145:421-433. Brown, A. H. D., M. F. Feldman, and E. Nevo. 1980. Multilocus structure of natural populations of Hordeum spontaneum. Genetics 96:523-536. Cardwell, T. N., R. J. Sheffer, and P. W. Hedrick. 1998. Differences in male development time among populations of the endangered Gila topminnow. Journal of Heredity (submitted). Caughley, G., and A. Gunn. 1996. Conservation biology in theory and practice. Blackwell Science. Cambridge, Massachusetts. Constanz, G. D. 1979. Life history patterns of a livebearing fish in contrasting environments. Oecologia 40:189-201. DeMarais, B. D. 1986. Morphological variation in Gila (Pisces: Cyprinidae) and geologic history: Lower Colorado river basin. M.S. Thesis. Arizona State University, Tempe, Arizona. Hedrick, P. W. 1971. A new approach to measuring genetic similarity. Evolution 25:276-280. Hedrick, P. W. 1994. Evolutionary genetics of the major histocompatibility complex. American Naturalist 143:945-964 . 16 Hedrick, P. W. 1996. Conservation genetics and molecular techniques: a perspective. Pages 459-477 in T. B. Smith and R. K. Wayne, editors. Molecular genetic approaches in conservation. Oxford University Press, New York. Hedrick, P. W., and T. Kim. 1997. Genetics of complex polymorphisms: parasites and maintenance of MHC variation. in R. S. Singh and C. K. Krimbas, editors. Evolutionary genetics from molecules to morphology. Cambridge University Press, New York. Hedrick, P. W., and K. M. Parker. 1998. MHC variation in the endangered Gila topminnow. Evolution (in press). Hedrick, P. W., and O. Savolainen. 1996. Molecular and adaptive variation: a perspective for endangered plants. Pages 92-102. in J. Machinski, D. H. Hammond, and L. Holter, Editors. Southwestern rare and endangered plants: proceedings of the second conference. General technical report RM-GTR-283. U. S. Department of Agriculture, Forest Service, Fort Collins, Colorado. Kumar, S., K. Tamura, and M. Nei. 1993. MEGA: molecular evolutionary genetics analysis, version 1.01. Pennsylvania State University, University Park, Pennsylvania. Marsh, P. C., and W. L. Minckley. 1990. Management of endangered Sonoran topminnow at Bylas Springs, Arizona: description, critique, and recommendations. Great Basin Naturalist 50:265-272. Meffe, G. K., D. A. Hendrickson, W. L. Minckley, and J. N. Rinne. 1983. Factors resulting in decline of the endangered Sonoran topminnow, Poeciliopsis occidentalis (Atheriniformes: Poeciliidae), in the United States. Biological Conservation 25:135-159. Meffe, G. K., and R. C. Vrijenhoek. 1988. Conservation genetics in the management of desert fishes. Conservation Biology 2:157-169. 17 Minckley, W. L. 1973. Fishes of Arizona. Arizona Game and Fish Department. Phoenix Arizona. Minckley, W. L., G. K. Meffe, and D. L. Soltz. 1991. Conservation and management of short-lived fishes: the cyprinodontoids. Pages 247-282. in W. L. Minckley and J. E. Deacon, editors. Battle against extinction: native fish management in the American west. University of Arizona Press, Tucson, AZ. Moritz, C. 1994. Defining "evolutionary significant units" for conservation. Trends in Ecology and Evolution 9:373-375. Moritz, C., T. E. Dowling, and W. M. Brown. 1987. Evolution of animal mitochondrial DNA: relevance for population biology and systematics. Annual Reviews of Ecology and Systematics 18:269-292. Nei, M. 1972. Genetic distance between populations. American Naturalist 106:283-292 . Nei, M. 1987. Molecular evolutionary genetics. Columbia University Press, New York. Parker, K. M., K. Hughes, T. J. Kim, and P. W. Hedrick. 1998. Isolation and characterization of microsatellite loci from the Gila topminnow (Poeciliopsis o. occidentalis) and examination in guppies (Poecilia reticulata). Molecular Ecology (in press). Quattro, J. M., P. L. Leberg, M. E. Douglas, and R. C. Vrijenhoek. 1996. Molecular evidence for a unique evolutionary lineage of endangered Sonoran desert fish (genus Poeciliopsis). Conservation Biology 10:128-135. Quattro, J. M., and R. C. Vrijenhoek. 1989. Fitness differences among remnant populations of the endangered Sonoran topminnow. Science 245:976-978. Rinne, J. N. 1976. Cyprinid fishes of the genus Gila from the lower Colorado River basin. Wassman Journal of Biology 34: 65-107. Sheffer, R. J., P. W. Hedrick, W. L. Minckley, and A. L. Velasco. 1997a. Fitness in the endangered Gila topminnow. Conservation Biology 11:162-171. 18 Sheffer, R. J., P. W. Hedrick and C. Shirley. 1997b. No bilateral asymmetry in wild-caught, endangered Gila topminnows. Heredity (in press). Simons, L. H., D. A. Hendrickson, and D. Papoulias. 1989. Recovery of the Gila topminnow: a success story? Conservation Biology 3:11-15. Slatkin, M., and L. Excoffier. 1996. Testing for linkage disequilibrium in genotypic data using the Expectation-Maximinization algorithm. Heredity 76:377-383. Smith, T. B., and R. K Wayne (editors). Molecular genetic approaches in conservation. Oxford University Press, New York. Takezaki, N, and M. Nei. 1996. Genetic distances and reconstruction of phylogenetic trees from microsatellite DNA. Genetics 144:389-399. U. S. Fish and Wildlife Service. 1993. Draft Gila topminnow recovery plan. U. S. Fish and Wildlife Service. Albuquerque, New Mexico. Vrijenhoek, R. C., M. E. Douglas, and G. K. Meffe. 1985. Conservation genetics of endangered fish populations in Arizona. Science 229:400-402 . Waples, R. S. 1995. Evolutionary significant units and the conservation of biological diversity under the Endangered Species Act. American Fisheries Society Symposium 17:8-27. Wilson, A. C., R. L. Cann, S. Carr, M. George, U. B. Gyllensten, et al. 1985. Mitochondrial DNA and two perspectives on evolutionary genetics. Biological Journal of the Linnean Society 26:375-400. 19 Table 1. Allelic frequencies for the five polymorphic microsatellite loci in the four populations. Frequencies in boldface are for "diagnostic" alleles, those only in a single population (except those in low frequency) or those in substantially higher in a given sample than all others. Locus Allele Bylas Spring Cienega Creek Monkey Spring Sharp Spring Average Pooc-G49 149 --- --- --- 0.038 0.009 159 0.250 1.000 1.000 0.962 0.803 161 0.750 --- --- --- 0.188 Pooc-6-10 287 --- --- 1.000 --- 0.250 297 1.000 1.000 --- 1.000 0.750 Pooc-C15 202 --- --- --- 0.050 0.012 204 --- --- 0.025 0.200 0.056 208 --- --- 0.088 0.012 0.025 210 --- --- 0.012 --- 0.003 214 --- --- 0.612 --- 0.153 216 --- --- 0.225 --- 0.056 218 --- 0.012 --- --- 0.003 222 --- --- 0.012 --- 0.003 224 --- --- 0.012 --- 0.003 226 --- --- --- 0.012 0.003 228 --- --- --- 0.100 0.025 230 --- --- --- 0.012 0.003 232 --- 0.362 0.012 0.025 0.100 234 --- --- --- 0.100 0.025 236 --- --- --- 0.338 0.084 238 --- 0.612 --- 0.112 0.181 240 1.000 0.012 --- 0.025 0.259 246 --- --- --- 0.012 0.003 Pooc-OO56 143 --- 0.200 --- --- 0.050 145 1.000 0.800 0.762 1.000 0.890 149 --- --- 0.238 --- 0.060 Pooc-LL53 142 --- --- 0.988 --- 0.247 144 --- 0.488 --- --- 0.122 146 1.000 0.512 --- --- 0.378 150 --- --- 0.012 0.425 0.109 154 --- --- --- 0.550 0.138 164 --- --- --- 0.025 0.006 20 Table 2. The observed heterozygosity (HO), expected heterozygosity (HE), and observed number of alleles (n) for the five polymorphic microsatellite loci in the four populations. Locus Measure Bylas Spring Cienega Creek Monkey Spring Sharp Spring Average Pooc-G49 HO 0.350 0.000 0.000 0.075 0.106 HE 0.375 0.000 0.000 0.075 0.112 n 2 1 1 2 1.5 Pooc-6-10 HO 0.000 0.000 0.000 0.000 0.000 HE 0.000 0.000 0.000 0.000 0.000 n 1 1 1 1 1.0 Pooc-C15 HO 0.000 0.600 0.600 0.800 0.500 HE 0.000 0.493 0.625 0.812 0.476 n 1 4 8 12 6.2 Pooc-OO56 HO 0.000 0.300 0.325 0.000 0.171 HE 0.000 0.320 0.362 0.000 0.171 n 1 2 2 1 1.5 Pooc-LL53 HO 0.000 0.625 0.025 0.500 0.288 HE 0.000 0.500 0.025 0.516 0.260 n 1 2 2 3 2.0 Average HO 0.070 0.305 0.190 0.275 0.210 HE 0.075 0.263 0.202 0.281 0.205 n 1.2 2.0 2.8 3.8 2.45 21 Table 3. The number of heterozygous loci observed per individual for the five microsatellite loci and the mean over loci (H) and the mean observed and expected variance of heterozygosity per individual, OsH(2 ) and EsH(2 ), respectively, and the probability (P) that the observed variance is greater than that expected at random. Heterozygosity Population 0 1 2 3 4 H OsH()2 EsH()2 P Bylas Spring 26 14 - - - 0.35 0.23 0.23 0.41 Cienega Creek 4 17 13 6 - 1.52 0.77 0.70 0.21 Monkey Spring 12 18 10 - - 0.95 0.56 0.50 0.20 Sharp Spring 3 21 14 2 - 1.38 0.50 0.49 0.42 22 Table 4. Probability of a unique allele for a given population at the top of the table when compared to the population at the side, averaged over the five polymorphic microsatellite loci. Bylas Spring Cienega Creek Monkey Spring Sharp Spring Average Bylas Spring - 0.335 0.600 0.522 0.486 Cienega Creek 0.150 - 0.597 0.370 0.372 Monkey Spring 0.550 0.567 - 0.470 0.529 Sharp Spring 0.350 0.242 0.572 - 0.388 Average 0.350 0.381 0.590 0.454 23 Table 5. The genetic distance (D) and FST between the population pairs for the four types of loci where the number of loci is given in parentheses. Comparison Microsatellites (10) Allozymes (25) mtDNA (1) MHC (1) D FST D FST D FST D FST Bylas-Cienega 0.188 0.479 0.080 1.000 0.0 0.0 0.35 0.33 Bylas-Monkey 0.440 0.712 0.000 0.000 0.0 0.0 ∞ 0.87 Bylas-Sharp 0.174 0.506 0.016 0.300 0.0 0.0 ∞ 0.46 Cienega-Monkey 0.294 0.494 0.080 1.000 0.0 0.0 ∞ 0.51 Cienega-Sharp 0.098 0.223 0.030 0.431 0.0 0.0 ∞ 0.24 Monkey-Sharp 0.271 0.466 0.020 0.300 0.0 0.0 2.35 0.37 Overall FST - 0.591 - 0.750 - 0.0 - 0.55 24 Table 6. The expected heterozygosity (HE) and observed number of alleles (n) for the different types of loci and the four populations where the number of loci in given in parentheses. Loci (number) Bylas Spring Cienega Creek Monkey Spring Sharp Spring HE n HE n HE n HE n Microsatellite (10) 0.038 1.10 0.132 1.50 0.101 1.90 0.140 2.40 Allozyme (25) 0.000 1.00 0.000 1.00 0.000 1.00 0.037 1.08 mtDNA (1) 0.000 1.00 0.000 1.00 0.000 1.00 0.000 1.00 MHC (1) 0.000 1.00 0.500 2.00 0.141 3.00 0.729 5.00 25 Table 7. Characteristics that are potentially relevant to understanding the amount of genetic variation within and the differentiation between the populations. Characteristic Bylas Spring Cienega Creek Monkey Spring Sharp Spring Population size (number of adults) Small (<100) Large (>>10,000) Intermediate (<10,000) Large (>>10,000) Age of population (years before present) Recent (<100) Old (>>10,000) Old (~10,000) Old (>>10,000) Relative historic immigration from other populations Low High Low High Number of founders Very low High Low High 26 Table 8. Characteristics of the four populations divided into those of the physical habitat, the biota, and the life history of Gila topminnows. Characteristic Bylas Spring Cienega Creek Monkey Spring Sharp Spring Physical Location Mainstem Gila Tributary to Santa Cruz Above natural dam, tributary to Santa Cruz Headwaters of Santa Cruz Temperature* Variable Variable Constant, warm spring Variable Habitat size Very small Large Small Large Flooding frequency Very low High Low Intermediate Biota Endemic vertebrates None None Pupfish, distinct chub None Mosquito fish Present Absent Absent Present Food supply Seasonally variable Seasonally variable Constant Seasonally variable Life history Male development Fast Fast Slow Fast Fecundity* High High Low High Female reproduction Seasonal Seasonal Year-round Seasonal * All these habitats are thermally ameliorated, either by low-volume spring inflow or groundwater exchange. Under laboratory conditions (Cardwell et al. 1998). *In the field. 27 Figure Legends Figure 1. Map showing the location of the four samples used. Figure 2. The phylogenetic tree for the four populations using the UPGMA approach based on the allelic frequencies at the five polymorphic microsatellite loci. 28 29 Appendix Table 1. Locus Allele Bylas Spring Cienega Creek Monkey Spring Sharp Spring Pgd a 1.00 - 1.00 0.45 f - 1.00 - 0.55 Est-4 a 1.00 - 1.00 0.70 b - 1.00 - 0.30 MHC - DRB Pooc-1 1.000 0.500 - - Pooc-2 - - - 0.325 Pooc-3 - - - 0.175 Pooc-4 - - 0.025 - Pooc-5 - 0.500 - - Pooc-6 - - 0.925 0.050 Pooc-7 - - 0.050 - Pooc-8 - - - 0.350 Pooc-9 - - - 0.100

you wish to report:

Your comment:

Your Name:

Submit

Cancel

...

Find results with:

Searching collections:

Home Molecular variation and evolutionarily significant units in the endangered Gila topminnow

Molecular variation and evolutionarily significant units in the endangered Gila topminnow

Description