Hierarchical population-genetic structure.
Both direct and indirect measures of gene flow indicate that dispersal is limited in many plant species and leads to spatial genetic differentiation (Ehrlich and Raven 1969; Levin and Kerster 1974; Brown 1979; Handel 1983; Crawford 1984; Loveless and Hamrick 1984; Slatkin 1985; Hamrick and Loveless 1986; Hamrick and Godt 1990; Heywood 1991). Comparisons of allozyme variation among many unrelated plant taxa have shown that the amount of gene flow, and hence the extent of genetic differentiation among population subdivisions, is broadly correlated with several ecological and life-history characteristics, particularly breeding system, life form, and mode of seed dispersal. However, seed dispersal appears to exert a much weaker influence than the breeding system on population genetic structure. Moreover, mode of seed dispersal has not proved a strong predictor of the relative degree of population differentiation compared among plant species. In particular, species with animal-mediated seed dispersal, via ingestion by or external attachment to birds and mammals, tend to have high levels of genetic differentiation similar to those with gravity-dispersed seeds. Wind-dispersed species tend to exhibit the least genetic differentiation among population subdivisions (Loveless and Hamrick 1984; Hamrick and Godt 1990).
Broad taxonomic comparisons such as those cited above may present a biased picture of the effects of seed dispersal on population genetic structure for at least two reasons. First, the effects of correlated life-history traits may obscure the influence of seed dispersal on genetic structure. For instance, many species with wind-dispersed seeds also tend to be wind-pollinated, outcrossing trees, whereas many species with seed dispersal by gravity or attachment to animals are self-fertilizing herbs (Loveless and Hamrick 1984; Uma Shaanker et al. 1990). Second, comparisons across taxa often do not differentiate among the scales at which genetic structure is measured in different studies, such that their results may not be directly comparable. To examine the effects of seed dispersal on genetic structure, independent of variation in other ecological and life-history characters, it is necessary to study patterns of variation within and among populations of similar species (Loveless and Hamrick 1984). Only a few such comparative studies have been made (Levin 1978; Schoen 1982; Sun and Ganders 1988; Van Dijk et al. 1988; Holtsford and Ellstrand 1989; Loveless and Hamrick 1988; Karron et al. 1988; Karron 1989; Pleasants and Wendel 1989), and none explicitly have examined seed dispersal.
To investigate the effects of seed dispersal on population genetic structure, we compare three ecologically similar species of forest herbs in the family Apiaceae whose fruits vary widely in morphological adaptations for dispersal. The fruits of Sanicula odorata and Osrnorhiza claytonii have appendages that should facilitate their attachment to animals (ectozoochory), whereas Cryptotaenia canadensis fruits are smooth and do not appear to be dispersed by animals. Based on differences in fruit morphology, we predict that seed dispersal should be the most restricted in Cryptotaenia and least restricted in Sanicula. These differences in seed-dispersal ability are expected to result in the most genetic differentiation among population subdivisions in Cryptotaenia and the least in Sanicula. To test this prediction, we use a hierarchical sampling design to examine patterns of genetic differentiation at several spatial scales. We address the following questions. (1) Do differences in fruit morphology among species correlate with their adherence to mammal fur, an analogue of dispersal ability? (2) Do species with more restricted seed dispersal have greater genetic differentiation among population subdivisions? (3) How does the apportionment of genetic variation within each species differ across spatial scales?
MATERIALS AND METHODS
Cryptotaenia canadensis (L.) DC., Osmorhiza claytonii (Michx.) C. B. Clarke, and Sanicula odorata (Raf.) Pryer and Phillippe (formerly Sanicula gregaria Bickn.; Pryer and Phillippe 1989) (Apiaceae) are three common forest understory herbs that grow sympatrically in southern Wisconsin, where they have probably cooccurred since the last glacial retreat ca. 10,000 yr ago. They are found in a wide variety of forest habitats, but tend to occur most frequently in disturbances such as treefall gaps, forest edges, and streamsides. Their distribution is thus often patchy (Curtis 1959; Fassett 1976).
All three species are polycarpic perennials with little or no vegetative spread. Cryptotaenia reproduces from seed and by (usually single) vegetative offshoots from monocarpic ramets (Baskin and Baskin 1988 pers. obs.). Osmorhiza reproduces only by seed (Baskin and Baskin 1984, 1991), whereas Sanicula appears capable of limited vegetative spread as evidenced by occasional multiple flowering ramets from the same short rootstock (Shan and Constance 1951 pers. obs.).
Several lines of evidence suggest that pollen dispersal distances are similarly restricted in each species. They each have protogynous flowers and inflorescences that are and romonoecious (Constance and Shan 1948; Shan and Constance 1951; Hiroe and Constance 1958; Gleason and Cronquist 1963; Baskin and Baskin 1988). All are self-fertile and capable of mechanical self-pollination in the absence of insect visitation, as determined using pollinator exclosures (Williams 1991). Like other members of the Apiaceae, these species have relatively unspecialized flowers visited by a suite of small, generalist pollinators (Knuth 1908; Bell 1971; Lindsey 1984). Common flower visitors to all species in this study were small solitary bees (Halictidae, Andrenidae), bee flies (Syrphidae), and beetles (Mordellidae). These visitors have been characterized as ineffective long-distance pollinators in an earlier study of related species of Apiaceae (Lindsey 1984). Limited observations suggest that patterns of pollinator visitation do not differ greatly among these plant species. Most pollinator movements were among flowers on a single plant or two adjacent plants. Estimates of mating system parameters are available only for Cryptotaenia, where outcrossing rates vary from 5%-40% among populations (C. F. Williams unpubl. data).
The fruits of these species vary widely in morphology and apparent adaptations for dispersal. The fruits of Sanicula and Osmorhiza have appendages (numerous recurved hooks and short spines, respectively) that should facilitate attachment to animals, whereas Cryptotaenia fruits are smooth and apparently gravity dispersed. Each hermaphroditic flower can produce two single-seeded, dry fruitlets comprising the schizocarp. On average, Cryptotaenia produces more seeds per plant ([Mathematical Expression Omitted], SD = 27.3, N = 21) than Osmorhiza ([Mathematical Expression Omitted] SD = 22.5, N = 96) or Sanicula ([Mathematical Expression Omitted], SD = 18.0, N = 37) (C. F. Williams unpubl. data). None of these species' seeds appear to be eaten by potential vertebrate dispersers, although herbivory and invertebrate seed predation are common (Williams 1991).
Study Sites and Sampling Design
We sampled 18 different sites (= populations) in Wisconsin and northeast Iowa. Thirteen sites are in the Baraboo Hills of Sauk County, Wisconsin, and five in outlying areas. All three species are found at most, but not all sites (Cryptotaenia, 17 sites; Osmorhiza, 16 sites; Sanicula, 14 sites). The vegetation at most sites is relatively undisturbed, southern mesic or dry-mesic forest (Curtis 1959). The Baraboo Hills and Kickapoo River sites are part of large, contiguous blocks of forest in which direct interpopulation gene flow can potentially occur.
The sampling design includes four hierarchical levels: geographic regions (65-375 km between population subdivisions), drainages within regions (2.5-20 km apart), populations within drainages (0.5-7.0 km apart), and subpopulations within populations (5-100 m apart). The four outlying regions are represented by one or two populations each. In Sauk County, four different drainages containing two to four populations each were sampled. Populations are further subdivided into varying numbers of sub-populations as follows. In most populations, groups of individuals were sampled at 5-to 15-m intervals along a 50-m transect. In some very low density populations, plants were sampled as encountered along 100- to 200-m transects, or from scattered, distinct groups. Subpopulation boundaries coincide with those of naturally occurring groups of plants except in dense, continuous populations in which five to six plants were sampled every 5 m, and two such points are combined to form a subpopulation sample. Subpopulations encompass areas ranging from approximately 50 [m.sup.2] to 250 [m.sup.2]. Subpopulation size in these transect sampled sites ranges from 3-28 ([Mathematical Expression Omitted]; Cryptotaenia, 13.9 [+ or -]6.6; Osmorhiza, 12.8 [+ or -] 4.0; Sanicula, 10.7 [+ or -] 4.6) plants. At two sites used for a study of microspatial genetic structure, 158-435 individuals of each species were sampled from mapped locations on a grid. The sampled area covers 30 m x 60 m at site BHG (70 m x 80 m for Cryptotaenia) and 10 m x 20 m at site HDB. At BHG, the sampled area is divided into 15 m x 15 m quadrats and those quadrats containing plants are used as subpopulations. At HDB, the grid was divided into 5 m x 10 m quadrats (10 x 10 for Osmorhiza) plus two segments of an adjacent transect sampled earlier. Subpopulation sample size in these mapped populations ranges from 3-195 (Cryptotaenia, 31.2 [+ or -]40.7; Osmorhiza, 25.4 [+ or -] 23.3; Sanicula, 35.5 [+ or -] 46.8) plants.
Leaf material was collected for isozyme analysis at the different sites from 1987 to 1989. Because these are perennial species, sample genotype frequencies within populations should not change greatly over this period. To confirm this expectation, different individuals were sampled from two sites (BHG and HDB) in repeated years. At these two sites, no significant differences in allele or genotype frequencies are observed between years; thus data from individuals and populations sampled in different years are analyzed together without regard to year of collection.
We estimated the relative adherence of different fruits to dispersers by comparing the length of time they remained attached to mammal fur. We used museum skins of deer (Odocoileus virginianus), raccoon (Procyon lotor), and squirrel (Sciurus carolinensis), which represent three common, potential dispersal agents in the study area. Fruits were gently rubbed onto the dorsal surface of the skin with several passes of a wooden tongue depressor. Twenty fruits were initially TABULAR DATA OMITTED applied in each replicate using deer and raccoon skins, and 10 fruits were applied on the squirrel skin. The specimen was then inverted, causing some seeds that did not adhere to the fur to drop off (zero shakes). The number of remaining seeds dislodged after 1, 3, 5, 10, and 20 shakes of the study skin were recorded. Shakes were standardized by holding the specimen by the anterior end and dropping the posterior end approximately 12 inches to a table edge. Ten replicates were made per each fruit and mammal species. Differences between plant species in the proportion of fruits remaining attached to mammal skins were tested with ANCOVA using the number of shakes as the covariate. Cryptotaenia seeds did not adhere at all; therefore, only differences between Osmorhiza and Sanicula were tested. Significant heterogeneity of slopes existed among the three mammal species, such that comparisons between fruits were made for each disperser separately. Differences between plant species adjusted for the number of shakes were tested using the model: (ln) proportion adherence = constant + fruit species (categorical) + number of standard shakes (continuous). Analyses were run using the multivariate general linear hypothesis (MGLH) module of Systat 5.0 (Wilkinson 1989).
Leaf samples ranged from 50 to 500 mg fresh weight, and except for small seedlings, sampling was nondestructive. Immediately after collection, samples were placed in labeled plastic bags and kept on ice or in a 5 [degrees] C refrigerator until processed, usually within 24 h. Electrophoresis was performed using crude leaf homogenate ground with an ice-cold mortar and pestle. The grinding buffer was a slight modification of that described by Wendel and Parks (1982) and Marry et al. (1984). Samples kept on ice after grinding TABULAR DATA OMITTED then stored at - 80 [degrees] C retained good activity for [is greater than or equal to] 2 yr for most enzymes. We prepared 11.2% or 12.0% w/v starch gels using the microwave technique of Marty et al. (1984). Homogenate was absorbed into 3 x 14 mm, Whatman #3 filter paper wicks. Gels were run in a 5 [degrees] C refrigerator at [is less than or equal to] 50 mA and 150 V for 4-6 h.
We used four gel buffer systems to resolve the different enzymes studied in each species: (1) AC, Morpholine-citrate pH 6.7 (modified from Clayton and Tretiak 1972, and Marty et al. 1984); (2) S4, Tris-citrate pH 7.5 (#4 of Soltis et al. 1983); (3) S6, NaOH-boric Acid pH 8.6 electrode/Tris-citrate pH 7.8 gel (#6 of Soltis et al. 1983); and (4) S11, Histidine-citrate pH 7.0 (# 11 of Soltis et al. 1983). Stain recipes are those found in Marty et al. (1984) except LAP (Soltis et al. 1983), and TPI (Stuber et al. 1988). Optimal buffer x enzyme combinations for each species are reported in Williams (1991).
A total of 15 enzyme systems coding for 31 putative loci were scored for Cryptotaenia, 14 systems coding for 24 loci were scored for Osmorhiza, and 16 systems coding for 35 loci were scored for Sanicula (Williams 1991). Twelve polymorphic loci were resolved for Cryptotaenia, 7 for Osmorhiza, and 13 for Sanicula. All polymorphic loci were used for estimating pairwise genetic distances. Only those loci with mean frequencies of the most common allele [is less than] 0.95 overall, or a frequency of [is less than] 0.95 in at least one population, were used for estimating F-statistics (9 loci for Cryptotaenia, 5 for Osmorhiza, 11 for Sanicula; table 2). Complete allele frequency data are reported in Williams (1991).
Analyses of Genetic Structure
We calculated several pairwise genetic distance measures between proportions for all polymorphic loci (Nei 1972, 1978; Rogers 1972; Cavalli-Sforza and Edwards 1967; Edwards 1971) using the BIOSYS-1 program (Swofford and Selander 1981). All measures produced similar groupings of populations using unweighted pair-group arithmetic averaging (UPGMA) clustering (Sneath and Sokal 1973). Results presented here use the geometric (chord) genetic distance of Edwards (1971), because of its large range of observed values and similarity to Wright's [F.sub.st], (Hartl and Clark 1989).
We calculated average pairwise genetic distances between populations for three levels of spatial organization: (1) within drainages, (2) between drainages within geographic regions, and (3) between regions. Statistical comparisons of average genetic distance estimates both within hierarchical levels between species, and between hierarchical levels within species employed boot-strapped confidence intervals (CIs) (Efron 1982) calculated from all pairwise distances at a given level of the hierarchy. Ninety-five percent CIs (one-tailed) were calculated as the central 90% of observations from the distribution of 2500 bootstrapped estimates. Two estimates are regarded as statistically different at the P [is less than] 0.05 significance level if their one-tailed 95% CIs do not overlap.
Hierarchical F-statistics (Wright 1951) were calculated following the methods of Weir and Cockerham (1984) and Weir (1990). This method facilitates the calculation of multilocus estimates and their variances using a jacknife procedure. F-statistics were calculated for genetic variance among individuals in the total sample ([Mathematical Expression Omitted]), inbreeding within subpopulations ([Mathematical Expression Omitted]), and differentiation among subpopulations within populations ([Mathematical Expression Omitted]), among populations within drainages, ([Mathematical Expression Omitted]), and among drainages within the total ([Mathematical Expression Omitted]), as outlined in the Appendix. Negative estimates of [Theta] can occur because the variance components from which F-statistics are calculated are in fact covariances and hence not constrained to be positive. Small negative values may result from bias in the estimator and should be regarded as not significantly different from zero. Negative estimates may also arise because genes are more similar between than within population subdivisions (a negative intraclass correlation) (Weir 1990). As with average genetic distances, statistical differences in multilocus estimates of F-statistics were tested as nonoverlap of one-tailed 95% CIs (two-tailed 90% CIs for [Mathematical Expression Omitted] and [Mathematical Expression Omitted]). Each bootstrap estimate was generated by drawing with replacement from the set of variance components used for each single locus estimate. The central 90% of the distribution of 1500 multilocus bootstrap estimates was used to define the upperland lower bounds of the CIs. Hierarchical F-statistics were estimated using the complete data set in which populations from outlying regions are treated as separate drainages (i.e., there were nine total drainages, four from Sauk County and five from outlying regions). F-statistics were also estimated for just the 13 populations in four drainages from Sauk County to be used to estimate rates of gene flow, Nm.
Estimates of effective rates of gene flow among population subdivisions, Nm, were calculated using the formula,
Nm = (1/[Theta] - 1)/4
(Wright 1931). The effects of differences in sample size among population subdivisions on estimates of [Theta] are accounted for by the method of Weir and Cockerham (1984), and hence Nm can be estimated directly from [Theta] (Slatkin 1985). Nm was estimated for each of the three levels of population subdivision used in the F-statistics analysis from [Mathematical Expression Omitted], [Mathematical Expression Omitted], and [Mathematical Expression Omitted], respectively. Standard errors of Nm were estimated using a jacknife procedure. Confidence intervals around each Nm estimate were calculated by bootstrapping 1500 times over the variance components of single locus estimates of each F-statistic and then calculating Nm from each bootstrapped multilocus estimate of [Theta].
Marked differences exist in adherence of fruits to mammal skins and among the three mammal species tested. Fruits of Sanicula adhere more firmly than those of Osmorhiza when adjusted for the number of "shakes" as the covariate (deer, F = 11.9, P = 0.001; raccoon, F = 99.3, P [is less than] 0.001; squirrel, F = 56.5, P [is less than] 0.001; ANCOVA). The fruits of Cryptotaenia do not adhere at all. The initial rate of seed loss, over the interval 0-1 shakes, is much steeper for Osmorhiza than Sanicula, suggesting that a large proportion of attached Osrnorhiza fruits are lost over a relatively short time or distance in natural situations. In both Osmorhiza and Sanicula, there is a long tail to the distribution of the number of fruits remaining attached indicating the potential for long-distance dispersal.
Within each species, average genetic distance between populations increases from within drainages to between drainages. These differences are significant at P [is less than] 0.05 as indicated by the nonoverlap of the one-tailed 95% confidence intervals around each estimate. Populations from outlying regions are not, on average, significantly more genetically dissimilar than populations between drainages within a region for Cryptotaenia and Sanicula. Genetic distance between populations increases significantly at each level of the spatial hierarchy in Osmorhiza.
Significant differences exist (P [is less than] 0.05) among species in average pairwise genetic distance at each level of the hierarchy as indicated by non-overlap of one-tailed 95% confidence intervals (CIs). Within drainages, Cryptotaenia populations are significantly more differentiated than Sanicula populations, whereas Osmorhiza is intermediate but not significantly different from the other species. Between drainages, Cryptotaenia has significantly larger average genetic distances among populations than either of the other species, which do not differ significantly from one another. Between regions, average genetic distance varies significantly among all species, with Cryptotaenia [is greater than] Osmorhiza [is greater than] Sanicula.
Populations do not form clusters based on pairwise genetic distances as predicted by geographic proximity alone. In general, populations sharing a drainage form fairly well-defined clusters, but outlying populations (outside Sauk County region) are not genetically distinct. Few novel alleles or large shifts in gene frequencies are found in the outlying populations. Also, no strong correlation exists between genetic and geographic distances among the Sauk County populations (Williams 1991).
Considerable variation exists among single locus estimates of each F-statistic. More than half of all estimates fall above or below the bounds of the 95% CIs of the multilocus estimates. However, the standard errors of the multilocus estimates are relatively small. Single locus estimates that vary most widely from the mean are those with low frequencies of the less common allele(s) or with alleles occurring in only one or a few populations (Williams 1991). Such variation in single locus estimates have a relatively small effect on the multilocus estimates of each F-statistic. Jacknife multilocus estimates, sequentially omitting each locus, vary only slightly ([is less than] 0.06 absolute) from the mean estimate in all cases. However, this deviation can amount to up to 50% of the mean for some very small ([is less than] 0.10) estimates of [Theta].
Significant differences appear among species in multilocus estimates of each F-statistic. The species-specific patterns of differentiation among population subdivisions are completely consistent with predictions based on their relative seed-dispersal abilities. At each level of the spatial hierarchy, the rank order of species' estimates are the reverse of predicted dispersal ability, Cryptotaenia [is greater than] Osmorhiza [is greater than] Sanicula. The only exception to this pattern is differentiation among Sauk County populations, where Osmorhiza has a somewhat higher estimate of [[Theta].sub.p] than Cryptotaenia. In both Sauk County and the larger sample of populations, Cryptotaenia has a significantly larger value of [Mathematical Expression Omitted] than Sanicula, whereas Osmorhiza has an intermediate value TABULAR DATA OMITTED of [Mathematical Expression Omitted] with a confidence interval overlapping those of both other species. Although there is no statistically significant variation among species in differentiation at the population level, [Mathematical Expression Omitted], they again occur in the predicted rank order, except in the Sauk County sample as noted above. Cryptotaenia has significantly greater genetic differentiation among drainages, [Mathematical Expression Omitted], than Sanicula. Osmorhiza is again intermediate and has a confidence TABULAR DATA OMITTED interval overlapping those of the other species in the total population sample. Although the rank order among species is the same, differences in estimates of [[Theta].sub.d] are not statistically significant in the Sauk County analysis. Moreover, wide variation exists among species for estimates of [F.sub.it] (Osmorhiza [is greater than] Cryptotaenia [is greater than] Sanicula) and [f.sub.is] (Osmorhiza [is greater than] Cryptotaenia = Sanicula).
The relative degree of genetic differentiation among spatial levels within species shows the expected increase at greater geographic scales. In both Osmorhiza and Sanicula, differentiation increases monotonically from subpopulations to populations to drainages. In Cryptotaenia, there is somewhat greater differentiation among subpopulations within populations than among populations within drainages. However, in no case is this pattern statistically significant at P [is less than] 0.05 except between the levels of populations and drainages in Cryptotaenia in the total population sample.
Differences among the three species' multilocus estimates of effective gene flow, Nm, further support predictions based on their relative seed dispersal abilities. In the sample of 13 Sauk County populations, rates of gene flow decline twofold to threefold over the range of distances sampled. As expected, the highest rates of gene flow are among subpopulations ([is less than] 100 m apart) and the lowest gene flow among drainages (2.5-20 km apart) in each species. Within species, estimates of Nm are not significantly different (P [is less than] 0.05, nonoverlap of 95% CIs) among spatial scales, although estimates differ at P [is less than] 0.10 between populations and drainages in Cryptotaenia, and subpopulations and drainages in Sanicula. The species maintain their predicted rank order, Sanicula [is greater than] Osmorhiza [is greater than] Cryptotaenia, at both the smallest and largest scales. Estimates of Nm differed significantly only at P [is less than] 0.05 between Sanicula and Cryptotaenia at the subpopulation level. Sanicula has the largest rank-order estimate of gene flow at all spatial scales. Standard errors of the estimates are largest in the two animal dispersed species, Sanicula and Osmorhiza, at all spatial scales.
The initial prediction that the morphology of the fruits of these species would be reflected in their dispersal ability was supported by studies of seed adherence. The hooked fruits of Sanicula remain attached to mammal fur longer and hence are potentially dispersed farther than the barbed fruits of Osmorhiza. The smooth fruits of Cryptotaenia do not adhere at all and appear primarily gravity dispersed. Seed and fruit morphology are often used to infer the general mode of dispersal or define dispersal syndromes (van der Pijl 1982). However, morphological adaptations alone cannot be used to predict the dispersal process (Howe and Smallwood 1982). Although the fruits of both Osrnorhiza and Sanicula have obvious morphological structures for attachment to animal dispersers, the process of ectozoochory can be highly stochastic. Many, if not most, ectozoochorus fruits produced are not removed by a dispersal agent but fall near the parent plant (Bullock and Primack 1977; Shmida and Ellner 1983; Sorenson 1986). Thus, even in plants with apparent adaptations for long-distance dispersal, such as Osmorhiza and Sanicula, the distribution of dispersal distances may still be highly leptokurtic and lead to local genetic differentiation (Levin and Kerster 1974; Levin 1984). Differences also may exist among animal species in the "quality" of dispersal they afford. Differences appear in the adherence of fruits to different types of mammal fur. In general, fruits remain better attached to animals with long, thick underhairs and lacking stiff guard hairs (e.g., raccoon). Other factors besides pelage characteristics, such as the movement patterns, grooming behavior, and size of different animal species will affect their potential as seed dispersers as well (Agnew and Flux 1970; Lacey 1982; Sorenson 1986).
If differences in seed dispersal among these three ecologically similar species are the primary source of variation influencing gene flow, then the relative amounts of genetic differentiation among population subdivisions should inversely reflect their dispersal abilities (i.e., Cryptotaenia [is greater than] Osmorhiza [is greater than] Sanicula). Likewise, if rates of gene flow are a monotonically decreasing function of distance between population subdivisions (Wright 1943, 1946; Kimura and Weiss 1964), then differentiation should increase at greater spatial scales. Two measures of genetic differentiation, average genetic distance, and multi-locus F-statistics, gave concordant results that support both of these predictions.
Variation among species in estimates of the inbreeding coefficient, [Mathematical Expression Omitted], indicates that differences in pollen dispersal may also contribute to variation in rates of gene flow among these species. Osmorhiza has significantly higher estimates of [f.sub.is] than the other species. This suggests that Osmorhiza has a substantially higher level of self-fertilization than the other species. This unexpected difference in three species' mating systems may retard gene flow through pollen in Osmorhiza relative to that of Cryptotaenia and Sanicula, which will lead to a more genetic differentiation among population subdivisions in Osmorhiza relative to the other species than would occur because of differences in seed dispersal alone. This may help explain why the amount of differentiation among populations and drainages in Osmorhiza is more similar to Cryptotaenia than Sanicula. More importantly however, the rank order of differentiation among species predicted by seed-dispersal ability was maintained despite the large differences in [Mathematical Expression Omitted], and presumably mating systems, at all spatial scales; Osmorhiza, with the highest values of [Mathematical Expression Omitted], still has less genetic differentiation than Cryptotaenia. This further supports the assumption that variation in seed dispersal is the primary factor leading to differences in population genetic structure in these species.
Cluster analysis of pairwise genetic distances and correlations of genetic and spatial distance indicate that geographic distance alone does not explain the pairwise migration rates between populations of these species. Long-distance migration may be stochastic, and/or other paths of "connectedness" between populations may better reflect gene flow than the linear distance between them. Ritland (1989) found similar patterns of genetic distance among montane populations of Mimulus caespitosus. In Mimulus, populations along the same streams were genetically very similar, but there was only a weak pairwise relationship between spatial distance and genetic distance at larger spatial scales. Animal dispersers may tend to follow the natural topography, following streambeds or contours. Passively dispersed seeds may be washed downhill. Within drainages, populations in the same stream channel (PGA and PGB, HDA and HDB) or at the top and bottom of a slope (BHG and BHQ, BHX and BHY) tended to be more genetically similar than to populations along another stream or on adjacent ridges. Similar migration patterns along stream drainages have been found in other studies of plants (Waser et al. 1982; Ritland 1989), snails (Selander and Whittam 1983; Arter 1990), and humans (Smouse and Wood 1987).
Stochastic rates and patterns of gene flow may also help explain the observed heterogeneity of single-locus estimates of F-statistics. Populations most likely do not receive all alleles at the same rates or from the same locations. Such historical artifacts may persist for long periods if migration rates are small. Somewhat surprisingly, it is in the animal-dispersed species, with the highest rates of gene flow, that the greatest variation in estimates of Nm are found. This suggests that it is occasional long-distance migration that is most important for gene flow, but such migration is very stochastic.
When the relationship between genetic and spatial distance is complicated by stochastic variation in migration rates or nonlinear paths of gene flow, it may be more relevant to examine the average genetic differentiation among population subdivisions at different levels of the spatial hierarchy. This as well as other studies of spatial genetic structure at different scales have shown that genetic differentiation among population subdivisions ([F.sub.st] = [Theta]) tends to increase with increasing spatial distance between sampling sites (Loveless and Hamrick 1984; Ritland and Ganders 1987; Bos et al. 1986; Van Dijk et al. 1988; Heywood 1991). On average, rates of gene flow (and hence genetic differentiation) should be most similar among populations at a given level of spatial subdivision, and less gene flow should occur among more distant populations in increasing hierarchical levels. These are the assumptions underlying what has been termed the hierarchical island model of gene flow (Slatkin and Voelm 1991). Such a model best explains the patterns of genetic differentiation observed in the three species studied here.
Surveys across a diverse array of plant taxa have not shown a strong relationship between modes of seed dispersal and levels of genetic differentiation among population subdivisions (Loveless and Hamrick 1984; Hamrick and Godt 1990). By comparing three ecologically similar, taxonomically related species we have shown that a reduction in seed dispersal is associated with the predicted increase in genetic differentiation among population subdivisions at spatial scales ranging from a few meters to hundreds of kilometers. The effects of seed dispersal therefore outweigh those of mating system variation or other conflicting ecological and evolutionary factors that may exist among these three species.
We thank The Nature Conservancy, University of Wisconsin Arboretum, and the Bureau of Endangered Resources, Wisconsin Department of Natural Resources for allowing us to work in the Baraboo Hills and elsewhere. Help in the field and laboratory were provided by B. Johnson, B. Franklin, P. Ode, M. Kuchenreuthcr, C. Umbanhowar, and S. Will-Wolf. Financial support was provided to C.F.W. by the Graduate School, University of Wisconsin-Madison, a National Institutes of Health traineeship in genetics, and the O. N. Allen Memorial Fellowship from the Department of Botany. Research support was provided to C.F.W. by Sigma Xi, the Lois Almon Small Grants Program, the J. J. Davis Fund (Department of Botany), and a National Science Foundation Dissertation Improvement Grant # BSR-8914599. Invaluable support was provided by the staff of the Department of Botany. D. Waller, N. Waser, Y. Linhart, and two anonymous reviewers provided insightful comments on earlier drafts of this paper.
Agnew, A. D. Q., and J. E.C. Flux. 1970. Plant dispersal by hares (Lepus capensis L.) in Kenya. Ecology 51:735-737.
Antonovics, J. 1968. Evolution in closely adjacent plant populations. VI. Manifold effects of gene flow. Heredity 23:507-524.
Antonovics, J., and A. D. Bradshaw. 1970. Evolution in closely adjacent populations. VIII. Clinal pattern at a mine boundary. Heredity 25:349-362.
Arter, H.E. 1990. Spatial relationship of gene flow paths between populations of the alpine snail Arianta arbustorum (Pulmonata: Helicidae). Evolution 44:966-980.
Baskin, J. M., and C. C. Baskin. 1984. Germination ecophysiology of the woodland herb Osmorhiza longistylis (Umbelliferae). American Journal of Botany 71:687-692.
-----. 1988. The ecological life cycle of Cryptotaenia canadensis (L.) DC. (Umbelliferae), a woodland herb with monocarpic ramets. American Midland Naturalist 119:165-173.
-----. 1991. Nondeep complex morphophysiological dormancy in seeds of Osmorhiza claytonii (Apiaceae). American Journal of Botany 78:588-593.
Bell, C.R. 1971. Breeding systems and floral biology of the Umbelliferae or evidence for specialization in unspecialized flowers. Pp. 93-107 in V. H. Heywood, ed. Biology and chemistry of the Umbelliferae. Botanical Journal of the Linnean Society 64: Supplement 1. Academic Press, London.
Bos, M., H. Harmens, and K. Vrieling. 1986. Gene flow in Plantago. I. Gene flow and neighborhood size in P. lanceolata. Heredity 56:43-54.
Brown, A. H. D. 1979. Enzyme polymorphisms in plant populations. Theoretical Population Biology 15:1-42.
Bullock, S. H., and R. B. Primack. 1977. Comparative experimental study of seed dispersal on animals. Ecology 58:681-686.
Cavalli-Sforza, L. L., and A. W. F. Edwards. 1967. Phylogenetic analysis: models and estimation procedures. Evolution 21:550-570.
Clayton, J., and D. Tretiak. 1972. Amine-citrate buffers for pH control in starch gel electrophoresis. Journal of Fisheries Research Board of Canada 29:1169-1172.
Constance, L., and R. H. Shan. 1948. The genus Osmorhiza (Umbelliferae): a study in geographic affinities. University of California Publications in Botany 23:111-156.
Crawford, T.J. 1984. What is a population? Pp. 135-173 in B. Shorrocks, ed. Evolutionary ecology. Blackwell, Oxford.
Crow, J. F., W. R. Engels, and C. Denniston. 1990. Phase three of Wright's shifting-balance theory. Evolution 44:233-247.
Curtis, J.T. 1959. The vegetation of Wisconsin. University of Wisconsin Press, Madison.
Edwards, A. W.F. 1971. Distances between populations on the basis of gene frequencies. Biometrics 27:873-881.
Efron, B. 1982. The jacknife, the bootstrap and other resampling plans. CBMS-NSF Regional Conference Series in Applied Mathematics, Monograph 38. SIAM, Philadelphia.
Ehrlich, P. R., and P. H. Raven. 1969. Differentiation of populations. Science 165:1228-1232.
Fassett, N. C. 1976. Spring flora of Wisconsin, 4th ed. (revision, O. S. Thompson). University of Wisconsin Press, Madison.
Gleason, H. A., and A. Cronquist. 1963. Manual of vascular plants of Northeastern United States and adjacent Canada. Van Nostrand, New York.
Hamrick, J. L., and M. J. W. Godt. 1990. Allozyme diversity in plant species. Pp. 43-63 in A. H. D. Brown, M. T. Clegg, A. L. Kahler, and B. S. Weir, eds. Plant population genetics, breeding, and genetic resources. Sinauer, Sunderland, Mass.
Hamrick, J. L., and M. D. Loveless. 1986. The influence of seed dispersal mechanisms on the genetic structure of plant populations. Pp. 211-224 in A. Estrada and T. H. Fleming, eds. Frugivores and seed dispersal. Dr. W. Junk, Dordrecht, The Netherlands.
Handel, S. N. 1983. Pollination ecology, plant population structure, and gene flow. Pp. 163-211 in L. Real, ed. Pollination biology. Academic Press, Orlando.
Hartl, D. L., and A. G. Clark. 1989. Principles of population genetics, 2d ed. Sinauer, Sunderland, Mass.
Heywood, J. L. 1991. Spatial analysis of genetic variation in plant populations. Annual Review of Ecology and Systematics 22:335-355.
Hiroe, M., and L. Constance. 1958. Umbelliferae of Japan. University of California Publications in Botany 30:1-144.
Holtsford, T. P., and N. C. Ellstrand. 1989. Variation in outcrossing rate and population genetic structure of Clarkia temloriensis (Onagraceae). Theoretical and Applied Genetics 78:480-488.
Howe, H. F., and J. Smallwood. 1982. Ecology of seed dispersal. Annual Review of Ecology and Systematics 13:201-228.
Jain, S. K., and A. D. Bradshaw. 1966. Evolutionary divergence among adjacent plant populations. I. The evidence and its theoretical analysis. Heredity 21:407-441.
Karron, J. D. 1989. Breeding systems and levels of inbreeding depression in geographically restricted and widespread species of Astragalus (Fabaceae). American Journal of Botany 76:331-340.
Karron, J. D., Y. B. Linhart, C. A. Chaulk, and C. A. Robertson. 1988. Genetic structure of populations of geographically restricted and widespread species of Astragalus. American Journal of Botany 75:1114-1119.
Kimura, M., and G. H. Weiss. 1964. The stepping stone model of population structure and the decrease of genetic correlation with distance. Genetics 49:561-576.
Knuth, K. 1908. Handbook of flower pollination. Translation by J. R. A. Davis. Clarendon Press, Oxford.
Lacey, E.P. 1982. Timing of seed dispersal in Daucus carota. Oikos 39:83-91.
Levin, D.A. 1978. Genetic variation in annual Phlox: self-compatible versus self-incompatible species. Evolution 32:245-263.
-----. 1984. Immigration in plants: an exercise in the subjunctive. Pp. 242-260 in R. Dirzo and J. Sarukhan, eds. Perspectives on plant population ecology. Sinauer, Sunderland, Mass.
Levin, D. A., and H. W. Kerster. 1974. Gene flow in seed plants. Evolutionary Biology 7:139-220.
Lindsey, A. H. 1984. Reproductive biology of Apiaceae. I. Floral visitors to Thaspium and Zizia and their importance in pollination. American Journal of Botany 71:375-387.
Loveless, M. D., and J. L. Hamrick. 1984. Ecological determinants of genetic structure in plant populations. Annual Review of Ecology and Systematics 15:65-95.
-----. 1988. Genetic divergence and evolutionary history in two North American species of Cirsium. Evolution 42:254-265.
Malecot, G. 1948. The mathematics of heredity. Translation by D. M. Yermanis, 1969. W. H. Freeman, San Francisco.
Marty, T. L., D. M. O'Malley, and R. P. Guries. 1984. A manual for starch-gel electrophoresis: new microwave edition. University of Wisconsin Department of Forestry Staff Paper Series 20.
McNeilly, T. 1968. Evolution in closely adjacent populations. III. Agrostis tenuis on a small copper mine. Heredity 23:99-108.
Nei, M. 1972. Genetic distance between populations. American Naturalist 106:283-292.
-----. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583-590.
Pleasants, J. M., and J. F. Wendel. 1989. Genetic diversity in a clonal narrow endemic, Erythronium propullans, and its widespread progenitor, Erythronium albidum. American Journal of Botany 76:1136-1151.
Pryer, K. M., and L. R. Phillippe. 1989. A synopsis of the genus Sanicula (Apiaceae) in eastern Canada. Canadian Journal of Botany 67:694-707.
Ritland, K. 1989. Genetic differentiation, diversity, and inbreeding in the mountain monkeyflower (Mimulus caespitosus) of the Washington Cascades. Canadian Journal of Botany 67:2017-2024.
Ritland, K., and F. R. Ganders. 1987. Covariation of selfing rates with parental gene fixation indices within populations of Mimulus guttatus. Evolution 41:760-771.
Rogers, J.S. 1972. Measures of genetic similarity and genetic distance. Pp. 145-153 in Studies in genetics VII. University of Texas Publication 7213. University of Texas, Austin.
Rohlf, F. J., and G. D. Schnell. 1971. An investigation of the isolation-by-distance model. American Naturalist 105:295-324.
Schoen, D.J. 1982. Genetic variation and the breeding system of Gilia achilleifolia. Evolution 36:361-370.
Selander, R. K., and T. S. Whittam. 1983. Protein polymorphism and the genetic structure of populations. Pp. 89-114 in M. Nei, ed. Evolution of genes and proteins. Sinauer, Sunderland, Mass.
Shan, R. H., and L. Constance. 1951. The genus Sanicula (Umbelliferae) in the old world and the new. University of California Publications in Botany 25:1-78.
Shmida, A., and S. Ellner. 1983. Seed dispersal on pastoral grazers in open Mediterranean chaparral, Israel. Israel Journal of Botany 32:147-159.
Slatkin, M. 1985. Gene flow in natural populations. Annual Review of Ecology and Systematics 16:393-430.
Slatkin, M., and L. Voelm. 1991. [F.sub.st] in a hierarchical island model. Genetics 127:627-629.
Smouse, P. E., and J. W. Wood. 1987. The genetic demography of the Gainj of Papua New Guinea: functional models migration and their genetic implications. Pp. 211-224 in B. D. Chepko-Sade and Z. T. Halpin, eds. Mammalian dispersal patterns: the effects of social structure on population genetics. University of Chicago Press, Chicago.
Snaydon, R. W. 1970. Rapid population differentiation in a mosaic environment. I. The response of Anthoxanthum odoratum populations to soils. Evolution 24:257-269.
Sneath, P. H. A., and R. R. Sokal. 1973. Numerical taxonomy. Freeman, San Francisco.
Sokal, R. R., G. M. Jacquez, and M. C. Wooten. 1989. Spatial autocorrelation analysis of migration and selection. Genetics 121:845-855.
Soltis, D. E., C. H. Haufler, D. C. Darrow, and G. J. Gastony. 1983. Starch gel electrophoresis of ferns: a compilation of grinding buffers, gel and electrode buffers, and staining schedules. American Fern Journal 73:9-27.
Sorenson, A. E. 1986. Seed dispersal by adhesion. Annual Review of Ecology and Systematics 17:443-463.
Stuber, C. W., J. F. Wendel, M. M. Goodman, and J. S. C. Smith. 1988. Techniques and scoring procedures for starch gel electrophoresis of enzymes from Maize (Zea mays L.). Technical Bulletin 286, North Carolina Agricultural Research Service, North Carolina State University, Raleigh, N.C.
Sun, M., and F. R. Ganders. 1988. Mixed mating systems in Hawaiian Bidens (Asteraceae). Evolution 42:516-527.
Swofford, D. L., and R. B. Selander. 1981. Biosys-1: a computer program for the analysis of allelic variation in population genetics and biochemical systematics, release 1.7 (1989). Illinois Natural History Survey Champaign, Ill.
Tonsor, S.J. 1990. Spatial patterns of differentiation for gene flow in Plantago lanceolata. Evolution 44: 1373-1378.
Turkington, R., and J. L. Harper. 1979. The growth, distribution, and neighbourhood relationships of Trifolium repens in a permanent pasture. IV. Fine-scale biotic differentiation. Journal of Ecology 67:245-254.
Turner, M. E., C. Stephens, and W. W. Anderson. 1982. Homozygosity and patch structure in plant populations as a result of nearest neighborhood pollination. Proceedings of the National Academy of Sciences, USA 79:203-207.
Uma Shaanker, R., K. N. Ganeshaiah, and T. R. Radhamani. 1990. Association among the modes of pollination and seed dispersal--ecological factors and phylogenetic constraints. Evolutionary Trends in Plants 42:107-111.
Van der Pijl, L. 1982. Principles of dispersal in higher plants, 3d ed. Springer, New York.
Van Dijk, H., K. Wolff, and A. De Vries. 1988. Genetic variability in Plantago species in relation to their ecology. 3. Genetic structure of populations of P. major, P. lanceolata and P. coronopus. Theoretical and Applied Genetics 75:518-528.
Wade, M. J., and C. J. Goodnight. 1991. Wright's shifting-balance theory: an experimental study. Science 253:1015-1018.
Waser, N. M., and M. V. Price. 1985. Reciprocal transplant experiments with Delphinium nelsonii (Ranunculaceae): evidence for local adaptation. American Journal of Botany 72:1726-1732.
Waser, N. M., R. K. Vickery, Jr., and M. V. Price. 1982. Patterns of seed dispersal and population differentiation in Mimulus guttatus. Evolution 36: 753-761.
Weir, B.S. 1990. Genetic data analysis. Sinauer, Sunderland, Mass.
Weir, B. S., and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38:1358-1370.
Wendel, J. F., and C. R. Parks. 1982. Genetic control of isozyme variation in Camellia japonica L. Journal of Heredity 73:197-204.
Wilkinson, L. 1989. SYSTAT 5.0: the system for statistics. SYSTAT, Inc., Evanston, Ill.
Williams, C. F. 1991. Spatial genetic consequences of seed dispersal: a comparative study of three sympatric forest herbs (Umbelliferae). Ph.D. diss. University of Wisconsin, Madison.
Wright, S. 1931. Evolution in Mendelian populations. Genetics 16:97-159.
-----. 1932. The roles of mutation, inbreeding, cross-breeding, and selection in evolution. Proceedings of the Sixth International Congress of Genetics 1:356-366.
-----. 1943. Isolation by distance. Genetics 28:114-138.
-----. 1946. Isolation by distance under diverse systems of mating. Genetics 31:39-59.
-----. 1951. The genetic structure of populations. Annals of Eugenics 15:323-354.
-----. 1988. Surfaces of selective value revisited. American Naturalist 131:115-123.
Computational Formulae for Hierarchical F-Statistics
Hierarchical four-level F-statistics (Wright 1951) were calculated by the method of Weir and Cockerham (1984) using the formulae given in Weir (1990). This approach is analogous to an ANOVA using allele frequencies and produces unbiased single- and multi-locus estimators of each statistic weighted by allele frequency and different population subdivision sample sizes. Estimates were first computed for each allele, then the appropriate components of variance were summed over alleles for single-locus estimates. Multi-locus estimates were calculated by summing variance components over all alleles at all loci.
Two typographical errors shown in Weir (1990) should be noted in the formulae for calculating four-level hierarchical F-statistics. On page 162, the first formula for calculating the sums of squares should read
[Mathematical Expression Omitted],
where [u.sub.i..] replaces the [s.sub.1] given. Moreover, the following formula should be
[Mathematical Expression Omitted],
in which [u.sub.ij.] replaces the [t.sub.ij] given in Weir (1990).
In the hierarchical four-level analysis, the five components of variance [V.sub.g], [V.sub.i], [V.sub.s], [V.sub.p], and [V.sub.d] (variances in genes within individuals, individuals within subpopulations, subpopulations, populations, and drainages respectively) were calculated from their corresponding mean squares and correction terms [calculated per Weir (1990)] as follows
[V.sub.g] = MSG.
[V.sub.i] = (MSI - MSG)/2.
[V.sub.s] = (MSS - MSI)/[2[n.sub.c6].
[V.sub.p] = (MSP -MSI)/2[n.sub.c5] - [n.sub.c4](MSS - MSI)/2[n.sub.c5][n.sub.c6].
[V.sub.d] = [(MSD - MSI)/2[n.sub.c3]] + [([n.sub.c2][n.sub.c4] - [n.sub.c1][n.sub.c5])(MSS - MSI)/2[n.sub.c3][n.sub.c5][n.sub.c6]] - [[n.sub.c2](MSP - MSI)/2[n.sub.c3][n.sub.c5]].
Estimates of the hierarchical F-statistics were then calculated from these variance components as
[Mathematical Expression Omitted],
[Mathematical Expression Omitted],
[Mathematical Expression Omitted],
[Mathematical Expression Omitted], and
[Mathematical Expression Omitted].
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Genetic Consequences of Seed Dispersal in Three Sympatric Forest Herbs, part 1|
|Author:||Williams, Charles F.; Guries, Raymond P.|
|Date:||Jun 1, 1994|
|Previous Article:||A multilevel approach to the significance of genetic variation in alcohol dehydrogenase of Drosophila.|
|Next Article:||The relationship between inbreeding depression and prior inbreeding among populations of four Mimulus taxa.|