• asexual reproduction;
  • bryozoan;
  • Cristatella mucedo;
  • gene flow;
  • genetic variation;
  • metapopulation;
  • microsatellites


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices

Theoretical models of the impact of a metapopulation structure on the genetics of a species have resulted in multiple predictions that have seldom been empirically evaluated. Here we present microsatellite data from 14 populations of a freshwater bryozoan, Cristatella mucedo, collected along a waterfowl migratory route in north-western Europe. C. mucedo is facultatively sexual and has the unusual tactic of dispersing via asexually generated propagules. These propagules are likely to be dispersed by waterfowl and therefore the populations that we sampled were expected to maintain some degree of connectivity. Our data illustrate a metapopulation comprising well-differentiated populations connected by low levels of ongoing gene flow, patterns that agree with predictions based on theoretical work. However, contrary to expectations of a metapopulation, particularly one in which asexual reproduction predominates, genetic variation within populations was often high. This diversity seems to be at least partially attributable to the gene flow that results from ongoing dispersal.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices

An explicit model of a metapopulation was first developed 30 years ago by Levins (1969, 1970), and describes what is now known as the classical metapopulation structure. Levins’ model describes a ‘population of populations’, all of which exist in a balance between extinction and recolonization, and are linked to one another by dispersal and gene flow. As the concept of a metapopulation has become more commonplace in literature and research, the term has often been broadly adapted to refer to a series of conspecific populations that are connected to one another by dispersal ( Hastings & Harrison, 1994). In accordance with this modified view, in which localized extinctions are not necessarily a prerequisite, a number of alternative metapopulation models have been created ( Hanski & Gilpin, 1991; Harrison, 1991; Hastings & Harrison, 1994).

Regardless of the model, a metapopulation structure will significantly impact the population genetics of the species in question. For example, the repeated localized extinctions and recolonizations, which are typical of the classical metapopulation, can profoundly affect genetic structure ( Slatkin, 1985, 1987; Wade & McCauley, 1988; Hastings & Harrison, 1994). These processes will often be accompanied by population bottlenecks, which serve to enhance genetic drift. Depending upon the balance between genetic drift and gene flow, levels of genetic differentiation among local populations may be either enhanced or diminished ( McCauley, 1989). Models have also demonstrated that reduced genetic variability and heterozygosity levels can result from metapopulation dynamics ( Gilpin, 1991), although empirical evidence for this is lacking ( Hanski, 1998).

Much of the research on the genetic consequences of a large-scale metapopulation structure remains theoretical owing to logistical constraints; however, here we present the results of a genetic study of an expansive metapopulation of the freshwater bryozoan Cristatella mucedo (Phylum Bryozoa, Class Phylactolaemata). C. mucedo inhabits discrete lakes and ponds, engages in sex relatively infrequently and disperses among sites via asexually produced propagules (statoblasts). Reproduction is predominantly by asexual means, including budding during colony growth, statoblast production and colony fission. There can be a brief period of sexual reproduction near the beginning of each season, although this is not apparent in all populations ( Okamura, 1997a). Sexually produced larvae are short-lived and provide a limited degree of within-site dispersal. In contrast, asexually produced statoblasts possess highly resistant chitinous valves and can survive throughout the winter. Their buoyancy, due to internalized gas-filled cells, allows wide dispersal within a site. In addition, small marginal hooks affect statoblast attachment to feathers or fur and allow dispersal to more distant sites via animal vectors ( Okamura, 1997a). Statoblasts of freshwater bryozoans have also been shown to resist desiccation, and can remain viable after passing through the digestive tracts of ducks and amphibians ( Brown, 1933; I. Charambilidou, personal communication).

This strategy of dispersing via asexually produced propagules is relatively unusual. Facultatively sexual freshwater taxa more commonly undergo sexual reproduction at the end of the summer, thereby generating genetically variable individuals for the re-establishment of populations in the following year ( Hughes, 1989; Lynch & Spitze, 1994). However, since bryozoan statoblasts are clonally produced, they can be generated in greater abundance than sexually produced propagules and this will maximize the chances of successful passive dispersal and ultimately of (re)colonization of other sites. The clear potential for dispersal by statoblasts suggests the possibility of C. mucedo conforming to a metapopulation, and indeed populations in southern England seem to adhere to such a structure. The first evidence for this was based on observations of apparent extinction and recolonization events, and the widespread availability of suitable but unoccupied sites ( Okamura, 1997a, b). Further evidence came from genetic work using RAPD-PCR which showed a high degree of relatedness among populations ( Hatton-Ellis et al., 1998 ). However, the implications of the study were limited by small sample sizes and the choice of marker. Use of RAPD-PCR does not permit straightforward calculations of Wright’s F-statistics (and their analogues), and precludes measures of heterozygosity and assessment of deviations from Hardy–Weinberg equilibrium.

Here we use microsatellite data to characterize C. mucedo populations collected from along a major waterfowl migratory route in north-western Europe ( Scott & Rose, 1996). Since waterfowl are likely vectors for transporting freshwater invertebrates among sites ( Weider et al., 1996 ; Taylor et al., 1998 ), we expected that the sampled populations would maintain some level of connectivity. We were specifically interested in: (1) determining whether C. mucedo conforms to a wide-scale metapopulation and (2) investigating the broad-scale population genetic patterns of C. mucedo. Microsatellites are codominant, locus-specific markers with relatively high rates of mutation ( Tautz, 1989; Weber & Wong, 1993), and they therefore provide us with new insights into the population genetics of C. mucedo. The data we present here supply a comprehensive view of the metapopulation genetics of an organism with an unusual life history, and provide an empirical evaluation of the role of metapopulation dynamics in the population structure of predominantly clonal organisms.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices

Collection of specimens

We collected samples from 14 C. mucedo populations found in seven different countries ( Fig. 1). C. mucedo occurs as loosely attached colonies on substrata such as stems and leaves of aquatic plants, submerged tree branches, roots, rocks and stones. At each site we collected approximately 30 colonies (Table 1) by either searching accessible substrata around the edges of the ponds and lakes, or by using a grappling hook to retrieve aquatic plants, branches, etc. As C. mucedo can reproduce by growth and fission, we took only one colony from clumped distributions on a single substratum, as in these situations the entire clump was likely to be derived from a single clone. Therefore, all of the colonies that we collected should have been the result of separate statoblasts or larvae. The colonies were stored separately in 99% ethanol at 4 °C.


Figure 1 Map showing location of populations used in this study. Sites are: 1=Loirston Loch, Scotland, 2=Tufty’s Corner, England, 3=Etang du Gros Caillou, France, 4=Le Lac de Grand Lieu, France, 5=Recreational Park De Brabantse, Nationaal Park de Beisbosch, Netherlands, 6=Herpen, Netherlands, 7=Ry Mølle Sø, Denmark, 8=Søndersø Nørresø, Denmark, 9=Rössjön, Sweden, 10=Sommen, Sweden, 11=Littoistenjarvi, Finland, 12=Enärjärvi, Finland, 13=Keitele, Finland, 1. 4=Konnevesi, Finland.

Download figure to PowerPoint

Table 1.   Measures of genetic diversity (D* = number of unique genotypes/N), total number of alleles, number of private alleles (alleles found in a single population; Slatkin, 1985), number of individuals, observed heterozygosity (Ho plus standard deviation), and expected heterozygosity (He plus standard deviation, calculated after Levene, 1949) for each population. Population numbers in parentheses after site name correspond to the population identification numbers in Fig. 1. Thumbnail image of

Laboratory methods

DNA was extracted from approximately 1 mm3 of tissue in 5% Chelex (BioRad) in ddH2O following the manufacturer’s protocol. Five loci were amplified from each individual using primer pairs 1.1, 2.2, 5.9, 6.7 and 9.4 ( Freeland et al., 1999 ). PCR reactions were performed in a Techne thermocycler using a 3’Cy5 fluorescently labelled forward primer (Pharmacia). The PCR reactions included 0.1–1.0 ng DNA, 1× PCR buffer (Promega), 1.5–2.5 m M MgCl2, 200 μM each of dGTP, dATP, dTTP and dCCT, 0.1–0.5 m M of each primer, and 1 U Taq DNA polymerase (Promega). The cycling parameters are given in Freeland et al. (1999 ). Ten femtomols of the amplified DNA from each sample was then run out on a 6% polyacrylamide gel in an ALFexpress DNA Sequencer. Internal and external size markers (Pharmacia) were included on the gel, together with four spaced samples of an allele of known size. The resulting profiles were analysed using AlleleLinks version 1.00 (Pharmacia). In total we analysed five microsatellite loci from 408 individuals (Table 1: individual data).


Both FST ( Wright, 1951) and RST ( Slatkin, 1995) have been used in studies of population genetics as measures of population differentiation. The former assumes a low mutation rate under an infinite alleles model, while the latter uses variances in allele sizes and thus can allow for a relatively high mutation rate and a stepwise mutation model. RST is a more appropriate statistic than FST for analysing microsatellite data, as its assumptions may be more representative of microsatellite mutation patterns ( Slatkin, 1995). However, Slatkin’s method for calculating RST assumes that populations of equal sample sizes have been included, and that all loci have equal variances. We therefore used RstCalc to calculate Rho, an unbiased measure of RST that corrects for both of these potential sources of bias ( Goodman, 1997). We based our analyses on both individual and clonal data, i.e. with and without clonal replicates, respectively, as including all individuals may bias estimates of genetic identity and divergence if there are multiple colonies of a single clone ( Stoddart, 1984; Ayre & Willis, 1988).

We used three parameters to infer patterns of gene flow: (1) instances of identical clonal genotypes occurring in more than one population; (2) calculations of Nm (number of migratory individuals); and (3) discriminant function analyses. Nm values were calculated in RstCalc, using the formula Nm=1/4[(1/Rho) – 1]. While this equation can provide estimates of gene flow among populations, it has been argued that F-statistics (and their analogues) are not necessarily appropriate for inferring gene flow ( Whitlock & McCauley, 1999) and should be supplemented with other data. We therefore performed discriminant function analyses to assign individuals to their most likely population of origin, thereby obtaining additional evidence of recent dispersal events. This involved calculating distance measurements based on the absolute difference in number of repeat units between each pair of individuals for each locus, and then summing over all loci. This value was then squared to obtain a more realistic representation of the number of mutational steps involved. As the data were not normally distributed (Kolmogorov–Smirnov: < 0.01 in all populations), we used a nonparametric discriminant function analysis. The discriminant function was evaluated using a crossvalidation method in which a single individual was removed from the data set for each calculation ( Manly, 1994).

The proportion of novel clones within each population was calculated as D* (=number of clones/number of individuals; Hunter, 1993). Assessment of Hardy–Weinberg equilibrium, and calculations of expected and observed heterozygosity values (Ho and He) were performed in Popgen ( Yeh & Boyle, 1997). This program uses Levene’s algorithm ( Levene, 1949) to calculate expected genotypic frequencies and expected heterozygosities, and χ2-tests to determine significance of departures from Hardy–Weinberg equilibrium. FIS statistics ( Wright, 1951) were calculated in Genepop ( Raymond & Rousset, 1995) following the estimates of Weir & Cockerham (1984). We considered the possibility that the Wahlund effect had depleted levels of observed heterozygosity in our data, as this may occur when individuals from genetically dissimilar backgrounds are mixed. We calculated the maximum potential contribution of the Wahlund effect to observed heterozygosity deficits by comparing the ratio of expected heterozygosity minus observed heterozygosity to the sum of the variation in allele frequencies ( Li, 1976). Instances of two-locus linkage disequilibrium were estimated in Popgen using Burrow’s composite measure between pairs of loci, and χ2 tests were used to determine significance ( Weir, 1979). A Mantel’s test ( Mantel, 1967) was employed to test for isolation by distance ( Wright, 1943) using pairwise Rho values and the shortest geographical distances between sites. Neighbour-joining trees reflecting the genetic distances among populations were reconstructed in Phylip ( Felsenstein, 1995) using a distance matrix calculated from pairwise population measures of Nei’s unbiased genetic distance D ( Nei, 1978).

In order to investigate factors that may be affecting levels of genetic diversity within populations, as measured by the total number of alleles, we performed Pearson product moment correlations between the number of alleles and Nm, and the number of alleles and Ho. The correlations involving Nm were done in two ways, and were based on both individual and clonal data sets. First, the pairwise values of Nm were compared to the mean number of alleles in each pair of populations (n=91). Second, the mean Nm values for each population averaged from all pairwise comparisons were compared to the number of alleles within each population (n=14). The correlation between Ho and the number of alleles was done using single values for each population (n=14).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices

Genetic variation, population subdivision and linkage disequilibrium

Table 2. Thumbnail image of
Table 3. Thumbnail image of
Table 4. Thumbnail image of

Figure 2 Total number of clones identified by one (n=5), two (n=9), three (n=10), four (n=5) and five (n=1) microsatellite loci. Respective n-values represent all possible combinations of the loci when 1, 2. , 3, 4 and 5 loci are considered.

Download figure to PowerPoint

Table 2.   Matrix of Nm values (above) and Rho values (below) of all individual data. Nm = 1/4[(1/Rst) – 1]. All pairwise comparisons show significant (< 0.05) departures from the null hypothesis, which states that there is no genetic differentiation between populations. Column and row numbers correspond to the population numbers in Fig. 1. Thumbnail image of
Table 3.   Matrix of Nm values (above) and Rho values (below) of clonal data. Nm = 1/4[(1/Rst) – 1]. All pairwise comparisons show significant departures (< 0.05) from the null hypothesis, which states that there is no genetic differentiation between populations, with the exception of the 17 values that are shown in bold. Note that all populations have been included for the sake of completion, but some sample sizes are very small (see Table 1). The column and row numbers correspond to the population numbers shown in Fig. 1. Thumbnail image of

There were 64 cases of two-locus linkage disquilibrium at the P=0.05 significance level when all individual data were analysed, which translated into only five cases of linkage when clonal replicate data were removed. We therefore conclude that the effect of linkage disequilibrium on our analyses is negligible.

Observed heterozygosity deficits

Observed heterozygosity was lower than the expected values at all loci and in all populations (Table 1), and therefore is unlikely to be an artefact of null alleles ( Callen et al., 1993 ). There was no correlation between the number of alleles and the overall observed heterozygosity within a population (P=0.894, r=0.039, n=14 for individual data; P=0.514, r=0.191, n=14 for clonal data). Levels of inbreeding were high, with nearly half (33/70) of the FIS values equal to 1.00 (calculated for each locus in each population). However, a comparison of heterozygosity levels and among-population allelic variation shows that the Wahlund effect could conceivably account for 58% of the observed heterozygosity deficit (Table 4). Given the low levels of inferred spatial gene flow, this is undoubtedly an overestimation; however, the potential role of temporal gene flow is unknown (see Discussion) and therefore a possible Wahlund effect should not be dismissed at this time.

Table 4.   Comparison of heterozygosities and allelic variances, showing that up to 58% of heterozygote deficiencies may be attributed to the Wahlund effect. Thumbnail image of

Gene flow

Despite sampling only a small proportion of potential sites, we found direct evidence of long-distance gene flow when the same multilocus genotype was recovered as individual colonies from both Rössjön in Sweden and the Nationaal Park de Biesbosch in The Netherlands. These two sites are separated by 700 km of land and sea. It is possible that the same genotype arose in these two populations as a result of chance recombination of alleles. To test this possibility, we performed a computer simulation that reconstructed the genotypes of the two populations in question by randomly assigning individual genotypes based on the observed allele frequencies found within each population. Twenty-five thousand bootstrap replicates demonstrated that, assuming neutrality, the probability of the same clone being found in these two populations as a result of chance alone is extremely low (< 0.001). There was also one instance in which the same clone was recovered as individual colonies from neighbouring sites. This clone occurred in both Konnevesi and Keitele in Finland, two lakes that are separated by 50 km but are connected by a waterway.

When we looked for clonal replicates in multiple populations based on only four of the microsatellite loci (n=5) we found between two and eight instances of the same multilocus genotype appearing in more than one population. When data from two microsatellite loci were excluded (n=10), we found between two and 21 instances of the same multilocus genotype appearing in more than one population. These estimates are valid inferences of relatively recent dispersal, as the rapid mutation rate of microsatellites means that clones with common alleles at three or four (out of five) microsatellite loci have recent coalescent dates.

In keeping with these inferred patterns of ongoing dispersal, population dendrograms (not shown) based on Nei’s genetic distance (D) showed that, with the exception of Finland in the clonal analysis, the populations do not cluster by country of origin. Mantel’s tests of isolation by distance (1000 permutations) showed a nonsignificant comparison of Rho and geographical distance for both individual (P=0.477) and clonal (P=0.066) data. A nonsignificant Mantel’s test for isolation by distance was also found in the C. mucedo populations in the Thames basin in southern England ( Hatton-Ellis et al., 1998 ).

The estimated numbers of migrants per generation between populations (Nm) are shown in Tables 2 and 3. These must be treated as approximate values because estimates of Nm from FST or its analogues (RST/Rho) are based on the assumption that the rate of gene flow is faster than the mutation rate ( Neigel, 1997), and this is not always the case for microsatellites ( Rayboud et al., 1998 ). As a result, calculations of Nm from microsatellite data are likely to be underestimates ( Neigel, 1997). Nevertheless, calculating Nm from pairwise FST or its analogues provides an indication of the amount of gene flow between pairs of populations ( Slatkin, 1993). While the overall Nm values are low (average Nm=0.49 for individual data and Nm=0.62 for clonal data), nine out of the 14 populations receive more than one immigrant per generation from at least one other population (Tables 2 and 3. The Nm values were positively correlated with the total number of alleles within each population based on both pairwise Nm comparisons (P=0.000, r=0.432, n=91 for individual data, and P=0.001, r=0.353, n=91 for clonal data) and mean Nm values (P=0.011, r=0.656, n=14 for individual data, and P=0.009, r=0.668, n=14 for clonal data), suggesting that gene flow contributes to levels of genetic diversity within populations.

The discriminant function analysis assigned 14% of individuals and 15% of clones to populations other than those from which they were collected (Table 5). The highest number of ‘misclassifications’ involved the Nationaal Park de Biesbosch, Netherlands. This is an important stopover site for migratory waterfowl ( Scott & Rose, 1996), and statoblasts could be dispersed to and from this site with a relatively high frequency. In the three cases in which more than one individual of the same multilocus genotype was assigned to another population, we cannot discern the true number of misclassifications, as clonal replicates could either have arisen in the source population and been transported separately, or could have clonally replicated in the destination population following a single dispersal event. Furthermore, these estimated levels of gene flow must be interpreted with caution as sampling a greater number of populations and/or increasing the sample size for each population would likely alter the picture. In addition, our analysis assumes a stepwise mutation model (SMM) which, although plausible ( Bell & Jurka, 1997), is not the only possibility. Nevertheless, when taken in conjunction with the other evidence for gene flow, the proportion of misclassifications is high enough for us to conclude that recent dispersal events have occurred among populations.

Table 5.   Classification of individuals to populations according to the discriminant function analysis. Numbers in bold represent individuals assigned to a population other than the one in which they were found. When clonal replicates are included in this number, the number in parentheses written immediately afterward refers to the number of clones. Rows correspond to the site from which the colony was collected, columns correspond to the sites to which the colony was classified. Column and row numbers correspond to the population numbers in Fig. 1. Thumbnail image of


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices

Gene flow and population differentiation

The occurrence of the same clonal genotypes in multiple populations, the Nm estimates derived from Rho, and the discriminant function analyses, all demon- strated dispersal among populations. This in itself answers our first question by providing evidence for a C. mucedo metapopulation existing over a broad spatial scale in north-western Europe. The lack of isolation by distance reinforces this assertion, and conforms to expectations of a classical metapopulation model in which dispersal from one population is equally likely to result in colonization of any other ( Levins, 1970; Hanski & Simberloff, 1997). Although our estimates of gene flow were generally low, they would undoubtedly have been much higher had we characterized more than 14 populations from a 3000-km-long transect. In addition, measures of dispersal will be underestimated if gene frequencies have not reached equilibrium following founder effects ( Crow & Aoki, 1984; Boileau & Taylor, 1994).

Despite evidence of ongoing gene flow across both short and long distances, populations showed high degrees of genetic differentiation. This finding agrees with theoretical evidence that has demonstrated the influence of founder effects and genetic bottlenecks on the creation of population differentiation within a metapopulation ( Wade & McCauley, 1988), a process that will be exacerbated by clonal reproduction ( Wright, 1978). Additional genetic patterns associated with this broad-scale metapopulation are discussed below.

Genetic diversity within populations

The extent to which processes of extinction and recolonization reduce the effective genetic size of a metapopulation is unclear, although theoretical models generally suggest that a reduction does occur ( Amos & Harwood, 1998). A decline in genetic variability due to population bottlenecks should be particularly pronounced in clonal species within a metapopulation, as asexual reproduction does not generate novel genotypes. However, our data have revealed high levels of genetic diversity. While at one extreme there was a site dominated by a single clone, at other sites the ratio of the number of multilocus genotypes to the number of individuals sampled was as high as 87%.

The brief period of sex at the onset of the growing season may play only a limited role in the maintenance of genetic variation within populations. First, many populations apparently forego sexual reproduction in any given year ( Uotila & Jokela, 1995; Okamura, 1997b). In addition, populations of facultatively sexual organisms whose genetic diversity results primarily from sex are generally in Hardy–Weinberg equilibrium ( Lynch & Spitze, 1994; Grebelnyi, 1996), but none of the C. mucedo populations we assayed was in Hardy–Weinberg proportions. Moreover, there is no correlation between genetic diversity and observed heterozygosity which is anticipated if a relatively large number of genotypes results from sex. A more likely explanation for high levels of genetic diversity in some C. mucedo populations is relatively frequent immigration from other sites. This is supported by the correlation between the number of migrants and the number of alleles within populations. However, it should be noted that, because the estimates of gene flow in this correlation were derived from levels of population differentiation, this relationship may be inflated by population bottlenecks which would simultaneously increase population differentiation and decrease diversity.

Observed heterozygosity

The overall levels of observed heterozygosity were extremely low in all populations examined. Initially this may seem surprising, as species that reproduce by both sexual and asexual means should have approximately equal numbers of observed heterozygote excesses and deficits, assuming that all genotypes have an equal probability of being cloned. However, as mentioned previously, up to 58% of the observed heterozygosity deficit may be accounted for by the Wahlund effect. Although levels of spatial gene flow are low, there is also the possibility for temporal gene flow. Statoblasts of at least some freshwater bryozoans may remain viable in sediment for several years ( Wood, 1991), and the subsequent hatching of these statoblasts could result in a population comprising individuals from different generations, another scenario that can create a Wahlund effect ( Eanes & Koehn, 1978).

A second likely factor contributing to low observed heterozygosity is inbreeding. Although outcrossing has been documented in C. mucedo ( Jones et al., 1994 ), and can also be inferred from the presence of some heterozygotes in this study, selfing or mating between closely related genotypes cannot be dismissed ( Okamura, 1994). This could be exacerbated by the limited number of clones that exist in some populations, and means that sexual reproduction will not necessarily lead to a marked increase in genetic variation. Selfing and inbreeding result in increased homozygosity at all loci ( Wright, 1969; Muirhead & Lande, 1997), as seen in C. mucedo in this study, and many other taxa including plants ( Eckert & Barrett, 1994; Del Carmen Mandujano et al., 1996 ), tapeworms ( Lymberry et al., 1997 ) and snails ( Viard et al., 1997 ). However, it should be noted that inbreeding depression will not necessarily result, as a highly inbred population may have had much of its mutational load purged by natural selection ( Charlesworth & Charlesworth, 1987).

A third factor that may contribute to the low observed heterozygosity levels is frequent bottlenecks within a metapopulation structure. Apparent extinction and colonization events, and large fluctuations in abundance, have been observed within C. mucedo populations in the UK ( Vernon et al., 1996 ; Okamura, 1997a) and Austria ( Wöss, 1994). The likelihood of a genetic bottleneck resulting in a decrease in observed heterozygosity is inversely proportional to the size of the bottleneck, and directly proportional to its duration. It will also depend on the genetic makeup of the founding individual(s) ( Nei et al., 1975 ; Cabe, 1998). At least some of the C. mucedo populations in this study have recently experienced bottlenecks, for example Loirston Loch (Scotland), which contained only one clone (Table 1). We therefore conclude that the consistently low observed heterozygosity levels that we see in this study result from a combination of the Wahlund effect, inbreeding and periodic bottlenecks, the last being an integral part of metapopulation dynamics.

Reproduction, dispersal and the metapopulation biology of C. mucedo

The genetic patterns associated with populations of C. mucedo across a broad spatial scale are consistent with a metapopulation structure (reviewed in Hanski, 1998; Hastings & Harrison, 1994). These include pronounced population subdivision, ongoing gene flow and low observed heterozygosity. However, the high levels of diversity in some populations are counter to expectations based on many metapopulation models and we suggest that dispersal plays an important role in the maintenance of within-population genetic variation. This may relate to the unusual tactic of producing asexual propagules at the end of the growing season, as opposed to the more common tactic of producing sexually generated overwintering propagules ( Hughes, 1989; Lynch & Spitze, 1994). Ladle et al. (1993 ) proposed that asexual reproduction may be favourable under a number of conditions, including when populations conform to a metapopulation structure. Asexual production of statoblasts will result in a relatively large number of propagules and, in combination with clonal reproduction, will increase the likelihood of achieving successful dispersal and (re)colonization. This could result in enormous fitness benefits for a particular clone, as seen for example in Loirston Loch (Scotland) where a single clone dominated. Thus, a clone may reduce its risk of extinction by spreading over broad spatial and temporal scales ( Cook, 1979; Cooper, 1984). Since many clonal organisms occur as subdivided populations, we predict that further instances will be revealed in which dispersal of asexual units influences levels of genetic variation within demes.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices

We thank D. Buss, K. Conrad, P. Hammond, B. Hirst, C. Jones, L. Pontin, V. Rimmer, C. Sorin, M. Taylor, L. Uotila and J. Viitala for technical assistance. We are also grateful to K. Conrad and R. Sibly for useful comments on the manuscript, M. Pagel for advice on the computer simulation, K. Conrad for writing the computer simulation program and C. Romualdi for conducting the discriminant function analyses. This work was supported by a Natural Environmental Research Council grant (GR3/11068) to B.O. and L.R.N.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices
  • 1
    Amos, W. & Harwood, J. 1998. Factors affecting levels of genetic diversity in natural populations. Phil. Trans. Roy. Soc. Lond. 353: 177 186.
  • 2
    Ayre, D.J. & Willis, B.L. 1988. Populations structure in the coral Pavona cactus: Clonal genotypes show little phenotypic plasticity . Mar. Biol. 99: 495 505.
  • 3
    Bell, G.I. & Jurka, J. 1997. The length distribution of perfect dimer repetitive DNA is consistent with its evolution by an unbiased single-step mutation process. J. Mol. Evol. 44: 414 421.
  • 4
    Boileau, M.G. & Taylor, B.E. 1994. Chance events, habitat age and the genetic structure of pond populations. Arch. Hydrobiol. 132: 191 202.
  • 5
    Brown, C.J.D. 1933. A limnological study of certain fresh-water Polyzoa with special reference to their statoblasts. Trans. Am. Microsc. Soc. 52: 271 314.
  • 6
    Cabe, P.R. 1998. The effects of founding bottlenecks on genetic variation in the European starling (Sturnus vulgaris) in North America . Heredity 80: 519 525.
  • 7
    Callen, D.F., Thompson, A.D., Shen, Y., Philips, H.A., Richards, R.I., Mulley, J.C., Sutherland, G.R. 1993. Incidence and origin of ‘null’ alleles in the (AC) n microsatellite markers. Am. J. Human Genet 52: 922 927.
  • 8
    Charlesworth, B. & Charlesworth, D. 1987. Inbreeding depression and its evolutionary consequences. Ann. Rev. Ecol. Syst. 18: 237 268.
  • 9
    Cook, R.E. 1979. Asexual reproduction: a further consideration. Am. Nat. 113: 769 772.
  • 10
    Cooper, W.S. 1984. Expected time to extinction and the concept of fundamental fitness. J. Theor. Biol. 107: 603 629.
  • 11
    Crow, J.F. & Aoki, K. 1984. Group selection for a polygenic behavioral trait: estimating the degree of population subdivision. PNAS 81: 6073 6077.
  • 12
    Del Carmen Mandujano, M., Montana, M.C., Eguiarte, L.E. 1996. Reproductive ecology and inbreeding depression in Opuntia rastrera (Cactaceae) in the Chihuahuan desert: why are sexually derived recruitments so rare? Am. J. Bot. 83: 63 70.
  • 13
    Eanes, W.F. & Koehn, R.K. 1978. An analysis of genetic structure in the monarch butterfly Danaus plexippus L. Evolution 32: 784 797.
  • 14
    Eckert, C. & Barrett, S.C.H. 1994. Inbreeding depression in partially self-fertilizing Decodon verticillatus (Lythraceae): population-genetic and experimental analyses. Evolution 48: 952 964.
  • 15
    Felsenstein, J. 1995. PHYLIP(Phylogeny Inference Package) v. 3.5c. University of Washington. Distributed by the author.
  • 16
    Freeland, J.R., Jones, C.S., Noble, L.R., Okamura, B. 1999. Polymorphic microsatellite loci identified in the highly clonal freshwater bryozoan Cristatella mucedo. Mol. Ecol. 8: 341 342.
  • 17
    Gilpin, M.E. 1991. The genetic effective size of a metapopulation. Biol. J. Linn. Soc. 42: 165 175.
  • 18
    Goodman, S.J. 1997. R-st calc: a collection of computer programs for calculating estimates of genetic differentiation from microsatellite data and determining their significance. Mol. Ecol. 6: 881 885.
  • 19
    Grebelnyi, S.D. 1996. Influence of parthenogenetic reproduction on the genotypic constitution and evolutionary success of populations and species. Hydrobiology 320: 55 61.
  • 20
    Hanski, I. 1998. Metapopulation dynamics. Nature 396: 41 49.
  • 21
    Hanski, I. & Gilpin, M.E. 1991. Metapopulation dynamics: brief history and conceptual domain. Biol. J. Linn. Soc. 42: 3 16.
  • 22
    Hanski, I. & Simberloff, D. 1997. The metapopulation approach, its history, conceptual domain, and application to conservation. In: Metapopulation Biology. Ecology, Genetics, and Evolution (I. Hanski & M. E. Gilpin, eds), pp. 5–26. Academic Press, San Diego.
  • 23
    Harrison, S. 1991. Local extinction in a metapopulation context: an empirical evaluation. Biol. J. Linn. Soc. 42: 73 88.
  • 24
    Hastings, A. & Harrison, S. 1994. Metapopulation dynamics and genetics. Ann. Rev. Ecol. Syst. 25: 167 188.
  • 25
    Hatton-Ellis, T., Noble, L.R., Okamura, B. 1998. Genetic variation in a freshwater bryozoan. I: Populations in the Thames basin. Mol. Ecol. 7: 1575 1585.
  • 26
    Hughes, R.N. 1989. A Functional Biology of Clonal Animals. Chapman & Hall, London.
  • 27
    Hunter, C.L. 1993. Genotypic variation and clonal structure in coral populations with different disturbance histories. Evolution 47: 1213 1228.
  • 28
    Jones, C.S., Okamura, B., Noble, L.R. 1994. Parent and larval RAPD fingerprints reveal outcrossing in freshwater bryozoans. Mol. Ecol. 3: 193 199.
  • 29
    Ladle, R.J., Johnstone, R.A., Judson, O.P. 1993. Coevolutionary dynamics of sex in a metapopulation: escaping the Red Queen. Proc. Roy. Soc. Lond. 253: 155 160.
  • 30
    Levene, H. 1949. On a matching problem in genetics. Ann. Math. Stat. 20: 91 94.
  • 31
    Levins, R. 1969. Some demographic and genetic consequences of environmental heterogeneity for biological control. Bull. Ent. Soc. 15: 237 240.
  • 32
    Levins, R. 1970. Lecture Notes on Mathematics in the Life Sciences. In: Some Mathematical Questions in Biology (M. Gerstenhaber, ed.), pp. 75–107. The American Mathematical Society, Providence.
  • 33
    Li, C.C. 1976. First Course in Population Genetics. Boxwood Press, California.
  • 34
    Lymberry, A.J., Constantine, C.C., Thompson, R.C.A. 1997. Self-fertilization without genomic or population structuring in a parasitic tapeworm. Evolution 51: 289 294.
  • 35
    Lynch, M. & Spitze, K. 1994. Evolutionary genetics of Daphnia. In:Ecological Genetics (L. A. Real, ed.), pp. 109–128. Princeton University Press, Princeton.
  • 36
    Manly, B.F.J. 1994. Multivariate Statistical Methods, a Primer, 2nd edn. Chapman & Hall, London.
  • 37
    Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27: 209 220.
  • 38
    McCauley, D.E. 1989. Extinction, colonization, and population structure; a study of a milkweed beetle. Am. Nat. 134: 365 376.
  • 39
    Muirhead, C.A. & Lande, R. 1997. Inbreeding depression under joint selfing, outcrossing, and asexuality. Evolution 51: 1409 1415.
  • 40
    Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583 590.
  • 41
    Nei, M., Maruyama, T., Chakraborty, R. 1975. The bottleneck effect and genetic variablility in populations. Evolution 29: 1 10.
  • 42
    Neigel, J.E. 1997. A comparison of alternative strategies for estimating gene flow from genetic markers. Ann. Rev. Ecol. Syst. 28: 105 128.
  • 43
    Okamura, B. 1994. Variation in local populations of the freshwater bryozoan Cristatella mucedo. In: Biology and Palaeobiology of Bryozoans (P. J. Hayward, J. S. Ryland & P. D. Taylor, eds), pp. 145–149. Olsen and Olsen, Fredensborg, Denmark.
  • 44
    Okamura, B. 1997a. The ecology of subdivided populations of a clonal freshwater bryozoan in southern England. Arch. Hydrobiol. 141: 13 34.
  • 45
    Okamura, B. 1997b. Genetic similarity, parasitism, and metapopulation structure in a freshwater bryozoan. In: Evolutionary Ecology of Freshwater Animals (B. Streit, T. Stadler & C. M. Lively, eds), pp. 293–320. Birkhaüser Verlag, Basel.
  • 46
    Rayboud, A.F., Mogg, R.J., Aldam, C., Gliddon, C.J., Thorpe, R.S., Clarke, R.T. 1998. The genetic structure of sea beet (Beta vulgaris ssp. maritima) populations. III. Detection of isolation by distance at microsatellite loci . Heredity 80: 127 132.
  • 47
    Raymond, M. & Rousset, F. 1995. GENEPOP (Version 1.2) – Population genetics software for exact test and ecumenicism. J. Heredity 86: 248 249.
  • 48
    Scott, D.A. & Rose, P.M. 1996. Atlas of Anatidae Populations in Africa and Western Eurasia. Wetlands International, The Netherlands.
  • 49
    Slatkin, M. 1985. Gene flow in natural populations. Ann. Rev. Ecol. Syst. 16: 393 430.
  • 50
    Slatkin, M. 1987. Gene flow and the geographic structure of natural populations. Science 236: 787 792.
  • 51
    Slatkin, M. 1993. Isolation by distance in equilibrium and non-equilibrium populations. Evolution 47: 264 279.
  • 52
    Slatkin, M. 1995. A measure of population subdivision based on microsatellite allele frequencies. Genetics 139: 457 462.
  • 53
    Stoddart, J.A. 1984. Genetical structure within populations of the coral Pocillopora damicornis. Mar. Biol. 81: 19 30.
  • 54
    Tautz, D. 1989. Hypervariability of simple sequences as a general source for polymorphic DNA markers. N.A.R. 17: 6463 6471.
  • 55
    Taylor, D., Finston, T., Hebert, P.D.N. 1998. Biogeography of a widespread freshwater crustacean: pseudocongruence and cryptic endemism in the North American Daphnia laevis complex. Evolution 52: 1648 1670.
  • 56
    Uotila, L. & Jokela, J. 1995. Variation in reproductive characteristics of colonies of the freshwater bryozoan Cristatella mucedo. Freshwater Biol. 34: 513 522.
  • 57
    Vernon, J.G., Okamura, B., Jones, C.S., Noble, L.R. 1996. Temporal patterns of clonality and parasitism in a population of fresh-water bryozoans. Proc. Roy. Soc. Lond. 263: 1313 1318.
  • 58
    Viard, F., Justy, F., Jarne, P. 1997. The influence of self-fertilization and population dynamics on the genetic structure of subdivided populations: a case study using microsatellite markers in the freshwater snail Bulinus truncatus. Evolution 51: 1518 1528.
  • 59
    Wade, M.J. & McCauley D.E., 1988. Extinction and recolonization: their effect on the genetic differentiation of local populations. Evolution 42: 995 1005.
  • 60
    Weber, J.L. & Wong, C. 1993. Mutation of human short tandem repeats. Human Mol. Genetics 2: 1123 1128.
  • 61
    Weider, L.J., Hobæk, A., Crease, T., Stibor, H. 1996. Molecular characterization of clonal population structure and biogeography of arctic apomictic Daphnia from Greenland and Iceland. Mol. Ecol. 5: 107 118.
  • 62
    Weir, B.S. 1979. Inferences about linkage disequilibrium. Biometrics 35: 235 254.
  • 63
    Weir, B.S. & Cockerham, C.C. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358 1370.
  • 64
    Whitlock, M.C. & McCauley D.E., 1999. Indirect measures of gene flow and migration: F-st ≠ 1/ (4Nm + 1). Heredity 82: 117 125.
  • 65
    Wood, T.S. 1991. Bryozoans. In: Ecology and Classification of North American Freshwater Invertebrates (J. H. Thorp & A. P. Covich, eds), pp. 481–499. Academic Press, New York.
  • 66
    Wöss, E.R. 1994. Seasonal fluctuations of bryozoan populations in five water bodies with special emphasis on the life cycle of Plumatella fungosa (Pallas). In: Biology and Palaeobiology of Bryozoans (P. J. Hayward, J. S. Ryland & P. D. Taylor, eds), pp. 211–214. Olsen and Olsen, Fredensbörg, Denmark.
  • 67
    Wright, S. 1943. Isolation by distance. Genetics 28: 139 156.
  • 68
    Wright, S. 1951. The genetical structure of populations. Eugenics 15: 323 354.
  • 69
    Wright, S. 1969. The Theory of Gene Frequencies. University of Chicago Press, Chicago.
  • 70
    Wright, S. 1978. Evolution and the Genetics of Populations, Vol. 4, Variability Within and Among Natural Populations. University of Chicago Press, Chicago.
  • 71
    Yeh, F.C. & Boyle, T.J.B. 1997. Population genetic analysis of co-dominant and dominant markers and quantitative traits. Belgian J. Botany 129: 157157.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  9. Appendices


Allele frequencies of five microsatellite loci in 14 European populations. Alleles are identified by the size of the amplified product. Loci are further described in Freeland et al. (1999). Population 1 = Loirston Loch, Scotland, 2 = Tufty’s Corner, England, 3 = Etang du Gros Caillou, France, 4 = Lac de Grand Lieu, France, 5 = Recreational Park De Brabantse, Nationaal Park de Beisbosch, Netherlands, 6 = Herpen, Netherlands, 7 = Ry Mølle Sø, Denmark, 8 = Søndersø Nørresø, Denmark, 9 = Rössjön, Sweden, 10 = Sommen, Sweden, 11 = Littoistenjarvi, Finland, 12 = Enärjärvi, Finland, 13 = Keitele, Finland, 14 = Konnevesi, Finland. 2


(contd.) 3


(contd.) 4