The predictions of neutralist and selectionist hypotheses have been tested many times in the past, but mostly using data only from organisms such as vertebrates, with generally low to average heterozygosities. The more recent discovery of particularly high levels of genetic variation in marine sponges and coelenterates provides an opportunity to use data from such species to contribute further to the understanding of the determinants of heterozygosity in natural populations. Therefore, 23 species of sponges and coelenterates from temperate, tropical and boreal waters were analysed by gel electrophoresis for an average of 14.3 enzyme loci per species.
Mean heterozygosity values for each species were unusually high, ranging between 0.106 and 0.401. The means and variances of the heterozygosity estimates showed reasonable correlation with neutralist predictions (with both the stepwise mutation and the infinite alleles models). Population sizes were generally difficult to estimate with any confidence, but, for one sponge species for which this was possible, levels of heterozygosity again were similar to neutralist predictions, although the same was not apparently true for three species of sea anemone. No differences were found between heterozygosity levels of tropical and temperate species of sponges and coelenterates, thus apparently contradicting the selectionist ‘trophic resource stability’ and ‘temporal environmental variation’ hypotheses. Conversely, however, the consistently high levels of genetic variation found in coelenterates and sponges may be argued to be related to common biological characteristics, such as sessile life, great evolutionary ‘age’, limited ability to disperse and probable low homoeostatic capability.
Our results seem, overall, to agree well with neutralist expectations for species with large, stable population sizes. Also, the mean heterozygosities, their variances and the observed and expected proportions of polymorphic loci seem to fit well with predictions based on the neutralist hypothesis. However, the selectionist ‘environmental grain’ and the ‘shifting balance’ hypotheses fit the data equally well. As with much earlier work, the problems in distinguishing between the various predictions of selectionist or neutralist ideas make it both difficult and unwise to draw definite conclusions.