Another reply to my question regarding significance testing of allelic richness: Doing a straightforward permutation test by randomising alleles among samples is not useful when you want to test a diversity index such as the allelic richness. The reason for that is that the alleles may or may not be comparable between populations, and this influences the test. A test between populations with richnesses 10 and 20 that share a lot of the alleles, will likely give a different result than a test between populations with richnesses 10 and 20 that have no alleles in common. In the latter case, the average richnes after randomisation is likely to be even higher than 20, the maximum of the two populations. It is however possible to do a bootstrap test, though this also has some (smaller) complications. See Thomas MG, Weale ME, Jones AL, et al. (2002) Founding mothers of Jewish communities: Geographically separated Jewish groups were independently founded by very few female ancestors. American Journal of Human Genetics 70, 1411-1420. I wrote a small app (GenoDive) to test for differences in diversity between pairs of populations, it does not calculate allelic richness (though that probably depends on your definition of allelic richness) but some other indices of diversity like Nei's (=Simpson's) index and the Shannon (=Wiener/Weiner) index. The program was not made for allelic data, but for clonal genotypes, though it can handle alleles if you just code every allele as a genotype (with an added assumption of HW-equilibrium as that breaks down the conenction between the two alleles at a locus). My program can be found at http://staff.science.uva.nl/~meirmans/ Good luck with your analysis! Patrick Meirmans: meirmans@science.uva.nl Tim Jones