ML versus Bayes: 100 data sets of 1 single artificial locus each. 4 population exchange migrants round robin from 1 to 4 to 3 to 2 to 1. In each population a sample of 20 individuals with 10000 bp each was taken. The data sets are in MIGRATE format. (data here)(paper here) [It would be very interesting for me and practical researchers, if these data set would be run with programs that compete with migrate: please let me know about such trials.]

Homburg: Dataset from my master thesis discussing the strange occurrence of a species (Rana ridibunda) of waterfrogs in a parent-hybrid system (Rana lessonae and R. esculenta). The data set is a mixture of categorical and quantitative data.(data here)(Master thesis here)

Anatolia (Data set of 5 populations of water frogs): Allozyme data of 31 loci (data here), this data set contains comments (version of migrate 2.1.8 or newer can read this with problems, for older versions or other converters you need to excise all lines that start with #. (data here)(paper here)

Simulated and real data sets (data here)(paper here)