Hello again. Thanks greatly for replies: David Remington, Martin Lascoux, Alan Meerow, Guy Hoelzer, Carlos, Maciek, Steve Keller, Anna-Karin, Miguel Angel Toro Ib��ez. Here I compile all replies and, as suggested by Alan Meerow and Steve Keller, I give some additional information important for final conclusion. Picea omorika is tertiary relict and endemic, widely distributed before, now app 40 scattered natural populations, up to several hundred trees each. Pollen and seed dispersed by wind, probably limited (distances not known). Outcrossing but self-fertile, inbreeding depression in later life stages. I don�t have any information if non-random matting is present, vegetative propagation present but not significant (Alan Meerow). It is not morphologically uniform and there is no lack of genetic diversity (surprisingly, but true). Scores for SSRs as well as for one mt marker highly repeatable, at least 2 alleles/locus sequenced (Maciek suggested technical artefact). There are for sure at least 2 subpop, because when I run STRUCTURE with K20 (absurd), one group of trees stick together. They are recent immigrants (based on presence of certain alleles), but I can not pinpoint source population because of mt haplotypes (mt DNA transferred by seed). Probably I do not have this population(s) in my sample set. Another fact � mt haplotypes in whole pop are population specific (!). David Remington suggested that if source pop is almost always found in admixed individuals, substructure is probably some sort of artifact even if it seems well-supported statistically. Doubts that I still have are because of the rest of the population � should I consider the rest of the population as 1 group or 2. As suggested by Martin Lascoux, Guy Hoelzer, Carlos, maybe the rest of the population is family (due to inbreeding), one or 2 sibships with LD between loci and deviations from HWE proportions. Solution � InStruct instead of STRUCTURE and Colony to check that. On the other hand, if there are 2 more subpop, in each of them LD in only between 1 or 2 pairs of loci and deviations from HWE proportions are almost insignificant (STRUCTURE actually did clustering based on this). There is excess of heterozygotes in each of them! Maybe there is non-random matting after all, but I�m reluctant to assume that. Steve Keller�s suggestion is recent admixture event with colonists which were drawn from 3 previously isolated populations (for 3 subpop). Maybe that is the case � several reinforcements, but only from certain source populations (due to presence of mt haplotypes) and only now and that, due to certain environmental conditions eg storms (spatial pattern of mt haplotypes is in waves!). Finally, I still have to check few more things, but if you have more suggestions, please sent them again. THANK YOU!!! Ref from Anna-Karin: Extensive spatial genetic strucutre revealed by AFLP but not SSR molecular markers in the wind-pollinated tree, Fagus sylvatica. A S Jump and J Penuelas, Mol Ecol 2007, 16, 925-936. Ref from Miguel Angel Toro Ib��ez: Heredity 99: 374-380 (2007). Jelena BFW Vienna aleksic_jelena@yahoo.com.au Jelena aleksic_jelena@yahoo.com.au