this study was conducted to test the null hypothesis that tamarack (larix laricina (du roi) k. koch) populations are random assemblages of genotypes, and that this absence of pattern can be observed on sites with differing ecological and demographic characteristics. a total of 1715 trees in four populations with distinct ecological and demographic characteristics were surveyed and sampled for isozyme analysis. these trees were mapped for later plotting and spatial autocorrelation analysis. seven variable and three monomorphic allozyme loci were resolved. visual examination of the distribution of single alleles over space revealed pattern in 22.2 percent of the 36 plots. spatial autocorrelation analysis resulted in calculation of 313 moran's i autocorrelation coefficients, 8.9 percent of which were significant (95 percent confidence level). in addition, results of tests of the correlograms constructed from these 313 coefficients revealed that 38.9 percent of the correlograms were significant using bonferroni's criterion (90 percent confidence level). these results suggested that the null hypothesis be rejected for a modest, but still important proportion of the tests. the spatial pattern that was observed, both by eye and through statistical tests, was manifested mainly as a tendency for similar genotypes to occur in trees that were near to each other.