A NOTE ON THE SPEED OF GENE FREQUENCY CHANGES IN REVERSE DIRECTIONS IN A FINITE POPULATION

this same issue. We should be interested to see what kind of results his method would give on more complex data (Lakovaara et al., 1972b). In principle the method of Dr. Farris is correct and his phylogenetic tree is the best fitting one for our distance matrix. Another question is how well his biochemical tree agrees with morphological and chromosome data. As Dr. Farris notes, his phylogenetic tree differs from ours primarily in resolving the old world species into two distinct groups, rather than chaining them together in a linear series. In his tree D. bijasciata is on the same main branch with the American species. Its deviation from the Eurasian branch is well within the limits of the standard error of D (d. our Table 2, Lakovaara et al., 1972a) . As for the American species, D. persimilis, D. pseudoobscura, D. miranda and D. lowei, the two proposed phylogenetic trees are identical and agree well with chromosome structure and morphological data. Morphology and the chromosome structure of the Old World species, however, fit poorly the phylogenetic tree of Dr. Farris. D. subobscura and D. alpina have acrocentric chromosomes, whereas the long metacentric chromosomes of the rest of the Old World species represent an apparently monophyletic condition, which is probably derived from D. subobscura-like ancestors. Morphologically these species (D. bifasciata, D. tristis, D. ambigua and D. subsilvestris) also form a group clearly different from D. alpina and D. subobscura. These two species are morphologically quite different from each other. The phylogenetic tree proposed by us is more in accord with the available morphological and chromosomal evidence. Considering the standard errors given in our Table 2 the conclusions of Dr. Farris appear too rigorous, for many more phylogenetic trees could be drawn within these errors. While it is obvious that the rates of evolution for the species of the D. obscure group have been different, the phylogenetic tree proposed by Dr. Farris hardly gives a correct picture of these rates for our data. Comparing phylogenetic trees constructed by different methods will provide valuable insight on the processes of evolution and speciation.