presvanab ly contains vety Friv biologically related persons, Elsewhere wee have argued that Eis appropriate to compare the average nunbar of Variants per sys Ken per thousand determinations ina populatiuas ascorta tic in th-se different fashions, but mot the number of different variants (2), and we maintaia that position here. The logie is that a rather -considerable’ shuffling of tribal populations, such as occurred in the dutribalization of the ancestors of modern Japanese or Europeans, should not alter the total number of variants present. . However, conveational : ' : statistical contrasts of total frequencies in these various ‘populations seem inappropriate: In particular, estimates of frequencies from studies of populations such as Micronesians or Amerindiangs are quite suscenotible to 2 "jackpot! ef£ect--one igland with a high frequency variant could markedly alter the picture. However, some 2/7 islands have been sampled ta dake, conprising a rather representative group. Taken at face value, there fs a 3-fold range among ethnic groups in the frequency of rare variants as defined, these varlants occurring in ticronesians with approximately half the frequency in which they have been encountered in several other ethnic groups. However, this apparent difference hinges, in part, on the definition of rare variant, which in turn is intimately related to the size and nature of the sample. thus, ££ as the study of Micronesians is extended it becomes clear that the PCM, ~3-I ike variant encountered in 49 persons (12) variant encuimtered in 39 persons (12) and the PGK-2-like are sharply localized and no or few additional examples of the variant are encountered, then either or both of these might drop below the arbitrary 2 percent frequency level add have tu be classed as rare variants, with a marked Impact on the average af 9009534