For Japanese, English, Brazilian and Austrian populations, estimates of genetic isolate size based on different types of consanguineous marriage range over more than one order of magnitude, being much greater for uncle-niece and aunt-nephew marriage than for first-cousin marriage, with second cousins giving estimates about 10 times greater than first cousins. These discrepancies agree with the probable non-randomness of consanguineous marriage, and are so great that little reliance can be placed on estimates of isolate size from consanguineous marriage.

It is known that for rare partially sex-linked recessive traits the ratio of affected females to affected males depends on the proportions of four types of first cousin matings. Data from Austrian and Japanese populations reveal three significant kinds of non-randomness which, however, would have very little effect on the expected frequency of females among partially sex-linked recessives, if such genes exist in man.

For the populations sampled, non-randomness is shown to persist in the different mating types of first cousins once removed, second cousins once removed, and second cousins.