Though so recognised an authority as Bates is responsible for the statement that the fecundity of the Amazonian Indians is of a low degree, because as many as four children in one family are rarely found, it is open to doubt whether he and his successors have not in this instance confounded effect and cause. It is certainly true that the normal number for a family is but two or three, yet that this is not a question of fertility the high percentage of pregnant women would seem to disprove. The numbers are remarkable in view of the fact that husbands abstain from any intercourse with their wives, not only during pregnancy but also throughout the period of lactation—far more prolonged with them than with Europeans. The result is that two and a half years between each child is the minimum difference of age, and in the majority of cases it is even greater.

The main reason why there are these limited families is, in my opinion, not a diminishing birth-rate, but an enormously high percentage of infant mortality. The test of the survival of the fittest is applied to the young Indian at the very moment of his birth, for the infant is immediately submerged in the nearest stream, a custom that easily leads to infanticide in the case of an unwanted child, or one with any apparent deformity.