must perish and leave no offspring, yet some of his descendants, because they have varied extremely favourably, must, when there is a struggle for existence, other things equal, survive in greater numbers, and cause the ultimate elimination of the offspring of the second. Now, of two individuals, one of which produces offspring without collaboration with another individual, and the other in collaboration with another and somewhat dissimilar individual, the offspring of the latter must tend to vary more than the offspring of the former, and therefore in the struggle for existence to bring about their ultimate extinction, whereby only the descendants of the individual that reproduced sexually, and which inherited this peculiarity, would be left to continue the race. Weismann's hypothesis therefore appears to be reasonable, and it is moreover supported by other considerations, for if we assume, as we must, that nonliving chemical compounds, under conditions we are ignorant of, did in the beginning of life pass over the border space and become living beings, it is difficult to imagine how they could possibly have multiplied sexually from the very first—how sexual union can have been the rule in the beginning. It seems more reasonable to suppose that these earliest forms multiplied asexually by fission, and that conjugation only occurred later, the purpose of it being to produce a greater amount of variability among the offspring. Possibly conjugation had its origin in attempts at cannibalism, which, among such excessively simple types as the first must have been, perhaps resulted in coalescence instead of assimilation. The extermination of those that multiplied asexually appears therefore to be the cause of the almost universal prevalence of sexual reproduction among low as well as among high organisms. If it be asked, Why, if sexual reproduction produces variability, and variability is of such importance, do low
