Association of amphimixis with reproduction--Origin of amphimixis--Its lowest forms--Amphimixis in Coccidia--Chromosomes in unicellular organisms--_Coccidium proprium_--'Amœba-nests' as a preliminary stage to amphimixis--Plastogamy of the Myxomycetes--Result: a strengthening of the power of adaptation--Strengthening of the power of assimilation--Use of complete amphimixis--Proof of its constant efficacy to be found in the rudimentary organs of Man--Allogamy--Means taken to prevent the mingling of nearly related forms--Amphimixis is not a 'formative' stimulus--Attraction of the germ-cells--Effects of inbreeding compared with those of parthenogenesis--Nathusius's case of injurious inbreeding--Hindrances to fertilization in the crossing of species--Probable reason for the injurious effect of inbreeding. We have endeavoured to understand why amphimixis should have been established among the processes of life, and we have now to turn to the question when and how, that is, in what form, it was first introduced. But first I should like to refer for a little to the association of amphimixis with reproduction, which we find in all multicellular organisms, and among the higher types so unexceptionally that, until not very long ago, amphimixis and reproduction were looked on as one and the same thing, and all multiplication was believed to be associated with 'fertilization.' We have seen that this is not the case, that on the other hand the two processes are quite distinct, and may be called contrasts rather than equivalents, for reproduction always means an increase in the number of individuals, while amphimixis implies--originally at least--their diminution by a half. Accordingly we found that, in unicellular organisms, amphimixis is not associated with reproduction, but interpolated between the divisions, and not even in such a manner that amphimixis precedes every multiplication by division, but so that the conjugation of two animals occurs only from time to time, after numerous divisions, sometimes hundreds, have occurred. It is obvious that this must be so, since, if amphimixis occurred regularly between every two divisions, no increase in the number of individuals would be brought about, at least if the fusion of the two conjugating individuals were complete. Why, then, is there such an intimate, and in the case of the higher types, such an indissoluble, association between reproduction and amphimixis that 'fertilization' appears to be a _sine qua non_ of reproduction, and not very long ago seemed to us to be the 'quickening of the ovum,' the 'burning spark' which causes the powder-barrel to explode?