A chemical mechanism cannot, then, be imagined, much less described, and the only other mechanism so far suggested is the Roux-Weismann one, involving the disintegration of the determinants supposed to be present in the egg nucleus. Let us suppose (in spite of the incredible difficulty in so doing) that there is such a mechanism. It must usher the nuclei containing the determinants of the embryonic structure into their places: those for the formation of the nerve-centre go forward; those for the mouth, gut, and anus go backwards and downwards; those for the arms go forwards, ventrally, and posteriorly, in a very definite way; and those for the complicated skeleton are distributed in a variety of directions which defy description. These nuclei are, in short, moved up and down, right and left, backwards and forwards, and become built up into a complicated architecture. Suppose we prevent this. Suppose we compress the segmenting egg between glass plates so that the nuclei are compelled to distribute themselves in one plane only: to form a flattened disc in which the only directions are right and left and anterior and posterior. This has been done by Driesch and others. On the Roux-Weismann original hypothesis a monstrous larva ought to result, for the first nuclei separated from each other have been forced into positions altogether different from those which they should have occupied had they developed normally. Yet on releasing the pressure readjustment takes place. New divisions occur so as to restore the normal form of larva. The Roux-Weismann subsidiary hypothesis is that the stimulus of the pressure has compelled the nuclei to divide at first in such a way as to compensate for the disturbance. Let us remove some of the blastomeres. On the original hypothesis the determinants for the structures which the nuclei of these blastomeres contained have been lost. These structures should, therefore, be missing in the embryo. But nothing of the sort is the result. Other nuclei divide and replace the lost ones, and the embryo develops as in the normal mode. The reply is that in addition to the determinants which were necessary for their own peculiar function, these nuclei contained a reserve of all others. On disturbance these determinants, “latent” in all other conditions, became active and restituted the lost parts.