But though it is designed to complete and to deepen our analysis, the present considerations may yet be said to mark a point of rest in the whole of our discussions: we have followed one single line of argumentation from the beginning until now; this line or this stream of thought, as you might call it, is now to break into different branches for a while, as if it had entered from a rocky defile into a plain. It seems to me that such a short rest will be not unconducive to a right understanding of all we have made out; and such a full and real conceiving again, such a realising of our problems of morphogenesis and their solutions, will be the best preparation for the philosophical part of these lectures. HARMONIOUS-EQUIPOTENTIAL SYSTEMS FORMED BY WANDERING CELLS All of the harmonious-equipotential systems which we have studied so far were the bases of histological differentiation; that is to say, the processes of their differentiation consisted in specifically localised elements of theirs becoming different *in situ*. Now we know at least one type of systems which also may be called harmonious-equipotential, but the differentiation of which does not simply relate to elements at a fixed place. An additional phenomenon enters here into the sphere of the others. The elements not only become different where they are, but a specific changing of locality, a specific kind of wandering, goes hand-in-hand with differences relating to the prospective value to be attained. I am speaking of the formation of the larval skeleton of our well-known Echinus. We know that the mesenchyme cells, which have left the blastoderm and are arranged in a sort of ring of bilateral structure, are the starting-point of this skeleton: it indeed originates in a sort of secretive process on the part of the cells; the cells are moving about and are secreting carbonate of lime during their wandering. The experiments now have shown, as we know, that a whole, though smaller, skeleton may also be formed, if only a half or a quarter of the mesenchyme cells are present, as happens to be the case in all experiments with isolated blastomeres of the two or four-cell stage of cleavage. It is clear that in these cases the performance of each single cell must be different from what it is in the normal case, and that the same sort of differences in the morphogenetic performances appears again, if the two- and the four-cell stage are compared with each other. And there are still some other phenomena showing the possibility of different performances being carried out by the individual cells. Peter has shown that the number of mesenchyme cells may vary enormously under certain conditions; but, in spite of that, the skeleton always will be complete. It may be said that this line of research is only of a relative value to our own questions, as, of course, variability relates to different individuals: but it seems to me that it adds a very good