We thus see that while the stem inhibits the growth of the leaves connected with it, the latter accelerate the growth in the stem. Both facts can probably be explained on the same basis; namely, on the assumption that it is the flow of substances from the leaf to the stem which inhibits the growth of the notches and accelerates the growth of the buds in the stem. On this assumption it would also follow that the leaves send root-forming substances towards the basal and shoot-forming substances towards the apex of the stem. It also seems to follow from recent as yet unpublished experiments by the writer that the root-forming substances are associated or identical with the substances which cause geotropic curvature in the stem. These observations show that the phenomena of correlation or of the influence of the whole over the parts is due to peculiarities of circulation or the flow of sap; and that the isolation prevents the sap from flowing away to other parts of the plant. There is no need for assuming the existence of a mysterious force which directs the piece to grow into a whole. [Illustration: FIG. 21] 3. Phenomena of inhibition or correlation such as we have described in _Bryophyllum_ are not lacking in the regeneration of animals, as experiments on _Tubularia_ show.[154] _Tubularia mesembryanthemum_ (Fig. 21) is a hydroid consisting of a long stem terminating at one end in a stolon which attaches itself to solid bodies such as rocks, at the other end in a polyp. The writer found that if we cut a piece from a stolon and suspend it in an aquarium it forms as a rule a polyp at either end (Fig. 22), but the velocity with which the two polyps are formed is not the same, the polyp at the oral end of the piece being formed much more rapidly--a day or one or two weeks sooner--than the aboral polyp. The process of polyp regeneration at the aboral pole could, however, be accelerated and its velocity made equal to that of the regeneration of the oral polyp by suppressing the formation of the latter. This was accomplished by depriving the oral pole of the oxygen necessary for regeneration, _e. g._, by merely putting the oral end of the piece of stem into the sand. It was, therefore, obvious that the formation of the oral polyp retarded the formation of the aboral polyp. This inhibition might have been due to the fact that a specific organ-forming material needed for the formation of a polyp existed in sufficient quantity in the stem for the formation of one polyp only at a time. This idea, however, was found to be incorrect since when the stem was cut into two or more pieces each piece formed a polyp at once at its oral pole and regenerated the aboral polyps also, but again with the usual delay. It seemed more probable then that the cause of the difference in the rapidity of polyp formation at both ends lay in the fact that certain material flowed first to the oral pole and induced polyp formation here but that this flow was reversed as soon as