Another series of experiments in grafting, similar to one of those made by Joest and myself on the earthworm, has been made by Harrison on the tadpole. I have also later made similar experiments. Two tadpoles are united by their posterior ends, as shown in Fig. 54, _A_, and a day or two after union one of the tails is cut off near the line of union. There is thus left attached to the end of the tail of one tadpole a part of the tail of the other united in a reverse direction, so that the exposed cut-end is the anterior end of the small piece. There grows out from this cut-end a structure that resembles a tail (Fig. 54, _B_, _C_, _D_). It contains a continuation of the notochord and nerve-cord, that taper in a characteristic way to the end of the new structure. The tail is flat and has a central band of muscle tissue, and a dorsal and ventral fin. The muscles of the normal tail have a characteristic V-shaped arrangement with the apex of the V’s turned forward, but unfortunately in the new tail the muscles are so irregular that it is impossible to make out their arrangement (Fig. 54, _D_). If the new part is in reality a tail, the V’s ought to stand in the same way as do those in the major component, and opposed to the V’s on the part from which the new material arises. If the new structure is not a tail at all, but a new growth, or even a suppressed trunk, then the V’s should stand as in the small part itself. It has not been possible as yet to obtain a decisive case. Harrison obtained one case in which the arrangement of the muscles in the new part seemed to be more as it should appear if the new part is a heteromorphic tail (Fig. 54, _D_). Even if this could be shown to be the case, it may be that under the conditions of the experiment the arrangement of the muscles is determined by the use of the tail, although this does not seem very probable. Harrison, after a careful analysis of the question, left it undecided, but seemed more inclined to the view that the result is due to the development of something new rather than a heteromorphic growth. On the contrary I am strongly inclined to believe that the latter is the true explanation. In another way I have been able to bring about the development of the same structure. A small triangular piece is cut from the upper part of the tail, as indicated in Fig. 56, _A_, one point of the triangle passing through the notochord, or even through the aorta. If the cut-surfaces are kept apart for a few hours, until the exposed end has been covered over by ectoderm, they may not unite afterward, and two exposed surfaces are left,--one at the distal end of the base of the tail, and the other at the proximal end of the outer part of the tail. The latter surface corresponds to that in the grafting-experiment. Regeneration may take place from the two surfaces; both new parts seem to be exactly alike, and both resemble a regenerated tail. The one from the proximal surface