importance in the economy of the Medusa that its marginal tentacles should be very sensitive to contact with minute particles, so that a very slight stimulus applied to them should start the co-ordinated movements of localization, it is not surprising that the tentacular rim should present nerve-endings so far sensitive that only by their excitation can the reflex mechanism be thrown into action. But if such is the explanation in this case, it is curious that in Tiaropsis indicans every part of the bell should be equally capable of yielding a stimulus to a precisely similar reflex action. "In pursuance of this point, I tried the experiment of cutting off _portions_ of the margin, and stimulating the bell _above the portions of the margin which I had removed_. I found that in this case the manubrium did not remain passive as it did when the _whole_ margin of the bell was removed; but that it made ineffectual efforts to find the offending body, and in doing so always touched some part of the margin which was still unmutilated. I can only explain this fact by supposing that the stimulus supplied to the mutilated part is spread over the bell, and falsely referred by the manubrium to some part of the sensitive--_i.e._ unmutilated--margin. "But to complete this account of the localizing movements, it is necessary to state one additional fact which, for the sake of clearness, I have hitherto omitted. If any one of the four radial tubes is irritated, the manubrium will correctly localize the seat of irritation, whether or not the margin of the bell has been previously removed. This greater case, so to speak, of localizing stimuli in the course of the radial tubes than anywhere else in the nectocalyx, except the margin, corresponds with what I found to be the case in Tiaropsis indicans and probably has a direct reference to the distribution of the principal nerve-tracts. "On the whole, therefore, contrasting this case of localization with the closely parallel case presented by Tiaropsis indicans, I should say that the two chiefly differ in the fresh-water Medusa, even when unmutilated, not being able to localize so promptly or so certainly, and in the localization being only performed with reference to the margin and radial tubes, instead of with reference to the whole excitable surface of the animal.