Previous work from this laboratory demonstrated a relation between ability to regenerate a limb and number of nerve fibers available per unit area of amputation wound. Animals with fewer fibers than the newt, Triturus, (for example, Rana, Anolis, and Mus) do not regrow the limb. An exception appeared in Xenopus whose limb is sparsely innervated and yet which regenerates. The present work demonstrates that the nerve fibers of Xenopus make up in individual size what they lack in number. When the average cross-sectional area of the axons was multiplied by the number of fibers per unit area, the results showed that the Xenopus limb is about as well supplied with axoplasm as that of Triturus. This was not true for Rana. The results also demonstrate for the first time a direct relation between caliber and trophic effectiveness of the axon, the evolutionary significance of which is discussed.In previous studies from this laboratory on the dependence of limb regeneration upon the nerve, a direct relation was drawn between the number of nerve fibers available at the amputation surface of the forelimb and the ability of the stump to regrow a new limb (see review, Singer, '52). In the newt, Triturus, the number of nerve fibers normally available per unit area of amputation surface is more than sufficient to cause regrowth. However, if this number is reduced below a critical threshold, by selective destruction of the brachial nerves or nerve roots, regeneration does not ensue. In forms which do not regenerate the limb, for example the frog, Rana, the lizard, Anolis, and the mouse, the number of nerve fibers per unit amputation area is substantially less than that found in the limb of the newt; indeed, it is within or is less than the range of nerve fibers required for regeneration of the newt's limb (Rzehak and Singer, '66, '66a). One of us (Singer, '51. '54, '61) also found for Rana and Anolis that, if the number of nerve fibers in the limb is augmented by deviating an extraneous large nerve into the limb, regeneration occurs presumably because the nerve fiber threshold for limb regeneration is now satisfied (see also Simpson, '61 ; Kudokotsev, '62).An exception appeared to the rule that quantity of innervation is directly related to ability of the limb to regenerate (Rzehak J. EXP. ZOOL., 166: 89-98.and Singer, '66a). We observed that the number of fibers normally available at the amputation of the upper arm is much less in the adult African clawed frog, Xenopus, than in Triturus; and, indeed, substantially less than the threshold number required for regeneration in Triturus (see fig. 2, left). Yet, the limb of Xenopus did regenerate. Initial histological observations on the nerves of Xenopus suggested an answer to this paradox. The peripheral fibers are in general much larger than those of any other form we studied. This being the case, it must follow that the amount of axoplasm per nerve fiber is much greater at the amputation surface in Xenopus than in Triturus. If trophic activity of the axon is directly relate...