The island of Tahiti has been scored by erosion so that a series of separated valleys radiate to the coast. From four successive valleys Mayer collected the species, and found that in the first (Tipaerui) valley all the shells were dextral (115, containing 73 young); in the second valley (Fautaua) 54 per cent. of adults and 55.5 per cent. of the young contained were sinistral; in the third valley (Hamuta) 69 per cent. of adults and 73 per cent. of young contained in them were sinistral; and lastly, in the fourth valley (Pirae) all the shells (131, containing 62 young) were sinistral.[18] In connection with these observations I may mention the fact that in a certain pond in the North of England[19] the sinistral form of _Limnaea peregra_ has been known to occur for about fifty years. Visiting it lately I found the left-handed shells to be about 3 per cent. of the population. The species is the commonest British freshwater shell, but left-handed specimens are exceedingly rare. Will anyone ask us to suppose that the persistence of a percentage of this rarity in the same place is an indication of some specially favouring circumstance in the waters of that pond? It is a horse-pond to all appearances exactly like any other horse-pond; and I believe that in perfect confidence we may accept the suggestion of common sense, which teaches us that there is nothing particular in the circumstances which either calls such varieties into existence or contributes in any direct way to their survival. Had the phenomenon of local variation been studied in detail before Darwin wrote, the attempt to make selection responsible for fixity wherever found, could never have been made. The proposition that not only the definiteness of local forms but their variability also is sporadic, can be established by countless illustrations taken from any group of either the animal or the vegetable kingdoms. Only exceptionally can the fixed differences be even suspected of contributing to adaptation, and sporadic variability, which is a no less positive fact, must manifestly lie outside the range of such suspicions. It is open to any one to suggest speculatively that the persistence of special varieties or of special variability in special places is an indication that in those places the conditions of life are such that the forms in question are tolerated though elsewhere the same types are exterminated; but that consideration, even if it could be proved to be well founded, is not one which lends much force to the thesis that definiteness of type is a consequence of Natural Selection. On the contrary, recourse to such reasoning implies the inevitable but very damaging admission that the stringency of Selection is frequently so far relaxed that two or more equally definite forms of the same species can persist side by side. There is no doubt that this is the simple truth, but when once that truth is perceived it is useless to invoke the control of Selection as