support a frog fauna, as the hearty croakings of the introduced Hyla aurea testify; second, a zone from about 1,000 to 1,500 feet high, where, in summer, the upper courses of the streams are dry, as also are the hill-spurs, an area ill-suited to the well-being of frogs; and third, the damp mist-swept hill-tops moist enough all the year round for frogs to live there in comfort.”

At Tokatea in December, 1938, I found L. archeyi restricted to the topmost zone, where it was breeding. At a lower altitude on the eastern side of the range, but still only several hundred yards from the ridge-top, L. hochstetteri was common in wet gold-mining drives and stream-beds. Thus here, although L. archeyi and L. hochstetteri live near each other, the narrow barren zone between the two species must be a material barrier if only during the dry period of the breeding season in November and December.

Archey assumes that the water-breeding ancestor of L. archeyi was widespread over the Coromandel Peninsula during a period of greater elevation with greater precipitation, and that the mist-swept, waterless topmost zone, which favoured the evolution of direct development and the loss of webs, was separated by a subsequent lowering of elevation.

It is yet to be discovered whether the larva of L. hochstetteri develops directly as in L. archeyi or passes through a more primitive aquatic stage comparable with that of Ascaphus (Noble, 1927, 63; Noble and Putnam, 1931).

The discovery of the former in L. hochstetteri would indicate that this species became distinct while the ancestral stock was still widespread, but at some time after the replacement of aquatic by direct development. The adult L.hochstetteri would in this case have reverted secondarily to the edges of streams, at some stage before the complete loss of webs. L. hochstetteri and L. archeyi (and probably L. hamiltoni) would then be closely related and equally close to the stage in leiopelmid evolution at which larval independence of surface water was first attained. This conclusion would also be in accordance with their general morphological resemblance when adult.

Should an aquatic larva be discovered in L. hochstetteri, this species would represent closely the adult and larval stages respectively of a widespread water-breeding Leiopelma ancestral to L. archeyi (and probably to L. hamiltoni). The association of the adult L. hochstetteri with surface water and its distribution far beyond the range of L. archeyi, which is known to have direct development, are two facts which would support this view.