The middle ear, its sound-transmitting mechanism, and the tympanum, well developed in most Anura, are readily derived from those of early labyrinthodonts, and are presumably effective for hearing airborne sounds whether on land or while floating in the water. Reduction of these organs in Urodela may be correlated with their customary restriction to subsurface habitats and inability to maintain a floating position while in water. Some light may be shed on the significance of the tadpole of Anura by considering the early stages of the ribbed frogs, Liopelmidae. _Leiopelma_ and _Ascaphus_ are so closely similar in the adult that there is no doubt that they belong in one family, primitive in some respects (including articulated ribs; pyriformis and caudalipuboischiotibialis muscles) but not in others (absence of tympanum and middle ear). In both genera the eggs are large, 5 mm. in _Leiopelma_, 4.5 mm. in _Ascaphus_, and unpigmented; but at this point the resemblance ends. [Illustration: Fig. 10. _Leiopelma hochstetteri_ larva, lateral and ventral (after Stephenson, 1955), ×4.] Stephenson (1955) showed that embryos of _L. hochstetteri_ develop equally well on land (in damp places) or in the water, and that embryos prematurely released from egg capsules develop successfully in the water. The larvae possess both pairs of legs (Fig. 10) and a broad gular fold similar to that of larval salamanders. In _L. hochstetteri_ the fold grows back over the forelegs temporarily, but remains free from the body and presently the legs reappear, whereas in _L. archeyi_ the forelegs are not covered at any time. No branchial chamber or spiracle is formed. Of course direct development, without a tadpole, occurs in several other groups of Anura, but in each case terrestrial adaptations are obvious. This is not true of _Leiopelma_, which Stephenson regards as more nearly comparable with Urodela in its development than with other Anura, and he sees in it a "primary and amphibious" mode instead of a terrestrial specialization. The _Ascaphus_ tadpole bears no outward resemblance to the larva of _Leiopelma_, but is a normal tadpole in form, although sluggish in activity. Its greatly expanded labial folds bear numerous rows of horny epidermal "teeth," which, with the lips, serve to anchor the tadpole to stones in the swift water of mountain brooks. Noble (1927) noticed that particles of food were taken in through the external nares, and that a current of water passed through these openings and out by way of the median spiracle. It appears that any action by the teeth and jaws in scraping algae from the rocks (which were bare in the stream where I have collected _Ascaphus_) would be quite incidental, and that the lips and teeth must be primarily a clinging mechanism. Certain other mountain brook tadpoles (for example, _Borborocoetes_) show similar devices, but these are developed independently, as specializations from the usual sort of tadpole.