The development of the inner ear of Xenopus laevis.

S ummary .1 From a study of the progressive development of the membranous labyrinth through various larval stages of Xenopus laevis , it is noted that, in essentials, the parts resemble those of the Phaneroglossa. 2 The semicircular canals are separated after the constriction of the auditory vesicle into a pars superior and a pars inferior. The horizontal canal is formed before the anterior vertical, the posterior vertical being the last to appear. This sequence agrees with that of some of the Phaneroglossa (Herter, 1921). The pars inferior produces the sacculus, the pars neglecta, the lagena and the pars basilaris, there being no tegumentum vasculosum as in Rana (Gaupp, 1904). 3 Several membranous structures of somewhat doubtful physiological significance were observed to be associated with the foramen utriculo‐sacculare, the sacculo‐endolymphatic orifice and the aditus partis neglectae. Probably, like the utriculo‐saccular valve of fishes (Pearson, 1936) and the utriculoendolymphatic valve of mammals (Bast, 1928), they serve to maintain an approximately constant pressure of endolymph within the labyrinth. 4 The perilymphatic system appears to differ from the Phaneroglossan condition mainly in having no true ductus reuniens, as a consequence of which the foramen perilymphaticum superius is lacking, and the single perilymphatic foramen is comparable to the foramen perilymphaticum inferius of Rana . An additional tympanal area is to be observed between the anterior part of the pars basilaris and the spatium sacculare. 5 The sacci endolymphatici are paired throughout and are not connected by anterior ascending or ventral processes. Only two pairs of calcareous sacs were observed in the partes spinales. In post‐metamorphic specimens the endolymphatic system could only be traced as far as the region between the fourth and fifth intervertebral foramina. This reduction of the spinal portion of the endolymphatic system, in conjunction with the well‐defined auditory receptor areas in the labyrinth, is not suggestive of primitiveness. Xenopus possesses several structural features which may be regarded more fittingly as neotenic than as primitive. Others, again, may be purely adaptive, and while it is admitted that it may be fallacious to draw conclusions from a few isolated morphological facts, it could be surmised that, anatomically, the membranous labyrinth and its associated endolymphatic and perilymphatic systems indicate a possible reversion of this form from a semi‐terrestrial to an aquatic habitat. At the same time, the same features could equally well be interpreted as an attempt in a purely aquatic form to approach the condition found in semiterrestrial Anura.