94 MB. AV. a. EIDEWOOD ON THE HTOBEANCHIAL 



hyobrancbial skeleton undergoes but little change, and that very 

 slowly ; and because the later stages, at which thie metamorphosis 

 is actually taking place, I have been unable, in spite of strenuous 

 endeavours, to procure. I have already delayed the publication 

 of this paper several months in the hope of being able to obtain 

 the latter stages ; but as there seems to be no prospect of 

 achieving this in the near future, I submit such results as I have 

 already obtained. For most of the specimens examined I am 

 indebted to the generosity of Mr. G, A. Boulenger, F.R.S. 



In making comparison of different tadpoles of the same species, 

 I find that actual size is no criterion of the degree of develop- 

 ment, since so much depends on the favourable or adverse con- 

 ditions of life. The carefully recorded measurements of Parker's 

 J^enopus larvae were useless to me for correlating his tadpoles 

 with mine. The only reliable data are afforded by the extent of 

 development of the paired limbs. The oldest of the three chosen 

 stages (Stage III.) is characterized by the presence of all four 

 limbs, the fore as well as the hind being well developed, and by 

 the possession of a large tail which shows no signs of absorption. 

 The larvae of the second stage have the hind limbs powei'fully 

 developed; but the fore limbs are only half as large as in 

 Stage III. The first stage is a most comprehensive one, in- 

 cluding tadpoles with hind limbs well developed and with the 

 fore limbs recently extruded ; tadpoles with hind legs not yet 

 markedly flexed and with no trace of fore limbs ; tadpoles with 

 the hind legs just appearing on the surface of the body ; and tad- 

 poles with no traces of paired limbs at all. There is no appre- 

 ciable difference beyond one of size in the hyobranchial skeleton 

 of any of these tadpoles of Stage I. ; but a trifling variation can 

 be discerned in the relative size of the larynx. 



In the figures of the developing hyobranchial skeleton 

 (Plate 11) the mandible is included, because it serves to give an 

 idea of the size of the mouth at each stage — a most desirable item 

 of information, — and because it demonstrates the relative huge- 

 ness of the hyoid arch, and shows that while the hyoid and 

 mandibular arches are closely approximated in Xenopus, they are 

 widely separated in the corresponding stages of Pipa. The 

 ossifications in the mandible are not indicated in the figures, 

 Parker shows (33. PI. 56. fig. 6, and PI. 58. fig. 4) a pair of " mento- 

 meckelian " or " inferior labial " cartilages in the symphysial 

 region of the mandible of the early larva of Xenopus. In this 



