GILL-FILAMENTS IN SAUROPSIDA 209 



plate 2 (see figs. 12 to 17 inclusive) show that the mound is an 

 evagination of the thickened, vesiculated ectoderm covering the 

 third arch, and contains a mesodermal core, thus almost repro- 

 ducing the early formation of external gills in the amphibia. 

 The lower end of it has already begun to develop filaments and 

 later the upper end will give rise to tufts of cells (fig. 18). In the 

 same embryo the ectoderm of the fourth arch forms an evagina- 

 tion, which however, is solid, and in this specimen much smaller, 

 tending to fuse with that from the third arch (fig. 15). Owing to 

 the rapidity with which the region behind the third arch is being 

 flattened out, the evagination on the fourth arch, which has been 

 observed in several embryos, has only a transitory existence of 

 its own. As the operculum extends backward all of the third 

 arch except the mound becomes covered over, while the mound 

 itself gradually assumes the shape of a wedge, with filaments at 

 its downward directed point, as indicated in figures 6 and 19. 

 As the hyoid arch continues its backward growth during the fifth 

 day it fuses with the third arch in such a way as to carry with 

 it on its under surface the tufted epithelium at the lower end of 

 the wedge, so that from now on, the filaments of this region of 

 fusion will appear to come from the under surface of the opercu- 

 lum (figs. 19 and 23). By the beginning of the sixth day the 

 whole edge on each side has become differentiated into a ridge 

 coextensive with the lateral margins of each operculum (fig. 7) . 

 Serial sections (fig. 24) show that the individual filaments borne 

 by the ridge are solid outgrowths of the epithelium, honey- 

 combed with degeneration vesicles. With the appearance of this 

 pair of ridges the first half of the life-history of the filaments may 

 be said to have been completed. 



While this differentiation of filaments has been going on at the 

 margins of the hyoid arches, the ventro-medially directed por- 

 tions of the two opercular processes have united to form a single 

 band of tissue slightly overlapping the pectoral body-wall and 

 extending across the ventral surface of the neck from side to 

 side, — the homologue of the 7nembrana hranchiostega according 

 to Rathke. From now on, the fused hyoid arches may there- 

 fore be referred to as a single structure, the plica opercularis, pos- 



