280 J. T. CUNNINGHAM. 



pushing in forms the glandular lamella, the uninvaginated part 

 forming the cover-lamella. In the glandular lamella a diflfer- 

 entiation takes places into conical masses of gland-cellsj the 

 apex of the cone turned to the cavity of the gland, and 

 ordinary ciliated cells. In the advanced larva of Ammocoetes 

 two ciliated grooves run transversely in the wall of the 

 pharynx, in front of the gill-sacs, and converge on the median 

 ventral line to meet in the opening of the thyroid. These 

 grooves Dohrn has ascertained to be derived from the endo- 

 dermal sacs which represent the spiracular clefts, and which in 

 Ammocoetes never acquire an opening to the exterior. 



Now the endostyle or hypobranchial groove of Ascidians, 

 e. g. Cione intestinalis or Salpa, is closely similar in histo- 

 logical structure to the thyroid of Ammocoetes. There is the 

 same differentiation into bulbous agglomerations of gland- 

 cells, and a more even layer of ciliated cells. Moreover, in the 

 Ascidian there is a pair of ciliated grooves immediately behind 

 the mouth, which ventrally converge to the hypobranchial 

 groove, dorsally to the ciliated pit (hypophysis). These 

 grooves of the Ascidian must be homologous with those of 

 Ammocoetes, and must therefore represent in the Ascidian the 

 spiracular clefts. And it follows that Tunicates must be 

 derived from fishes, not vice versa. The reason suggested 

 for the transformation is that the thyroid and spiracular clefts 

 have been converted into mucous-secreting organs to aid in 

 the conveyance of nourishment to the oesophagus. 



In Amphioxus there is not a hypobranchial groove, but a 

 hypobranchial ridge, but this ridge has the same histological 

 character as the thyroid in Ammocoetes and the hypobranchial 

 groove in Ascidians. A homologue of the peripharyngeal 

 ciliated grooves is not mentioned as occurring in Amphioxus, 

 and the development of the hypobranchial ridge has not been 

 studied. 



The conclusion drawn from all this is that both Tunicates 

 and Amphioxus are degenerate fishes derived from ancestors 

 more or less similar to the Cyclostomata. A difficulty which 

 arises in considering Dohrn's arguments is that no reason is 



