Scharff — On DoArn's Theories on the Origin of Vertebrates. 31 



with the spiracular artery of elasmobranchs. As there are two- 

 cephalic aortae in Petromyzon, one on each side of the notochord, 

 each of these two arteries opens into the aorta of its own side. The 

 backward growth of the mesodermic velnm causes the original 

 opening of the thyroid to be pushed back, so that it comes to lie at 

 the level of the second pair of branchial diverticula. 



An ingrowing lamella of connective tissue now divides the 

 thyroid into two parts. Subsequently another lamella breaks into 

 the anterior part of the thyroid on each side and in so doing pushes 

 in the outer wall. Dohrn now calls the invaginated part "the 

 glandular lamella" and the uninvaginated "the opercular lamella." 

 In the glandular-lamella, a differentiation of the cells takes place 

 into masses of long, conical gland-cells and shorter ciliated cells. 



At a later stage, says Dohrn, we find two ciliated grooves run- 

 ning along the floor of the pharynx, and converging on the median 

 ventral line. They meet at the opening of the thyroid. 



Developmental history shows that these grooves are the last 

 remains of the most anterior gill-pouches, which are homologous 

 with the spiracular clefts of elasmobranchs. 



If we compare the fully developed endostyle or hypobranchial 

 groove of an ascidian — say Ciona intestinalis — with the thyroid of 

 Ammocoetes, we find, in spite of a series of minor dissimilarities, 

 the same characteristic fundamental structure in both. Both 

 organs are on the ventral side of the pharynx and in the same 

 position, and have a similar length. Two primary branches of the 

 main branchial artery pass along the sides of the thyroid ; similar 

 blood-vessels accompany the endostyle. However, it is chiefly the 

 part of the glandular cells and that of the ciliated cells in the 

 thyroid which corresponds very closely with what we find in the 

 tunicate endostyle, except that the former is somewhat more 

 complicated. No one can doubt, however, that these are homologous 

 organs, especially if we add that the ascidian possesses a pair of 

 ciliated grooves immediately behind the mouth, which converge 

 ventrally towards the endostyle. Dohrn pointed out before that 

 the two grooves in Ammocoetes arose from a transformation of the 

 spiracular clefts. If therefore these are identical with the ciliated 

 grooves of tunicates, it follows that the latter also represent 

 spiracular clefts. 



Tunicates must therefore be derived from fishes, not vice versa. 



