ASCIDIACEA — FIXATION AND METAMORPHOSIS. 377 



number being then further increased by the intercalation both of new- 

 slits and new rows between tlmse already formed. The gill-slits are 

 at first elongated in the transverse direction, but their shape changes 

 Later, as they lengthen longitudinally. 



According to van Beneden and Julin (No. ( .») the formation of 

 the additional gill-slits in Phallusia (Ascidiella) scabroides follows an 

 entirely different rule. A longitudinal row of six primary slits here 

 first forms (Fig. 174 J, 1-6). Of these, the fourth (-1) in the row is 

 said to appear first, the first in position (1) being the second to form. 

 The fifth (5) forms next and then the second (2), while those that 

 occupy the third and sixth positions (3 and 6) form last. These six 

 primary gill-slits are markedly elongate transversely (Fig. 174 />') 

 and each becomes divided by projecting outgrowths of the trabeculae 

 lying between them. In this way the six primary slits give rise to 

 six rows of secondary slits. In later stages, new slits are said to 

 break through between these. 



In Clavelina, according to Seeligkr, two transverse rows of gill-slits are 

 found even in the free-swimming larva (Fig. 173 A). A new row forms in 

 front of these and behind them another is added after fixation. In this case 

 the number of slits is increased by the appearance of new independent perfora- 

 tions (and thus not by the division of those already present). This is also the 

 case, according to Garstaxg (No. 21<i) with the buds of Botryllus, while, in 

 the larva, the number of slits is increased through the division of those 

 formed first. According to this author also, Pyrosoma, with its transversely 

 elongated slits in a single row, represents a specially primitive condition. 



The rise of the six primary gill-slits (Fig. 174 A, 1-6) has i*ecentlv 

 been carefully examined in ('in, hi by Willey (No. 54a). Slits 1 ami 

 1 arise first, simultaneously and apparently independently of one 

 another. Willey is inclined to regard these two as parts of a single 

 slit separated as in Amphioxus by a tongue-bar. The next slits to 

 appear (2 and 3) form by abstriction from 1 and 4, while 5 and 6 

 arise independently. Willey thus regards slits 1-4 as parts of a 

 single primary slit separated by abstriction. The stage depicted in 

 Fig. 171 .1 would then possess only three actual primary slits. As 

 the six slits present divide in the manner stated by van Beneden 

 and Julin, six transverse rows of slits are formed, these all at first 

 being elongated transversely but later lengthening longitudinally 

 (parallel to the endostyle). Further division of the apertures leads 

 to the intercalation between these six rows of other transverse rows 

 (Willey). 



Even at an early stage, an entodermal fold can lie seen projecting 

 inward on the inner side of the branchial region between every two 



