324 ARTHUR WILLET. 



In fig. 2 an individual is represented at a stage shortly after 

 fixation, when the tail has been entirely drawn inside the body 

 of the Ascidian, but before any change of axis has set in. It . 

 is here seen that a line drawn through the two slits at right 

 angles would be parallel to the direction of the endostyle. The 

 two slits arise already in the older free-swimming larvse, but 

 are first clearly recognisable just after fixation on account of 

 the fact that the spaces in the body, namely, the lumen of the 

 intestine and the body-cavity, undergo a considerable expansion 

 at that period of the development, as will be seen on comparing 

 fig. I with fig. 2. 



As to the actual mode of formation of the first two slits 

 which arise simultaneously, it is necessary to have recourse 

 to transverse or horizontal sections. It will suflice to describe 

 the conditions observed in transverse sections, as the only 

 essential difference to be noticed between a transverse and 

 horizontal section of a larva at the time of fixation is that in 

 the latter direction the endostyle is cut transversely (cf. PI. 

 XXX, fig. 13), while in the former it does not occur at all 

 in sections through the region of the branchial sac which lies 

 behind the mouth. The primary position and relations of the 

 endostyle will be described below. 



Figs. 11 and 12 represent transverse sections through the 

 region of the branchial tract of an individual of about the stage 

 shown in fig. 2. The sections have the advantage of being 

 slightly oblique, so that on the left side of the section the wall 

 of the branchial sac immediately in front of the branchial 

 tract of that side is shown, and can be compared with the wall 

 on the right side of the section in which the latter passes 

 through the branchial tract. It will thus be seen at a glance 

 in fig. 11, on comparing the two sides of the branchial sac 

 together, that what we have to do with in the branchial tract 

 is not two outgrowths from the wall of the branchial sac, but 

 one ingrowth or involution of the wall projecting into the 

 branchial cavity and involving with it a space which com- 

 municates freely with the body-cavity and contains blood-cor- 

 puscles. This ingrowth necessarily gives rise to two angles 



