254 GILBERT 0. BOURNE. 



was able to detect a single delicate flagelUim projecting from 

 tlie free extremity of each ; these flagella ^Yere most often 

 recognisable in cells from tlie epithelium of the tentacles. In 

 tlie majority of cellsj however, no flagellum could be detected, 

 and in many cases, especially in those cells in which they 

 were longest, the free processes are club-shaped, having 

 rounded extremities attached to the body of the cell by a 

 relatively narrow pedicle. This feature suggests that the 

 cells, like the endoderm-cells of Hydra, may withdraw their 

 flagella and become amoeboid. Fig. 18, h, is a careful camera 

 drawing of some of the best preserved of these cells, very 

 highly magnified. The resemblance to a collar-cell is 

 certainly very striking, especially in the tallest cell on the 

 right hand of the figure ; but it would be rash to make any 

 positive assertion about it. It must be remembered that my 

 single specimen of Oligotrema was preserved in the tropics 

 far from a laboratory, and that it had been in spirit for a 

 long time before it reached me. It was therefore very un- 

 likely that the details of such delicate structures as collar- 

 cells could be faithfull}^ preserved, and it says much for Dr. 

 Willey's skill and care in preserving his collections that the 

 specimen reached me in such good histological condition as 

 it is. Though I have looked through sections of various 

 Ascidians belonging to the genera j\[olgula, Ascidia, Phal- 

 lusia, Ciona, and Styelopsis, I have been unable to find any 

 trace of an epithelium comparable with that of the branchial 

 sac of Oligotrema, and, so far as I can ascertain, no such 

 epithelium has been described in any other Ascidian. The 

 nearest approach to it that I can find is a band of elongated 

 claviform cells, described by E. Hertwig as occurring in the 

 ventral groove (Bauchrinne) of Cynthia canopus. In some 

 places, where they are not well pi-eserved, the cells of Oligo- 

 trema resemble those figured by him ; but his drawings show 

 the nuclei near the free ends of the cells, whereas in Oligo- 

 trema they are situated at the basal ends. 



This modified ej>ithelium clothes the tentacles and lines 

 nearly the whole of the branchial sac, l)ut does not occur in 



