30 - J, H. OETON. 



mainly as a water pump and a feeding organ, and it may indeed be 

 doubted whether on the whole oxygenation of the blood occurs in the gill 

 at all. On the other hand, the expenditure of such a large amount of 

 energy as is necessary to produce the main and food currents may well 

 leave the blood as it issues from the gill in a less pure condition than that 

 in which it entered. Moreover, it is to be remembered that the gill-fila- 

 ments of Amphioxus are very compact organs in which only the coelomic 

 blood-vessel lies at all near the surface (see Benham, 15, PI. 6). 



It would therefore seem more probable that oxygenation of the blood in 

 Amphioxus takes place in the relatively vast coelomic spaces adjacent 

 to the atrium, as, for instance, the endostylar coelomic canal in which 

 is contained the branchial artery, and the various coelomic spaces in the 

 dorsal regions of the atrium. Thus it is highly probable that the gill of 

 Amphioxus, like that of Lamellibranchs (see later, pp. 44 and 45), is 

 mainly a feeding organ and a water pump. 



THE MODE OF FEEDING IN ASCIDIANS. 



The similarity in general structure in the gills of Amphioxus and 

 Ascidians renders the mode of feeding in the Ascidians a matter of much 

 interest in comparison with that of Amphioxus. 



The mode of food-collection in Ascidians has been investigated by 

 many zoologists, and correctly described by Fol (17) in various Ascidians, 

 and later by Roule (8, pp. 66, 67) in Ciona. It is curious that English 

 writers, including Willey (1, p. 185) and Herdman (9, pp. 15 and 16, p. 46), 

 describe the process differently, apparently accepting the earlier work of 

 Fol (4, 1872). Fol, in a summary of his later work (17, p. 240), states 

 that "Die Rinne (the endostyle) ist kein Ernahrungs-abschnitt des 

 Kiemenkoibes sondern ein Driisenorgan. Die Nahrungsaufnahme 

 findet gkichzeitig mit der Athmimg im ganzen Kiemenkorbe statt." 

 Roule describing this process in Ciona states : " Un mucus, secrete 

 par le raphe ventral, est deverse en fins filaments sur la face interne 

 de la parol branchiale ; la, ces filaments agglutinent tous les petits cor- 

 puscules, amenes par I'eau, qui passent a leur portee ; puis, entraines par 

 les mouvements des cils vibraiiles des papilles, ils se dirigent obliquement 

 en haut et en arriere, de telle sorte que les filaments les plus anterieurs 

 se rassemblent en une seule masse qui suit le raphe dorsal pour penetrer 

 dans la bouche oesophagienne, tandis que les posterieurs y parviennent 

 directement . ' ' On the other hand, English writers describe food-collection 

 as occurring in the endostyle and peri-pharyngeal grooves ; the former 

 groove conducting mucus with contained particles forwards to the latter, 



