INTRODUCTION. \i. 



its place in our system ; but, unfortunately, the very meagre 

 descriptions of some of the older species do not give us the 

 necessary information. No Ascidian, then, can he properly 

 described without dissection and microscopic examination, 

 and at the least the branchial sac, and, in some cases, the 

 tentacles and the dorsal tubercle, should be figured, as well 

 as the exterior of the animal. 



The classification of the Tuuicata which I followed in my 

 " Challenger " Reports (1882), is one which has been pretty 

 widely adopted. I gave a Revision of it, with some slight 

 modifications and expansions, in my " Revised Classification 

 of the Tunicata," &c., in 1891.* I am well aware of several 

 other classifications which have appeared since, such as 

 Lahille's in 1887, Garstang's modification of Lahille in 1895, 

 and Sluiter's also in 1895, but after giving them all very 

 full consideration, I am supported in adhering to the " Chal- 

 lenger ' classification by finding that it is the one which 

 Seeliger seems to receive with most approval in his great 

 work on the Tunicata now appearing in Bronn's " Klassen 

 und Ordnungen des Thier-Reichs." I am aware that the 

 Ascidia3 Composita? do not form a natural group, but can be 

 broken into several sets which are related to different kinds 

 of Simple Ascidians. I believe I was myself the first to 

 point out this polyphyletic nature of the group ; and yet I 

 consider it a practical convenience to retain the term Com- 

 pound Ascidians in our classification, especially in a Museum 

 Catalogue, where it is impossible to represent the complexi- 

 ties of a phylogenetic arrangement of lines branching in 

 the three dimensions of space. 



Lahille, in his system of classification, makes use only of 

 the condition and details of structure of the branchial sac, 

 an organ that is physiologically of such dominant import- 

 ance in Tunicata that it becomes liable to the loss or 



* Journ. Liun. Sue., Zool., vol. xxiii., p. 558. 



