•“>o 
THE VOYAGE OF H.M.S. CHALLENGER. 
to distinguish Simple from Compound Ascidians. Eeproduction by gemmation and the 
formation of colonies in the latter group will not hold, since it is possible to pass from 
dona — a typical Simple Ascidian — to Distoma and the very centre of the Compound 
Ascidians through the following series of forms which show a perfect gradation of these 
characters : — dona, Rhopalcea, Ecteinascidia, davelina, Diazona, Chondrostachys, Oxy- 
corynia, Distoma. 
The formation of common cloacal cavities, canals, and apertures cannot be considered 
as a diagnostic feature of the Compound Ascidians. Although recent investigations 
by Giard and others upon living material have demonstrated their presence in some 
genera in which they were previously unknown, yet there are some forms considered by 
all authorities as Synascidise, such as Chondrostachys, Diazona, Distoma, and others, 
in which the atrial apertures of the Ascidiozooids open independently on the surface of 
the colony, and no common cloaca is formed. 
Lastly, we come to characters taken from the condition of the test, but these break 
down like the others. In the first place, in passing along the series of forms mentioned 
above as connecting dona and Distoma, we encounter all stages between a distinct test 
or tunic for each individual and a common mass in which a number of Ascidiozooids 
are imbedded. And, secondly, the remarkable group of Polystyelidse, which were briefly 
characterised by Giard in 1874, present many of the characters of highly differentiated 
Simple Ascidians (the Cynthiidse) along with the supposed Synascidian feature of a 
colony composed of many Ascidiozooids completely buried in a common test. 
Thus all the diagnostic features usually employed fail utterly, and we find that, as 
our knowledge of the two groups extends, it becomes more and more difficult to 
distinguish even in an artificial and arbitrary manner between Simple and Compound 
Ascidians. The condition of the test is probably the best distinguishing feature to 
employ, but it must be remembered that, as I have just shown, it will not hold good in 
all cases. As a general rule, however, in the Ascidise Compositse the Ascidiozooids 
forming the colony have not separate tests, but are imbedded in a common investing 
mass, while in the Ascidise Simplices, when colonies are formed, the Ascidiozooids 
remain partially distinct from one another and each retains its own test. 
The mutual relation of the different families of the Ascidise Compositse is a very 
difficult matter to determine on account of the peculiarity of their evolution. They 
seem to form, when expressed diagrammatically, a network rather than a tree-like 
figure, and this is due, I believe, to the group being formed of several distinct branches 
which have arisen at different times from different groups of the Simple Ascidians and 
have become modified so as to form, in some cases, divergent and in others convergent 
lines. This will l^e discussed more fully further on in the Report in considering the 
phylogeny of the group, and the arrangement of the various families according to their 
natural affinities may Ije postponed till then (see Summary and General Remarks). 
