EEPOET ON THE THNICATA. 
325 
branchial sac in all, and still more the condition of the reproductive organs, show such 
a close resemblance to the corresponding parts in species of Polycarpa and Styela, that the 
two groups must be closely related, as was supposed by Giard and by Heller. But still 
it is impossible to regard the Polystyelidee as Simple Ascidians. Thylacium, Oculinaria, 
and Goodsiria were all originally described as Social Ascidians allied to Clavelina, which 
produces buds. Carus and Giard both came to the conclusion that reproduction by 
gemmation probably took place in the genera they described, and although, like them, 
I have not been able to settle the matter conclusively, I think there can be very little 
doubt that the specimens I have examined are true colonies produced by a process of 
gemmation from the vascular appendages of a single Ascidiozooid. 
It is well knovTi that some species of the Styelinse {e.g., Styela grossularia) under 
certain circumstances produce aggregations which have a superficial resemblance to 
colonies. When individuals are closely crowded together their tests unite to form a 
continuous mass, and young individuals of the same species attach themselves to the 
tests of the older specimens. These aggregations, however, are not colonies. There is 
no reproduction by gemmation, the individuals in the mass have all been produced from 
ova, and have no relation to one another except as near neighbours, and finally they 
are imbedded in a common test or investing mass. 
The specimens of the Polystyelidae in the Challenger collection are not mere aggrega- 
tions of individuals, they are colonies of Ascidiozooids imbedded in a common test which 
is penetrated by a system of vessels — consequently they must be regarded as Compound 
Ascidians. They are not, however, closely allied to most of the other Compound 
Ascidians, but have, I am inclined to believe, been evolved separately from the Simple 
Ascidians, and not from the end of that group occupied by Clavelina and Ecteinascidia, to 
which some of the Compound Ascidians are closely allied,^ but from near the genus Poly- 
carpa amongst the Cynthiidse. The only other family of Compound Ascidians to which 
the Polystyelidse seem to be closely allied is the Botryllidse. The general shape of the 
Ascidiozooids, the appearance and course of the alimentary canal, and the structure of the 
branchial sac, are so similar in the two families, that I believe the Botryllidse to be more 
nearly related to the Polystyelidse than they are to the rest of the Compound Ascidians. 
The chief characteristics of the Polystyelidse are the large and usually rounded 
Ascidiozooids, with their four-1 obed apertures, and the total absence of common cloacal 
cavities, the possession of branching vessels like those of the Botryllidse in the test, the 
presence of numerous strong internal longitudinal bars, and sometimes of folds, in the 
branchial sac, and the simple condition of the dorsal lamina. 
The shape of the colony varies greatly. It is massive and pedunculated in Goodsiria 
placenta, massive and sessile in Goodsiria coccinea, incrusting in Synstyela incrustans, 
and broken up into small pieces united by irregular stolons in Chorizocormus. 
1 See Summary and General Remarks at the end of this Report. 
