138 GEODIA AGASSIZII. 



Lanibe. Thoy are much more numerous in the former than in the latter. 

 Fully ileveloped sterroids occur sparingly in all adult specimens. In other 

 Get)didae, where I have found them, they also occur in small numbers. These 

 facts lead me to consider the sterroids as spicules produced, like the milk-teeth 

 of mammals, in the immature state only. If this assumption is correct, the 

 sterroids might be similar to the ancestral form of the normal sterrasters, and 

 represent a link connecting the latter with the sphaerastei-s from which I should 

 be inclined to derive them. 



To simplify the references to these sponges, I will, in the following discussion 

 of their relative systematic position, designate: — 



To avoid confountUng characters inunature in nature witli the systematically 

 important peculiarities of the full-grown sponges, the young specimen from 

 Station 4228 and the immature specimen described by Lambe are not taken 

 into account in the following discussion. 



In the character of their shape these sponges are very similar, their differ- 

 ences in this respect not exceeding the individual variations usually met with in 

 the species of geodine sponges. In the structure of their canal-system, their soft 

 parts, and the general arrangement of their skeleton they are also uniform. In 

 their colour and in the shape, relative frequency, and dimensions of their spicules 

 however, only the specimens collected at one and the same station agree. The 

 differences in the colour of specimens from different stations are certainly very 

 considerable. But since it is very likely that these differences have been pro- 

 duced post mortem through differences in the external influences to which the 

 different lots were exposed after capture, they are without systematic signifi- 

 cance. The differences in the spicules, on the other hand, are systematically 

 important, and it is therefore necessary to study them with care if we wish to 

 decide in what systematic relation these otherwise similar sponges stand to each 

 other. 



The differences in the shape and relative frequency of the spicules of the 

 several lots of these sponges are as follows : in E and F both thick club-shaped 

 and thin cylindrical styles are met with in small numbers; in A, D, and G thin 

 ones only; and in B, C, H and / no styles at all. In E, F, and G the ortho- 

 plagiotriaenes with clades either abruptly bent down at the end or terminally 



