232 
the primordial interambulacral plate retained as in the Lepidocen- 
trids, only more or less reduced in larger specimens. Also in the 
Diadematids the primordial interambulacral plate is certainly present 
in young specimens — as it is upon the whole in all young 
Echinoids — the only difference being that in the Diadematoids 
it is resorbed at an early point of the development, in the Echi- 
mothurids only partially, while in the Lepidocentrids it was retained 
in its original shape during life. There is here a gradation agreeing, 
indeed, most beautifully with what must be required on the assump- 
tion of a direct genetic connection between these three families, 
and there is not required a.…series of unknown forms to bridge 
over the gap; in fact, there can murere be said to be a gap in 
the series Lepid i Echinothuridæ—Diadematoid As for 
Bather's other argument, that it is.impossible to decide whether 
such forms as the Triassic Mesodiadema and Diademopsis are 
really Cidarids or Diadematoids, because they are intermediate 
between. these two groups, I would venture to think that the 
difficulty is due to the imperfect knowledge of these forms; as 
Bather himself says: "were the complete test, and still more the 
living animal, preserved to us, the diffieulty might not occur” (Op. 
cit. p. 254). To base so "violent” a hypothesis on some few 
fragments of tests and some doubtful radioles against all other 
evidences afforded by both recent and, much better known, fossil 
forms, I really find unjustifiable. 
While thus I cannot agree with Jackson, Tornquist and 
Bather in assuming that the Diadematoids were derived from the 
Cidaroids, I am inclined to think on the other hand, that the 
Saleniids, and then probably all the Stirodonta, are derived from the 
Cidaroids, not from the Diadematids. A main argument for this is 
the structure of the spines in the Saleniids, viz. the presence of 
an outer layer, ostracum, corresponding to what occurs in Cidarids. 
It does not appear very probable that this structare has developed 
from Diadematid spines. The Arbaciids present traces of the 
same outer layer and thus appear to represent a further stage of 
