Esystamniie Rauales, 2 oeu 
very depressed shape of the test. A further result of this shifting of the plates 
(if such it be) is to bring the normally alternating interambulacrals of the adoral 
surface into an opposite and paired position, and to increase the number of main 
tubercles about and below the ambitus. Thus the long radioles with their serrated 
edges form a framework supporting the animal on the surface of the ooze. 
The one point of resemblance that survives criticism is the persistence of the 
primordial median interambulacral in the adult of both genera. But this is a character 
shared byePygmaeocidaris with other Arbaciid genera, and in its other Arbaciid 
characters that genus by no means approaches Tiarechinus. The chief differences, 
admitted by DoeDERLEIN, are: the far greater number of interambulacrals in Pygmaeo- 
cidaris; and the persistence of primary ambulacrals in Ziarechinus, as opposed 
to the formation of diademoid and arbacioid majors in Pygmaeocidaris. DOEDERLEIN 
considers the alleged absence of external gills in the Tiarechinidae to be an arbitrary 
assumption, and holds that a contrary conclusion may be drawn from the published 
figures of the fossil forms; but in this he seems to be carried away by his hypo- 
thesis. Whether Pygmaeocidaris agrees with other Arbaciidae in the development 
of knobs and sockets on the joint-faces of its interambulacrals, has not been stated. 
Professor DorDERLEIN’s hypothesis manifestly demands that Pygmaeocidaris 
shall be regarded as the most primitive of known Arbaciidae. Now the Family 
contains about six genera, of which four are known only from Recent seas and 
the remaining two, Arbacia and Coclopleurus, do not ascend beyond the Eocene. It is, 
as anyone would naturally expect, these two genera of greater antiquity that more 
nearly approach the Jurassic and Cretaceous Hemicidaridae, from which they are 
in all probability descended. In just those features that have given rise to the com- 
parison with Tiarechinus, Pygmacocidaris appears to be the most specialised rather 
than the most primitive of the Arbaciidae. To prove the contrary, one demands 
evidence from a series of intervening forms. But all the evidence from those 
Mesozoic and Cainozoic genera that have been regarded as ancestors of the modern 
Arbaciidae points in the contrary direction. 
The Tiarechinidae became specialised at an early period, when the ambulac- 
rals were still simple primaries, before external gill-notches had developed, and possibly 
before the additional interambulacral columns found in Palaeozoic Echinoids had 
been suppressed. But in their specialisation they proceeded much further than any 
recent Arbaciid. Their interambulacrals were more reduced, their plates more firmly 
fused, their apical system relatively enlarged, and the radioles still more restricted 
to the oral surface. Pygmaeocidaris is far advanced in all the characters of the 
Arbacina, but has not fully attained the peculiar Tiarechinid characters. Therefore 
I regard these latter characters as independently acquired by it, and in no way as 
ancestral. Between the earlier Arbacina and the Tiarechinidae there is no particular 
resemblance; and while study of recent Arbaciidae may elucidate the mode of 
life of their Triassic homoeomorphs, | fail to see that it can cast any light on the 
affinities of the latter. 
The genus Triadocidaris is rediagnosed (p. 69), with 7. subsimilis (Mtnsv.) 
as genotype, and of it three new species are described: 7. persimilis related to 
T. subsimilis, T. praeternobilis related to T. subnobilis, and T. immunita which 
shows resemblances to both Anaulocidaris and Mesodiadema. 
Miocidaris is discussed at length (p. 83) and rediagnosed with M. Cassiani 
17* 
