254 Triassic Echinoderms of Bakony. 
at the apex and the restrained plates at the peristomial margin seems to contain 
much truth, though other causes need not be excluded. 
Here for a moment we may pause to consider a possible alternative. It might 
be a less violent hypothesis to deduce the Ectobranchiata from another line of 
Cidaroida represented in Devonian and Carboniferous times by the Lepidocentridae. 
In this Family the interambulacral plates do not pass on to the peristomial membrane, 
so that one of the chief features differentiating the Diademoida from the Cidaridae 
is already present. But in the Lepidocentridae this absence of interambulacrals 
from the peristomial membrane appears to be due to a different cause. In this 
Family the initial single interradial plate is retained in the adult, whereas the Diadem- 
oida are clearly derived from forms in which the adoral interambulacrals have 
disappeared. A considerable series of genera would therefore be required to bridge 
the gap between the Lepidocentridae and the Diademoida, but of this postulated 
series no representatives are known. This hypothesis therefore would present more 
rather than fewer difficulties. 
Returning to the hypothesis of the origin of the Diademoida from the Cida- 
ridae, we find the state of affairs just the contrary. More than once in the prece- 
ding pages the difficulty mentioned has been that of deciding whether a certain fossil 
belonged to the Cidaridae or the Diadematidae. Were the complete test, and still 
_more the living animal, preserved to us, the difficulty might not occur; but when 
we have only the corona or fragments of it to deal with, the impression gained 
is that there was a very gradual change in all its elements, and that the difficulty 
is merely an expression of this fact. Let us then proceed to discuss the change 
in the ambulacra. 
The essential differences between the Cidaroid and Diademoid types of ambul- 
acrum are these: In the Cidaroid the primary ambulacral plates are equal in size, 
equituberculate, distinct, in simple series, those of the a column alternating with 
those of the b column, with the pore-pairs also in simple series (unigeminal). In 
the Diademoid the primary plates are united by threes, in which the middle plate 
is the largest and bears the largest tubercle; the sutures are obscured, especially on 
the outer face where the base of the large tubercle encroaches on the adjacent 
ambulacrals; the pore-pairs of each triad form an arc, with the large tubercle as its 
approximate cehtre. In the ultimate development of this type, the triads of primary 
ambulacrals are so closely united as to form a single large plate or major; and it 
is these majors, and not the primaries, that alternate along the zigzag suture between 
the a and } columns. A slighter difference generally obtains in the disposition of 
the podial pores: in the Cidaridae it is usual for those of a single pair to be 
approximately equidistant from the adoral margin of the ambulacral, and to be 
unenclosed by a distinct raised rim or peripodium; in the Diademoida, as in many 
other Ectobranchiata, there is unsually a peripodium, and its long axis lies at an 
angle to the long axis of the ambulacral, so that the inner or radiad pore is the 
more adoral. Finally, the width of the ambulacra tends to be relatively greater in 
the Diademoida than in the Cidaridae. 
The ambulacra of the Triassic Cidaridae entirely conform to the Cidaroid 
type. That of Triadocidaris persimilis has been described minutely (pp. 72, 73), 
and a fragment apparently belonging to a Miocidaris is also described (p. 93). 
The Triassic ambulacra that are here referred to the Diademoida, on the other 
