120 Triassic Echinoderms of Bakony. 
tubercle. Other plates are less simple, for they have a double curvature, the curve of 
the adapical margin being concave in one part and convex in another; such are e, 
g,1,n, 0, and r (figs. 199, 202, 208, 204). This double curve exists in various com- 
binations. In one type the adapical concavity is near the tubercle, and, according as 
the tubercle is remote from or near to the adradial margin, so will that concavity 
be in the inner or outer half of the plate. Thus e, which is from the a column, and g 
and o, which are from the J column, all have the tubercle at about one-third of the dis- 
tance from the adradial margin; and in them the adradial half of the plate has the adapical 
margin concave and the adoral convex, while in the interradial half the conditions are 
reversed (figs. 199, 202). Inv, on the other hand, which like o comes from ab column, 
the tubercle is more interradial, and this portion of the adapical margin is concave, 
while the adradial half is convex (fig. 204). In the second type of double curvature 
the adapical margin near the tubercle is convex instead of concave, while the correspon- 
ding part of the adoral margin is concave. Thus in / and u, which are from a and b 
columns respectively, the tubercle is remote from the adradial margin, and the adapicai 
margin is convex in its interradial half (fig. 208). Among the various specimens there 
does not happen to be one in which the tubercle is adradial and the adapical margin 
convex in that half and concave in the other half; but there seems no reason why 
such a form should not occur. The varieties of marginal curvature may be tabulated thus: 
Examples 
(adepicel margin \\ 1. Single curve sini eh ORS) a GAGs kato 
A. ' concave j a. tubercle interradial . + = | 
\ at the tubercle j = Lame Te ae b. tubercle adradial . e, g,0 = 3 
asae Sas Bok So 1. Single curve dbs dupctke cs hs Aiea fed Gs = 2 
B.} convex a. tubercle interradial. J, 1 = 2 
las the tubercle J aL OUe ne Sa b. tubercle adradial . none = 0 
Total = 13 
From this it appears that in each column six varieties are possible, of which all 
are actually found except B, 2, b. 
The distribution of these varieties in the various regions of the column presents 
a problem that cannot be solved with the evidence at present available. It can only be 
pointed out that the variation does not appear to be related in any way to the 
size of the plates. 
The combination of the bending of the plates with the transverse curvature of 
each plate, and with the normal meridional curvature of the interambulacrum, results 
in a slight apparent torsion of each plate. This is easily seen when the plate is 
looked at from one or other of the transverse margins (fig. 205). 
Each plate bears a single main tubercle, and the position of this with refer- 
ence to the adradial margin is variable (see diagnosis and table of measurements) 
and seems unrelated to the width of the plate. Thus in the plates c, m, and 0, 
each with a width of 7.7 mm., the distance of the centre of the tubercle from the 
adradial margin is respectively 3.9, 4.5, and 3.1 mm. The mamelon is depressed; 
the neck, short and undercut;. the platform almost flat, but a very slight parapet 
with a gently rounded edge’ may sometimes be seen; from this the boss slopes 
steeply, sometimes with almost vertical sides, and passes into a broad, slightly 
depressed scrobicule. The ‘scrobicular circle is clearly defined at the sides, but 
broken at the transverse margins, indicating that the scrobicules were markedly 
