81 
Kchinoiä Tests, Ci dar oi da. 
adradial margin as it nears the peristome ; and this appearance is due partly to the 
great obliquity of the adradial margin, partly to the diminished angularity of the inter¬ 
radial suture causing the centre of each plate itself to come nearer to the adradial margin. 
Each main tubercle has a mamelon of fair relative size, its transverse diameter 
being from 0.14 to 0.2 of the transverse diameter of the plate, while its meridional 
diameter tends to be rather greater and is never less. In the adoral region the relat¬ 
ive size of the mamelon increases towards the peristome, owing to the rapid less- 
ening of the diameters of the plates, while its absolute size increases away from 
the peristome. The Perforation is circular but not always central, thus in e (hg. 144) it is 
shifted towards the adambital margin, and in c away from that margin ; that is to 
say, if our previous inference with regard to c be correct, the shifting in both cases 
is adapical. The neck is undercut slightly or not at all. 
The boss has a well-marked platform, without parapet in specimens c to g, 
with very slight parapet in b, and with slightly more defined parapet in a\ this 
occasionally shows very faint traces of crenelation. From this the boss slopes at a 
moderately high angle, with straight, or slightly concave, or waved sides into the 
flush or feebly sunk scrobicule. The scrobicular circle is not defined, nor is there 
a scrobicular ring; but the scrobicule merges, often with faint radiating ridges or 
folds, into the extra-scrobicular surface, which bears miliaries of two sizes. These 
may be quite irregulär, as in a, where about 16 are contained in (2.5 mm.) 2 ; or 
may tend to continue the radiating folds just mentioned, as in e ; or may be rather 
more regulär, especially along the adradial margin, as in b, where in the second 
plate a definite row is formed by 6 equal miliaries. In no case do the scrobicules 
appear confluent, but they are always separated by miliaries. The scrobicules may 
be almost circular in the relatively high plates, for example c, or elliptical in the 
lower plates, for example e, in which the intertubercular miliaries are reduced to a 
single thin row. 
The adradial margin of each plate is almost straight, if anything with a slight 
convex curve ; it is bevelled on its inner surface, and provided with denticles which 
die out against a slight, rounded meridional ridge (figs. 143, 146, 147). The number 
of denticles to a plate, which indicates the number of ambulacrals corresponding with 
an interambulacral, is about 11 to plates 3 and 4 of a, plate 3 of b, the larger 
plate of e, and in c. Plate 2 of b has about 9 1 /., ; and the smaller plate of e, 8 or 
9. .The miliaries on the adradial margin do not correspond with the denticles 
in either number or position. 
In specimens a, c, and e, where the denticles are clearly seen, the structure 
is essentially the same as in other species of Triadocidaris: the denticles on the 
inner surface are continuous with the projections of the extreme margin; at the 
adoral end of the suture the denticles die out, while the scolloped margin thickens. 
In b, however, the marginal projections appear to die out soon and to be replaced 
by depressions, as though there were fa^nt denticles alternating with the marginal pro¬ 
jections (fig. 147). This appearance might possibly be held to Support Professor Doeder- 
lein’s view that the denticles correspond with the ambulacrals instead of alternating with 
them ; but this it can only do by contradicting the view that the denticles of the 
inner surface are homologous with the projecting angles of the margin. Other species, 
e. g. Mesodiadema latum, show this structure more clearly, and prove that the depressions 
on the inner surface are not depressions between denticles but depressions in the denticles. 
Resultate der wissenschaftl Erforschung des Balatonsees. I Bd. l.T. Pal. Anh. 6 
