Dr. F. A. Bather — Studies in Edrioasteroidea. 9 



here both ends may he incomplete. If allowance be made for 

 imperfections and crushing, we may legitimately imagine a normal 

 height of about 14 mm., and a diameter of 9 - 8 mm., or say two-thirds 

 the height. The diameter seems to have been the same all the way 

 up, and there is no indication that the turret was curved. 



This turret is built up of a large number of separate thin plates, 

 imbricating in such a way that their outer or free ends are directed 

 upwards, i.e. towards the adoral face (PI. II, Tig. 4). At the base 

 of the turret these plates lie almost horizontally, but the inward 

 dip gradually increases with each successive layer. The thickness 

 of the turret wall is therefore at the base equal to the diameter of 

 a plate ; in A and C this amounts to at least 1*3 mm., leaving a lumen 

 of about 7 mm. diameter (PI. II, Fig. 3). Higher up the thickness 

 does not decrease, but is formed rather by the combined thicknesses 

 of the imbricating plates. The disc or body of the animal seems 

 thus supported by the turret as a bird on its nest (PL II, Fig. 6). 

 The arrangement of the imbricating plates is not very regular, but 

 those of one tier tend to alternate with those of the tier below. In 

 A the number of plates exposed along a single vertical line from 

 top to bottom is about twenty-four ; this means that the number 

 of tiers is about forty-eight. About ten plates (twenty tiers) occur 

 within 5 mm., but the plates are more widely separate above than 

 they are below. Near the base four plates may be exposed along 

 a vertical line of 1 mm., but this is in part due to the fact that the 

 alternation is less regular, so that in these regions the number of tiers 

 within the same distance would be only five or six, not eight. In 

 the absence of a vertical section it is not possible to estimate precisely 

 the proportion of each plate exposed, i.e. the extent of overlap ; but 

 the actual depth of the exposed portion in the upper part of the 

 turret is about *5 mm. It seems safe to say that not more than 

 one-third of the plate is ever exposed. 



No one of the turret plates can be entirely isolated, but in general 

 outline those near the base may be compared to a salad-plate, with 

 a slight concave curve next the lumen of the turret, and a stronger 

 convex curve on the outside (PI. II, Fig. 3). The convex curve 

 is more pronounced in plates from the upper part of the turret. 

 In specimen A a plate at the base has a width of about 3 - 75 mm. 

 In other plates the width seems to have exceeded this, and to have 

 about thrice the diameter. Probably the plates below are wider than 

 those above. 



Specimens A and C are laterally compressed, so that the cross-section 

 of the turret may be said to have flat sides and rounded ends. 

 Whether it bear any causal relation to this compression or no, it 

 is the case that the tiers of turret-plates sink towards the base on 

 the sides and rise towards the ends. This is visible at the very 

 base in A. 



In specimen C, which is generally of a yellowish- grey colour, the 

 turret-plates are spotted with black in their stereom, and the spots 

 are chiefly conspicuous on the free borders, where they seem rather 

 regularly spaced. The meaning of these spots is suggested by A, 

 for there^ on the free borders of the turret-plates, are what appear to 



