CARADOCIAN CYSTIDEA FROM GIRVAN. 455 



than the Canadian. Here the periproct has passed up between plates 12 and 14, which 

 meet either not at all or possibly, in rare cases, only by their extreme points. As a 

 rule they are widely separated, and 12 abuts on plates 17 and 23 ; 14 on plates 19 and 18. 

 This is the natural result of the widening of the periproct below them, and the down- 

 ward pressure on plates 18 and 23 by the narrowing of the tegmen. An earlier stage 

 of the latter development is seen in E. Billings' figure of P. squamosa (1858, pi. 1, 

 f, le), where a tongue from plate 18 almost reaches the periproct. A later stage is not 

 found, nor is it to be expected, for in the British species the limits of strength had been 

 reached, so that in the fossils it is rare for the assemblage of plates covering the oeso- 

 phagus on the anal face (viz. 23, 18, 24) to be preserved in position. 



§ 372. Plate 13, then, lay between two opposed forces, and its disappearance is not 

 to be wondered at. But as to the manner of its going, there is still a difficulty. If it 

 were pressed downward, its place would have been filled by plate 18 ; therefore, if on 

 no other ground, I do not incline to Carpenter's suggestion. If, on the other hand, 

 plate 13 was absorbed to make more room for the periproct, it is not clear why plates 

 12 and 14 should ever have met; and this seems opposed to Dr Kirk's explanation. 

 The remaining view, that plates 13 and 14 have fused, does not present any difficulties 

 of that nature ; and if we inquire why they should have fused, we find an obvious reason 

 in the need for greater rigidity in the periproctal frame, especially in this region near 

 so many vital organs. Dr Kirk, however, writes as though he had more primitive 

 fossils proving his view. 



§ 373. The homologies of the Deltoids (20-24) are, owing to the constriction and 

 the bilateral modification of the tegmen, by no means clear. Plate 23 is fixed for us 

 by its pore ; and there is no difficulty in recognising the adjacent plate 24. On the 

 antanal face, the plate resting between 15 and 16 is presumably 21. If the plate to 

 the right of this is 22, and the plate to the left of it 20, then we observe that, whereas 

 20 curves round the edge of the flattened theca so as to abut on 24, and so as to form 

 a complete facet for the brachiole, plate 22, on the other hand, is confined to the 

 antanal face, so that the brachiole rests on 21, 17, and 23. Sometimes, however, plate 

 22 seems to bend round the theca on to the anal face, or it may be that a proximal 

 brachiolar is enlarged. 



§ 374. The deltoids, of course, are interradial elements, and it is still not quite clear 

 to which rays the Brachioles belong. The position of plate 22 implies that one of 

 them springs from the groove primitively between 22 and 23, i.e. the left posterior. 

 The other brachiole may be either that between 20 and 2 1 (anterior), or that between 

 20 and 24 (right anterior). In Erinocystis, according to Jaekel (1899, p. 250), the 

 brachioles are probably I. and IV., i.e. left posterior and right anterior; but the con- 

 ditions in that genus are so different that no homology can be postulated. There are 

 at present no anatomical grounds on which to base a decision ; but one might be 

 inferred from the relations of the theca to the sea-floor. If the anal side was downwards, 

 as Dr Kirk (1911, top of p. 20) seems to suppose, then the brachiole retained would 



TRANS. ROY. SOC. EDIN., VOL. XLIX. PART II (NO. 6). 60 



