XT. BaJanocrinus. 391 



while between the forks is an intercalated crenella, making 

 five denticles at the periphery. The change from these 

 peripheral crenellae to those of the radial ridge-gronps is 

 rather rapid, but the outermost adradial crenellae show a 

 similar disposition to increase by the same methods. Thus 

 the crenelhie of the ridge-groups, so long as they can be 

 distinguished, display a slight alternation in size. In a radius 

 4'3 mm. long the crenellae can be distinguished for about 

 2*6 mm. from the periphery, after which they coalesce into 

 the parallel or bowed ridges. These last do not unite to 

 form a ring round the central area. The floors are much 

 less depressed than in normal joint-faces, and may even be 

 flush in some specimens. In the Epizygal the radial ridge- 

 groups leading to the cirrus-facets (of which there are two in 

 E 22027) are gradually raised towards the j)eriphery from 

 the point where the crenellae cease to be distinguishable. 

 Iq the Hypozygal (which in the same specimen shows the 

 indent of only one cirrus-facet) the corresponding tract is 

 depressed, and the radial canal opens outwardly into a strong 

 V-shaped notch (Fig. 4). 



A similar tendency to multiplication of crenellae on the 

 syzvgial joint-face is well shown in Isocrinus hercuniae Bather 

 (1909, 'Trias. Echin. Bakony/ pi. iv. figs. 108-110), a 

 species which in some respects approaches Balanocrinus. 



We pass now to the other variations seen in the Biariitz 

 stem-fragments. Most of these have been alluded to by one 

 or another of the previous writers, and have occasionally 

 given rise to doubts concerning the homogeneity of the 

 species. 



The transverse section is rarely quite symmetrical, but, 

 apart from the modification due to the cirri, the variation 

 from symmetry is clearly dne to Mechanical Crushing. 

 This has in some cases flattened the specimens into bands : 

 in E 21943 the diameters are 10 mm. and 4*4 mm. This 

 effect is not without interest, as indicating a very weak 

 internal stereom. A further result of the crushing is the 

 formation of five longitudinal cracks, so that such a stem is 

 at first reminiscent of those Palaeozoic stems that are split 

 into pentameres. The cracks in the Biarritz s[)ecimens, how- 

 ever, are not confined to the radii, where any division into 

 natural pentameres would occur, but are also on the inter- 

 radii, or in other words at the outer apices of the sectors 

 (E 22026, cf. d'Archiac's fig. 16 a). Both positions are 

 taken by the cracks in E 22025, and the fracture on the 

 interradii contrasts with the clean jointing T)n the radii. 



