. Dr. F. A. Bather — Studies in Edrioasteroidea. 319 



to the cover-plates of Edrioaster, as may best be seen by a comparison 

 with Text-figure 2 of Study IV (1914, p. 164). The difference lies 

 in the greater development of articulating muscles and their surfaces 

 of attachment in Arcliaster, especially in the fact that the adam- 

 bulacrals are united to each other by a longitudinal muscle. It may 

 also be noted that in Archaster each adambulacral is attached to two 

 ambulacrals (floor-plates) ; Mr. Spencer, however, points out that the 

 attachments are not really equal, as so frequently implied for this 

 and other modern asteroids, but that one of the attachments pre- 

 ponderates and " connects the ambulacral to the adambulacral with 

 which it corresponds in the series". In the early Palaeozoic Asteroidea 

 each ambulacral is associated only with its own adambulacral, just as 

 in Edrioaster each floor-plate has its cover-plate. In these early 

 forms also, as appears from the descriptions of Schondorf, Hudson, 

 and Spencer, the articulating sculpture and muscle-connections were 

 much less developed. 



The cover-plates of the Agelacrinidae, recently so well described by 

 Dr. Foerste (1914, §§ 16 and 22) are in some respects even more like 

 Asteroid adambulacrals. The inequality of the admedian sides 

 ('facets' W.K.S., loc. cit.) is generally exaggerated, so that in plan 

 the visible portion of the cover-plate " has a spinous prolongation on 

 the proximal side" (Foerste), or is "shaped like a bent finger" 

 (Spencer, 1904), or is what J. W. Gregory calls 'boot-shaped', with 

 the sole towards the mouth. Further, as Foerste has shown, in 

 many of the Ordovician species the vertical section of a cover-plate is 

 roughly sickle-shaped (or boomerang-shaped), the blade corresponding 

 to the part which arches over the groove, and the handle to a narrower 

 extension beneath the adjacent interambulacrals. Where blade and 

 handle join is a slight groove between two ridges, forming an 

 articulation with the floor-plate. Similar passage of the adambu- 

 lacrals beneath the adjacent interambulacral (ventro-lateral) plates 

 may be seen in many modern starfish. The chief difference is that 

 the cover-plates of Agelacrinidae are not one to each floor-plate ; but 

 this doubtless is due to the fusion of the floor-plates into groups, as 

 already explained. 



The adambulacrals of starfish do not close down over the groove in 

 the same way as did the cover-plates of Edrioaster, and the contents 

 of the groove are generally protected by groove-spines borne on the 

 adradial margin of the adambulacral plates. None the less, by the 

 approximation of the two sides of the groove, effected by the ventral 

 cross-muscles, the adambulacrals may be brought quite close together, 

 and in some species they may when thus closed be observed to 

 alternate just like the cover-plates of an Edrioasteroid. See, for 

 example, Ludwig, 1905, "Asteroidea of the Albatross," Mem. Mus. 

 Comp. Zool. Harvard, vol. 33, fig. 51, Pentagonaster ernesti, fig. 135, 

 Paulia horrida ; Koehler, 1910, "Shallow-water Ast. Indian Mus.," 

 pi. xi, fig. 3, Pentaceros indicus, pi. xii, fig. 2, P. reinhardti, and 

 others; Gregory, 1899, op. cit. fig. lb , Lindstromaster ; the ray in the 

 north-west position. 



The alleged invariable presence of groove-spines, and the frequent 

 presence of other spines on the adambulacrals of Asteroidea does not 



