64 BRITISH PALEOZOIC ASTEROZOA. 



nection with respiration. It is interesting to note that similarly shaped supero- 

 marginalia are also met with in the Cretaceous genus Stauranderaster. In this 

 genus the outer edge is usually also cut away, so that the ossicle appears breast- 

 plate-shaped (87, pi. xix, fig. 1, and text-fig. 28, pp. 120, 125). 



(2) Changes in the Ossicles of the Disc ami Median Regions of the Arms. 



Studies of the Promopalaeasteridse show that in the most advanced species the 

 accessory plates of the disc, as well as the radialia and the adradialia, tend to lose 

 their regular shape and arrangement, and even to break up into smaller pieces (see 

 Text- fig. 52, p. 93). The same change may take place in other lineages. Thus 

 Palaeaster appears to be a " Hudsonaster " which has undergone secondary 

 alterations of this type, as also do Neopalseaster crawfordsvillensis (p. 102) and 

 MesopaJseaster? hetleyi (p. 101). 



The arrangement of the lineages and their changes is perhaps most easily 

 followed from the table and diagram on p. 65. 



The most reasonable interpretation appears to be, that the primitive stock gave 

 rise at various times to offshoots. Some of these offshoots never went beyond the 

 " Hudsonaster " stage, and reached their highest elaboration in that stage. 

 Oirvanaster sculptus (see p. 70) is a British form which obviously reached a 

 high degree of specialisation in the " Hudsonaster " stage. Other offshoots went 

 through the " Hudsonaster " stage in pre-Middle Ordovician times, and appear 

 among the earliest known Asterozoa as the Mesopalseasterinae. Probably several 

 lineages are represented here. The English " Caractacaster " (p. 80) is, for 

 instance, apparently quite distinct from the true American " MesopaJseaster.''^ 

 Probably other lineages also are represented, but the imperfect condition of the 

 known specimens of many of the species does not allow the forms to be examined 

 in detail. Some such lineage must have given rise to the Devonian Xenasteridae. 



A remarkable offshoot sheaf gave rise to the Promopalaeasterinse. These 

 appear in the Middle Ordovician as large, highly specialised forms which have 

 already gone through both the " Hudsonaster " and " Mesopalaeaster " stages. 

 The lineage illustrates clearly the principle that the maximum of elaboration is 

 often quickly followed by extinction, for no " Promopalspaster " has been found 

 beyond the Upper Ordovician. Smith Woodward (88) has made observations on 

 fossil fishes which recall the secondary alterations in the plates of Asteroids, and 

 especially those changes to be noticed in " Promopalaeaster." He remarks : " There 

 was also a constant tendency for the primitive symmetry of the parts of the 

 skeleton in successive members of a group to become marred by various more 

 or less irregular fusions, suppressions, and subdivisions. Finally, some of the 

 successive species of each group gradually increased in bodily size until the 

 maximum was reached, just before the time for extinction had arrived." The 



