1896.] OF THE PAt/JBOZOlO OjHItJEOiDBA. l02S 



retained Job. Miiller's two orders, as he did also in his latest 

 paper. 



In 1892 a short and pregnant paper by Prof. F. J. Bell [1] 

 lifted the classification of the Ophiurids on to a different plane. 

 Bell recognized the great importance of the vertebral ossicles and 

 that they are of three main types : (1) the " streptospondyline," 

 where the vertebral ossicles articulate by saddle-shaped surfaces, 

 which do not bear lateral processes or pits ; (2) the " zygospondy- 

 line," where lateral processes and pits on the articular surfaces of 

 vertebral ossicles limit the power of movement ; (3) the " clad- 

 ophiuroid" (or astrophiuroid), where the vertebral ossicles articulate 

 by hourglass-shaped surfaces. 



Bell therefore proposed to divide the Ophiurids into three 

 groups : (1) the Streptophiurse, for those with streptospondyline 

 ossicles; (2) the OladophiursB, for those with hourglass-shaped 

 articulations ; (3) the Zygophiura), for those with zygospondyline 

 ossicles. 



The definition of these three orders was no doubt a great 

 improvement on any previous arrangement of the Ophiurids. There 

 is, however, considerable dilliculty in applying this system to 

 the fossil forms, especially in the case of the Streptophiura). It 

 appears doubtful whether even some recent genera, as Opliiohelus, 

 can be correctly described as having vertebral ossicles articulating 

 by ball-and-socket joints. But this statement certainly cannot be 

 made of many Palaeozoic Ophiurids, which represent a more 

 primitive condition than that of the recent species ; they are indeed 

 so primitive that they cannot be made to enter into any of Bell's 

 orders. 



The two most striking characters of these Pateozoic genera are 

 the absence of ventral arm-plates' and of true vertebral ossicles. 

 The latter are represented by free paired plates, hke the ambu- 

 lacral ossicles of Asterids. 



The ambulacral ossicles are the most important plates in the 

 arms of both Asterids and Ophiurids, so that it is d, ■priori probable 

 that they offer a better basis for classification than the external 

 arm-plates. As we descend from the Zygophiuree, first to the 

 Cladophiuras, and then to the Streptophiuraj, we notice a decrease 

 in the complexity and completeness of the vertebral ossicles. It 

 is not therefore surprising, when we go back to Palajozoic times, 

 to find Ophiurids with an arm-structure still simpler than anything 

 found in the Strei)tophiur8e. In these early forms the central 

 arm-ossicles occur as a double series of free plates, below which is 

 an open ambulacral groove. Hence the arms appear, at first sight, 

 to bo Aslcrid rather than Ophiurid in arrangement. 



Hence I propose to found a fourth order of Ophiuroidea to 

 include those without vertebral ossicles, but which have in each 

 arm a double series of free ambulacral plates, which articulate like 



' This cliai-acler is also found in the genua Ophioieresis of BeU, one of the 

 most primitive of living Ophiurids ; it has, however, vertebral ossicles with 

 streptospondyline articulations. 



