42 BULLETIN 88^ UNITED STATES NATIONAL MUSEUM. 



lay much stress upon this feature among the Ordovicic asterids, nor 

 does he regard it even necessarily as of specific value, but in the later 

 Asteroidea the arrangement is always opposite or nearly so. Ver- 

 rill (1914:20) states that he has noticed the alternate arrangement 

 in living Pycnopodia "as an abnormal variation in some of the 

 rays, and also that it may occur from lateral bending." Evi- 

 dently the alternate disposition is a fixed character among some 

 Ordovicic forms, as in Anorthaster, but more will certainly be dis- 

 covered in these old strata and probably rather among the Crypto- 

 zonia than the Phanerozonia. The writer is led to this view because 

 of the alternate arrangement in many Auluroidea, a subclass of 

 Stelleroidea that seems to have had its origin in some cryptozonian 

 asterid. Out of the Auluroidea \vith opposite ambulacraha came the 

 Opliiuroidea with their highly modified ambulacrals which are 

 coossified into ''vertebral ossicles." In the Auluroidea the arrange- 

 ment of the ambulacrals is either alternate or opposite, but they are 

 never coossified but are so modified as to suggest the Ophiuroidea 

 rather than the Asteroidea. The structure of these plates is de- 

 scribed elsewhere and need not be repeated here. 



The ambulacrals are very variable in number throughout the 

 genera. In Hudsonaster, Stenaster, and Tetraster there are as few as 

 10 in a column, while in Promopalxaster there are certainly 50 and 

 possibly as many as 60. The cryptozonian genus Vrasterella has 

 certainly over 100 in a column. In general, these ossicles abut 

 against each other, but in Vrasterella and other genera they overlap 

 proximally to a certain extent. As yet the writer has seen no 

 Paleozoic asterid with more than 2 columns of ambulacrals, though 

 in Promo palseaster magnijicus the proximal areas of the rays for a 

 short distance have 4 columns of podial openings. Here, however, 

 the ossicles have not yet wedged themselves into 4 columns of 

 ambulacraha. 



In shape and surface contour there is also considerable variation 

 among the ambulacraha. In Hudsonaster, Tetraster, and several 

 Devonic genera they are more or less rectangular, either squarish or 

 drawn out laterally. In other forms of Hudsonaster, along with 

 Stenaster and Vrasterella, the ossicles are 1— -shaped, with the lateral 

 shaft thin, thus allowing for large, elongate podial -openings. The 

 more primitive genera have the outer surfaces plain or shghtly ridged, 

 wliile in derived forms hke Promopalseaster the ridges are high and 

 straight or have their medial terminations orally directed. In 

 P. magnijicus, the ridges in the proximal region are arranged in pairs 

 having the shape of a tuning fork; in P. wyliojfi one sees that these 

 ridges approach one another in pairs transitional to their transforma- 

 tion into the tuning-fork form. 



