ECHIXIIXE. 233 



pedicellarisD of that family coincides with that which I have reached from the 

 study of all the regular Echini: i. e. that while the pedicellaria? often afford 

 good specific characters (doubtless in correlation with other features) they are 

 not, taken by themselves, rehable as a guide in seeking for the true interrela- 

 tionships of the species. It does not seem that a character of such uncertain 

 value in the Temnopleuridse can possibly become of prime importance in the 

 closely related Echinida3. Taken in connection with the other characters how- 

 ever, the globiferous pedicellarice undoubtedly assist in tracing the differentiation 

 of the species, and in grouping them in genera, and in a few cases they are the 

 most obvious, if not the most important, generic character. 



On turning to the test for the characters upon which to base a natural classi- 

 fication, it is apparent at once that in all the less specialized Echinidse as well as 



in the Temnopleuridse the outline of the test is circular ^ and the ambulacral 

 plates are made up of an adoral primary element and two secondary elements, 



the pairs of pores being placed in nearly vertical arcs of three. From this simple 



ancestral form, development has proceeded along at least four different lines. 



(1) The Temnopleurids have developed more or less sculptured tests, with 

 the coronal plates often united by dowelling, while undergoing little if any modi- 

 fication of the ambulacral structure; until we reach the highly speciahzed 

 condition of Holopneustes where the lateral spreading of the poriferous areas, 

 associated probably with some sort of vertical pressure, has greatly increased 

 the number of ambulacral plates, with accompanying displacement, but without 

 increasing the number of their elements. 



(2) A similar development of the ambulacra, by great increase in the number 

 of ambulacral plates, without altering their tripartite structure has occurred in 

 Tripneustes and its allies in the EchinidiC proper. In the Strongjdocentrotidae 

 and Echinometrida3, development of the ambulacra has been along a different 

 line, for there is rarely any great crowding of the ambulacral plates or displace- 

 ment of their elements but instead there has been a more or less marked increase 

 in the number of component elements in each plate. The connection between 

 these families and the Echinidae is obvious and it is very interesting to note that 

 in the case of each one, there is a genus which might properly be assigned either 

 to the parent or the derived family. 



(3) The Echinomctridse differ from the Echinidie, besides the difference in 

 the ambulacral plates, in the elongated form of the test. Yet in Parasalenia 



1 The case of Microcyphus anmdatufi Mortensen appears to be no exception, as a series of eight 

 specimens shows that the outHne of the ambitus is circuhir or shghtly pentagonal. 



/ 



