Herbert L. Hawkins — Studies on the Echinoidea, etc. 253 



in Cu7iulus might be called " parallel variants ", on the supposition 

 that there is no direct phyletic sequence between Conulus and 

 Discoides. 



(2) Variant 2. (PI. XVI, Figs. 4 and 5 ; cf. also Figs. 6 and 9.) 



One specimen from Broadwindsor (Fig. 5) and two from Bridport 

 (Fig. 4) show the development of a supernumerary plate more or 

 less midway between genitals 2 and 5. In both the Bridport forms 

 this is small and pentagonal, having sutural connexion with all the 

 genitals except plate 4. (Both specimens have the " Broadwindsor" 

 arrangement of the genital plates.) In the Broadwindsor specimen 

 the supernumerary is large and hexagonal, and, being in contact with 

 all five genitals, acts as a kind of " centrale ". There is no evidence 

 to prove that this included plate is not really genital 5, while the 

 plate in the posterior region is the additional one ; but it seems 

 more reasonable to suppose that the appearances are not deceptive, 

 and that the interna! plate is actually the supernumerary. 



Additional internal plates within the true cycles of the apical 

 system occur in the Calycina among Kegular, and in some of the 

 Ulypeits—Nucleolites series of the Irregular, Echinoids. This distri- 

 bution of supernumeraries opens up an interesting speculation when 

 applied to the variants under notice. ^Acrosalenia is an ancient, and in 

 most respects, primitive member of the Diademoida (Centrechinoida), 

 and may well have originated collaterally with the Holectypoids in 

 Liassic times, even if it is not on their line of descent. So that 

 a comparison of the aberrant Holectypus with Acrosalenia would class 

 the variation in the former as either arrested (or regressive) or 

 perhaps "arrested parallel". In the case of "Nucleolites" (as illustrated 

 by " JV." orbicularis (PL XVI, Fig. 9), a considerable number of more 

 or less symmetrically placed supernumeraries occurs posteriorly to the 

 large madreporic genital. In Fig. 9, as I interpret it, plate c (the 

 most posterior) represents genital 5, and either b or a might be 

 correlated with the additional plate in Holectypus. In view of the 

 fairly certain primitiveness of Holectypus, and the equally probable 

 lack of direct sequence from that genus to the Xucleolitidas, a com- 

 parison of the two would class the variant as an imperfectly 

 "progressive parallel". This comparison will appear less far-fetched 

 when the next section has been read. 



(3) Variant 3. (PI. XVI, Fig. 6 ; cf. also Figs. 8 and 9.) 



This solitary specimen from Broadwindsor shows a very striking 

 abnormality. In the first place, it possesses the characters of variant 1 , 

 in that the madreporite is partly situated on genital 3s There is 

 a " centrale "-like, hexagonal, imperforate plate a little behind the 

 centre of the system, approximately similar in area to the two 

 genitals (1 and 4) that flank it. Behind this included plate the 

 two posterior oculars (I and V) meet. Both are enlarged and dis- 

 torted in shape, but plate 1 is by far the larger of the two. There 

 is thus no posterior genital (5) ; at least, as concerns the margin of 

 the system. At first I w r as of the opinion that the included hexagonal 

 plate was actually the fifth genital shifted anteriorly. But in view of 



