144 HAWAIIAN AND OTHER PACIFIC ECHINI. 



viduals. The length of stalk is also very variable, though it commonly 

 exceeds that of the head. The structure of the stalk is of the usual type 

 in the Diadematidse, a cylinder of irregularly but closely united coarse rods. 

 It is very interesting to note that in Kamptosoma the stalk is made up 

 of 'Mong threads, almost only united at the ends" (Mortensen); the stalk 

 in this genus, therefore, bears the same relation to the stalk in other Echino- 

 thurids that the stalk of the pedicellariae in Micropyga does to that of the 

 other Diadematids, — a very notable case of " parallelism." The valves of 

 the tridentate pedicellariae vary so much in size and form that a general de- 

 scription of them is impracticable, but certain kinds are very constant and 

 easily recognizable, and these deserve a few words. First of all are involute 

 valves, which are more or less curved and meet only at tip (PL 66, fig. I) ; 

 in these the margins of the blade are rolled in, forming a nearly cylindri- 

 cal middle part of greater or less length between the base of the blade and 

 the somewhat expanded tip ; secondly, there are contiguous valves, which 

 are straight and in contact for nearly their entire length, — these may be 

 very broad (PI. 65, fig. 18), or rather narrow (PL 64, fig. 5) ; a third sort 

 may be referred to as convercjent valves, which are straight or little curved, 

 little or not at all involute, meeting only at tip or for their distal half ; such 

 valves are shown on PL 65, fig. 5 ; see also PL 67, figs. ^ 12. Although 

 tliese three sorts of valves are usually recognizable with ease, they do inter- 

 grade more or less (PL 65, fig. I^ ; 66, fig. 2). The ophicephalous pedicellariaa 

 (PI. 65, figs. 1, 2) are much less common, as a rule, than tlie tridentate, 

 and have been found as yet in only a comparatively few species ; they are 

 strongly constricted at the middle, and the articular loops on the valves 

 (PL 65, fig. 3 ; 67, fig. 21) are well developed ; these pedicellariae show no 

 tendency to intergrade with the other forms. The fourth form of pedicelhiria 

 was first described by Wyville Thomson as a " tetradactyle " pedicellaria, 

 and Mortensen has retained the name. Unfortunately, however, the num- 

 ber of valves is often three or five, so that the '' tetra" — is very inappro- 

 priate ; wo would suggest that this form be called simply the dacft/hus 

 pedicelhiriiTD. They are known only from cerUiin species of Anrosoma, 

 and show considerable diversity in the form of the valves; these may be 

 long and slender, united only near the base and quite free distally (PL 66, 

 fig. 20), or they may be nuich shorter and well surrounded by tissue as 

 de Mcijere figures them, or the valves may be wholly imbedded in tissue 

 (PL 64, fig. i). The valves differ notably in shape (compare PL 64, fig. 4, and 



i 



