22 HAWAIIAN AND OTHER PACIFIC ECHINI. 



16-18) the primary spines are spatulate or swollen at the tip, at least abactinally. 

 In ravenelii, the primaries around the mouth (PI. 122, figs. 12, 13) are noticeably 

 modified. So far as the mihary spines are concerned, in audouini, japonicus, 

 leptostracon, rarispinus, and virescens, they are more or less swollen or club- 

 shaped at tip (PI. 122, figs. 8, 10), while in all the other species they are cylindri- 

 cal or terete (PI. 122, fig. 7), though in lamprus, and in individual cases in other 

 species, some of them may approach the club-shaped form. 



The quadridentate pedicellariae (PI. 123, figs. 10, 11) have so far been found 

 only in ravenelii and subdepressus , but they will probably be detected in some 

 other species. Their valves (PI. 123, fig. 7) are narrow, compressed and meet 

 only near the tip. The tridentatc have been found in all of the nineteen species 

 but they show considerable variety in form and in a few species (audouini, 

 australasiae, humilis, rotundus) they seem to be very scarce. The valves 

 may be broadly in contact (PI. 123, figs. 20, 21, 29) or meet only at the tip (PI. 

 123, fig. 1; PI. 124, fig. 1); they may be long and narrow throughout (PI. 123, 

 figs. 15, 16) or expanded at the tip (PI. 123, figs. 2, 3, 12, 19; PI. 124, figs. 2-6), 

 or broad and somewhat leaf shaped (PI. 123, figs. 26, 27). In lamprus, the blade 

 is almost tubular and only a little expanded at the tip (PI. 123, fig. 22). The 

 ophicephalous pedicellariae (PI. 123, fig. i 7) are less common than the tridentate 

 in most species and were not found at all in audouini, prostratus, and rotundus; 

 they were very scarce in australasiae and humilis. The opening of the blade 

 (PI. 123, figs. 4, 28) is broad and low and surrounded by conspicuous teeth. The 

 loops vary much in size and form in the three valves of the same pedicellaria 

 (compare PI. 123, figs. 5 and 6 or figs. 13 and H). In many cases, the largest 

 valve has a notable bihamate loop (PI. 123, figs. 8, 9, 30, 31), and these hooks 

 may even unite at the ends wdth the sides of the loop (PI. 123, fig. 23). The 

 triphyllous pedicellariae (PL 123, fig. 18) are so small they are difficult to find, 

 and while they probably occur in all the species, I failed to find them in several. 

 The valves (PI. 123, figs. 24, 25) are broad and fiat with finely serrate margins. 



Key to the Species of Clypeaster. 



Margin of test very thick, the upper surface rising so uniformly from ambitus to madre- 

 porite that a real margin can hardly be measured, but even in flattened individ- 

 uals it is rarely less than .30 test-length; height of test rarely less than ..3.5 test- 

 length; lower surface of test deeply concave. 

 Test not evenly convex above, the median area of petals being more or less markedly 

 elevated, not only above the poriferous areas but above the interambulacra also; 

 pore-pairs in petals numerous (more than forty-five on each side in an unpaired 



