118 COLLECTIONS FEOM MELANESIA. 



step. Coloration can only be safely used when the hypothesis that 

 the genus or group is of so late an origin as not to have yet deve- 

 loped definite structural characters by which its species may be 

 discriminated is not controverted by any one species ; where it is, 

 the character must be less freely used, probably to disappear more 

 and more, not so much with the increased evolution of the genus as 

 with our increased knowledge of it. 



The succeeding pages contain an account of, or tbe names of, 

 124 species, very fairly divided among the 5 orders. Of these there 

 are 30 new species, 15 of which belong to the Crinoids and the 

 other 15 to the remaining orders. There is no new Echinid ; but 

 I have had to suggest a new name for the form which by some has 

 been regarded as Salmacis globator. 



ECHINOIDEA. 



1. Phyllacanthus annulifera. 

 A. Agassiz, Rev. Ech. p. 387 *. 



A most valuable and important series for exhibiting the variation 

 of this species. 



Thursday Island ; Port Molle (14 fms.) ; Port Curtis ; Prince of 

 Wales Channel ; Albany Island ; Port Darwin. 



2. Diadema setosum. 

 A. Agassiz, Rev. Ech. p. 274. 



Port Molle ; coral-reef. 



3. Salrnacis bicolor. 

 A. Agassiz, Rev. Ech. p. 471. 



A good series, with spines well preserved on most from Port 

 Molle. 



4. Salmacis sulcata. 

 A. Agassiz, Rev. Ech. p. 476. 



Port Dcnison (4 fms.) ; Port Molle (14 fms.). 



5. Salmacis alexandri. 



Salmacis globator, Alex. Agassiz {not L. Agassiz), Rev. Ech. p. 473 ; 

 'Challenger' Reports, iii. p. 113. 



Salmacis globator, form a, Bell, P. Z. S. 1880, p. 433, pi. xli. fig. 1. 



In the third part of my " Observations on the Characters of the 



* As a rule, I shall give only one bibliographical reference, and that to the 

 4 Revision of the Echini ' of Prof. Alex. Agassiz. 



