Collected by the late Dr. J. I. Northrop. 185 



Mediterranean. The distinction between Ragactis and Heteractis 

 concerns practically the tentacles only, the evaginations upon 

 these structures being in the former genus small and appearing as 

 tubercles, while in the latter they involve the entire circumference 

 of the tentacle, which thus becomes moniliform. How far this 

 distinction is valid can only be determined by further examina- 

 tion of the anatomy of the various species assigned to the two 

 genera, and for the present it seems to me more convenient to 

 retain for the species under discussion the name proposed hy its 

 sponsors. 



An interesting fact which has developed from the examination 

 of H. lucida is that it is unquestionably a Sagartid. The occur- 

 rence of acontia (to which my attention was first called bj^ Dr. 

 Northrop) and of cinclides, the existence of only six perfect 

 mesenteries and the mesogloeal sphincter point indubitably to 

 this relationship, and it will be interesting to know whether 

 Ragactis pulchra and the other species assigned to the genus 

 Heteractis are likewise Sagartids. Andres has grouped together 

 in the family Heteractidse several genera, e. g., in addition to the 

 two already mentioned, Eloactis, Rhopalactis and Stauractis. Of 

 these, as I have already suggested ('93), the genus Eloactis, as 

 judged by the species E. Mazeli, is probably with greater pro- 

 priety referable to the family Halcampidae, and I have now 

 shown that one at least of the members of the genus Ragactis 

 must also be transferred to another family. Families based upon 

 miere external characters have no firm foundation, tentacles may 

 vary enormously in shape, warts or tubercles may be absent or 

 present within the limits of the same famil}', and cases like the 

 present show, if proof be still wanting, that the criteria for a 

 proper classification of the Actiniaria are to be obtained only by 

 the careful study of the internal anatomy of the various forms. 



We must add then, provisionally, to the genera included in the 

 family Sagartidse, the genus Heteractis, since one of the forms 

 assigned to it has proved to be a typical Sagartid., Whether, 

 however, the type species of the genus, H. aurora, will also prove 

 to be a Sagartid, or whether it will be necessary to remove H. 

 lucida from the genus Heteractis, remains to be determined. And 

 furthermore the affinities of Ragactis pulchra must also be ac- 

 curately studied before the proper generic term for H. lucida can 

 be finally determined. 



