376 



TORREY 



however, that Edwardsia^ as it exists to-day^ cannot be the ancestral 

 form of the Actinians whether the mesenterial arrangement is primi- 

 tive or not. This statement would apply equally well to Ed-wardsia 

 even though the rudimentary hexactinian mesenteries were absent. 



On page 3S3 I have presented a scheme of the possible interrela- 

 tionships of the different species of Edwardsia on the basis of mesen- 

 terial formulas merely. Whether such a scheme truly expresses the 

 phylogenetic development of the group remains to be seen. 



Family EDWARDSIID^ Andres. 

 Genus Edwardsia Andres. 



Physa well developed, retractile, adhesive. Number of tentacles 

 variable, from fourteen to thirty-six, arranged in two cycles. Mesen- 

 teries equalling or surpassing the tentacles in number. Capsules con- 

 taining nematocysts numerous, arranged in rows, or scattered. 



From time to time since 1S84 when Andres established this genus 

 for the reception of four species of Edwardsia with twenty tentacles 

 or more, including E. sipunculoides^ the adequacy of the number of 

 tentacles as a criterion of the genus has been questioned. At first sight 

 it seems to separate forms with sixteen tentacles from those with 

 twenty. It seems even more so when it is remembered that the species 

 with sixteen tentacles typically may have but fourteen as in E. bcaji- 

 tempsi or fifteen as in E. adenensis. It was an appreciation of this 

 difficulty which led Carlgren to suggest, in 1893, that t\iQ arrangement 

 of the tentacles was of more fundamental importance than their num- 

 ber. With this idea in mind, he formulated three types of tentacular 

 arrangement. At present we are concerned with but two of these, the 

 so-called Edwardsia and Edwardsiella types. According to the 



Fig. 5 . Ediuardsia type 

 (Carlgren). 



Fig. 6. Diagram showing ar- 

 rangement of mesenteries in 

 Ed-wardsia sp. (San Pedro). 



