ANEMONES 397 



keeping specimens of JSIetridium under observation for a few weeks. 

 In a few days most of them will show signs of laceration at the edge of 

 the pedal disk, and some surrounded by groups of young. Occa- 

 sionally the entire edge may be torn away at once, later breaking up 

 into several pieces each of which usually gives rise to at least one indi- 

 vidual. 



Basal fragmentation is especially interesting as a factor in the varia- 

 tion of the mesenteries and siphonoglyphs, so remarkable in M. 

 dianthus. Parker, in 1897, discussed the variation of these structures 

 in a very interesting paper, and suggested that the variation in the 

 number of siphonoglyphs might be correlated with the methods of re- 

 production. He appears to have thought that all monoglyphic individ- 

 uals possibly arose as small buds which were cut off from the margin 

 of the animal between its aboral disk and column ; while all diglyphic 

 polyps may have been sexually produced. In a later paper he 

 slightly modified this suggestion by correlating mesenterial variation 

 also with method of reproduction, expressing the opinion that regular 

 hexamerous diglyphic specimens are possibly the result of sexual re- 

 production only. Whether monoglyphic or irregular diglyphic 

 individuals are, in his opinion, ever produced sexually he does 

 not say. 



The great numerical importance of basal fragmentation among the 

 modes of reproduction in Metridiu7n and the irregularity of the mes- 

 enteries in the fragments which grow into new polyps, render a statis- 

 tical method of inquiry into the suggested correlation of type and 

 reproductive process necessarily inadequate. The problem will 

 approach complete solution only when we have overcome the difficul- 

 ties in the way of raising polyps from the Q^%' However, I have de- 

 cided to set forth the results of my observations on the variability of 

 Metridium^ because they not only differ from those of Parker and 

 McMurrich in some details but offer grounds for believing that the 

 variation in Metridiiun is not necessarily correlated with the method 

 of reproduction. 



It will be Avcll to consider first the methods of reproduction in 

 Aletridiufn^ and the relative importance of each in producing varia- 

 tions. 



Besides the sexual method Metriditi?n reproduces non-sexually by 

 means of (a) fission, {b) budding, and (c) basal fragmentation, al- 

 ready considered. 



Fission is longitudinal, cither equal or unequal, more often the latter. 

 Usually the plane of division passes through one siphonoglyph, though 



