DKNDKUCOMETES PxVRADOXUS. 331 



difficulties of the investigation, but the establishment of the 

 fact that variations of such an important character do occur 

 is, in my opinion, a result of considerable interest. 



At some time during the last three stages (H, J, K) the old 

 meganucleus becomes very large, and is bent on itself in the 

 form of a loop or horseshoe. One extremity of this figure 

 passes into the conjugative process, and approaching the 

 limiting membrane traverses it and fuses with the correspond- 

 ing extremity of the meganucleus of the other individual. 

 The exact phase at which this meganuclear conjugation takes 

 place seems to vaiy considerably; all that can be said at 

 present is that, so far as my experience goes, it usually 

 occurs between Stages J and K. The number of cases of 

 actual contact of megauuclei observed by Mr. Wadsworth 

 and. me is small, and this may be interpreted to mean either 

 that the period of meganuclear conjugation is very short or 

 that it does not always occur. Until some satisfactory 

 method is invented of watching the nuclear phenomena of the 

 conjugation of living Dendrocometes, it is impossible to 

 prove that the meganuclear conjugation never fails. I am 

 inclined to believe that it always occurs. Similarly I have no 

 proof to offer of the length of time occupied by this process; 

 but I am inclined to believe, on the circumstantial evidence 

 at my disposal, that it is very brief. 



Soon after the nieganuclei have conjugated they sepai'ate 

 and begin to degenerate. 



'J'he usual phenomena of nuclear change during conjuga- 

 tion in Dendrocometes may be represented by the following 

 diagram, in which the circles above M represent the stages in 

 the meganucleus, and the black dots above m, m, m, the stages 

 in the history of the micronuclei. Conjugation of the mega- 

 nuclei usually occurs in Stage J, as explained above, and in 

 Stage K these bodies disintegrate. 



In the following diagrams I have endeavoured to interpret 

 certain phenomena Avhich appear to be variations of the more 

 usual stages. In Diagram B, which starts with Stage G, 

 the important variation is that two of the micronuclei formed 



