XII.] CONJUGATION AND SEXUAL REPRODUCTION. 187 



Objections. 



It will be urged against my views that they are based upon 

 a method of formation of the reproductive nuclei, which, although 

 common among Infusoria, is by no means the only one. As 

 a matter of fact, Maupas, whose researches form the only 

 foundation for this part of ni}'- argument, describes another 

 method in the Oxytrichidae. If I neglect the fact that in this case 

 two micronuclei are found in the animal preparatory to conju- 

 gation, it is because this difference is merely due to the fact that 

 two of the grand-daughter-nuclei (instead of only onej undergo 

 a second division. Thus two pairs of reproductive nuclei 

 arise, of which only one is functional, while the other disin- 

 tegrates. But the theoretical explanation is in no way affected 

 by these observations. 



The only facts which do not at once harmonize with my 

 view is the behaviour of the micronucleus in male Vorticdlidae. 

 In this case the period of growth of the micronucleus (stages 

 A^ — A-) is preceded by its division. I cannot at present 

 explain this, unless it simply means that instead of four dif- 

 ferent combinations of idants out of which one functional 

 reproductive nucleus is to be chosen, eight are in this case 

 afforded. A glance at the figure given by Maupas (op. cit. 

 p. 364) at once renders this suggestion clear. In any case, the 

 extra division must be an ' equal division.' 



Thus the departures from the ordinary modes of division of 

 the micronucleus raise no definite objection to my explanation. 



Evidence that the processes which I have explained as 

 ' reducing divisions ' are really such, is afforded by some of the 

 figures given by Maupas, as in figs. 9-13 on Plate XVIII, in 

 which the development of the spindle for the nuclear division of 

 Onychodromus grandis is represented. The rod-like chromato- 

 somes lie longitudinally in the spindle, and appear to be dividing 

 transversely. Since we must imagine that the ids are arranged 

 lengthwise, the transverse division of the idants must lead to 

 a diminution in the number of ids in each rod to half their 

 original number. Complete certainty cannot, however, be 

 attained by an examination of these figures ; the matter must 

 be settled by fresh observations, especiall}^ directed to the point. 

 The whole mechanism of nuclear division differs in essential 



