19G 



is assumed to be a bivalent chromosome, and the tetrad formation is 

 therefore interpreted as follows : — abed— I (spireme) ah — cd — JcJ (segmen- 

 ted spireme)" etc. is incorrect in so far as the formules are concerned. 

 The number of letters from a to I inclusive is 12, I distinctly main- 

 tained that the number of chromosomes in the spireme stage is 24. 



When I proposed the formula — t—j for a tetrad in Caloptenus, my only 



purpose was to use a formula which would indicate that all the four 

 chromosomes of a tetrad are unlike, and hat I thought it necessary, 

 to avoid misunderstanding, to give letters for all the chromosomes, 

 the complete formula would plainly have been not abed — I but abed — x, 

 which latter series amounts to 24 in number. As a matter of fact I 

 have nowhere used the formula abed— I. This formula is, therefore, 

 not my interpretation of the spireme-stage in the formation of tetrads 

 and I can scarcely be held accountable for the difficulties which this 

 interpretation entails. 



So much für the arithmetical side of the matter. 



A much more interesting question is raised by Wilson's assump- 

 tion that a tetrad must not be made up of four unlike chromosomes. 

 To quote again from Wilson, ''his primary tetrad must therefore be 



not — T—, as he assumes, but either y- or (if we assume that the 



e \ d 



normal number of chromosomes undergoes a preliminary doubling) 



j-rj • Underlying this statement there is manifestly the assumption 



that a doubling of the chromosomes involves a division of the chromo- 

 somes which is qualitatively and also quantitatively an exact halving. 

 This assumption simply begs the whole question under discussion. As 

 far as I can discover, the only essential point on which we differ is 

 just the point which Wilson settles by an assumption. Is there in 

 all cases a preliminary longitudinal splitting of the chromatic thread 

 in the spireme stage? Wilson believes that there always is such a 

 longitudinal splitting, relying upon the work of vom Rath, Hacker, 

 RücKERT and others. I have maintained that in Caloptenus there is 

 no longitudinal splitting. To assume that even in Caloptenus there 

 must be a longitudinal splitting is certainly no argument against my 

 position, nor is it at all apparent how this assumption shows my ac- 

 count to be self-contradictory. But if in the place of my formula 



for the tetrad — --, we should substitute Wilson's proposed formula 

 f~rr , the account would then be plainly self- contradictory. It 



