136 MENDELISM chap. 



that found in the tetrad stage. As is shown in 

 F'g- 39> we may suppose that the first division is 

 merely quantitative, resulting in two cells each 

 containing a representation of the two maternal 

 and of the two paternal chromosomes. The second 

 division, however, we must suppose to be qualitative, 

 the maternal and paternal portion both of the, long 

 and of the short chromosomes separating from oiie 

 another. This may occur in two different ways. 



The maternal part of the long gamete may be 

 associated with the maternal part of the short 

 gamete (as shown in a), in which case the -paternal 

 part of the long gamete must be associated with the 

 paternal part of the short gamete ; or (as shown in 

 0) the maternal part of the long gamete may be 

 associated with the paternal part of the short one, 

 in which case the paternal part of the long one 

 must be associated with the maternal part of the 

 short one. As to which of these two things happens" 

 the chances are equal. Consequently the gametes 

 produced, as shown in the bottom line of Fig. 39, 

 are of four different, kinds which are found in equal 

 numbers. 



Let us now suppose that our original maternal , 

 gamete contained two Mendelian factors A and B, , 

 one in each chromosome, and that the paternal 

 gamete contained neither. The zygote must be of 

 the constitution Aa, Bb, and should produce equal 

 numbers of gametes of the four kinds, AB, Ab, aB, 

 ab. Clearly this is in accordance with the chromo- 

 some theory, provided we assume that each different 

 factor is carried in a different pair of chromosomes. 



In the earlier days of Mendelian studies the close 



