STUDIES ON HIGH AND LOW NON-DISJUNCTION 109 



2. The normal percentngo of exceptions is 4,3 but in the so-called 

 »eosin line of high non-disjunction» this percentage is about 22,o. 



3. It is shown that A'AT-females homozygous for the X-chromo- 

 somes froui the eosin line always produce the high |)ercentage whereas 

 XÄ'y-females heterozygous for all the chroinosonies from the eosin hue 

 produce the low i)ercentage. From this fact it is also conchided that 

 the production of a high percentage of exceptions is not dependent on 

 a cytoplasmic »factor» only inherited from the females. 



4. The method which has been used for the study of the genetical 

 basis of these different percentages of exceptions consists in the sepa- 

 ration of the X-chromosomes from the eosin line and in making up 

 new ZXy-females which are homozygous for different parts of tiiese A':s. 



5. In this way a number of percentages are found which all lie 

 between the ordinary low percentage and the high percentage typical 

 to the eosin line. 



6. It has not been possible to find any genes responsible for all 

 the various percentages which have been found, but the data indicate 

 that the longer the part of the A'^-chromosomes from the eosin line is, 

 with respect to which the A'A'F-females are homozygous, the higher is 

 the percentage of exceptions among their offspring and the shorter this 

 part is, the lower is the percentage of exceptions. 



7. From this it is concluded that the most probable explanation 

 is that it is the length per se of the part of the A:s (A = A'^-chro mo- 

 some from the eosin line) with respect to which the AAY-females are 

 homozygous which determines the percentage of exceptions. (Cf. the 

 partial summary at p. 103). 



8. Denoting by y the percentage of exceptions and by z the length 

 of the part of the X with respect to which the A'AY-females are homo- 

 zygous it is shown that the approximate relation y = 4,3 + 0,2 z fits the 

 experimental data fairly well. With the aid of this relation the follow- 

 ing three values for the total length of the A-chromosome are found 

 from different experiments: 76,n, 85,5 and 85,7 units. 



9. It has been pointed out that this explanation is to be looked 

 upon as only an alternative one, the other alternative being a compli- 

 cated interaction of a number of different genes for the percentage of 

 exceptions, and that in order to decide which of the explanations is 

 the correct one it is necessary to make a more detailed study of the 

 problem. 



10. If however it is correct that it is the length of the part of the 

 X as such, with respect to which the A'A'Y-females are homozygous 



