114 



S. IKENO : 



From the results given iu the Table VII, A we expect to have 1 

 CCBRBB, 2 CCBi'Bb and 1 CCt-rbb iu the i^-generation. In the 

 Table VII, B showing the i^^-ofeprlug we have the jK-parents of mageuta 

 colour, No. 1-23. Of these No. 1 has produced only 1 magenta plant 

 and is not entered in the total, because it indicates naturally nothing for our 

 experiment. Nos. 2-4 belong to CCBBBB, and Kos. 5-23 to CCBrBb 

 wliilst each of 12 oranges belongs to CCrvbb, because it has produced, 

 nothing but oranges. We have thus 3 CCBBBB: 19 CCBrBb: 

 12 CCrrbb, theoretically 8-5 : 17-0 : 8-5. Of the heterozygous magentas 

 CCBrBb (i. e. Nos. 5-23) each of Nos. 5-12 has segregated into magentas 

 and oranges, just as did the i^, -hybrid iu is, their approximate ratio teiug 

 3:1 in total, and e'sidently this is so, because iu thesa magentas the factoids 

 B and B remain in absolute linkage. On the conti-ary, in each of magentas, 

 No3. 13-23, we observe the appaarauce of the reds, and this is clearly due 

 to the breaking down of the complete linkage between these factoi-s. Since 

 we have in all 181 magentas, 20 reds aud 64 oranges, i.e. !'&% reds, there 

 are too few reds to consider that the free assortment has taken place between 

 B .ind B, because we should have iu the latter case neaily 149 magentas, 

 50 reds aud C(j orauges (9:3: 4), i.e. lS-2% reds. It is quite evident that 

 iu our present case the linkage has been chamjed from complete to paiiicd, or 

 at least has changed its intensity from very high to low. What is then the 

 ratio of "couphng" or "linkage" in the latter case? For its determinatiou 

 the following calculations were made : — 



If we calculate the closeness of fit In- the method of Peaeson we find for the 

 fii"st aud the second case ;^- = 0'1099 and 0'4006 respectively, each of which 

 should indicate an almost perfect agi-eemeut of the actual number with the 



