358 HANS AND OLOF TEUIN 



to me that T stands not for thickness of stem, but as a factor for 

 large number of internodes.» In fact, this view is strongly supported 

 by the results in our crosses I + II and III. 



There might be a correlation between stem thickness and inter- 

 node number, but certain facts make this very little probable. 

 White's statement (1. c.) that all dwarfs known to him have few 

 internodes but many of them thick stem, shows that there can be no 

 general physiological correlation between the two characters, but it 

 does not exclude gametic coupling. Certain facts (e. g. the total 

 absence of correlation betw^een stem thickness and the quotient total 

 length/internode length) which cannot be in detail discussed here, 

 make it, as far as we can see, rather improbable that such coupling 

 occurs in the crosses here recorded. 



Beyond the already mentioned weak and doubtful correlation 

 between stem thickness and internode length, caused by the internode 

 lengthening effect of T, there is another, far more marked, correlation 

 between the two characters. If the coefficient of correlation between 

 stem thickness and internode length is determined for the whole F2 

 in crosses I + II and III it will prove to be: in crosses I + II r ^ 0,223 

 + 0,040, in cross III r = 0,255 + 0,043. Thus r is 5,5 or 6 times its own 

 standard error and the existence of a correlation can not be doubted. 

 This correlation may also be demonstrated as a difference in the 

 average stem thickness between the Le and the U part of Fo. In 

 crosses I + II the average stem thickness in the U part is 3,68 + 0,04-i 

 and in the L^ part 4,08 + O,04o and thus the difference is 0,4o + 0,062. 

 In cross III the corresponding numbers are 4,46 + 0,o82; 4,94 + 0,046 

 and 0,48 + 0,095. The differences being 5 or 6 times their own standard 

 errors they must be of real significance. As the crosses are Th X '^e 

 the correlation here demonstrated cannot be due to gametic coupling, 

 between T and Le, which would have caused a negative correlation 

 in this respect. The explanation must be either that L-» is coupled 

 with another factor that increases the thickness of stem, or that 

 Le itself has this effect. 



The results hitherto obtained as regards the inheritance of stature 

 in peas may, in our opinion, be summarized thus: 1) The length of 

 internodes is mainly determined by one factor. Le whose dominant 

 state produces long internodes. Other factors, however, genetically 

 modify the length of internodes, but it is not known how many or 

 which those factors are. That the factor (factors?) for thick stem 

 may influence the length of internodes, cannot be denied. This in- 



