CONTRIBUTIONS TO THE GENETICS OF PISUM 3G1 



different numbers of sterile nodes, from lines with 6 — 7 and up to 

 lines with 24 — 25. v. Umscii (1921) has recordt'd an analogous case 

 in Hordciiin, where a great many lines with small dilierences in the 

 length of beard are lound, but where a cross between two extreme 

 lines gives monohybrid segregation. In order to explain this she 

 suggests a series of multiple allelomorphs with small diflcrences 

 between different members of the series in the power of »lengtiiening 

 of l)eard». An analogous explanation might be used in the case here 

 discussed. The existence of polymeric factors (Nilsson-Ehle) must, 

 however, be remembered, and moreover the fact, often pointed out 

 e. g. by JoHANNSEN, that a factor does not work isolated but together 

 with the whole complex of factors in the zygote. Therefore it seems 

 somewhat unnecessary to suggest multiple allelomorphs and varia- 

 tion of valence in cases, where the known segregation phenomena may 

 be explained in another way. 



CORRELATIONS BETWEEN NUMBER OF STERILE 

 NODES AND OTHER CHARACTERS. 



The correlation between earliness of flowering and ripening and 

 a low number of sterile nodes has already been pointed out (Tedin 

 1897). It may only be added that in no line cultivated here at 

 Svalöf a breaking of this correlation has been observed. The state- 

 ment of great number of sterile nodes as a dominant character coin- 

 cides well with the fact, pointed out by different authors, that lateness 

 dominates earliness in pease. 



In cross III (the only cross in which each plant may be classi- 

 ficated in respect both to Lc and to Sn) the segregation in SnLe : SJe' 

 SnLe '• sJe is 290 : 106 : 96 : 30, corresponding to an expected segrega- 

 tion — after 9:3:3:1 ratio — in 293,62 + 11,33 : 97,88 + 8,92 : 97,88 + 

 8,92 : 32,62 ± 5,53 and thus in all four combinations the difference is 

 less than the standard error. Hence there is no sign of coupling 

 between Le and Sn- 



Between thickness of stem and number of sterile nodes there is, 

 however, a marked correlation. In crosses I + II the coefficient of 

 correlation between those two characters is r = 0,375 -f 0,036. In cross 

 III r-= 0,208 + 0,042, in cross IV r ^ O.ssi + 0,08i. Thus, r being from 

 4 V2 lip to 10 times its own standard error, there is no doubt that low 

 number of sterile nodes is correlated with thin stem. As to the nature 

 of this correlation nothing is with certainty known, it may be a 



