R N. Salaman and J. W. Lesley 33 



" prostrates " formed a distinct class combining the creeping habit with 

 the extensive deficiency of interfascicular secondary growth character- 

 istic of the type. 



The families are derived from nine distinct Pi " upright " and three 

 distinct Pj " prostrate " parents. Seven out of the twenty-one families 

 contain a small proportion of " prostrates," while the remaining fourteen 

 contain only upright plants. The proportions certainly do not point to 

 a single factor difference in any. On the contrary, all the larger families 

 clearly show that the " prostrate " and " upright " differ in more than one 

 factor. The five F^ families from the same cross (1235) are especially 

 remarkable. One contains " prostrates " in a proportion approaching ^, 

 whilst the others contain none. If we put them together they give : 



" Upright " 409 to " Prostrate " 6, 



nearly a 63 : 1 ratio, but on this basis the figures in 1285, 2 are not 

 within the limits of simple sampling thus : — 



Actual Numbers Calculated on 63 : 1 So- ' 



81:6 85 -64: 1-36 3-47 



SO that the deviation is about four times the standard error. Of the 

 other families 1455, 13 suggests a 63 : 1 ratio, as do the two families 

 from 1275. 



Taken all together the figures are : 



"Upright" 826, 

 " Prostrate " 13, 



or approximately 63 : 1, but of course it is not certain that this summa- 

 tion is legitimate. 



Pa plants from crosses between " uprights " and " prostrates " were 

 selfed either in the first year or when grown on from tubers. All the 

 seed obtained was from protected flowers, in fact the entire succession 

 of Pi and Pa families in this series were from similarly bagged seed. 

 Table VI shows the results obtained in these P3 families. 



Now if, as the proportion in Pg led us to suppose, the " upright " and 

 "' prostrate " parents difier in three factors, any one of which in a single 

 dose gives "uprightness," Pg "uprights" when selfed should prove to be of 

 four different kinds ; they will consist of pure " uprights " and of impure 

 " uprights " of three distinct kinds, each yielding " uprights " and " pros- 

 trates " in different proportions. More than half (f |) will be pure 

 " uprights," the remainder (f|) impure ; of these latter nearly half (^) 



Jouru. of Gen. x 3 



