57 



Totals 354 12 347 



A long crossed with the legitimate pollen of a mid, or reciprocally, 

 gives equality of the parental forms, with the occasional presence of 

 the third non-parental type. Amongst a progeny of 695, the short- 

 styled form occurs 6 times, or 1 per 117. Whether this is due to 

 possible causes of error, or has a real significance, I do not at present 

 feel qualified to say, though I incline to the latter opinion. 



Again, a long crossed legitimately with a short, and i-eciprocally, 

 gives a predominance of the parental forms, with the mid-styled 

 plant occurring 12 times in a total of 713, or 1 per 59. 



The crosses between mid-styled plants and short-styled plants give 

 an entirely different result. No regularit}' is visible if the numbers 

 are given in an unanalysed tabular form. 



TABLE 111. 



Mid-styled female x Short-styled male Short-styled female x Mid-styled male 



Longs Mids Shorts Longs Mids Shorts 



Totals 146 355 546 218 395 558 



Analysis of F., Mid x Short and Short x Mid, flowering 1910. 



In Table IV, the first group is the addition of all cases where there 

 is at least one Long-styled plant amongst the four grandparents. We 

 get 118 Longs : 121 Mids : 270 Shorts — a possible approximation to 

 the ratio 1:1:2. 



In the second and third groups there is complete absence of grand- 

 parental Longs. The second gives 23 : 23 : 49 — probably the 1:1:2 

 ratio. But the third gToup gives Mids and Shorts equally, with an 

 occasional Long, the numbers being 6 : 102 : 101. 



Clearly all Mids, or all Shorts, are not similarly constituted. The 

 above families do not give clear evidence as to which form bi'ings in 

 the difference. There are certainly two, and probably more than two, 



