GABRIELLE L. C. HOWARD. 59 



and the tall plants form exactly one quarter of the whole. The 

 large range of variation in the tall forms indicates that the ratio 

 is not a simple 3:1. Cultures 27 and 231 show a similar 

 result, an accumulation below 95 cm., a total break, or very 

 few plants about 100 and a rise beyond this. These three 

 cultures Avere derived from three F^ plants of very similar height, 

 83, 88 and 90 cm. The curve formed by combining these 

 cultures is shoA\ai in Plate XIV. Culture 8 is very similar to 

 these three cultures, but the curve appears to be shifted to a 

 higher value. The cultures 9, 35 and 163 are all similar 

 and seem to point to at least two homozj^gotic combinations at 

 about 120 and 170 to 180 cms., with possible intermediates. 

 The combined curves from these three cultures is given in 

 Plate XIV ; the individual cultures all agree with this. A 

 large number of plants have been self-poUinated and the Fi 

 should give sufficient uniform cultures to identify all the factors. 

 The values obtained in some of the cultures seem somewhat 

 high for a rudimentary height of 50cm. common to both parents. 

 The fact that in all the cultures growTi in the Fo and F3 

 notliing shorter than 50 cm. has been throAMi out would point 

 to nothing smaUer than this combination existing in either ; yet 

 if this factor or combination be common to both, the mean of the 

 highest combination possible can only be (85 — 50)+ (106 — 50) + 50 

 cm. The limits of the tallest homozygotic combination 

 should also throw light on another interesting point. Since 

 every plant must have height, it is reasonable to suppose that all 

 types possess a common factor which represents the smallest 

 height which the tobacco plant can possess. This, either by the 

 superposition of numerous small factors or by combination with 

 various independent factors, gives the height of the different 

 plants. The first supposition of small superposed factors is 

 entirely disproved by the cross (Type 16 x Type 35), in which 

 plants have been produced almost equivalent to the combined 

 height of both parents, showing that the constituents of the 

 height of both parents must be almost all different. Should a 



