No. 639] IXHEBITANCE IN NICOTIANA TABACUM 



331 



tions give totals of 113 pink: 44 red: 42 white, whilst the 

 9:3:4 expectation, disregarding fractions, is 112 pink: 37 

 red : 50 white. In the back-crosses to both the white and 

 the red parents the data are obviously in satisfactory- 

 agreement with the 1 : 1 expectations. These figures do 

 not establish conclusively the validity of a bigenic for- 

 mulation for this case, but taken together with the data 

 from the Alba-Mackophylla series which we have pre- 

 sented elsewhere*' it seems most reasonable to interpret 

 them in this manner. An alternative mono-hybrid inter- 

 pretation might be argued, but it would not fit the 

 totals as well as the dihybrid ratio. The growing of F.^ 

 populations would, of course, soon settle the question, 

 but the results so far secured indicate essential genetical 

 identity of alha and "Cuba" in their flower color factors. 



In order to demonstrate the ditiference in behavior of 

 red of macrophylla and carmine of purpurea we have 

 made parallel crosses between them and a innnl)or of 

 other Tahacum varieties. The flower coloi.- of these 

 varieties and of their F^ hybrids with it/ti< rnj)li//llu and 

 purpurea are listed in Table II. In each case tlie l\ with 

 macrophi/Ua was pink but with purpurea it was always a 

 full, intense carmine. Among two hundred plants of the 

 CuBA-PuKPUREA scries one plant appeared which bore 



TABLE II 



I\ Eesults of Parallel Crosses of Macrophylla and Purpurea with a 



angustifolia (U. C. B. G. 68/07) . 

 "Cavala" (U. C. B. G. 72/05).. 

 "Cuba" (U. C. B. G. 200/14). . , 



kvers on one side and light pink ones on 

 other. Further studies on this, the most striking 



6 Setchell, Goodspeed, and Clausen, loc. cit. 



