298 Journal of Agricultural Research voi. iv, No. 4 



oped. The same is true for all other series containing I or i. Futhermore, when 

 neither B nor R is present, as in Plate C, figure G, red pigment develops in the eye, 

 probably owing to the factor I, bringing in the red, and the factor causing a dark eye. 

 E acts upon I to produce the pink (red) eye. 



I seems to have affected the reds (R), but not the blues nor the red associated with 

 the blue making purple, represented by factor B. For example, the plants with 

 the phenotypic formula eeBBrrii (PI. C, fig. K) seem to be the same, the I having 

 nothing to act upon, E or R, to cause more red to appear. The same is true of EEbbrrli 

 and eebbrrii (PI. C, fig. H); I has no effect, and both appear as white. 



The Observed and Calculated Ratios 



The 543 plants that comprised the Fj generation were divided as 

 accurately as possible into groups, each group containing plants which 

 were similar in appearance. This was done long before any explana- 

 tion was found to account for their inheritance. Therefore, the writer 

 was not prejudiced in the slightest degree in making his selections. 

 This division of this p9pulation into groups was exceedingly difificult. 

 The boundaries of the groups were not clear-cut, and very many border 

 plants were found which were not easy to classify. This difficulty of 

 classification may account for some of the differences that are seen 

 between certain observed and calculated ratios. These are not so 

 serious, however, to the man who has worked with the plants and tried 

 to classify them. He is surprised not so much at the differences but 

 at the nearness to theoretical ratios in most cases. For example, the 

 writer foimd it very difficult to divide a certain number of bluish or 

 purplish plants into their proper groups. The extremes of these groups 

 could be readily recognized, but there were very many plants which 

 might, so far as observation was concerned, go as well into one group as 

 another. But the writer made the grouping as correctly as possible, 

 not knowing what the interpretation would be. This accounts for the 

 small number of plants in type B (PI. C) and the large number in type 

 A (PI. C). There were many plants which might be equally well classed 

 in either group. Other differences may be accounted for in this way. 



The writer is by no means certain that all of the differences between 

 observed and calculated ratios can be thus accounted for. There may be 

 present linkage or repulsion, but neither has been detected as yet. The 

 interpretation of the observations is considered merely tentative. Future 

 crossings may modify it in many ways. 



The dearth of whites mxay perhaps be accounted for. The plants 

 with white flowers in this series were noticeably the weakest, and inas- 

 much as many plants died during the prosecution of the experiment, it is 

 likely that a considerable proportion of these were whites. 



The weakness of white-flowered plants is a common observation among 

 gardeners. Hottes, for example, says that white varieties of the gladiolus 

 are so weak that it is almost impossible to propagate them. 



If the classes are grouped together, the individual differences tend to 

 disappear. For instance, if all of the plants with dark-eyed flowers (E) 



