R. RugCtLES Gates 357 



closer approach to equality than Mendelians frequently regard as 

 sufficient to prove their point. In the reciprocal cross, yrandiflora 

 X ricbr{calyx\ the F^ contained 147 plants, of which 58 bloomed. The 

 latter were in the ratio 34 red-budded to 24^ green-budded. Anyone 

 can decide for himself whether this ratio is nearer 1 : 1, as it should 

 be on my view, or 3 : 1 as it should be on Shull's view. The full data 

 from this femily, however, prove the situation still more conclusively, 

 for all the plants were carefully examined in the rosette stage and 42 

 were found to show clearly red on their ventral surfaces, 71 were 

 without any trace of red, and 37 were doubtful, i.e. showing the red 

 faintly. We now know that the latter must be classed with the first, 

 for all alike develop red buds, and plants with very little red on the 

 rosette have the usual deep red buds. In this family there were there- 

 fore 79 with the ruhricalyx character {R) to 71 withotit it ()•), a close 

 approach to equality. • 



These ratios are exceptionally close to expectation and, taken 

 together, they furnish abundant evidence as will, I think, be agreed 

 by all Mendelians, to establish the very high probability that a single 

 unit-difference was concei'ned in the origin of ruhricalyx from rubri- 

 nervis. On the other hand, the evidence is sufficient to establish 

 absolutely the impossibility that two Mendelian fixctors could have 



From this Table it will be seen that there is not an atom of 

 evidence for the presence of duplicate factors in Oe. ruhricalyx, while 

 the evidence for a single factor is of the kind which is universally 

 regarded by Mendelians as conclusive. 



In addition to the records presented above, several other ratios 

 were given in my paper (1914, Pedigi'ee 1). These are of families 



' A full account of these crosses is given in my former paper (1914). 



^ In a previous paper (Gates, 1914, Pedigree 1) this was given as 23. I have not the 

 original records at hand for determining the cause of this discrepancy, but the difference 

 is immaterial to the argument anyway. 



