240 Gates. 



the parent plants were entered in making up the final records. It 

 is that cultures number 55, 63 and 64 were each derived from a 

 different, self-pollinated flower of the same parent plant (No. VII. 2). 

 This shows (i) that the ratio 5 : i is significant and represents what 

 we may perhaps call the degree of prepotency of this rubricalyx plant; 

 and (2) that the ratios obtained are not modifications of a 3 : i ratio 

 brought about by unconscious selection or in any other way, but 

 represent the ratio of types which this plant is capable of producing 

 as a result of its germinal constitution. The three flowers from which 

 these cultures were produced, were self-pollinated respectively on 

 July 17, 18 and 17, 1911, the seeds from each capsule being sown sepa- 

 rately to produce the above results. 



We are thus forced to conclude that different plants of the F^, though 

 ■externally alike, yet in F^ produce individuals having the dominant and 

 recessive characters in widely varying ratios, though this ratio is constant 

 for each parent individual. 



This ratio may be approximately 3 : i. 5 : i, or 33 : i. It is not 

 ■certain that the exact numerical ratios are significant as such, but 

 rather, they may represent merely the degree of prepotency of the indi- 

 vidual. This will be referred to again later. 



In II. D Table III are shown the results of crosses between red- 

 and green-budded plants of the Fj. A 3 : i or i : i ratio would be 

 anticipated, according to whether the red-budded plant was homo- 

 zygous or heterozygous. But culture 56 yields about 2 : i, and only 

 culture 58 the expected i : i. In section III of the table, F, red- 

 "budded plants of rubricalyx x grandiflora have been crossed back 

 with pure grandiflora from Alabama. The resulting R : r ratio is 

 approximately the same in all three families, but their average, i. 33 : i, 

 is significantly greater than 1:1. 



In section IV an Fj red-budded plant from rubricalyx x grandi- 

 flora is crossed back with rubricalyx. The buds in this culture are 

 all red without exception, and darker red than those of other cultures. 

 The significance of the latter fact wiU be discussed later. Sections V 

 and VII give results of ^ the two double-reciprocal crosses between 

 ■grandiflora and rubricalyx. In culture 72 the result is very different 

 irom the expected 3:1, while in 75 and 76 the results agree approxi- 

 mately with each other, and the totals are consonant with a 3 : i ratio. 

 In section VI, in which grandiflora x rubricalyx red-budded Fj is 

 crossed back with grandiflora, the ratios for the three cultures (two 

 of which are from the same plant. No. VI. 6) agree fairly well with 



