252 



Gates. 



dition is all the more evident because in another series of crosses, not 

 here described, between Lamarckiana and a different race of grandi- 

 flora, there is in many cases rather sharp alternation, even in F^, 

 between the grandiflora and the Lamarckiana types of bud. 



Referring again to table III (p. 235), in the F2 families from. 

 grandiflora x ruhricalyx there was no segregation as regards either 

 shape or pubescence, but a completely blended condition. 



The same is true of the reciprocal cross, the fact being specially 

 noted in cultures 55 and 61 that rw buds, i. c, r buds which are as 

 squarish and pubescent as rubrinervis, did not occur. The exact bud 

 characters of many of the plants were catalogued, so far as this 

 could be done by using simple descriptive formulae. In these F^ 

 cultures very few if any buds were completely like either parent, and 

 the terms ru and gr of table III mean only "resembling rubrinervis 

 or grandiflora'" . The ratios, therefore, in section II B and II C of the 

 table have no definite significance and the same is true of the ratio 

 rii : gr in section II D. 



In back-crossing with grandiflora a new and interesting phenomenon 

 appears, namely spotting of the sepals with red. To attempt to account 

 for this by a latent "spotting factor" would, I believe, be wholly un- 

 justified. It has been found that in various organisms, when the first 

 cross between different coloured races {e. g. whites and negroes) is 

 crossed back with either pure race, piebaldism, or a tendency for the 

 pigment to be unequally distributed in patches, not infrequently 

 makes its appearance. I think undoubtedly the spotting of the sepals 

 in (ruhricalyx x grandiflora) x grandiflora, which occurred in all the 

 r buds of cult. 67, is a comparable phenomenon. 



The spotting of the sepals is of course not sharply separated from 

 a condition in which a small amount of pigment is found along the 

 median ridges of the sepals. The spotted fgs) buds do not occur in 

 all families of f ruhricalyx x grandiflora ) :< grandiflora. They were 

 specially noted as absent from Cult. 68, but this point was not observed 

 in Cult. 71 which bloomed very late. In Cult. 69 the buds could not 

 be classified, but they varied from gr i [i. e., grandiflora bud with a 



colour pattern of sepals 3 to 5, long hairs on the sepals — (/. e. about one-quarter 



4 

 the number in rubrinervis) while two of the side branches bore buds which were 

 slender, rounded, and with marginal red streaks on the sepals (1. e., like grandiflora^ 



but with - long hairs. 



2 



