C. Pelt.kw and F. ^1. Durham 1(5'; 



Evidence ov segregation in Kewensis. 



The types used, howevei' fertilized, in general give nothing but 

 plants resembling the seed parent. P. florihunda, floribunda isabellina, 

 and verticillata, except in three isolated cases enumerated below, which 

 we regard as errors, have given no offspring not identical with them- 

 selves. P. Kewensis moreover, though a hybrid, commonly throws 

 nothing like the species from which it is derived. Among Kewensis 

 families there may be some variation, notably in the size of the plants 

 or in the amount of meal present, and in respect of this last character 

 the evidence points to the occurrence of segregation. One exam})l(' of 

 variation in colour occurred, and a double form has also arisen. 



We have further to record a group of seven cases in which solitary 

 individuals of unexpected nature came from Kewensis seed. These may 

 not impossibly be due to errors, but for certain reasons we are disposed 

 to regard them as instances of segregation occurring sporadically. 



But the most remarkable departure from the usual genetic behaviour 

 of these plants was the appearance, on at least one occasion', of a diploid 

 individual as the offspring of the tetraploid type, and on breeding from 

 this individual we raised families in which segregation is habitual and 

 probably runs a normal course. 



These several statements may now be amplified. 



A. Aberrant cases probably due to error. 



(1) A family of verticillata plants was sent us from Kew, arising 

 from vert, xjivr. One of these plants on self-fertilization gave a 

 large number of verticillata plants and one plant oi florihunda. 



(2) A plant of florihunda var. isabellina (extracted from a cross 

 with a diploid plant which came from Kewensis, no. ||) on self- 

 fertilization gave 28 plants, one of which had yellow flowers, the rest 

 •with the pale flowers of isabellina. The plant with yellow flowers 

 differed from the others in several respects. This is the only case we 

 have met with of the isabellina type not breeding true. 



(3) A plant of floribunda, long styled or pin (the recessive form), 

 extracted from a cross with a di])loid form which came from Kewensis, 

 no. if, on selfing gave three thrums and four pins. This is the only 

 case we have had of a jjin plant not breeding true. 



1 Only in one of these cases liave the chromosomes been counted. In one other case, 

 ihe florihumla-Win habit of the plants in question suggest that they are diploid. 



