LEAKE AND RAM PRASAD. 53 



bearing the red colour in the veins and pink flowers and from the 

 single plant with a white flower. All the three former have given 

 the two forms with pink flowers and also white-flowered plants, 

 while the whito-flowered plant has given whites only. The similar- 

 ity between the behaviour of this series and that of the cross between 

 Type 3 and Type 9 (red-flowered G. arboreum and a white-flowered G. 

 negledmn) is so striking that it is impossible to avoid the conclusion 

 that the original plant was derived from a cross between two plants 

 having similar characters. It is needless to follow this series through 

 the details of the leaf factor. That of the original plant was inter- 

 mediate and the subsequent generations have split up in a manner 

 similar in every respect to the case noted above. From its occurrence 

 in a packet of seed of G. arboreum, it is probable that this plant was a 

 first cross and it may safely be concluded that one parent was G. 

 arboreum while the other must have been a white-flowered plant with 



leaf factor/^ 2 — probably a plant of Type 6. 



It is unnecessary to enter in detail into each of the numerous 

 cases of impurity which have been observed to arise in like manner 

 in cultures raised from the seed of naturally fertilised fruits. In 

 the few cases in which they have been investigated they have been 

 proved by the character of their progeny to be the result of cross- 

 fertilisation. For the present purpose it will be sufficient to describe 

 the three series of observations which most clearly illustrate the point. 



(1) Type 2 {G. herbaceum, Linn.) has been grown for a series of 

 years ; during the first two seasons only seed from protected flowers 

 was sown and the culture so produced proved to be pure. In 1907 

 this seed was limited in amount, and a second culture was, therefore, 

 raised in 1908 from seed of the same parents, but gathered from 

 unprotected flowers. Of these two cixltures the former consisted 

 of twenty-one plants all of which were true to type, while of the lat- 

 ter eleven plants were pure and three undoubtedly crosses. In 

 two of these three there was a marked development of sympodial 

 secondary branches and a consequently shortened vegetative period, 

 while in the third case the leaf factor was also intermedia te(2' 5) 



