sxx. Timehri. 



exist ; and with these it is possible that by the application of Mendelian 

 principles in raising seedlings, new varieties of high value may be 

 obtained. Up to the present however, this has not taken place. 



At the inception of the sugar-cane breeding work here Jeninan was 

 enthusiastic over the possibilities of raising new varieries of high promise 

 by controlled methods of cross-fertilisation ; but in 1892-93 our hopes in 

 that direction received a severe shock. Using a variety of sugar-cane, 

 the Kara-kara-wa cane, which our experience in three preceding years 

 had shown to produce seedling-canes having usually somewhat close 

 resemblance to the parent variety, and placingit under conditions by which 

 it was impossible for its arrow or flowering shoot to be either cross- 

 fertilised by any other variety or fertilised by any other flower shoot of 

 its own kind, we got seedling canes from the one arrow of 267 different 

 sorts. The parent cane in its own seedling stage was hence possibly 

 derived from 14 diverse ancestral strains. 



Supposing for example that we take two kinds of sugar-cane, one X 

 having as ancestral kinds the varieties A, B, C, D, E and F and the 

 other Y derived from its ancestors A, B, G, H, I, and J, it is evident 

 that 406 different combinations can arise from the interbreeding of the 

 two kinds, instead of a single blend or cross X x Y. 



By Mendelian segregation the inheritable properties of this diverse 

 progeny will fall into three groups. We do not know which properties 

 are inherited ; but assuming that the general characteristics as a whole 

 are heritable the segregation of the seedlings from the cross X and Y 

 may give rise in the tirst generation to 1,218 groups of varieties. 



Now either X or Y by interbreeding with its own kind could pro- 

 duce only 15 x 3 groups or 45 general strains of sugar-canes. The com- 

 plexity introduced by the cross-fertilisation of existent complex hybrids 

 is well illustrated by this example. 



Up to 1902 we had not made any systematic attempt at raising canes 

 of controlled parentage. We now do it as a matter of regular routine — 

 not with any idea of getting seedlings having definite and desired charac- 

 teristics but as a means of greatly widening the range of their variation. 

 We have complete proof of the success of the method in this line. Un- 

 fortunately there is no chance in British Guiana of controlled cross- 

 fertilisation of the sugar-cane proving a short and certain way to success 

 in the production of new varieties of high saccharine value. 



Probably a more disappointing investigation has never been pursued 

 than has been the search for improved varieties of sugar-cane. There are 

 now many stations at work at it in the tropics and sub-tropics ; their re- 

 sults appear to be very similar ; — in the earlier years working with natural 

 varieties of sugar-cane several kinds of high promise are almost invaria- 

 bly obtained ; in later years, when the mass of material for parental pur- 

 poses has rapidly and enormously increased, the production of really good 

 yarieties appears to become increasingly difficult and results satisfactory 



