a oe ee ~ 
Sd j st etait as 
THE IMPROVEMENT OF THE SUGAR-CANE BY HYBRIDISATION. 321 
that more than 95 per cent. of the hybrids resemble the male parent.” 
The hybrids in Barbados, as might be expected, also show that certain 
external characters resemble those of one of the parents. 
In the experimental work carried on at Barbados on these lines, only 
those varieties that have stood the stringent tests on a large scale fora 
considerable time under varying conditions of soils and climate were 
chosen, as many of the newer seedlings show fluctuating variations when 
submitted to adverse conditions. Care in securing good parent varieties 
is of the greatest importance, because the number of varieties which may 
be kept under trial is limited. After having chosen the variety, it is 
essential that only the choicest individuals are taken for experimental 
purposes, for in Java it has been found that the amount of sugar ina 
cane varies directly with the weight of the cane, and also as a rule heavy 
plants give rise to heavy offspring. 
Once haying obtained the desired type of seedling, it will be easy to 
multiply it to any extent without the necessity of fixing the type by 
further breeding, as the sugar-cane, on a large scale, is propagated by 
cuttings, and not by seed. 
CLASSIFICATION AND AN APPEAL FOR UNIFORMITY. 
Having reviewed the method of obtaining seedlings and some of the 
problems for future work, it becomes necessary to discuss some of the 
results already obtained. Perhaps the best way is to describe the advances 
made by the various sugar-producing countries separately. Before doing 
this, however, it will he advisable to notice in passing the methods of 
naming and classifying the different varieties of canes. 
Most cf the older writers classified canes according to the countries 
of their origin; in many cases their true origin was unknown and there- 
fore new names were provided. Subsequently local names were assigned 
to the same variety, and shortly a confusing number of synonyms was 
established. In 1890 Harrison and Jenman !° recorded that, in their collec- 
tion of the world’s canes on the Experiment Stations in British Guiana, 
the Bourbon cane (one of the oldest varieties) was represented under six 
distinct names, and the White Transparent under four. (See fig. p. 328.) 
They therefore suggested that a system of classification should be 
universally devised, and finally concluded that the best and easiest method 
was to arrange them in groups according to their outward characters. 
Five classes were formed :— 
(1) Yellowish-green or green, often blotched with red. 
(2) White, vinous or brown tinged canes. 
(3) Grey or pink tinged canes. 
(4) Ribbon or striped canes. 
(5) Claret or purple canes. 
Stubbs"! in Louisiana, however, only recognises three classes, as 
distinguished from the five of Harrison and Jenman, viz. :— 
(1) White, yellow, or green canes. 
(2) Striped canes. 
(3) Solid colours other than in (1). 
In comparing these two independent classifications and looking at the 
synonyms established, it is seen that differences occur, but they show 
x 
