C, A. BARBER gg 



but, in the rest, there is a marked tendency for the thinner forms to have 

 greater branching. For safe deductions on this point, a much larger number 

 of clumps must be available for examination. 



(3) Deaths during growth and the period of xAiaximum tillering. 

 In considering these and other tables on the tillering power of different 

 varieties of canes, founded on the number of canes produced at harvest, it is 

 necessary again to sound a note of warning at the somewhat loose use of the 

 term in general practice. The total number of canes at crop time is not in 

 reality a safe guide to the shooting power, or tillering capacity in its narrower 

 sense, because a large number of shoots die during the life of the plant. This 

 is a necessary result of cultivation, where a tufted grass is forced to grow within 

 narrow limits, so as to obtain as many matured stalks as possible. There is not 

 room for the development of a number of the shoots formed and hence the 

 mortality among them is very considerable. Stubbs,i in his careful experi- 

 ments on the Purple and Striped Louisiana canes in 1894-95, calculated that the 

 deaths of shoots during gro^\-th were 58-9 per cent, in 1894 and 53-9 per cent, in 

 1895. MuUer von Czernicki, in Java, counted the number of shoots appearin^^ 

 above ground at varying periods between 60 and 150 days, and showed con- 

 clusively that the numbers were far greater at the earlier than the later period. 

 Thus, in Cheribon 120-180 shoots were counted in different plots at 60 days from 

 planting and only 60-70 at 150 days ; the figures for J. 247 were 160-240 and 

 90-100, and, for J. 100, 100-170 and 82-86 respectively. Striiben, in a series of 

 experiments on J. 247, found that the better grown plots in the first two or 

 three months gave 300-400 shoots per row, in one case the number reaching 

 415, whereas at eight months all of the rows gave only about 110 shoots. No 

 data are as yet available as to whether Indian canes suffer this great mortality 

 during the earlier period of growth, but there is some reason to suspect, from 

 shoot-counting observations, that it is a much less serious factor than in the 

 thick canes, and further countings have been commenced to settle the 

 question. 



Another point to be held in view is the relative rate of germination and 

 tillering in different varieties. This of course does not refer to the effect of 

 cold and drought, as for instance in North India where the early growth is so 

 much slower than in the Indian Peninsula, but only includes comparisons 

 where the conditions are altogether as similar as it is possible to make them. 



^ References to these papers will be given later. 



