88 TILLERING IN INDIAN SUQAROANES 



crop. A note will then be added on the richness of the juice in branches of 

 different orders in the clump. 



(a) Effect of spacing on tillering as judged by the number of shoots pro- 

 duced p>er clump. Stubbs, in 1892-93, showed that, by planting the sets at 6", 

 12" and 18" apart, the number of shoots produced differed a good deal. At 

 three months after planting the 6" plants had, on the average, 4-1 shoots each, 

 those at 12" had 5-8 shoots, while those at 1 8" had 6-4 shoots per plant. Observa- 

 tions have not as yet been made on this point at the Cane-breeding Station. 

 The following figures have been deduced from those published by Muller von 

 Czernicki and referred to above. We have obtained them by dividing the 

 maxinuim number of shoots in his countings by the number of sets in the row. 

 The cases selected are the extremes and an intermediate one, namely, where 

 sets were planted 10, 14 and 20 in the row. The following are the maximum 

 numbers of shoots for these spacings : Cheribon (tops), 12-4, 11 "1 and 9-0 ; J. 100 

 (tops), 12-5, 10-6, 8-4 ; J. 247 (tops), 16-8, 14-5 and 120 ; J. 247 (sets), 15-9, 12-5, 

 and 106. The extreme differences in these spacings are roughly as 3 to 2 

 shoots per plant for the wider plantings. 



ib) Effect of spacing on the total nu7nber of canes per clumji at harvest. 

 We are able to get more cases in which this has been observed, in that countings 

 of canes at harvest appear to have been made regularly for many years in Java. 

 Stubbs gives the figures for the canes at crop time (seven months from planting), 

 in Louisiana, and from these we find that the number of canes per clump 

 at 6" is 2-2, at 12", 3-7 and at 18", 4*9. Comparing these figm-es with those in 

 section {a) we see that, although a number of shoots had died, the ultimate 

 differences had increased. 



Kilian gives the number of canes at harvest per bouw (TTS acres) when 

 the rows were 3|', 4' and 5' apart, and we can obtain proportional figures for 

 the number of canes per clump by multiplying these two sets of figures together. 

 It is to be noted that the differences in spacing were not nearly so great as in 

 Stubbs' experiments, but the results are still very definite. Taking the table 

 wiven on p. 82, we get the proportional numbers as 4-6, D'O and 5-9 canes per 

 clump in the richer land and 3*9, 4*3 and 5'0 in the poorer. 



Muller von Czernicki does not give the numbers of canes at harvest, but 

 counts them at 5-6 months, using the Java convention of taking two or three 

 thin canes to one thick. Selecting the rows as before with 10, 14 and 20 sets, 

 ■ we get the following figures : — 



Cheribon (tops), 6-1, 4-7 and 3-6 ; J. 100 (tops), 8-2, 5-9 and 4-3 ; J. 247 

 (tops), lO'l, 6*8 and 4*8 ; J. 247 (sets), 95, 6-6 and 5-0. Here again there is an 



