C. A. BARBER l09 



each shoot separately, we see that there are two very distinct periods of 

 growth in it, the first answering closely to the tillering period, when it is increas- 

 ing in thickness and length of joint and is busy in forming its branches, and 

 the second, when, after attaining its full thickness, it commences to form joints 

 of appreciable length, and rapidly shoots into the air, a stage comparable 

 with the elongating period in gra&ses. Thus the periods which characterize 

 the shoots of the grass plant at one and the same time are present in the cane 

 plant also, but in each shoot independently of the others. In spite of this 

 fundamental difference, it is possible to separate cane varieties as to their 

 general periods of gro\\'th. The term " cane formation " is well known in 

 the fields, and is used to indicate the first appearance of solid canes between 

 the bases of sheathing leaves just above the ground, And. this cane formation 

 differs a good deal in different varieties. It is, for instance, much more rapid 

 in the Saretha than in the Sunnabile group {see Mem. Ill, p. 159 and PI. IV), 

 and still more so than in the later Mungo group. The length of the tillering 

 period thus finds its expression in the rate of maturing of the canes in any 

 clump, and this has been carefully studied in all the varieties ; and a special 

 series of dissections has been made for the purpose, at three to four months 

 after planting. 



(3) Diagrams and formula of the branching system. 



In a previous paper a few pages were devoted to the branching system of 

 the Saretha and Sunnabile groups of canes, and to these were added a diagram 

 of the branching of a Pansahi plant ^nth several photographs of dissected 

 plants, showing the differences in length and thickness of joints in branches 

 of successive orders (Mem. Ill, pp. 156-160). The conclusions arrived 

 at were stated to be preliminary, as a much larger series of dissections was 

 projected during the approaching season ; but, from the work already done, 

 it was suggested that, by studying the branching typical of any group of cane 

 varieties, some idea might be obtained as to its relations with the more primi- 

 tive forms on the one hand and the tropical canes on the other, and its place 

 in the ascending series of evolution approximately gauged. Tentative formulae 

 were suggested to express the general course of branching in particular cases. 

 This work has now considerably progressed, and the larger series of dissections 

 has been completed, presenting us ^Anth a mass of interesting material for 

 study. 



The grouping of the cane varieties is that adopted in a short paper in the 

 Agricultural Journal of India (Vol. XI, Part IV, Oct., 1916). Six varieties 

 of each of the main groups, Saretha, Sunnabile, Pansahi, Nargori and Mungo, 



