106 TILLERING IN INDIAN SUGARCANES 



forest. In the latter case, we have frequently been able to detect, on looking 

 upwards, a hexagonal outline for the whole leafy top of an individual tree. 



We can now return to the consideration of the way in which the rate and 

 manner of development of the shoots, of different orders of branching, is 

 impressed permanently on the morphological character of the mature canes. 



We have measured the basal, short-jointed portion in each cane of each 

 dissection, and find it longer, w^th a greater number of closely packed joints 

 in the as than in the &s and cs. In the later formed canes, however, we find 

 the matter complicated by the incidence of curvature, and the length of this 

 basal portion again tends to increase in the branches of higher orders, often 

 indeed ultimately exceeding that in the main shoot. But there is no danger 

 of confusing the different classes of branches on this account. Besides the 

 actual curvature itself, which is absent in the as, there is usually a great thicken- 

 ino- in the curved portion in late branches, followed by a rapid thinning when 

 the curved portion has passed ; then the change in length is sudden, and quite 

 lono- joints are immediately reached, as contrasted with the extremely leisurely 

 increase in length in the earlier ones. 



In measuring the length and thickness of branches of different orders 

 we have confined our attention to the first two feet cf cane. There are several 

 reasons for this. In the first place the dissections would have been practically 

 impossible, if a great mass of leafy canes was attached to the base during its 

 manipulation, so it has been the custom, in the older canes, to cut the clump 

 at three to fovir feet from the ground before bringing it to the laboratory. 

 Then ao'ain, we have learnt, from our series of measurements of the length of 

 joints, that, in the plotted curves, the joints reach their maximum very soon 

 after the cane has emerged from the ground {cf. Chart II, p. 175, Mem. III). 

 The longest joints are almost always met with in the first two or three feet, 

 and afterwards there is a regular decrease until the end of the cane. Lastly, 

 we have come to the conclusion that, at two feet from the ground, the thickness 

 of the cane has reached a very fair average, although it is sometimes compli- 

 cated by a varietal thickening or narrowing after this region has been passed. 



In measuring the length of joints in the first two feet, we have omitted the 

 unformed, basal portion, and only started measuring when the first inch -long 

 joint has been encountered. This has also meant the omission of runners and, 

 usually, the curved portion in later formed shoots. In other words, we have 

 taken these measurements only in the fully formed cane. The results have shown 

 a very marked increase in length of joint in the successive branchings, the 

 joints in h being longer than those in a, and in c longer than b, and so forth. 



