C. A. BARBER 41 



and thus establish a connected series from the wild grasses to the thick canes 

 of the tropics. 



Lastly, differences were observed in the richness of the juice in the early 

 and late canes of a plant, and these did not altogether tally with the views 

 held regarding the richness of the thick and thin canes in tropical cane fields. 

 The literature of the subject is punctuated by references to the relative richness 

 in the juice of the " mother " cane and its branches, but, as no dissections 

 seem to have been made, it is difficult to understand how the various observers 

 distinguished these two classes of canes. There is obviously great confusion 

 on the subject, for one observer, after stating his opinion, admitted that the 

 mother shoot need not of necessity be the original main shoot of the clump 

 but was the " thickest and best grown " ! As will be seen, the result of our 

 study is exactly the reverse, in that the main shoot is thinner and less well 

 grown than its branches. This is indeed perfectly natural, when we 

 consider the available equipment of leaves and roots in the young cane, as 

 compared with that at the disposal of branches formed when the plant has 

 grown up. 



The dissection of the cane stool is a rather intricate and laborious piece 

 of work. Upon taking it out of the ground, each clump is seen to be covered 

 by a dense mass of tough roots, among which the soft buds are hidden, and these 

 roots have all to be carefully cut away before the nature of the branching can 

 be seen. The planting material with us consists of sets or pieces of cane 

 on which there are at least three healthy buds, and these buds, usually all of 

 them, develop into larger or smaller plants, which, however, are quite separate 

 and only influence one another as regards the space available for their indepen- 

 dent growth. It is usually impossible to make the dissection unless these 

 plants are cut out and dealt with separately. In the list of dissections given 

 on pages 99 and 100, the number of clumps and plants are therefore 

 enumerated for each variety. 



The main dissecting work was done in the 1916-17 and 1917-18 seasons, 

 and in each year clumps were examined at two stages of growth for entirely 

 different reasons. The first was at about four months, for the determination of 

 the rate of cane foimation ; and the second at about eight months, for the study 

 of the canes formed at crop time. It was soon found that, after this latter 

 date, no new canes could be started in time to mature. In 1910-17, the dissec- 

 tions were largely concentrated on the Saretha and Sunnabile groups, which 

 at that time had recently been separated and were being described. Six 

 varieties of each of these groups were examined, and to them were added a few 



