44 CHAPTER IV 



improving the cane is not well founded. The original experiments in Java 

 seem to have ignored the root branching system of the cane, later studied by 

 Barber. Other characteristics that have been examined in the hope of 

 finding a correlation with desirable features are : — Tillering, long joints 

 versus shoit joints, thick stalks versus thin stalks, conical joints versus 

 cylindrical joints, flowering and not flowering. No amelioration of the cane 

 and the estabhshment of no permanent strain has resulted from these ex- 

 periments and, generally, it was found that the characteristics themselves 

 were not inherited ; thus the asexual descendants of canes that had flowered 

 showed no greater tendency to flower than did the progeny of canes that had 

 not flowered. 



The Classification and Identification of Canes. — This section is to be 

 read as dealing onh' with the thick tropical canes which form almost the 

 entire mass of the cultivation outside of India, where canes of a different 

 type produce upwards of 2,000,000 tons of sugar annually. These last 

 varieties are discussed elsewhere. 



The older writers, who frequently were not systematic botanists, generally 

 adopted coloiu: of the stalk as the criterion of classification. This sx^stem 

 's used by Vieillard^^, Tawcett^", Harrison and Jenman^^, SoltwedeP^^ 

 Kriiger^s, Stubbs^^ Dahl and Arendrup.^^ Bouton^^, however, based his 

 classification of Mauritius canes on length of internode, and Debassyns^'^ in 

 Reunion as long ago as 1848 divided canes into such as flowered and did not 

 flower. The divisions adopted by Harrison and Jenman are (i) yellowish 

 green and green canes often blotched with red ; (2) white, vinous and brown 

 canes ; (3) grey or pink-tinged canes ; (4) ribbon canes ; (5) claret and 

 purple canes. Stubbs and also Kriiger only employ three classes : — (i) 

 yellow and green canes ; (2) ribbon canes ; (3) solid colours other than yellow 

 and green. While the colour of the stalk is at once seen to be a ready aid 

 to identification, it manifestly breaks down as a criterion of classification as 

 it would necessarily separate those closely allied sports where colour alone 

 forms the distinguishing feature. 



In addition to colour Kriiger^^ uses as means of identification the following 

 characteristics : — P'-esence or absence of wax, shape and arrangement of 

 internodes, shape of eye, presence or absence of channel above eye, rows of 

 roots, colour of pith, colour of leaf sheath, pilosity of sheath, colour of leaf 

 blade, shape of lobes at junction of sheath and blade, general characters. 



CowgilP^ uses all these characters and places most rehance on variations 

 in parts of the stalk. Sahasrabuddhe^^ suggests the use of the eye of the cane 

 as a means of identification, and distinguishes five types : — (i) White Trans 

 parent type — flat, broad, pointed, with point extending beyond ring with a 

 distinct channel ; (2) Bourbon type— flat triangular buds, more or less 

 pointed with an indistinct channel ; (3) White Tanna type — as in (2) but more 

 or less circular ; (4) MehgeU type — long, narrow pointed, extending weP 

 beyond the ring with distinct narrow channel ; (5) Mammary type — circular 

 buds with no channel. 



Very recently Barber^^ in India, Jeswiet*" in Java, and Fawcett*^ in 

 Argentina have niade detailed morphological studies of canes, including such 

 points as the venation of the leaf and the arrangement of the groups of hairs 

 on the eye and leaf sheaf. The groupings recognized by Jeswiet are indicated 

 in Figs. 14, 15, 16, 17 and 18 ; 14 is the upper side of an eye with apical 

 growing point ; 15 is the corresponding underside ; 16 is the upper side of a 



