PI,ANT BREEDING I267 



of the " Deli " and " Canarie " races. According to CoMES, the 3 races are 

 the result of crosses between the varieties havaniensis and macrophylla 

 {purpurea) (i). .The " Deli " race is finer and in a favourable climate yields 

 cigar wrappers which are much in demand. The " Kedoe " race is not so 

 fine, but is more aromatic and more vigorous, and does not require so much 

 attention as the " Deli " race. It also sometimes furnishes wrappers for 

 cigars of a less fine quality' . The "Kedoe" plant presents some points of 

 resemblance with one of the " Manilla " races, and that is alleged to be its 

 origin. The " Canarie " race will stand more drought and heavy soils. 

 In the East of Java it does not yield as good a tobacco as the " Deli " and 

 " Kedoe " races, while in the Sultanates of Djokjakarta and Soerakarta, 

 in the middle of Java, it is the only one cultivated and supplies a tobacco 

 of dull appearance for cigar wrappers. 



All the parent plants chosen in the autumn of 1912 in the different plan- 

 tations were analysed on the methods set out in Bulletin No. 9. The results 

 of this large task, however, are only pubHc in respect to the " Kedoe " race 

 which was the first to be studied. The analysis embraced the height of the 

 plant, the number, shape, uniformit}-, venation and nuttual position of the 

 leaves. The following points of the first 15 leaves, green and dry, were 

 studied for each plant : length, width, area, ratio between length and width 

 and fineness, i. e. ratio between area and weight of leaves. Finalty, for the 

 dried leaves, the weight of the midrib was indicated, the ratio between the 

 weight of the leaf and the rib, then the colour, determined by means of a 

 scale of colours, and finally the burning qualities and the colour of the ash. 

 In this way average values were obtained and also coefficients as to correla- 

 tion which may serve as a standard of comparison for all subsequent studies 

 in relation to the " Kedoe " race. 



There was found : an inverse correlation between the number of the 

 distance apart of the leaves ; a direct correlation between the number of 

 the leaves and the ratio of their length and width ; a third, direct, between the 

 weight of the leaf and the weight of the midrib ; a fourth, inverse, between the 

 weight of the leaf and the ratio between length and width ; a fifth, direct, 

 between length and width of the leaves and so on. 



The methods hitherto adopted in laboratories in studying the combust- 

 ion of tobacco have no practical bearing, in the writer's opinion, but may 

 on the other hand serve for comparison of the dift'erent races and lines. The 

 fineness of the tobacco, as well as its rapid and continuous combustion, de- 

 pend mainly (questions of race apart) on the ripeness of the leaves, and in 

 the second place on the drying and fermentation of the tobacco. I,eaves 

 which are too old and those which are too young exhibit the same faults : 

 want of fineness and defective combustion. The reasons for these defects, 

 however, are not the same in the two categories. In over-ripe leaves there 

 is an excess of cellulose and woodv substances ; on the other hand, the over- 



(i) The writer adopts iii its entirety the system of tobacco classification of ItaUau writers. 

 Thus the species Nicotiaiui Tabacum is subdivided into varieties, the varieties into races, the 

 races into genotypes or pure lines and the lines into individuals. 



