80 STUDIES IN INDIAN TOBACCOS. 



slowly in size from the base and then diminish at the top of 

 the plant, the diminution increasing very rapidly at the end. 

 The leaves in the centre of the plant are approximately uniform 

 in most of the types. Above the significant leaves come a few 

 very small leaves which vary greatly in shape. In some of the 

 broad-leaved types, they are small replicas of the large leaves; in 

 types with lanceolate leaves they are linear ; while in many cases 

 they may be linear even if the significant leaves are not lanceolate. 

 In Tables XVII, XVIII, and XIX detailed measurements are 

 given of all the leaves on three plants of each of the parents. 

 Unfortunately, by the time the plant is mature, the lower leaves 

 have become damaged. The large number of fairly uniform 

 leaves in the centre of the plant is, however, evident. Tjpe 9 

 (Table XVII and Plate XVIII) is an example of an extreme case 

 in which no two leaves are alike. The best material for all 

 leaf measurements is undoubtedly to be found in the central 

 leaves, which are practically uniform, and all the measurements 

 in this paper have been taken on such typical leaves. It is 

 impossible to choose any definite leaf (for example, the tenth leaf) 

 suitable to all plants owing to the variations in the number of 

 leaves per plant, but it is quite easy with practice to pick out the 

 middle portion of the plant by eye and to select a typical leaf 

 in this position. The tables show that in this region of the 

 plant there are a large number of leaves, all of which are suitable. 

 Measurements carried out on the parents by taking a typical leaf 

 from a large number of plants have shown that the method is 

 most satisfactory. The usual practice of obtaining the average 

 between the bottom, the middle and the top leaf is open to many 

 objections and would be quite unsuitable for the kinds of tobacco 

 grown in India. It presupposes a uniform change in the size of 

 the leaves on the plant, and takes no account of the large number 

 of uniform leaves in the centre of the plant, which vary in their 

 effect on the average with the number of leaves per plant, and 

 the type of the plant. Moreover, the determination of the top leaf 

 is necessarily arbitrary. The real top leaves are far too small 



