88 CROSS-FERTILISATION IN THE INDIAN COTTONS. 



sufficiently large scale to admit of wide distribution. 'J'lie work 

 of improvonicnt, therefore, falls into two .sectioiis^tbe juoduction 

 of such an improved t}'pe, or t}'pes, in a stale of purity and the mul- 

 tiplication of these types when established, until a stage is reached 

 when distribution may be attempted. On the extent to which 

 natural cross-fertilisation is found to occur, will depend the pre- 

 cautionary measures which must be adopted to prevent a gradual 

 deterioration of the crop after distribution has been commenced. 



In the first stage the number of individuals handled will be 

 Imiited, and it is possible to adopt artificial methods, by protec- 

 tion of the individual flowers or even of the whole plant, to jnevent 

 access to the flowers of foreign pollen of unknown origin. In this 

 stage, if natural crossing is shown to occur, the determination of 

 the hmits to which continued self-fertilisation can be carried is of 

 an importance greater than the determination of the extent of 

 3uch crossing, for on this will depend the methods to be adopted 

 in the isolation of improved types. In the second stage the num- 

 bers become too great for such artificial protection. The extent 

 to which crossing takes place, then assumes a position of major 

 importance, and, inasmuch as crossing between individuals of 

 any particular type has no harmful influence, the determination 

 of the conditions which render such crossing impossible is espe- 

 cially desirable. 



That it is possible to determine with any mathematical pre- 

 cision the extent of cross-fertihsation taking place under certain 

 conditions is hardly to be expected. The labour involved in at- 

 tempting even a few such determinations would be considerable 

 nor would it be safe to generalise from these, even when obtained, 

 for in this particular it is quite possible that two similar types, or 

 even individuals, might differ widely. For practical purposes a 

 general idea of the extent of cross-fertilisation and the degree of 

 sterility induced by repeated self-fertiUsation is sufficient to indi- 

 cate the methods and precautions it is necessary to adopt. The 

 following observations on the Indian cottons are not, therefore, 

 the outcome of a carefully planned series of experiments but rather 



