October, 1912.] 



293 



planting each batch of seedlings into a separate plot to gather the crop 

 and measure or weigh the grain obtained from 100 plants of each lot. Only 

 the largest crop is to be retained for further sowing. In this way the 

 parental plants are tested by their performance— by the power of their 

 grain when used as seed to yield a good harvest. This is the essential 

 quality which we desire to get in seed, and we can only get it if we select 

 for it. We cannot ensure the presence of this quality in the seed by 

 selecting the heaviest individual grains or even by selecting the heaviest 

 yielding plants, because a heavy yield from a particular plant may be 

 due to the accident of soil or situation. What we want to do is to pick 

 out those plants from the seed of which we shall get the best average 

 crop under ordinary conditions. We can only do this by sowing the 

 seed and finding out what crop is actually given. Although an ordinary 

 paddy field looks so uniform there is great variation among the plants 

 in cropping power. In quite a small experiment the crop obtained from 

 the seed of the best plants was 50 per cent, better than the average crop. 

 Finally the improvement thus obtained is comparatively permanent, 

 whilst the whole of the improvement possible is got at a single step if 

 the selection is carried out on a sufficiently large scale. The seed should 

 continue to give good crops for several seasons without any further selec- 

 tion. 



Transplanting in Relation to Seed Selection. 



The method here described is of special importance in connection with 

 transplanting in paddy cultivation, in fact the full benefit of it will only be 

 obtained if transplanting is adopted. By using seed which transmits the 

 tendency to tiller well, good crops can be got when transplanting is carried 

 out at distances of 12 by 12 inches or even more, whereas if the plants tiller 

 badly they must be transplanted much closer in order to furnish a satis- 

 factory yield. Where transplanting is adopted it will be found that good 

 tillering and a good yield are qualities which are closely associated in the 

 same plants, so that the selection of many-tillered plants will increase the 

 yield and vice versa, 



REMARKS BY THE DIRECTOR OF AGRICULTURE. 

 Mr. R. N. Lyne : Your Excellency, I think we owe a debt of gratitude 

 to Mr. de Silva for the interesting and instructive paper he has read on 

 this important question of rice and paddy cultivation in Ceylon, and to 

 Dr. Lock for his illuminating observations. There are one or two points 

 I would, however, wish to refer to. Mr. de Silva referred to the fact that 

 an iucretlsed yield had been obtained through selecting seed by a process 

 of floating off the light seeds, and Dr. Lock states that he has no experi- 

 ence of this method and he does not think it will lead to any permanent 

 importance. What I think Mr. de Silva must have meant was that this 

 was merely a mechanical selecting of seed for good germinating power, 

 and sowing only those seeds which have their vitality maintained. It is 

 obvious that if you take a bushel of seed and sow it, if only 50 per cent, 

 have any germinating power you will get a very much poorer crop than 

 if the whole of your seeds had germinating power, or say 80 or 90 per cent, 

 which is a reasonable percentage to expect from properly selected seed. 

 At the same time no doubt it is a fact, that strains of wheat 

 have been improved by such a system of selection only and not by 

 selecting from large and better producing plants as Dr. Lock would 

 first do and which, of course, is a much more direct and easier way. Then 

 Mr, de SUva refers to the advantage which we may expect from a 



