72 PLANT-BREEDING 



but with another significance to the word than that used by 

 Hallett, as we shall soon see. It opened the prospect of a 

 new manner of operating and that at a time when the results 

 of the previous methods had become such as to make all 

 further trials almost hopeless. 



As yet it was, however, only a presumption, resting on 

 the small evidence quoted. It was more an indication of 

 what could be expected than a proof of what really was. 

 It had first to be tested. This was done at once, and on as 

 large a scale as possible. It was in the summer of 1892 

 that the described rare uniform plots were seen, and in the 

 harvest of the same year a renewed search for starting-points 

 for new races was made. But this time each ear was kept 

 separate, and two or more heads were combined only when 

 they were gathered on the same individual plant. Some 

 ears were chosen as the best representatives of their varie- 

 ties, others as deviating from the type in one respect or 

 another. All in all, about two thousand ears and panicles 

 of different species and varieties, representing as many 

 divergent types as possible were selected. The grains of 

 each were sown on a separate plot, and next year all the 

 groups were descended from one single mother plant each. 



The results of tliis trial greatly exceeded all previous 

 expectations. Almost all the numbers were seen to be uni- 

 form, all the offspring of a single plant being wholly similar. 

 Exceptions there were, but they were exceedingly rare. For 

 instance, among the 422 cultures of oats, 397 were uniform 

 and only 25 multiple. But of course it could be expected 

 that among so large a number of cars, some hybrids would 

 be met with, and others which would be only partially self- 

 fertilized, but for the remaining part contaminated by the 

 pollen of their neighbors. In either case, the progeny 

 would be dissimilar, and especially in the former the splitting 

 up of the hybrids would give rise to quite a considerable 



