S May, 1908.] Imfrovemeut of Cereals by Selection and Crossing. 



plot in the field, so as to test its quality and purity. The samples chosen 

 from each variety were supposed to be uniform, but it was soon found 

 froaii practical exjDerience obtained from nearly 1,000 different lots, that 

 there were more or less different types among them. At this stage Dr. 

 Nillson was appomted Director in 1890 and he brought the necessary 

 .scientific knowl|edge to bear upon the practical problems. He saw that 

 the principle on which the selection was made must be somewhere at 

 fault, if the progeny did not agree with the mother-plants from Avhich 

 • thev were derived. An accidental observation just gave him the clue 

 which was needed. Among the thousands of cultures in the experimental 

 fields, there were a few which bore only one type and their uniform 

 .appearance readily marked them out from the surrounding plots. Among 

 the numerous samples selected, there were some common groups represented 

 -bv a number of ears, and those which w^ere rare had only been met with 

 iii a single head. From the perfect system of keeping records, showing 

 the number of ears used in each experiment, it was discovered that just 

 those cultures derived from a single ear were ijure and uniform and the 

 -conclusion was at once evident that cultures in order to be pure must be 

 started from a single ear. This discovery of ISillson threw a new light 

 on the principle of selection but still it w\as only based on a few examples 

 and required further prr.nf. In selecting samples for next season each 

 eai was kept separate and (inl\ two or more were combined when gathered 

 from the same individual plant. In this way over two thousand ears, in- 

 cluding wheat and oats, were selected, the grains of each were sown on a 

 separate plot and each group was descended from a single mother-plant. 

 The results amply justified the previous conclusion, for although there 

 were rare exceptions where the progeny was mixed, it could be accounted 

 for bv the original selection of hybrids. The selection of single individuals 

 as the only reliable .source of purity was thus raised to the rank of an ex- 

 perimental! v established fact. From this accidental discovery as it may 

 be called there are some very important deductions to be made. In the 

 first place it shows that repeated selection is unnecessary, for the types were 

 so uniform that further selection w^as simply impossible — the strain ^Tas 

 -established. It is the differences which exist in seemingly pure strains which 

 supplv the materials for selection, and there were no other differences from 

 the average in each type than such as might arise from plants grown on 

 the outside or in the" centre of a field or being accidentally crowded to- 

 gether. 



In the second place it .shows the sufficiency of one initial choice. The 

 type is isolated from the start, from its being the progeny of a single 

 mother-plant, and onlv requires to be kept free from accidental admix- 

 tures. It might seem at first sight as if the slow method of multiplication 

 might interfere with its practical usefulness, but it was the method of such 

 successful breeders as Le Couteur and Shirreff, and De Vries very per- 

 tinently remarks: — " In the beginning it was feared that the reduction of 

 the commencement of a race to one single head might protract_ its multipli- 

 ■cation so as to require more generations to reach the quantity necessary 

 for its ultimate distribution. Experience, however, soon showed this fear 

 to be unfounded, .since the elimination of all further selection soon over- 

 •comes the incipient deficiency. In reality the multiplication of a separate 

 ■culture mav go on as fast 'or even faster than that of an old-fashioned 

 methodical' selection. All the numerous new sorts which the station has 

 since introduced into the trade are derived each from a single individual." 

 From the farmer's point of view, however, the purity of the strain of the 

 -^vheat he is cultivating does not count for much unless it has better (juahties 



