REPORT OF THE CEREALIST 201 



SESSIONAL PAPER No. 16 



points of distinction. These selections within a single variety are usiially termed 

 strains. If, however, the original seed with which we commence our selection experi- 

 ments consists of a mixture of clearly distinct sorts, then selection if properly carried 

 on will enable us to isolate these various types, some of which may never before have 

 been grown in a pure state. Under such conditions we may obtain by selection new 

 varieties. 



Selection may be practised in two rather distinct ways. We may either carry on 

 repeated selection year after year with a view to obtaining slight but repeated advances 

 in some particular direction, or we may once for all pick out a number of specially 

 promisii^g plants or heads or seeds and propagate from each of these separately, with- 

 out further selection, but retaining, after a few years' study of them, only that strain 

 which has proved best. By one method we try to improve the whole mass as such; 

 by the other we merely seek to discover, from the study of its descendants, which was 

 the best plant . in the whole original mass and to retain ultimately only the pure 

 descendants of that plant, all the inferior strains being discarded. 



The method of repeated selection is the same in principle whether we choose each 

 year the largest or heaviest or hardest kernels or the largest or earliest heads — or on 

 whatever character we may base it. At first sight this system seems so full of promise 

 that one is not surprised at the number of experimentalists who have made use of it. 

 It fits in so well with the Darwinian ideas which have dominated the whole realm of 

 biology for so many years. If natural selection has done so much, why may not 

 artificial selection accomplish even more and in far shorter periods of time? While 

 no one can set the limits of what can be done by repeated selection in any direction, 

 the results which have been obtained have proved that the advancement is usually 

 much too slow for ordinary purposes. This method was most carefully tested in 

 Sweden, at the famous experiment station at Svalof, but was finally abandoned as 

 practically useless. Other workers elsewhere have been equally disappointed. While 

 it certainly is of value in some cases, one must beware of expecting too much from it 

 and must clearly recognize "its sharp limitations and the dangers which always attend 

 its use. 



Of late years some new ideas in regard to the origin of sjoecies have been made 

 prominent, especially by DeVries ; and we have been led to think less of the importance 

 of griidual changes in large masses of plants and more of the value of sudden changes 

 in individuals. We now recognize that each plant has a measure of individuality, 

 usually slight and unimportant, but occasionally so striking as to be easily observed. 

 As these points of individuality are often transmitted to all the descendants of the 

 original plant, we are generally able to obtain some strains of unusual value by 

 keeping separate the progeny of each selected individual which was chosen because 

 of some element of apparent superiority over the others. This second method of 

 selection may perhaps not inappropriately be termed DeVriesian. 



Personally I am a firm believer in the superiority of this method over the first, 

 or Darwinian, whenever the original seed with which the work is commenced is in 

 pure condition, true to name; because the method of repeated selection has some 

 serious disadvantages, quite apart from the fact that the improvements effected by it 

 are much too slow. In the first place one is obliged to decide every year, when the 

 time for selection arrives, which are the best seeds or heads or plants (as the case may 

 be), the decision being based on appearance or weight or some such characteristics; 

 while one is unable to take into consideration that quality which is perhaps the most 

 important of all, namely, the power of these selected individuals to transmit their 

 own peculiarities to their progeny. Animal breeders know that the best looking 

 animal does not always prove the most satisfactory parent. 



In the second method of selection the original choice of a number of individuals 

 is open to the same objections as were urged against the first method. But in the 

 second method the original choice is only of a provisional nature, the final selection of 



