102 MAINS AGRICULTURAL EXPERIMENT STATION. I916. 



METHODS. 



The general methods of pure line breeding in the case of 

 cereals may be briefly outlined as follows : In the first year 

 individual plants are selected from the fields or plots of the com- 

 mercial varieties. These plants are selected because in some 

 one or more respects they appear to be better than their neigh- 

 bors growing under the same conditions. Each of these plants 

 is harvested separately and various data recorded concerning l\. 

 On the basis of the notes made in the field and the data on the 

 threshed plants a further selection is made. 



The seed from the plants which are finally selected are sown 

 the next spring in short rows in a cereal garden. In our 

 work we usually plant 25 kernels in each row. The rows are 

 one foot apart and the plants three inches apart in the row. 

 This allows each plant sufficient room to develop and the 25 

 plants are enough to judge the character of each pure line. 

 The garden rows are subjected to a severe selection. Careful 

 notes are taken in the field and considerable data taken at the 

 time of threshing. Out of several hundred rows usually only 

 a few are judged good enough to be continued. 



The third year the seed from the selected rows is sown in 

 small multiplying plots. In our work these plots have been 

 1-2000 acre in area and usually in duplicate. In these plots the 

 grain is planted much closer together in fact, approaching the 

 number of grains per unit area, that are usually sown in the 

 field. These plots are subjected to a still further selection and 

 only the best retained. 



The fourth year there is sufficient seed from each selected 

 pure line to sow one or more field plots. According to our 

 method of testing varieties 6 these are sown in 1-40 acre plots. 

 Here they are tested in duplicate or quadruplicate plots fo r 

 several years and subjected to still further selection until only 

 those which are superior in some respects, at least, to already 

 existing varieties are retained. 



In the work with which this bulletin is concerned over 450 

 plants were selected in 1910. Of these something less than 200 

 were tested out in garden rows. Eighty were deemed good 

 enough for the 1-2000 acre test, but only 34 ever reached a field 



"Surface, F. M. and Barber, C. W. Loc. cit. 



