METHODS OF BREEDING SMALL GRAINS 115 



if the rows harvested are 16 and 20 ft. long, respectively, the 

 conversion factor will be 0.1. 



The rod-rows are about twice as far apart as the rows made by 

 a field grain drill. As from one and one-half to two times as much 

 seed is planted per nursery row as under field planting, the rate 

 of seeding per acre does not differ materially in the two methods. 

 These row trials have been shown to give results similar to those 

 from field tests, although the average yield of the crop is not 

 the same (Montgomery, 1913; Love and Craig, 1918a). 



As has been previously noted, 'there are two general methods 

 of work, i. e., the use of single- and three-row plots. Three-row 

 plots in which the central row only is used to secure yield are de- 

 sirable as they help to control mixtures at planting and harvest- 

 ing time, overcome competition between nearby varieties and help 

 in obtaining more dependable data on lodging. They require 

 more land and the cost is somewhat greater for planting and 

 cultivating. In sections where soil heterogeneity is very great 

 it is possible that the use of single-row plots and numerous repli- 

 cations may be somewhat better than three-row plots and fewer 

 replications. On land that is well suited for field plot work the 

 use of three-row plots and three replications is advised. 



After a strain has been grown for three years it may well be 

 removed from the row-yield trial and either increased if it shows 

 promise or discarded if it appears to be of no value. At Cornell 

 new sorts are introduced to the farmers for trial directly from the 

 rod-rows. In many cases the new sort is finally tried in variety 

 plots planted by the usual field-plot method. This gives an 

 expression of yield under normal methods of planting and favor- 

 able field conditions. 



Summary of Methods of Selection. 1. Determination of the 

 varieties which possess economic possibilities. These may be 

 commercial varieties or new introductions. 



2. Head selection of these promising varieties. 



3. Test of head selections in plant-rows. The very undesirable 

 strains are eliminated in the field by inspection. A few may be 

 discarded on the basis of yield or seed characters. 



4. Yield determinations of the selections, using three plots of 

 a single row each, systematically replicated, if seed is available. 



5. Continuation of the row test. When land is well suited it 

 is believed that four systematically distributed plots of three 

 rows each will give reliable results. Possibly the arrangement 



