62 



BREEDING CROP PLANTS 



central rows. The plants on each end of the plots to a depth of 

 at least a foot were cut and discarded. 



TABLE XIII. COMPARISON OF AVERAGE YIELD OF OATS, WHEAT AND BARLEY 



HARVESTED FROM BORDER Rows AND CENTRAL Rows 



PLOTS 132 BY 8.5 FT. 



It is clear from Table XIII that border effect may profoundly 

 influence yield. Long, narrow plots have a larger proportion of 

 their area in border than those which more nearly approach a 

 square. This would seem to indicate that square plots should 

 be given preference over oblong ones unless the borders are 

 discarded. As a matter of fact, most workers use long, narrow 

 plots because of the greater ease with which they may be seeded 

 and harvested with machinery. Furthermore, border effect 

 may be entirely removed by discarding the borders and ends 

 of the plot. 



The removal of borders becomes still more desirable when the 

 fact is considered that different strains and varieties may react 

 unequally to borders or ends. Evidence has been accumulated 

 which shows that some strains utilize the border to a greater 

 degree than others. Obviously those strains which gain least 

 from the alley space will not be given a fair trial unless border 

 rows are discarded from all the plots in the experiment. Con- 

 versely, those which have the greater ability to use the border 

 may be given a higher rating than they deserve unless the bor- 

 ders are removed. 



It would seem from the evidence presented that it is highly 

 desirable to discard ends and borders to a depth of at least a foot 

 in the case of rectangular plots and a foot at each end in the case 

 of rod rows growing side by side. 



Competition as a Factor in Plot Variability. Competition 

 between nearby strains, particularly under certain experimental 

 conditions, may seriously influence results. A tall variety may 



