198 



STATE BOARD OF AGRICULTURE. 



The adjacent horizontal lines represent pounds per plot. In selectino- 

 the land, we try to find a piece as nearly nniform as possible for this 

 class of work to begin with. The accuracy of our scales, in weighino- 

 yields is usually half a pound. Therefore by using pound lines we 

 can place a cross half way between two lines and proceed with as great 

 accuracy as the scales. 



An average is taken of the various yields of the check plots, and a 

 liorizontal line is drawn to represent that average. Dots are made on 

 the vertical plot lines having the same relation to the line of average 

 check as the crosses have to the curve. If a cross is five pounds above 

 the curve of soil fertility, the corresponding dot is placed five pounds 

 above the line of average fertility. If a cross is below the curve, the 

 dot is placed the same distance below the line of average check yield. 

 When these dots have been properly placed on all the plot lines, those 

 representing adjacent plots are connected by straight lines. This 

 gives us the yield cune, where the question of soil fertility has been 

 taken out of consideration. Averages can now be made between dupli- 

 cate plots to still further eliminate error. Three or more of these yearly 

 averages should be had to give us a fair comparison among the strains 

 in question. 



(Fig. 15 illustrates the way the seed plots of cross-fertilized plants 

 may be grouped.) 



STAKES AND LABELS. 



The stakes are two inches by one half-inch and two feet long. They 

 are painted white freshly each year. The labels are plain white card 



Fig. 1. A field of pedigree cowpea.s No. G0901, crop 1909. 



