COMMERCIAL LINT 131 



ment, and determine the out-turn and length for each 

 plant in a certain period of the season, we can draw curves 

 showing the distribution of the variations in each respect 

 through the family. If we now place these curves on 

 two adjacent sides of a square, we can make a target 

 diagram in the following way : Find the position of 

 plant No. 1 in the length curve, and draw a line into the 

 square at right angles to the side along which the length 

 curve is plotted. Then repeat the process with the out- 

 turn curve. At the point of intersection of these two 

 lines make a conspicuous dot. Repeat the process for 

 each plant, when it will be found that the group of dots 

 thus made will give a picture of the amount of " scatter " 

 in both characters at once, instead of only showing one 

 at a time as the curves did (Fig. 16, p. 134). 



If we are handling a pure strain in this way, the scatter 

 in either character will be definite and regular, and the 

 dots wuHiorm (with suitable plotting) a circular group, 

 the centre of which is densely dotted, while the dots 

 become fewer and fewer towards the margin of the group 

 (Fig. 16, Targets 5-7, 10). The centre of the group lies 

 at the mean for each character. 



If we now mix two strains together, which are dis- 

 tinguished in their average out-turns and lengths, the dots 

 will form two groups. If the two strains are very widely 

 different, the groups will be quite separate. If they are 

 only slightly different, the fact will still be recognizable 

 when the diagram is viewed from a distance, for instead 

 of blurring into a circular arrangement, with the darkest 

 spot in the centre, the blurred diagram will be more or 



