Edward N. Wentwortii 11;") 



that the standard deviation is the best method of comparison. The 

 interesting point is the relative constancy with which Line I breeds 

 true for fecundity. 



To show that this low type is absolutely distinct let us briefly 

 compare Table I and Table II. While the mean for the low line is 



TABLE II. 



Breeding Record of Line II. {Strongly Vigoroii.i.) 



Pair A Pair B Pair C Pair D 



F2 116 132 98 141 



I -^ -I I -^ 1 I ^ 1 -I -^ n 



I II II I I I 



F3 140 128 110 131 112 122 152 135 



I I I 



n ^ ^ h ^ 



I I I 



^4 161 127 134 117 144 136 128 137 120 129 116 129 147 160 118 151 



I I I I I I I I I I I I I I I I 

 Fa 142 138 127 141 116 134 139 151 138 114 101 145 126 154 136 113 



II II II II 

 ^'o 154 136 127 114 141 104 145 162 



I I I I II II 



F: 136 156 112 144 152 94 139 134 



I I I I I I I I 

 f, 163 147 141 152 127 96 136 157 



II II I I I I 

 Fg 157 138 134 133 141 97 132 172 



I' I II I I I I 



Flo 149 14* 129 156 132 100 141 ISO 



29-5 the mean for the high line is 135'86. The means of each genera- 

 tion when compared as in Line I are as follows : 



The mean for the entire line is 135-86 and the standard deviation 

 18 3. This gives a coefficient of variability of 13-47, lower than the 

 coefficient for the low line although the range of variation is obviously 

 higher as evidenced by the standard deviation. Suffice it to say, 

 however, that the two lines are quite distinct, they overlap in nowise 



