\'aRIATI()X AM) CoKKKI.ATlON IN TiMOTllY. 467 



TABLE VIII — Coefficients of Place Variation 



1905 and 1906 



1906 and 1907 

 1905 and 1907 



•474 ± -009 

 .550 ± .008 

 •382 ± .010 



Height 



•519 ± -008 

 .58=; ± .008 

 .486 ± .009 



An inspection of Table VIII shows that in the case of both characters 

 the highest correlation was between 1906 and 1907 and the lowest be- 

 tween 1905 and 1907. The low correlation between 1905 and 1907 

 indicates that instead of shifting back to former conditions in the latter 

 season there was a tendency of at least part of the sensitive individuals 

 to continue to vary in the same direction in which they had apparently 

 started in 1906, i. e., that the departure from the type of 1905 became 

 more pronounced each year. 



Another point of interest brought out by this table is that the different 

 characters under observation were not equally afifected by the causes 

 which produced place variation, weight being more susceptible than 

 height as shown by the lower coefficients. Moreover, certain portions 

 of the frequency distribution appear to respond more readily to environ- 

 mental influences than others. By reference to the correlation tables 

 (Figs. 145-150) it will be seen that for weight the higher values were 

 subject to the greatest variation while for height the fluctuation was 

 much greater among the lower values. 



When we consider the fact that in each case we are dealing with the 

 same character in the same population the correlations must be regarded 

 as comparatively low, thus suggesting the fallacy of basing our conclu- 

 sions regarding the condition of the population on the observations of 

 a single season. 



If our selections for weight (Figs. 145, 146 and 147) had been made 

 from the heaviest individuals of 1905 our choice would have included 

 several which were only mediocre, or in some instances far below the 

 average, in 1906 and 1907. On the other hand, if the selections had been 

 confined to 1907 a large per cent of the high-yielding individuals of 1905 

 would have been excluded. If height had been taken as a basis for selec- 

 tion it would have made but little difference whether the choice was made 

 in 1905 or in either of the subsequent years so far as the few extremely 

 tall individuals were concerned (Figs, 148, 149 and 150). This does 

 not hold true, however, for the remainder of the population, especially 



