REJECTION OF ABNORMAL PLOTS 405 



the limits shown in the figure, the rule has been applied in this country 

 to all plots whose heights classify them with a given site. The natural 

 variation in volume for plots within one site class is greater than 15 per 

 cent, independent of abnormalities — hence if all plots which vary 7| 

 per cent above or below the average volume for the site at that age are 

 rejected, about half of the plots, although normal, may be thrown out. 

 If this rule is to be correctly applied as a test of normality, the arbitrary 

 permitted variation of 15 per cent, if used at all, should first be corrected 

 by finding what the normal yield of the particular plot should be, based 

 on its actual height. If height for the plot is midway between quality 

 I and II, normal yield is also midway between the averages for these 

 qualities. The steps necessary would be as follows: 



1. Draw curves of average height as shown in Fig. 84, and curves 

 of average volume as shown in Fig. 85. 



2. Determine the per cent of variation above or below average height, 

 for each plot, and subtract or add the same per cent from the volume of 

 the plot. This gives the corrected volume of the plot based on. 

 average height for the site. 



3. Compare the corrected volume of the plot with the average volume 

 for the site. If it falls above or below the calculated normal by more 

 than the desired per cent of error the plot can be thrown out. 



4. After testing the normality of all plots, re-compute the average, 

 using only those plots accepted as conforming to the standard. 



If 15 per cent is a proper standard of variation for forests under 

 management, it is probable that even with the above method this per 

 cent is too small as a criterion of normality for natural stands. It 

 should be possible, by eye, to select plots of which at least 95 per cent 

 will be suitable for inclusion in obtaining the average results for a stand- 

 ard yield table. With a range of basal area increased to 25 per cent 

 for plots of the same height based on age as indicated, it is probable that 

 only distinctly abnormal plots will be rejected. 



In constructing volume tables it is not customary to reject trees 

 after they have been measured for volume, since rejection can take 

 place in the selection of the tree. With plots for yield tables, the desire 

 to secure a theoretically normal or uniform standard may easily lead 

 to too rigid a rejection of plots which are entirely suitable for the aver- 

 age sought. Maximum yields, on the basis of site alone, should never 

 be sought by these average curves of yield, since the best portions of 

 the site will exceed the average. Again, such tables, if made for natural 

 stands, should show what can reasonably be expected in stands repro- 

 duced naturally and not thinned, on the average acre for site. A con- 

 sistent average showing the probable progress of a fully or normally 

 stocked acre by decades, and not an abnormal maximum yield, is the 



