296 IMPROVING THE ACCURACY OP TIMBER ESTIMATES 



of the area or 75 per cent of a correction of 10 per cent is applied to the 

 whole area. 



Since the greatest danger in applying corrections to mechanical 

 averages lies in failure to obtain a proper weighted average, and since 

 it is better to let these mechanical averages stand rather than to intro- 

 duce an unknown factor, dependent merely upon a guess, observations 

 intended to demonstrate the need for a correction factor must be made 

 as systematically as the strips themselves are run. Fixed points should 

 be chosen at definite intervals along the strips at which to take these 

 observations. These may be taken for instance at points 20 rods apart 

 on the strip. At these points, the areas on either side of the strip 

 should be compared with the stand upon the strip. 



The final result is expressed in terms of a per cent, but if each sepa- 

 rate observation of a series is so expressed, the resultant per cent will 

 not be weighted by the volumes to which its components apply; e.g., 

 two successive observations may give the following result: 



The actual correction factor is +2J per cent instead of zero. 



This principle of weighting the observations by volume is very 

 simply applied. It consists of entering for each observation, not the 

 per cent of comparison, but a comparison based on an ocular estimate 

 of the stand per acre. The estimator puts down in two parallel columns, 

 first the stand per acre estimated to be on the strip at that point, second, 

 the stand per acre estimated to be on the remaining area. In arriving 

 at this he includes as large an area as comes under his observation 

 both on and off the strip. For double observations, i.e., taken on both 

 sides of the strip, it is necessary to record the stand on the strip twice, 

 once for each observation off strip. 



On the completion of the unit, these stands on and off strip are 

 totaled. By dividing the total off strip by the total on strip, the true 

 weighted volume correction factor is obtained. 



This factor is a percentage relation and therefore does not require 

 that the ocular estimates per acre on which it is based be correct, pro- 

 vided they are in the proper proportion. Each ocular guess may be 25 

 per cent too low, yet the resultant correction factor will be, identical 



