The possible saving by sampling can be estimated from the cost figures 

 which have been cited » Assuming that the amount of check scaling vjould 

 be the same in either sample tree measurement or total tree measure- 

 ment, the measuring cost of 12 cents a tree can be used for purposes 

 of comparison. From Figure 1, a comparison of number of trees in a 

 sample with number of trees in the sale shows that total tree measure- 

 ment might cost t?\fice as much as samplings About 10 cents per tree 

 can be saved by sample tree measurement o 



Sampling on timber sales is, naturally, more attractive to the seller 

 than the buyer, for two reasons: first, the seller pays for the 

 measuring job and is, therefore, interested in economy; second, for 

 a seller, such as the Forest Service, who has many sales, the possible 

 errors resulting from sampling tend to be compensating, although this 

 is not strictly true vjhen a wide range of stumpage values is involved. 

 For a single purchaser, however, who does not have the advantage of 

 many sales, the error, if any, could be to his disadvantage. For this 

 reason, a purchaser might lower his stumpage bid if sampling methods 

 are to be used on the sale, even though the possible error has a 50-50 

 chance of being in his favor. 



SmMARY 



The increased use of tree measurement on Forest Service timber sales 

 in the Northern Region and the desire to reduce the cost of routine 

 measuring jobs in sales administration prompted this study of sample 

 tree measurements The study xvas expected to determine the applica- 

 bility » work out methods, and establish sampling rates for the three 

 species most commonly tree measured in the Northern Region — ponderosa 

 pine, Douglas-fir, and vjestem larcho 



In the study, samples vjere drawn from scale books on eight national 

 forests. The samples included 24 sales in ponderosa pine stands, 22 

 in Douglas-fir, and 21 in vjestern larcho In addition, 41 samples of 

 check scale data vjere collected. As far as existing material per- 

 mitted, a range in timber size;, site, and defect conditions was 

 covered o 



The study showed that sampling in tree measured sales of the species 

 mentioned is a practical method of volume determination, both in 

 economy and accuracy. Although cost figures are inadequate, it is 

 estimated that a saving of about 10 cents a tree can be made by 

 sampling. The sample size and consequent sampling rate for a sale 

 of a particular number of trees are dependent upon several factors. 



-26- 



