310 IMPROVING THE ACCURACY OF TIMBER ESTIMATES 



their results is made on the basis of the assumption that accurate results on this 

 area were obtained by the strip method. The cost of the original estimate was $60.00 

 or 2.6jS per acre, 1.3jS per thousand. The cost of the subsequent strip estimate 

 was 8£ per acre or 4?i per thousand. The results clearly show that the average stand 

 per acre was successfully obtained for the pitch pine types in which the timber could 

 be seen, and where the area was carefully mapped in two degrees of density of stock- 

 ing and checked by strips and plots carefully selected there was no need of a subse- 

 quent estimate. 



The method of counting every tree was successful for white pine since all of the 

 tree tops were seen and the average tree was correctly guessed at, but for white oak, 

 the total count apparently failed. This was due not to a defect in the method or 

 its application, but to the fact that 123,000 feet of white oak was found later con- 

 cealed in the swamps. This reduced the error to 23 per cent for the portion seen 

 and counted. 



The estimate of spruce, hemlock and poplar broke down because of the funda- 

 mental difficulty of applying the sample plot method when based upon selection 

 and not on systematic arrangement. The swamp should have been crossed and all 

 parts examined. As it was, the sample plots were selected near the boundary where 

 the timber was one-half to two-thirds again as heavy a stand per acre as in the wetter 

 portions. This resulted in over-estimating spruce, hemlock and poplar. An 

 area or density correction here, or another day spent on that portion of the tract 

 would have greatly reduced this error. 



In extensive mapping and estimating of large areas for purposes 

 of classification as in the preliminary examinations for the establish- 

 ment of national forests, rough sketch maps of the areas of timber 

 types are made on the above principles by location of the cruiser on 

 a map and by triangulation. The estimate must depend upon the 

 location of occasional sample plots chosen with the best skill possible 

 to get average stands. 



In State work the construction of maps showing the timber resources 

 of the State or of various counties is usually carried on by similar 

 methods of mapping, using roads and the principle of the wheel or 

 odometer for distances and sample plots for average stands. In Massa- 

 chusetts a different principle is employed. Strips 4 rods wide are run 

 at J-mile intervals on which detailed measurements are taken of the 

 stand. No attempt is made to complete the map of timber in the inter- 

 vening areas, but the data are assumed to show the average for an entire 

 town, an assumption which is probably correct owing to the large 

 area involved. 



241. Estimating by Means of Felled Sample Trees. In the absence 

 of volume tables in earlier European practice, it was found that volume 

 of stands could be determined by calculating the diameter of the aver- 

 age tree, felling it and determining the cubic volume. This volume 

 multiplied by the number of trees in the stand .was supposed to give 

 the number of cubic feet in the entire stand. Since height and form 

 factor of individual trees both varied over a wide range, it was quite 



