Estimation of the parameters in the allometric equation by fitting a simple linear regression to the logarithmically transformed variables results in biased estimates of the arithmetic mean. This bias expressed as a percent of the mean approaches the limit −(1 − e−σ2/2) (100) as n increases. An adjusted estimator developed by Finney rather than the one given by Baskerville should be used when s2 is large and n is small. A change of measurement scale of the x or y variables presents no difficulty, but problems arise if variables are transformed to logarithms other than base e.

A method of two-stage forest sampling is presented. The first stage is Bitterlich sampling, and the second stage (subsampling) is by PPS sampling. The method is suitable for use in updating an earlier survey or in returning to the field to obtain intensive measurement for the selected subsample trees after the first-stage sampling has been completed. The necessary equations are given and their use shown by numerical example.

No abstract

The use of basal-area-factor (BAF)-5 and BAF-10 when point sampling in an Appalachian hardwood forest caused serious underestimation of sawtimber volume BAF's of 20 and 40 gave appropriate volume estimates and smallest mean-square errors. We recommend that BAF's 20 to 40 be used for sawtimber cruises. North. J. Appl. For. 2:23-24, June 1984.

No abstract