Dougl.), subalpine fir (Abies lasiocarpa [Hook.] Nutt.), 

 aspen (Populus tremuloides Michx.), and narrowleaf 

 Cottonwood (Populus angustifolia James). 



Each species was considered separately for comparison 

 of English with metric diameter measurements. 

 Diameter measurements taken by the two standards dif- 

 fered by only a few tenths of an inch. Because volume 

 prediction is the primary use of the diameter measure- 

 ment, the diameter data were analyzed in terms of effect 

 on volume prediction. A relative percent difference 

 statistic was derived to make comparisons: 



VDIFF = 



MV-EV 

 EV 



X 100 



where: 



VDIFF = Percent volume difference between volumes 

 computed with metric and EngUsh diameters. 



MV = Volume computed with diameter measured at 

 1.3 m. 



EV = Volume computed with diameter measured at 

 1.37 m. 



Gross cubic-foot volume equations (to a 4-inch top) 

 were used for cottonwood (Kemp 1958). Engelmann 

 spruce (Myers and Edminster 1972), subalpine fir (Hatch 

 1975), Douglas-fir (Hatch 1975), aspen (Edminster and 

 others 1981), and lodgepole pine (Myers 1972). All of 

 these equations were developed using EngUsh standards 

 with d.b.h. measured at 4.5 ft (1.37 m). 



Statistics were analyzed by constructing 95 percent 

 confidence limits on the volume difference statistic 

 (VDIFF) sorted into groups by species and 2-inch 

 diameter classes using both Student's t-test and the chi- 

 square test. The chi-square confidence hmits were deriv- 

 ed using the pivotal quantity method (Mood and others 

 1974). 



Student's t-test is conamonly used in making paired 

 comparisons, but Freese (1960) points out the chi-square 

 test is more appropriate in most forestry appUcations. 

 The chi-square test considers both bias and precision 

 when making a paired comparison, while Student's t-test 

 considers onty bias. Both tests require normally 

 distributed data. Figure 1 illustrates that this is a 

 reasonable assumption for subalpine fir. The other 

 species data had similar distributions. Table 1 Usts the 

 data with both Student's t-test and chi-square con- 

 fidence limits. If a confidence limit does not include zero, 

 there is a statistically significant difference between 

 Enghsh and metric points of measurement. The con- 

 fidence Limits do not include zero for the most part. 

 Those that do include zero are for a small sample size or 

 for small trees that were not compatible with the volume 

 percentage (VDIFF) transformation. Thus, these tests in- 

 dicate that statistically the diameters taken at 1.3 m 

 and 1.37 m are different, which is not unexpected. The 

 question then becomes what impact will the diameters 

 taken at the different measurement points have on 

 volume? 



30 r- 



20 



>- 

 o 



o 



LU 



10 



Mm HUL 



-1.6 



0.0 



1.6 



3.2 



4.8 



6.4 



8.0 



PERCENT VOLUME DIFFERENCE 



Figure 1. — Volume differences (VDIFF) for 

 subalpine fir computed witti metric and 

 Englisli diameters. 



2 



