556 JOURNAL OF FORESTRY 



lesser evil. It is an error, moreover, which is easily recognized and 

 eliminated. A "cull" factor must be applied to the results of any vol- 

 ume table to take care not only of rot, but also of the loss through 

 breakage, etc. It is a simple matter to increase this factor by the 

 trifling amount (rarely over 1 or 2 per cent) necessary to allow for 

 the unused top. 



It may be objected that height measurements are seldom taken with 

 precision in cruising and that the diflference in height to the fixed and to 

 the used top, being smaller than the accuracy of height measurement, 

 may be ignored. The fallacy in this argument lies in its comparison of 

 the compensating errors of ocular height estimates with the non-com- 

 pensating error of a lower or higher point of measurement. Even a 

 poor marksman wants to have his gun-sights properly adjusted. 



THE FACTOR OF FORM OR TAPER 



If it be accepted, then, that diameters should be measured at breast 

 height and heights to a fixed top cutting limit, at least for the common 

 tvpe of volume table, one question still remains to be asked. Is a 

 table based on diameter and height alone adequate, or must the third 

 factor of taper or form be added? The answer is: "Not if we can 

 help it," for there is no known method of calculating accurately this 

 third factor which does not involve a determination both much more 

 difficult and far less exact than those of height and diameter. And it 

 is doubtful if very serious errors result from our present practice in the 

 case of western conifers. An indication of the results of the neglect of 

 this third factor is given by the figures of Tables 1 and 2. The prob- 

 able variation of individual trees from group averages when form is 

 neglected was in these cases from 6 to 15 per cent. These figures 

 should be compared with the probable error in volume computation 

 by the much discussed form-point method of Jonson. Petrini has 

 shown ^ that this amounts to 5.4 per cent for Norway spruce. The 

 gain through estimating the form point of each tree is not sufficient to 

 conclusively prove its desirability. On the basis of the probable errors 

 calculated in Tables 1 and 3, it appears that the probable error (due to 

 volume tables) of "forty" estimates will be under 1 per cent, even with 

 scattered stands of but from 1 to 6 trees per acre. 



' Heddelanden f ran Statens Skogsforsoksanstalt, p. 267, Hafte 15, 1918. 



