--age at ground level (on trees greater than about 6 inches d.b.h., age was 

 measured at d.b.h. and increased by a constant 10 years to approximate age 

 at ground level) ; 



--site index (determined from site index curves and a 50-year base, USDA Forest 

 Service Northern Region Compartment Prescription Handbook, October 1965); and 



--basal area and trees per acre (for surrounding trees greater than 5 inches d.b.h., 

 basal area and trees per acre were measured using one prism plot (20 basal area 

 factor) having the sample tree at the center. For trees 5 inches and less in 

 d.b.h., a 1/300-acre plot was used). 



Many sample trees exceeded the d.b.h. range of Storey's and Fahnestock's data. A 

 large number of trees within the d.b.h. range of their data were also sampled to help 

 determine whether our data should be combined for analysis. Because Storey's and 

 Fahnestock's data lacked weights of dead branches occurring below the live crown and 

 lacked adequate information on size distribution of branchwood, some trees within the 

 range of their data were sampled only for dead crown weights and counts of branch basal 

 diameters for determining size distribution of crown material. For these trees, live 

 crown weights were not recorded. A listing of data is in appendix II. 



Analysis 



Crown We'lght and Bole Vle-lght Fev Tree 



For each crown section, fresh crown weights were reduced to an ovendry basis, using 

 an average moisture content that was determined by weighting sample moisture contents of 

 foliage and the branchwood size classes by their respective weights. The foliage and 

 branchwood weights were determined by means of step 2 in the section on foliage and 

 branchwood fractions. 



Live crown weights gathered in this study are probably slightly different from 

 those of Storey and Fahnestock. In the studies by Storey and Fahnestock, which consist 

 almost entirely of the same shared data, live crowns include dead branches found within 

 the live crown sections. In this study, live crowns contain only live branches. Most 

 of the dead branch weight occurred below the live crown; thus, the inconsistency, if it 

 exists, should be of minor consequence. Plots of crown weight over d.b.h. indicated 

 that data from Storey, Fahnestock, and this study fit together smoothly and could be 

 pooled. 



For d.b.h. greater than 1.0 inch, the relationships between ovendry crown weights 

 and tree characteristics were determined by first screening logical combinations of the 

 following variables using a multiple regression computer program called REX (Grosenbaugh 

 1967) : 



w = f(d, d2, d^, h, dh, d^h, c, dc, d^c, R, dR, d^R) 

 Inw = f(lnd, Inh, Indh, Inc, Indc, InR, IndR) 



where 



d = d.b.h., inch 

 h - tree height, ft 

 c = crown length, ft 



R = crovm ratio, (live crown length/tree height) 10 

 w = weight of crowns, lb. 



Unless otherwise stated, the definition of these terms applies throughout the paper. 

 Variables such as crown width and diameter at the base of live crown were omitted from 

 the analysis; however, all data are available for others to analyze if desirable. 



6 



