Two bark thickness measurements were recorded for each tree — one on 

 the north side and one on the south side at 41/^ feet above the ground — to 

 the nearest 1/32 inch (fig. l). Tree diameters were measured with a 

 diameter tape to the nearest 0.1 inch after all loose bark was removed 

 from the measurement area. These data were analyzed by linear regression 

 techniques, using average bark thickness values for each tree. This pro- 

 vided a total of 209 observations. 



Figure 1.— Bark thickness data were measured by using a depth gage. 



The following regression equation was developed from the data: 

 Y — .135 + 0.020X1 

 where: Y = average bark thickness 

 Xi = tree d.b.h. 



The coefficient of determination (R-) was found to be .51 for the 

 equation, indicating that 51 percent of the variation in bark thickness 

 was associated with d.b.h. 



An analysis of variance was computed, yielding an F- value for variation 

 due to regression of 433. 



This value is highly significant, indicating that the variation accounted 

 for by fitting the regression line to the data was significantly greater than 

 the remaining unexplained variation. Figure 2 is a graphic presentation 

 of the regression equation. 



2 



