Partiale density. — Density values were determined from 12 randomly picked samples 

 from each species and diameter size class. A mercury pycnometer furnished the density 

 values for the needles and 0- to 1-cm. size classes. For the other size classes, 

 volumes for calculating density were determined using average diameters and a length of 

 15 cm. for the samples. Average diameters were based on two measurements taken perpen- 

 dicularly at 1-cm. intervals. Weights were on an ovendry basis and volume on an air-dry 

 basis. Standard errors of the estimate were less than 4 percent of the mean density 

 values for each size class. 



Lew heats of combustion^ total ash content ^ and siZica-free ash content. — Duplicate 

 determinations for three samples picked randomly from the slash provided these values 

 for needles and the 0- to 1- and 1- to 5-cm. size classes. 



Moisture content. -Six samples for each size class, three near the top and three 

 near the bottom of the slash, were collected just prior to ignition. The branchwood 

 was ovendried and the needles subjected to xylene distillation. Standard errors for 

 each plot ranged from 3-1/2 to 5 percent of the mean moisture contents. 



Wind. — Wind velocity was measured at the expected midflame height close to each 

 plot but away from indraft influences of the fires. Anemometers (having a threshold of 

 0.5 m.p.h.) connected to portable battery-operated recorders furnished the measurements. 

 A wind vane also placed at midflame height recorded wind direction. 



Temperature and relative humidity. — Temperature and relative humidity were measured 

 just prior to ignition and continuously. They were recorded for reference but were not 

 required for calculations in the spread rate model. 



Measurement o£ Fire Variables 



The plots were ignited across one end and allowed to burn for 5 feet to attain a 

 quasi-steady state before spread rate measurements were begun. Time for the fire to 

 burn between 2-foot intervals was recorded over a 20-foot distance (fig. 3). 



Reaction intensity was determined from: 



^R = y^^O^-' 



1=1 



T = D/R (3) 



where : 



T = reaction time of flame front (time for propagating flame to pass a fixed 

 point), min. 



n = average fractional weight loss of fuel in flame front 



w = loading, Ib./ft.^ 



h = low heat yield, B.t.u./lb. 



i = index for particle size classes 



D = flame depth, ft. 



R = rate of fire spread, ft. /min. 



8 



