976 JOURNAL OF FORESTRY 



Slowly spreading fires, on the other hand, exhibit a straight-line 

 relation, as Table 5 shows. The intervals of elapsed time are 15 

 minutes. 



Table 5 



Average ... 113 



It is seen that increases in perimeter are very nearly constant for all 

 periods, except the last. 



INFLUENCE OF WIND VELOCITY ON RATE OF SPREAD 



No discussion is necessary to prove that w^ind has a profound influ- 

 ence on rate of spread of forest fires ; the statement is sufficient. But to 

 deduce a general law from experimental data is a very difficult prob- 

 lem, for wind is only one of several factors which combine to determine 

 the rate of spread, and its isolation and evaluation is not easily accom- 

 plished. In the 33 experimental fires, however, enough measurements 

 were secured so that at least some tentative conclusions can be offered 

 regarding the effect of this factor. 



Three series of fires, 8, 10, and 15 in number, were available. For 

 each series the same criterion of rate of spread, namely, perimeter 15 

 minutes after the start, was used. The other factors — slope, tempera- 

 ture, relative humidity, and elevation — were averaged for each series 

 and were found to agree very closely, especially for Series I and III. 

 The average figures only are shown in Table 6. 



Table 6 



Relative Average No. of 



Series Slope humidity temperature fires 



I 15° 26 82 8 



11 14° 30 72 10 



III 12° 22 78^ 15 



Average 14 26 77.3 



Series I and III were burned during periods of the warm, dry, early 

 fall weather, long enough after rains so that the litter was air dry. 



