Fire-Season Forecasts 597 



inclusive, will be compared. It might be stated here that only 

 maximum temperatiires are used, as it is considered that these have 

 a greater effect on the probable hazard by drying out the cover. 



There is not enough difference in the mean maximum tempera- 

 tiu*es of the simimer months to have any effect, and about all there 

 is to say is that dry winters are Hable to be the coolest. 



If precipitation has no real relation to the number of fires, and 

 if the temperature has no effect, at least a wet winter should mean 

 a fire season that wovild be late in beginning, or at least late in be- 

 coming dangerous. So a table is given showing precipitation for 

 the previous nine months, the nimiber of fires in May, June and 

 July, the acreage burned, and in what ten-day period the sea- 

 son began. A wet winter ought to be a safe basis for making a 

 prophecy, for the condition of the groimd in the early season has a 

 marked effect on fires, and to make the table stronger the precipi- 

 tation in March, April and May, and also the number of clear days 

 in April, May and June will be given to show how much chance the 

 sun had to dry out the ground. 



Clear Days Period No. Fires Acreage Precip. 

 April, May Fires June and in June in April 

 Year Precip. and June Began July and July and May 



1911 42.3 59 June 1 15 1,264 1171 



1914 38.4 53 June 20 28 114 6 20 



1915 37.8 43 July 1 6 6 1007 



1912 19.6 56 May 10 16 2,953 1065 



1913 17.5 61 July 1 8 409 6.43 



The two wettest seasons have the smallest acreage in June and 

 July, as compared with the two dryest, which is right. But the 

 number of fires for the two wettest was greater than for the two 

 dryest. The season that had the fewest clear days began late 

 and had the best fire record, but the one which had the most clear 

 days began just as late, and only had two more fires. The year 

 which had the^most rain in March, April and May began its sea- 



