70G JOURNAL OF FORESTRY 



2. The area burned over depends upon the number of fires that start 

 and the average acreage per fire. 



3. The number of fires that may start depends upon the presence 

 or absence of causes of fire at the time when the forest is in condition 

 to burn. This to some extent is modified by regional differences and 

 by differences in character of the forest. 



4. The average acreage per fire depends on differences in chmate, 

 in type of forest, in infiammabihty of forest as influenced by character 

 and condition of the stand and ground cover, and in the effectiveness 

 of suppression organization. 



5. The proportion of the total value of forest burned over that will 

 be destroyed by fire depends upon the climate, and upon the type, age, 

 and condition of the forest. Thus, the proportion of loss per burned 

 acre will be much less in an open, clean stand of mature longleaf pine 

 than in a heavy stand of young white pine with a heavy litter of dry 

 slash. 



PROPOSED METHOD OF CLASSIFYING RISKS 



Following these principles, forest risks may be classified as follows : 



(a) Divide the United States into a number of regions, chiefly on 

 the basis of climate and general forest conditions. There will probably 

 be between 20 and 30 of these regions. 



(b) Classify the principal types of forest in each region, grouping 

 types together where they are substantially similar in inflammability 

 and in situation. There may be from 5 to 10 types in each region, 

 probably nearer 5 than 10 in most regions. 



(c) Divide each type into about 4 age-groups, as, for instance, 

 reproduction, poles, immature merchantable, mature. The points of 

 division will of course vary somewhat with type and with region. 



(d) Subdivide each age group into several (perhaps five) grades 

 of comparative inflammability, using a scoring system that takes ac- 

 count of the character and condition of ground cover (duff, grass, 

 weeds, brush, reproduction, debris from lumbering, windfalls), stand- 

 ing snags, moss and dead lower branches on trees, density of stand 

 and of crown cover, scars or pitch streaks on trees, character of soil, 

 slope, and exposure. 



(e) Grade each of the above grades of inflammability into several 

 (perhaps 5) classes according to the comparative possibilities of sup- 



