i BULLETIN 1495, U. S. DEPARTMENT OF AGRICULTURE 
Obviously, as in setting up a criterion of successful protection 
measured by area burned, no strict mathematical justification can 
now be adduced for any figure. In most years of the decade under 
discussion, when the rate of one fire per 10-day period per 100,000 
acres of protected area was exceeded in spring, the need of an aux- 
iliary protection organization was manifest, and when it was not 
employed a considerable acreage was burned. The same thing was 
true in the fall whenever special protection was abandoned before 
fires fell short of this rate. One fire per 10-day period per 100,000 
acres has therefore been adopted as a reasonable criterion of need 
for special protection on a regional and major-type basis. In certain 
localities and under particular circumstances, placing of the full pro- 
tection organization before any fires have started may be good 
management. 
METHOD OF STUDY 
MATERIAL USED AND METHODS OF HANDLING 
The data used in this investigation are derived from reports made 
by forest officers on 10,476 fires that occurred from 1911 to 1920, 
inclusive, on and adjacent to 12 timbered national forests of Cali- 
fornia, as distinguished from the brush-covered protection forests. 
(Table 4.) It is the Forest Service practice to record as soon as 
possible after each fire the essential facts of the fire, including its 
origin, its history, the factors affecting its spread, and its cost. Pro- 
eressive changes in the recording forms have been made during 
these years to improve the quality, to facilitate the use, and to in- 
crease the scope of the data recorded. Of the hundreds of indi- 
viduals who have been connected with fire control and who are 
responsible for the data on these forms, many were unaccustomed to 
recording notes in the field and to assembling facts on paper. The 
pressure of work made it impossible for many of the field officers to 
prepare the reports for some days after a fire; and for the same 
reason detailed examination and survey of many large fires has been 
impracticable. It must be recognized, therefore, that in quality, 
accuracy, and completeness the data are far from perfect. But such 
a study as this is not contingent on perfection of data. Data clearly 
inaccurate have been discarded, and it is probable that ordinary 
errors are sufficiently compensating so that in the mass the data 
give approximately true values. 
Taste 4.—Dzistribution of fires by cause and cover type, 1911-1920, 12 timbered 
forests 
General-risk fires Speciai-risk fires 
Type TSR TAPP aeI Gy CRON! GFh aah a eas 
Incen- |} Light-} Un- Brush | Lum- | Miscel-} Rail- | causes 
Total 
Camper diary | ning | known burning| bering |laneous} road 
Western yellow pine_ 829 460 | 1, 564 237 | 3,090 87 208 144 1 3, 654 
Mixed conifer_______ 620 429 | 1,228 197 | 2,474 80 272 100 73 2, 999 
IDYowhellkes saps 94. 260 263 9 626 17 2 3 J 659 
Sugar pine-fir_______ 56 21 117 17 211 14 9 5 3 242 
TAN AE ot a es dome 176 26 512 37 751 17 il 20 1, 811 
Grasse eee ee cores 86 30 53 14 183 9 13 14 5 224 
Chaparral eas 72 95 30 27 224 15 6 22 14 281 
Woodland 225252202 62 126 93 15 296 12 4 Ii 7 330 
Brushes ois aes ae rsh 249 307 449 95 | 1,100 49 33 58 36 1, 276 
Motalecaeees 2,244) 1,754) 4,309 648 | 8, 955 300 558 387 276 | 10,476 
