WEATHER BUREAU. 61 



The survey in Cottonwood Creek was conducted under very fav- 

 orable conditions. In the first place, the entire watershed of but 26 

 square miles had been previously surveyed by the United States 

 Geological Survey, and in the second place, it was possible to measure 

 accurately the run-off from the watershed during the snow-melting 

 season. The results are not yet at hand, but in this connection it 

 may be remarked that a snow survey in this watershed during the 

 previous year brought out the fact that the run-off from snow was 

 remarkably small. This result, it may be remembered, agrees with 

 the conclusions reached at Wagon Wheel Gap, Colo., at the experi- 

 ment station being conducted jointly by the Forest Service and the 

 Weather Bureau. 



An extension of the snowf aU work to the watershed of Salt River 

 in Arizona is of sufficient importance to merit separate mention. 



SNOWFALL m SALT RIVER WATERSHED. 



The demand for information as to the probable amount of snow 

 water available for irrigation purposes on the watershed of the Salt 

 River of Arizona, above the Roosevelt Reservoir, made itself felt 

 several years ago. 



Considering that one of the most important reclamation projects 

 under the supervision of the Federal Government centers in the Salt 

 River Valley of Arizona, it was deemed advisable to spare no effort 

 toward acquiring the information sought. At that time it was known 

 in a general way that the most representative snowfall tract in the 

 watershed of the Salt River above the Roosevelt Reservoir was in the 

 Apache Indian Reservation, north of Fort Apache, in the mountain 

 district of the NortJa Fork of White River, a tributary of the Salt. 

 That country was, however, almost an absolute terra incognita as far 

 as topographic maps are concerned. The first endeavors were there- 

 fore in the nature of a reconnoissance survey in an effort to construct 

 a rough topographic map of the watershed. A first survey was made 

 in November, 1913, from Roosevelt, Ariz., as a base, the objective 

 being a four days journey by wagon from that point. A few snow- 

 stakes were set up on that journey, but while the original plan was not 

 carried out by reason of severe snowstorms in the latter part of the 

 month, a substantial contribution was made toward mapping the 

 region. In this preliminary survey valuable aid was extended by 

 officials of the United States Reclamation Service, the United States 

 Forest Service, and by officials and agents of the United States Indian 

 Service. 



Such stakes as were installed on the first expedition were visited by 

 a second expedition, also in the nature of a reconnoissance survey, in 

 April, 1914. On this expedition the altitude of the snow fine was 

 noted and a considerable amount of general information was secured 

 that will be useful for comparative purposes in future years. It was 

 also concluded as a result of the second expedition that the region about 

 Paradise Creek, a tributary of the North Fork of the Wliite River, was 

 the most accessible locality for making permanent measurements of 

 the depth and density of the snowfields in the vicinity of Fort Apache. 



The work of 1915 consisted in two attempts to reach Paradise 

 Creek. The last one, made in the labter part of March, 1915, was 

 successful only by completing the last 8 miles of tlio journey on snow 

 skis after having left the horses and camp outfit on the way. The 



