VARIATIONS IN LAKE LEVELS 403 



valuable commercial tree, is governed, first of all, so far as we 

 know at present, by fire. Had fires been kept out of these forests 

 in the last thousand years the fir which gives them their- dis- 

 tinctive character would not be in existence, but would be re- 

 placed in all probability by the hemlock, which fills even the 

 densest of the Puget Sound forests with its innumerable seed- 

 lings. I hasten to add that these facts do not imply any desira- 

 bility in the fires which are now devastating the West. 



These examples of the relations of fire and the forest are cited 

 because they are conspicuous among the few which have already 

 been worked out. Without question a number of relations of 

 vastly greater importance remain to attract and reward the 

 student of this branch, one of the most fruitful and fascinating 

 of all the fascinating and fruitful branches, of forestry in the 

 United States. 



VARIATIONS IN LAKE LEVELS AND ATMOSPHERIC 

 PRECIPITATION 



By Alfred J. Henry, 

 Chief of Division of Records, U. S. Weather Bureau 



A study of the fluctuations in the surface level of the Great 

 Lakes is always an interesting problem. It is especialty so at 

 the present time, owing to the near completion of the Chicago 

 drainage canal and the projection of various industrial enter- 

 prises, which, when completed, will divert large quantities of 

 water from present channels. The physical problems involved 

 in an adjustment of the situation are manifold and intricate, as 

 are also the commercial interests seeking recognition. The writer 

 desires at this time, however, merely to direct attention to the 

 possibility of determining the probable level of the lakes by 

 accurately gauging the precipitation over the various watersheds. 



The stage of water in a river or natural reservoir system, such 

 as the Great Lakes, is dependent primarily upon supply, dis- 

 charge, and evaporation. In small streams the correspondence 

 between precipitation and water stages is easily observed. The 

 drainage basins themselves are small, and when precipitation 

 has once begun but a short time lapses before all portions of the 

 basin are contributing to the stream-flow. In a small basin of 

 uniform surface and slope it is possible to calculate the exact 



