PHYSICS. 407 



to determine the hourly disturbance of level at each gage by the wind, 

 the fluctuations in level due to the wind having been found to be of 

 such a character that a consideration of mean daily values of elevations 

 and of the wind directions and velocities is not sufficient. These 

 successively discovered difficulties must apparently be overcome in 

 the follo\\ing order to secure a better formulation of the laws governing 

 the amount of stream flow: (a) The response of the water surface at 

 a recording gage, hour by hour, to winds of varying directions and 

 velocities must be determined. (5) The fluctuation of the level at the 

 gage from day to day, due to both the winds and the barometric 

 gradients, must be evaluated, (c) The evaporation from the lake 

 surface under the various conditions must be determined with all the 

 accuracy possible. When this full-scale determination of the laws 

 of evaporation from a large water surface becomes available it is 

 probable that progress on the stream-flow investigation can be made. 



The investigation had progressed to the stage indicated by the 

 general statement of the preceding paragraph in August 1916. 



During June, July, and August 1920 the investigation has been concen- 

 trated on the hourly wind effects. Decided progress has been made. 

 Nearly complete determinations have been secured at Buffalo and 

 Cleveland on Lake Erie. The success in this respect has shown that 

 this part of the problem can, with reasonable promptness, be conquered 

 for each station sufficiently to make the way clear thereafter for rapid 

 and continuous progress on the main problem of determining the 

 evaporation. It is now clear that the process of computing the evap- 

 oration can be greatly simplified and shortened in comparison with that 

 originally contemplated. It appears, therefore, that the month of 

 July 1920 has marked the change in the investigation from the stage 

 of successive discovery of new difficulties and complexities to a stage, 

 which it is hoped will last to the end, of steady progress along a reason- 

 ably clear road. It is believed that the principal difficulties and com- 

 plexities are now known and that effective methods have been devel- 

 oped for overcoming them. 



An intensive study is nearly complete of the record of hourly eleva- 

 tions of water surface on 22 days each at Buffalo and Cleveland. 

 The principal points of interest and importance are indicated briefly 

 in the paragraphs which follow. It is fitting that the detailed state- 

 ment of these points be reserved until the study is complete and has 

 been strengthened and confirmed by a similar study at three points 

 on Lake Michigan-Huron. 



The response of the lake surface to each change of wind is very 

 prompt. The kg is a few minutes only, certainly less than one hour. 

 Whenever the wind over any part of the lake changes in direction or 

 velocity, a corresponding change in the slope of the water surface in 



