HYDROLOGY OF NEW YORK 



109 



catchments of 40 to 50 inches and with a corresponding runoff 

 somewhat less than one half of the rainfall. The minimum runoff 

 for these streams is from one fourth to one sixth of the corres- 

 ponding rainfall, or from about 16 per cent to 25 per cent. 



A further class, the far western streams, may be mentioned, in 

 which the runoff is only a very small percentage of the rainfall, in 

 some cases not more than 1 per cent to 5 per cent, or at times even 

 less. Probably comprehensive study would further subdivide 

 These streams, but the intention at present is to merely call atten- 

 tion to some of the more marked peculiarities as. a basis for final 

 detailed study. 



If one takes the streams of the far west, as for instance Loup 

 river, in Nebraska, with a catchment area of 13,512 square miles, 

 where the rainfall in 1891, observed at 21 stations, was on an 

 average only 12.81 inches and the runoff of the stream did not 

 much exceed 1 inch, he will find entirely different conditions from 

 those above stated. In many cases streams in that locality run 

 much less than 1 inch. The South Platte, at Denver, Colo., in 

 1890, with a rainfall of 11.81 inches, ran 0.62 inch. The catch- 

 ment area at this place is 3810 square miles. At Orchard, Colo., 

 the South Platte, in 1898, with a rainfall of about 17 inches, ran 

 0.9 inch. The catchment area at this place is 12,260 square miles. 

 The Kepublican river, at Junction, Nebr., with a rainfall of about 

 26 to 28 inches, in 1898, ran 0.39 inch. The catchment here is 

 25,837 square miles in extent. 



The foregoing statements indicate the essential truth of the 

 proposition that, broadly, each stream is a law unto itself. Any 

 formula, for either maximum, average, or mean runoff, which 

 does not take this into account is incomplete. 



Estimation of runoff from rainfall diagrams. Can runoff of 

 si reams be estimated from diagrams of monthly rainfall? The 

 writer has spent considerable time on this problem without arriv- 

 ing at any very satisfactory conclusion. For some months such 

 a diagram may be made to fit quite closely, while for others, 

 differences of as much as 2 or 3 inches appear. The conclusion 



