HYDROLOGY OF NEW YORK 



197 



others not here published, that there is no definite relation between 

 evaporation and mean annual temperature. 



Exponential formula. On fig. 10 this relation is expressed by 

 an exponential formula, after the manner proposed by Mr Fitz- 

 Gerald in his paper, Flow of Water in 18-inch Pipes. 1 Such a 

 curve has the advantage that it is the best approximation possible 

 to obtain from the given data. It will be noticed that it differs 

 slightly from the curve of fig. 9. At 30 inches rainfall this differ- 

 ence amounts to about 1.3 inches of runoff. 



While on the subject of exponential formulas it may be re- 

 marked that their chief advantage lies in the possibility of taking 

 any set of data and deducing the curve which best suits the 

 conditions. 



Description of runoff diagrams. Fig. 11 is a runoff diagram of 

 the Hudson and Genesee rivers, Hudson river for 1888-1901, in- 

 clusive, and Genesee river for 1890-1898, inclusive. In preparing 

 this and the following diagrams it is considered that if both run- 

 off and precipitation were correctly measured the points would 

 fall in a regular curve approximate ly like those shown on figs. 9 

 and 10. Such diagrams may therefore be taken as a criterion of 

 the accuracy with which the observations have been made. It is 

 easier, however, to measure the runoff than it is to measure the 

 precipitation, and hence when large variation occurs, as it does 

 in these several diagrams, we may first look for it in the pre- 

 cipitation records. As regards the Hudson area, it has been the 

 writer's custom to take the rainfall of the northern plateau of 

 the State Weather Bureau as, on the whole, best representing the 

 - rainfall of the Upper Hudson area. With the exception of the 

 years 1899 and 1900 the points all fall within from an inch to 

 an inch and a half of the curve. Those two years have, however, 

 been computed by a less accurate method than the preceding 

 ones. It is concluded, therefore, that aside from 1899 and 1900 

 the curves represent the rainfall and runoff of the Hudson and 

 Genesee rivers with considerable accuracy. 



Fig. 12 shows in a similar manner a runoff diagram for Mus- 

 kingum river from 1888 to 1895, inclusive. 



3 Trans. Am. Soc. C. E., Vol. XXXV, p. 241. 



