120 



by the rains of the fall season (the replenisliinj,' jieriod). If, now, there 

 is a deficiency of rain in the replenishing,' season also, a greater or less 

 proportion of the rainfall of the winter and spring months must go to 

 fill up the ground, and the rnn-off of this jieriod will be correspondingly 

 decreased. The most nnfavorable condition, therefore, is a dry fall, 

 followed by a dry winter and si»ring. If, for example, such a fall as 

 that of 190S, with as low as li.2- inches of rain at several of the stations, 

 for the three months. September, October and November, should be com- 

 bined with such a winter and spring as tliat of 1900-1901 (a not impos- 

 sil)Ie contingency), the probable catch of water, on the basis of 50 per 

 cent, of the rainfall of the stor.-ige period, would be only ."• inches for 

 the entire year'. 



The available catch of water in a dry season, that is, one of 30 

 inches of rainfall, will be a considerably smaller proportion of the rain- 

 fall than the catch of a wet season. In the latter case the run-off may 

 lie from oO to GO per cent, of the rainfall, while in a dry season it is 

 likely to fall as low. in the region under consideration, as 25 per cent.. 

 ov even lower.- From these data it ai)i)ears that there will be years 



' It is not decmi'd necessary to enter here into tlie teehnieal discussion of the 

 relation of rainfall to run-olT. A very full discussion of these points may bo found 

 in the works of Vermeule and Kafter. cited aljove. Ordinarily, in this latitude the 

 run-off of the winter and si)ring months may vary from 50 to 75 per cent, of tlie 

 rainfall. For the remaining montlis of the year it will vary from 0.0 to 20 per 

 cent, of the rainfall. I'nfortunately there are no satisfactory run-off data for 

 the region. The gagings at Shoals from 1003 to 190G, inclusive, are the only ones 

 of a stream lying largely within tlie region under consideration. These indicate a 

 mean annual run-off of 12.53 inclies, wliich is about 30 per cent, of the rainfall of 

 tlic region for tlie same interval. (The mean annual rainfall for the nearest sta- 

 tion, Paoli, for this interval was 42.75 inches.) This interval includes two years 

 of less than 40 inches rainfall, namely, 1003, with 35.18 inches, and 1004, with 

 .'{0.00 inches. On the Musl<ingum Kiver in Ohio, a stream lying in a region of 

 similar topography and climatic conditions to tlio catcliment of the east fork of 

 White Hiver, and like the latter, mostly in the driftless area, the run-off has been 

 known to fall as low as 25 per cent, of the rainfall. 



- The run-off formube of Vermeule are of interest in this connection. While 

 designed to cover tlie conditions in New .Jersey and southeastern New York, they 

 are liasod on certain general considerations, such for cxanujle as mean annual tem- 

 I'erature, etc., which are applicable to other regions as well. Vermeulo's general 

 formula is: E=(l l-|-0.20 K) M, where E stands for annual evaporation, R for 

 rainfall, and M is a factor depending on mean annual temperature. Tlie values 

 of M are as follows for the mean annual temperatures noted in the present region: 

 52°, 1.14; 53". I.IS; 54°, 1.22; 55°, 1.2(;; 50 \ 1.30; and 57 \ 1.34. Thus for n 

 mean annual temperature of 52° tlie evaporation, with a rainfall of .30 Inches, 

 should be 22. 4(! inches, and this subtracted from the total rainfall would leave n 

 runoff for the year of 7.54 inches. For the hiirber temix-ratures the runoff would 

 lie lorr.'spoiKliiigly less, .'uiil iiiigbl, .•ucnrding lo the formula, fall as low as 2 

 inches. II is uiit probable, bowcNcr. thai it ever does fall as low as the, latter 

 llgure. 



