GREAT SIZE OF UNDERGROUND STREAMS. 585 



in southern California to be four times as great as the most rapid rate in 

 the Arkansas sands — that is, 20 meters a day — the overground rate would 

 be 12,000 times as fast as the underground rate. When it is remembered 

 that underground about three-fourths of the space is occupied by sand 

 and gravel, and that the openings present are more or less irregular, at 

 the maximum the available openings are not more than one-fourth the 

 entire space. It follows that, in order to accommodate the underground 

 water, the channel .would have to be 48,000 times larger than at the 

 surface. This number may be an overestimate, but certainly an under- 

 ground channel through sand and gravel to accommodate a certain amount 

 of water must be many thousand times larger than an overground channel 

 with the same slopes which accommodates the same amount of water. 



While many streams in the semiarid and arid regions normally have 

 an underground course as above described, at occasional times of flood, 

 which may be for a brief season each year or may occur only once in a 

 number of seasons, the mountain streams are of such increased volume- 

 that the underground circulation is not sufficient to' dispose of the water, in 

 which case there is for a time also an overground circulation, which usually 

 follows .approximately the same general course as the ground waters. It 

 is at these times of flood that the fans of the wash are built up and the 

 overground channels determined. Such circulations are ideally illustrated 

 by the Santa Ana, San Gabriel, San Antonio, and other streams of southern 

 California. 



Excellent illustrations of underground flowage are furnished by 

 artesian systems. For instance, in Wisconsin the Potsdam sandstone is an 

 artesian water-bearing stratum. This sandstone is 200 to 250 meters thick. 

 The annual precipitation in the district of its outcrop is approximately 50 

 centimeters. Supposing- that one-half the rainfall enters the sea of ground 

 water in the sandstone, this would give an additional increment every 

 year of 25 centimeters. If the pore space of the sandstone be supposed 

 to be 16f per cent, in order that the increment added shall find space in 

 the sandstone without raising the water level it is necessary to suppose 

 that the water of the previous year shall have moved downward 150 

 centimeters. In other words, the vertical flowage per annum is only 

 150 centimeters, or 1£ meters. Suppose the dip of the sandstone to be 2 

 meters per kilometer ; it would follow that the lateral movement of the 



