CONTINENTAL TERRACES AND SUBMARINE VALLEYS I47 



averages only i to iooo, the velocity of the bottom current has 

 registered as much as 1.5 miles per hour. The high speed is 

 explained by the unusual muddiness of the Puerco River at 

 time of flood. After the entry of its water into the reservoir, a 

 cubic foot of this water weighs two to three pounds more than 

 a cubic foot of clear water at the same temperature. Another 

 important observation: the bottom current was actually six de- 

 grees Fahrenheit warmer than the overlying water, so that 

 there can be no thought of explaining the flow by lower tem- 

 perature and resulting smaller degree of thermal expansion. 17 



Bottom currents, similarly motivated by suspended silt, have 

 been demonstrated in four other American reservoirs: Lake 

 Lee (North Carolina), Lake Murray and Saluda Reservoir 

 (South Carolina), and San Carlos Reservoir (Arizona). The 

 Government engineers have thus good ground for the con- 

 clusion that underflow of the kind described is a "general phe- 

 nomenon in reservoirs" subject to inflow of muddy water on 

 the grand scale. The silty water slides down the floor of a 

 reservoir "much as water itself flows under air." 



The studies at Lake Geneva, Lake Constance, and the six 

 artificial lakes in the United States have established two funda- 

 mental facts: first, a silty underflow persists, keeping much of 

 its initial velocity for many hours and even as much as seven 

 days; second, the speed of flow can exceed one mile an hour 

 although the slope traversed is only 1 to 1000, and although the 

 thickness of the bottom current is only a few feet. Since the 

 continental slope is about 1 to 15, and since the thickness of 

 the silty current of Glacial times must have been at least 100 

 feet and probably more, we already have from analogies in 

 Nature some good quantitative support for the preferred expla- 

 nation of the submarine canyons and furrows. But other quan- 

 titative criteria as to the worth of the explanation should be 

 found if properly controlled experiments are made in the 



