POSSIBLE CAUSES OF CIRCULATIOK 229 



and will continue as long as the pressure of. the fresh water column is not 

 in excess of that of the salt water. 



Under normal conditions, when the pressures are approximately in 

 equilibrium, a 50 per cent dilution of the salt water will result if the 

 passages are of the same size. It may be conceived, however, that the 

 fresh water channel is only half the size of that of the salt water, in 

 which case, although the pressures might be in equilibrium, the mixture 

 would no longer be composed of equal volumes of the two, but of f salt 

 water to -J fresh water. With a large fresh water passage and a small 

 salt water passage, the reverse would be true. 



The point of emergence has much to do with the nature of the circula- 

 tion. Strange as it may seem at first sight, an outlet above the level of 

 the sea is more favorable than one considerably below it. The conditions 

 may perhaps be made more clear by reference to figure 3. In this figure 

 the outlet is indicated as occurring at a point above the level of the sea 

 under conditions where the full benefit of the low weight of the diluted 

 column is obtained. If, instead of this arrangement, the outlet in the 

 same system were as low or lower than the bottom of the salt water inlet 

 column (S), the downward pressure of the sea on the outlet would be 

 -greater than that in the underground channel and no flow would result. 

 It is apparent, therefore, that, inasmuch as the weight of the fresh water 

 column (F) can not be greater than the salt water column (S) except 

 where the outlet is hydrostatically lower than either, the circulation sys- 

 tem at Cephalonia must in all probability terminate at or near sealevel. 



It has been suggested to the writer that an inflow of the Cephalonia 

 type should, under the dilution hypothesis, vary with seasonal changes 

 of the fresh water supply. In reality deep seated waters almost nowhere 

 show any material variation with seasons, the fluctuations being prac- 

 tically limited to the upper part of the unconfined surface body of 

 ground water. As long as the head of the salt water column equals or 

 exceeds that of the fresh, as it must in order that circulation may exist, 

 it will control the circulation and no variation will be noted. 



An alternative hypothesis which has also been suggested to the writer 

 assumes the influx of the salt water to be due to suction, the action being 

 analogous to the indraft of air produced by the water jet in the Eichards 

 suction apparatus of our chemical laboratories or to that produced by the 

 steam jet in boiler injectors. The former process is known to produce 

 a strong and continuous suction of air in some cases,* but the conditions 



* M. L. Puller : Indraft wells in southern Georgia. Science, new series, vol. 23, 1906, 

 p. 140. 



XIX — Bull. Geol. Soc. Am., Vol. 18, 1906 



