POSSIBLE CAUSES OF ClECULATlON 22? 



been lelt in very recent years, and, although full data are not at hand, it 

 is more than probable that fractures have been produced by these or 

 similar shocks in the past. These can not fail to be of importance in 

 connecting the solution passages with one another. The intimate con- 

 nection of the passages, making to all practical purposes a network, has 

 been brought out at several points in this country by the experiments 

 made for the United States Geological Survey by S. W. McCallie at 

 Quitman, Georgia; by E. H. Sellards at Ocala, Florida, and by G. C. 

 Matson at Georgetown, Kentucky, at each of which localities salt in- 

 serted into sinks or borings found entrance into wells some distance 

 away. In none of the instances, however, was the movement direct from 

 the point of insertion to the well, for the salinity, instead of increasing 

 enormously, as it would have done if such had been the case, showed only 

 relatively moderate fluctuations. The three limestones, although of 

 widely different types, showed the same phenomena in each case, suggest- 

 ing that it is a normal characteristic of this class of rocks. 



If such a network of passages exists at Argostoli, as in all probability 

 it does, it follows that it would be practically impossible for a single 

 passage bearing the water to maintain an independent existence to any 

 considerable depth. The encountering of fresh water passages under 

 such conditions and a mixing of the waters seems inevitable. 



Conditions of dilution. — When the salt and fresh waters meet, one or 

 the other of two things will happen : Either the salt water will be forced 

 upward by a longer column of fresh water or the fresh water will be 

 forced back by the salt. Which of the two occurs will depend on the 

 relative heads of the salt and fresh waters. In this connection it should 

 be noted that the phenomena of the sea mills is one occurring at the 

 borders of the sea toward which the underground waters are moving and 

 which marks their points of escape. From this it follows, in accordance 

 with well known laws of artesian circulation, that the head of the fresh 

 water will be very low at this point, even if the original source is in re- 

 gions of considerable elevation. Under such conditions and with the 

 juncture of the fresh and salt water passages occurring at a depth of a 

 few hundred feet, the weight of the salt water raay easily be such as to 

 counterbalance the head of the fresh water column. 



The supposed conditions of circulation can be better shown by the aid 

 of the following diagram (figure 2) than by extended description. In 

 the figure SM shows the situation of the sea mills; P, the trunk subter- 

 ranean channel fed by crevices ; the vertical lines at P, the well at Argos- 

 toli ; M, the mainland of the island of Cephalonia from which the water 



