POSSIBLE CAUSES OF CIRCULATION 225 



or below the sea, must be, from the hydrostatic standpoint, at a greater 

 elevation, since the head at the mills is 3 feet below sealevel. 



EFFECT OF TEMPERATURE 



In order that circulation may be established through the agency of the 

 interior temperature, unequal heating of the water in the circulatory 

 system must take place. This might result from a relatively strong 

 application of heat at a particular point, as in the vicinity of an igneous 

 intrusion, or, on the other hand, it might be due to the action of uniform 

 heat on an unsymmetrical passage, as in figure 1, in which the dotted 

 Ime represents a horizontal isogeotherm. In igneous intrusion we have a 

 possible cause, but of its actual existence there is no positive evidence. 

 The horizontal isotherm, on the contrary, is the normal condition. It is 

 evident that with the arrangement shown in the figure the water in the 



Figure 1. — Conditions of Thermal Circulation (Crosby and Crosby). 



long, gently inclined arm is much more exposed to the action of high tem- 

 peratures than that in either of the more nearly vertical arms, the result 

 being to decrease its density until movement is set up by virtue of the ex- 

 cess of pressure of the colder and heavier water in the left arm. As is 

 pointed out by Crosby and Crosby, this is certainly not an improbable ar- 

 rangement and provides the mechanism for a strong circulation of water. 



EFFECT OF DILUTION 



Amount of dilution necessary. — To establish circulation without the 

 agency of heat, a slight dilution of the water is all that is necessary. 

 The specific gravity of the sea water of the Mediterranean is 1.03 ; hence 

 a column of salt water 100 feet in length would support a column of 

 fresh water 103 feet high. Of course, the existence of an absolutely 

 fresh column balanced against a purely salt column is out of the question 

 in a circulation system of the Cephalonia type ; but fortunately such con- 

 ditions are unnecessary, for even with a dilution of only 10 per cent a 

 depth of 100 feet would give a working head of .3 foot, which would be 

 suflBcient to establish the conditions for circulation. With a 25 per cent 



