286 THE PHYSICAL GEOGRAPHY OF THE SEA. 



some great river, as the Amazon, or in the regions of constant 

 precipitation, or in other parts, as on the polar side of 40° in the 

 North Atlantic, where it rains more than it evaporates. Yet in 

 the case of the Red Sea, and all such natural salt-pans, as that and 

 other rainless portions of the sea may be called, there is, on ac- 

 count of currents which are continually bearing away the water 

 that has given off its vapors and bringing forward that which is 

 less concentrated as to brine, a moderate degree of saltness which 

 its waters can not exceed. We moreover find that, though the con- 

 stituents of sea water, like those of the atmosphere, are not for 

 every place invariably the same as to their proportions, yet they 

 are the same, or nearly the same, as to their character. When, 

 therefore, we take into consideration the fact that, as a general 

 rule, sea water is, with the exception above stated, every where 

 and always the same, and that it can only be made so by being 

 well shaken together, we find grounds on which to base the con- 

 jecture that the ocean has its system of circulation, which is well 

 calculated to excite our admiration, for it is as wonderful as the 

 circulation of the blood. 



464. In order to investigate the effect of the salts of the sea 

 Hypotheses, upou its currcuts, and to catch a glimpse of the 

 laws by which the circulation of its waters is governed, hypothe- 

 sis, in the present meagre state of absolute knowledge with regard 

 to the subject, seems to be as necessary to progress as is a corner- 

 stone to a building. To make progress with such investigations, 

 we want something to build upon. In the absence of facts, we 

 are sometimes permitted to suppose them ; only, in supposing 

 them, we should take not only the possible, but the probable; 

 and in making the selection of the various hypotheses which are 

 suggested, we are bound to prefer that one by which the greatest 

 number of phenomena can be reconciled. When we have found, 

 tried, and offered such a one, we are entitled to claim for it a re- 

 spectful consideration at least, until we discover it leading us into 

 some palpable absurdity, or until some other hypothesis be sug- 

 gested which will account equally as well, but for a greater num- 

 ber of phenomena. Then, as honest searchers after truth, we 

 should be ready to give up the former, adopt the latter, and hold 

 it until some other better than either of the two be offered. With 

 this understanding, I venture to offer an hypothesis with regard 



