36 
THE PHYSICAL GEOGRAPHY OF THE SEA. 
Suppose the excess of precipitation in these extra-tropical regions 
of the sea amounts to but twelve inches, or even to but two, it is 
twelve inches or two inches, as the case may be, of fresh water 
added to the sea in those parts, and which, therefore, tends to 
lessen the specific gravity of sea water there to that extent ; and 
for the simple reason, that what is taken from one scale, by being 
put into the other, reduplicates the difference. 
35. Now, that we may form some idea as to the influence which 
the salts left by the vapor that the trade-winds, northeast and south- 
east, take up from sea water, is calculated to exert in creating 
currents, let us make a partial calculation to show how much salt 
this vapor held in solution before it was taken up, and, of course, 
while yet in the state of sea water. The northeast trade-wind 
regions of the Atlantic embrace an area of at least three million 
square miles ; and the yearly evaporation from it is 33), we 
will suppose, fifteen feet. The salt that is contained in a mass of 
sea water covering to the depth of fifteen feet an area of three 
million square miles in superficial extent, would be sufficient to 
cover the British islands to the depth of fourteen feet. As this 
water supplies the trade-winds with vapor, it therefore becomes 
Salter, and as it becomes Salter, the forces of aggregation among 
its particles are increased, as we may infer from the fact 27), 
that the waters of the Gulf Stream are reluctant to mix with those 
of the ocean. 
Now, whatever be the cause that enables these waters to re- 
main on the surface, whether it be from the fact just stated, and 
in consequence of which the waters of the Gulf Stream are held 
together in their channel ; or whether it be from the fact that the 
expansion from the heat of the torrid zone is sufficient to com- 
pensate for this increased saltness ; or whether it be from both 
of these influences together that these waters are kept on the sur- 
face, suffice it to say, we do know that they go into the Carib- 
bean Sea (§ 34) as a surface current. The trade-winds, by their 
constant force, may assist to skim them off" from the Atlantic, and 
push them along into the Caribbean Sea, whence, for causes un- 
known, they escape by the channel of the Gulf Stream in prefer- 
ence to any other. 
36. In the present state of our knowledge concerning this won- 
derful phenomenon — for the Gulf Stream is one of the most mar- 
