THE GULF STREAM. 211 



diminished through increase of saltness ; still it is an 

 important motive force. A large portion of the evapo- 

 rated water is also precipitated over the equatorial 

 regions in the form of rain ; yet that a very large por- 

 tion is carried away from equatorial and tropical to 

 temperate zones is beyond dispute. 



But now, how are we to get over the arguments by 

 which Herschel seeks to show that the buoyant water 

 will not rapidly move off, and that the effect of evapo- 

 ration is merely to produce opposing inrushes of water 

 which destroy each other's effect ? Easily, I take it, if 

 we remember that the buoyancy of the water does pro- 

 duce a surface-flow from the equator, however slight, 

 and that this is sufficient to destroy the balance of 

 forces which might otherwise make it doubtful whether 

 the place of the evaporated water would be supplied 

 from below or from above. I apprehend that there is 

 a continual under-flow of cooler water, rushing in 

 towards the equator on both sides, to supply the place 

 of the water evaporated by the sun's heat. Now there 

 can be no question that under-currents arriving in this 

 manner, whether from the north or from the south, 

 would acquire a strong westerly motion (just as the 

 trade-winds do). Thus they would generate from below 

 the great equatorial westerly current. In this up-flow 

 of cool currents having a strong westerly motion, I find 

 the mainspring of the series of motions. The water 

 thus pouring in towards the equator is withdrawn from 

 beneath the temperate and arctic zones, so that room is 

 continually being made for that north-easterly surface- 



p 2 



