46 THE PHYSICAL GEOGRAPHY OF THE SEA. 



trough of the Gulf Stream, therefore, may be supposed to waver 

 about in the ocean not unlike a pennon in the breeze. Its head 

 is confined between the shoals of the Bahamas and the Carolinas ; 

 but that part of it which stretches over toward the Grand Banks 

 of Newfoundland is, as the temperature of the waters of the ocean 

 changes, first pressed down toward the south, and then again up 

 toward the north, according to the season of the year. 



65. To appreciate the extent of the force by which it is so press- 

 ed, let us imagine the waters of the Gulf Stream to extend all the 

 way to the bottom of the sea, so as completely to separate, by an 

 impenetrable liquid wall, if you please, the waters of the ocean on 

 the right from the waters in the ocean on the left of the stream. 

 It is the height of summer : the waters of the sea on either hand 

 are for the most part in a liquid state, and the Gulf Stream, let it 

 be supposed, has assumed a normal condition between the two di- 

 visions, adjusting itself to the pressure on either side so as to bal- 

 ance them exactly and be in equilibrium. Now, again, it is the 

 dead of winter, and the temperature of the waters over an area of 

 millions pf square miles in the North Atlantic has been changed 

 many degrees, and this change of temperature has been followed 

 by a change in the specific gravity of those waters, amounting, no 

 doubt, in the aggregate, to nxany hundred millions of tons, over 

 the whole ocean ; for sea water, unlike fresh (§ 31), contracts to 

 freezing. Now is it probable that, in passing from their summer 

 to their winter temperature, the sea waters to the right of the Gulf 

 Stream should change their specific gravity exactly as much in 

 the aggregate as do the waters in the whole ocean to the left of 

 it ? If not, the difference must be compensated by some means. 

 Sparks are not more prone to fly upward, nor water to seek its 

 level, than Nature is sure with her efforts to restore equilibrium 

 in both sea and air whenever, wherever, and by whatever it be 

 disturbed. Therefore, though the waters of the Gulf vStream do 

 not extend to the bottom, and though they be not impenetrable 

 to the waters on either hand, yet, seeing that they have a waste 

 of waters on the right and a waste of waters on the left, to which 

 (§ 2) they offer a sort of resisting permeability, we are enabled to 

 comprehend how the waters on either hand, as their specific grav- 



