10 F. L. Ekman, 



so mucli in the warming of the water at the equator as in its cooling 

 at the poles. The warm surface current is not driven forward by a 

 (.wis a tergo^)^ but by a (.wis a fronte))^ namely by the suction that arises 

 from the sinking of the cold polar water. On all these points, I, as the 

 reader will already have discovered, entertain a different opinion. The 

 great surplus of pressure at the poles vanishes with the uniformity of 

 the sea-level, as soon as we suppose the strata of water arranged ac- 

 cording to their specific gravity. The warm external stream is really 

 driven by a cwis a tergoi), which exists at every point of its course, in 

 consequence namely of the differences of pressure, arising from the higher 

 level of the lighter water behind, whereby the equilibrium is everywhere 

 in the upper strata disturbed to a certain depth. The cold under-current 

 would arise from another similar cause, namely the differences of pres- 

 sure caused by constant efflux of the superficial water in warmer latitu- 

 des and afflux in colder, whereby the equilibrium between the water- 

 strata would be constantly disturbed down to the bottom. Lastly I ob- 

 serve that the oceanic circulation, which could be brought about by diffe- 

 rences of temperature only, would be far less powerful than, and of a 

 character wholly different from that, which the actual phœuomena of the 

 ocean present, because other entirely dissimilar and very powerful causes 

 contribute to their production, of which more hereafter. 



Chilling of water to a certain degree produces freezing. If we 

 calculate what change the volume of sea-water undergoes in cooling, so 

 that pure water is separated from it in the form of free moving ice '), 

 we find that the last result will be a small diminution of volume, and 

 that accordingly water must flow from somewhere or other to the spot, 

 where the freezing is going on. The formation of ice therefore causes 

 no interruption of the effect, which the cooling can produce on the mo- 

 tion of the water. 



') According to my experiments, the maximum density and freezing point of 



soa-water (when not overcliilled and at the usual barometric pressure) take place at 

 the following temperatures: 



Saltness Freezing Max. density. 



2.09 »/„ P.i 0. OO44 C. 



2.54 » 1-4 » 1.68 » 



3.00 » 1.7 » 2.71 » 



3.50 » 1.9 » 4.00 » 



