20 . F. L. Ekman, 



that the river-water, streaming onward from behind, could constantly 

 supply the loss, caused by the outflow, and preserve the level unchan- 

 ged. Thus the cause of the under-current, which in the former case 

 was occasioned b};' the reduced hydrostatic pressure of the superficial 

 water in consequence of the sinking of the level, does not in this case 

 exist. Hence the difference of specific weight between river-and sea- 

 water ought here only to produce differences of level, which also should 

 be constant as long as the supply of river-water is continued and con- 

 stant. I look upon the under-current, which, in these cases, very di- 

 stinctly appears, as a reaction-stream, to be accounted for on the gene- 

 ral principles, above laid down for the explanation of such streams. But 

 the same causes, which produce an under-current at the mouths of rivers, 

 must also give rise to an imder-current in the channels, whereby separate 

 ocean-basins discharge their superfluous fresher water into the ocean. 



This seems to be the place to say a few words on the reaction- 

 streams in lakes as compared with those in the ocean. Our knowledge 

 of the former is indeed verj'' scanty, nevertheless the general experience 

 we have of streams in lakes would seem to justify our assuming, that 

 the under-currents, produced by the. reaction of rivers in lakes, are far 

 inferior in intensity to those produced iji the sea. What then, it will 

 be asked, is the cause of this difference, if the under-current in the sea 

 is not caused by difference of pressure, depending on the different spe- 

 cific weight of the water? 



In the first place the changes of level, occasioned by the pouring 

 of fresh water into fresh, can only produce such distiirbances of equili- 

 brium as extend all the way to the bottom; but the movements, that re- 

 sult from this disturbances, ought not, as I have just said, to be follo- 

 wed in any great degree by any reaction-stream. Secondly, Avhen a 

 river falls into a lake, the molecules of river-water have no reason to 

 change the direction of their motion, unless obliged by the contour of 

 the shore. The river-water therefore proceeds as a distinct stream in 

 the upper water-strata, the particles of which it sets in motion b}- fric- 

 tion. But the reaction-stream, thus caused, must show itself in an afflux 

 of water from the sides as well as from below, and in niau}^ instances, 

 in consequence of the lake's form and depth, the requisite quantity of 

 water can most easily be brought to the stream in the first mentioned 

 manner. Accordingly the reaction-streams, resulting from rivers, will 

 in fresh-water lakes chiefly take the form of eddy-streams on each side 

 of the almost strait stream of river-water. 



