FORCHHAMMEKS ANALYSES. 411 



spreads over this part of the Atlantic Ocean ; 

 and thus we see that water of tropical currents 

 will keep its character even in high northern 

 latitudes. 



" The water of different seas is much more 

 uniform in its composition than is generally 

 believed. In that respect my analyses agree 

 with the newer analyses of atmospheric air, in 

 showing that the differences are very slight 

 indeed. Sea-water, it is true, may contain 

 more or less salt; from a very small quantity, 

 as in the 'interior part of the Baltic, to an 

 amount of 37*1 parts in 1000 parts, which I 

 found in water from Malta, and which is the 

 greatest quantity I ever observed; but the re- 

 lative proportion of its constituent saline parts 

 changes very little. In the longitude of Green- 

 land, and more than 100 miles to the south of 

 the southernmost point of that large tract of 

 land, sea-water contains only 35'0 in 1000 

 parts. In going from this point towards the 

 north-west it decreases constantly; and in 

 Davis' Straits, at a distance of about forty 

 miles from the land, it only contains 32'5 parts 

 of salt in 1000 parts of sea- water. This cha- 

 racter seems to remain in the current which 

 runs parallel to the shores of : North America; 

 and at N. lat. 431 and W. long. 46^ the sea- 

 water contained only 33'8 parts of salt. Thus 

 tropical and polar currents seem not only to be 



