Dr. E. L. Moss. Observations on Arctic Sea-water and Ice. 545 



by extrapolation; but taking Hubbard's curve as an extreme, tbe 

 error possible in the small range of temperature involved cannot 

 extend beyond tbe fifth place of decimals. The uncertainty of the 

 process has, however, induced me to discard a column in which the 

 specific gravities were expressed at the natural temperatures. In both 

 tables Stampfer's coefficients have been used for the waters from 

 melted ice. 



But that an Oertling balance and a reserve of pure nitrate of silver 

 were supplied, the chlorine estimations must have been omitted, for 

 the standard silver solutions were destroyed by freezing — a disaster 

 which considerably curtailed the number of experiments. 



The water of the polar basin, to such depths as we reached, had 

 already acquired the low specific gravity characteristic of outflowing 

 polar currents. This low specific gravity was maintained during 

 winter and spring, after nine months' perpetual freezing. It is to be 

 remembered, however, that the samples were from no great depth, and 

 from a zone of the Polar Sea that annually receives, not only its own 

 precipitation, but also the precipitation of the neighbouring shores, 

 and that, too, at a comparatively high temperature. 



The channels between the Polar Sea and Smith's Sound contain two 

 strata of sea- water, not owing their temperatures to local causes — an 

 upper stratum of polar water overlies a warmer northward flowing 

 extension of the Atlantic. The specific gravity observations show that 

 the relative position of the layers is due to salinity influencing their 

 density more than temperature. To overlie the polar water, the 

 denser stratum would require a temperature above 10° C. The highest 

 deep temperature obtained was below the freezing point, but there is 

 sufficient range between the specific gravities of the deepest samples 

 obtained and of Atlantic water to permit warmer Atlantic water to 

 exist at a greater depth. 



In places where the deep water yet retains some temperature above 

 the freezing point, direct dilution at its own depth, with fresh water 

 from deep icebergs or subglacial streams, may so far reduce its salinity 

 as to let it carry warmth to the surface, and thus help to occasion the 

 well known "rottenness " of floes in the neighbourhood of icebergs and 

 glacier cliffs. 



In addition to low specific gravity and temperature, the waters of 

 outflowing polar currents possess a third characteristic in the disturbed 

 proportion between their chlorides and sulphates, pointed out by 

 Forchhammer.* His great range of observations gives an oceanic pro- 

 portion of 11*87 of sulphuric acid to 100 of chlorine ; while his polar 

 currents show an average increase in the proportion of sulphates 

 amounting to *52. The sealed samples of water brought home by me 



* " Phil. Trans.," 1865, Part I, p. 228. 



2 o 2 



