60 VOYAGE TO THE POLAR SEA. JULY 



the neighbourhood of H.M.S. " Alert's" winter-quarters, 

 and four sketched on sledge journeys, all showing 

 stratification. The lower part of the floes did not 

 exhibit stratification, and consequently a few apparent 

 exceptions occurred in overturned or much tilted floe- 

 bergs. Some authorities, such as Wrangell (" Wrangell," 

 edited by Sabine, appendix) and Belcher (" Last of the 

 Arctic Voyages," p. 101) have attributed the thickness 

 and the stratification of ice seen by them to the sliding 

 up of one floe over another ; but in our ice, the extent 

 and evenness of the stratification, and the invariable 

 progressive reduction in the depth of the strata from 

 above downward to their final disappearance below 

 precisely as in glacier neve, cannot be thus accounted 

 for. 



' The saltness of the Polar floes, notwithstanding 

 the (I think) irresistible evidence of their growth by 

 annual snow-fall, is to be accounted for by infiltration 

 and freezing of sea water as the spongy snow-ice sinks 

 season by season, and to a very large extent by the. 

 rapid diffusion of briny efflorescence from frozen sea- 

 water crushed up in cracks. We often had uncomfort- 

 able evidence of this diffusion in our sledging tea. 



' In April and May the passage of snow into ice was 

 experimentally determined to take place through the 

 growth of the deeper, and therefore colder, crystals 

 at the expense of the superficial. Later on an inverse 

 process helps the wind to harden the surface snow into 

 a layer which remains distinct from succeeding 

 snowfalls. 



' The birthplace and nursery of Polar floes is not, 

 in my opinion, near land, because in our experience 



