A CONTRIBUTION TO THE STUDY OF ICE-STRUCTURES ll 



of the new formation are much elongated in the vertical direction. When typically 

 formed in this way the prisms pass downwards, gradually increasing in size until they 

 merge with the prisms of the lake ice. In the upper portions a horizontal section 

 shows clear ice-kernels, the cross-sections of the prisms, separated by much white, 

 opaque, saccharoidal, and honeycombed interspaces {see Plate II, Fig. 1 ). The white 

 interspaces arise from the drawing apart of the grains to fuse with the prism kernels. 

 Successive horizontal sections below show less and less of the honeycombed interspaces, 

 eventually passing into clear ice with only occasional bubbles. The modus operandi 

 in the development of this structure is not as clear as it might be, and further information 

 upon the subject is wanting. Cross-sections of such surfaces often show radial patches 

 (cf. Plate II, Fig. 2). 



Abrasion by wind and snow during the winter produced on portions of the surface 

 of Blue Lake large horizontal polished areas of this kind of ice. An elegant figured 

 pattern was thus developed. Where the vertical prismatic arrangement appears, the 

 term "coralloidal," as used by David and Priestley, aptly describes it. In summer a 

 coralloidal surface tends to become exceedingly rough, for ablation proceeds very 

 rapidly in the spongy mass between the kernels, leaving the latter projecting in relief. 

 The more rapid ablation of the spongy matrix is owing to the fact that it offers less 

 bulk for evaporation, more surface relatively, and has a slightly lower melting-point 

 (for any traces of saline that may have been contained in the snow will be con- 

 centrated therein) than in the kernels. The individuals of all ice-formations originating 

 from neve, whether they be prismatic or not, always exhibit tortuous interlocking 

 boundaries. The normal prismatic ice of lakes is composed of individuals with straight 

 polygonal boundaries. 



The accompanying diagram (Fig. 3) illustrates the section laid bare by Priestley in 

 a shaft sunk in July near the centre of the southern half of Blue Lake. 



In descending order the types met with were as follows : 



(1 ) At the surface there was about one foot of hard snow with specific gravity of 

 0-46, compacted by the winter blizzards; actually a fine grained neve as the term is 

 used in Switzerland. 



(2) Three inches of a coarse crystalline spongy layer, developed by the recrystal- 

 lisation during the autumn of the bottom portions of the snow under the influence 

 of warmth below, conducted from the lake-ice, and increasing chill from the autumn 

 conditions above. 



(3) Six inches of coralloidal ice in which there were vertical prisms averaging an 

 inch in diameter at their base. These prisms frequently coalesced below in intricate 

 articulating boundaries. This formation doubtless originated from a fall of snow 

 during the summer, or possibly from a fall of the year before. 



(4) Eighteen inches of granular, crystalline ice loaded with bubbles. The individuals 

 were very regular in size, averaging a third of an inch in diameter. The texture was 

 hypidiomorphic granular. Occasionally a tendency to short prismatic structure was 

 revealed. Bubbles were distributed irregularly and abundantly throughout, those near 

 the top being largest. In the latter position they were noted to frequently attain a 

 diameter of a quarter of an inch. At a foot below the top of the band eighty-three 

 were counted in an area of seventy-five cubic centimetres, equal to an average of 

 1-1 to the cubic centimetre. At that spot the average diameter was a sixteenth of 

 an inch. 



It will be noted that this layer corresponds closely to the ice of old permanent 

 ice-slabs where there is little movement. It must have originated from the solidifica- 

 tion of snow-drifts accumulated upon the lake at least two summers before ; in all 

 probability even considerably earlier than that. 



