VEGA-EXPEDITIONEN S VETENSKAPLIGA AliBETEK. 



283 



II 



III 



The different behavior of these samples of ice before melt- 

 ing is obvious, and vet from an analytical point of view 

 they must be regarded as pure water, since no impurity can 

 be detected by chemical agencies in I & II and only slight 

 traces- of chlorine etc. . . . were found in III. 



If we admit, that minimal quantities of foreign substances 

 are the real cause of the peculiar behavior of the ice-samples 

 I, II and III, we will be justilied in concluding, that all kinds 

 of ice or snow occurring in nature will show this irregularity 

 in a still higher degree. The purest ice, that covers a high- 

 land lake in winter, will scarcely be as free from impurities 

 as ice from destilled water (Number III) but will probably 

 contain a minimal quantity of calcium-salts, sulphates etc, 

 which must occasion a greater irregularit}^ of expansion than 

 that shown in table III. A transparent block of bluish gla- 

 ■cier-ice will hardly surpass III in purity and in all cases be 

 inferior to II. Consequently we must acknowledge the fact, 

 that the ice-masses of the glaciers are liable to contraction of volinne 

 at temperatures heloiv their melting point, a circumstance, which 

 henceforth can not be left without consideration in the theory 

 of glaciers. 



Although the contraction of volume, which precedes the 

 melting of the ice, mav be new to science, there is no lack of 



