VHGA-EXrEDlTIONKNS V KT KN S K A PLIG A AliBKTEN, 289 



point of the ice sink O^io C below zero, while the presence of 

 a minimal quantity of impurity [tho iiiflucnce of whicli can 

 not 1)0 ignorod in ihat part of the clacier, whicli is in contact 

 vvith the rocky f^round and the dihris separated irom it by 

 erosion] is snflicient, to lower the meltinti; point considerably 

 and make the ice soften and shrink in volume long before. 

 It must, in short, be taken into consideration, that every kind 

 of ice in nature becomes plastic and tough in the vicinity of its 

 mélting point. ' 



I next proceeded to the determi nation of tlie volumes of 

 the liquid .water [sample I]. These nnmbers [see table I, Bj are 

 graphically represented by the red line in plate 23. The curve 

 is traced under the assumption of 



q = 0.00017905 

 g = 0.00002754. 



It will givc the reader an opportunity of comparing the 

 performance of the new dilatometer with the results obtained 

 formerly by means of otlier methods. 



CHAPTER 4. 



On ice and water of little saltness. 



From the reasons given in chapter 1, pages 250 — 253, I liave 

 ventured to draw the conclnsion, that the ice of the Siberian 

 sea is formed to a great part by the freczing of the brackish 



1 With this restrictiou I think wc may safely adopt tlie glacier theory 

 of Rendu or Forbes. 8ome important observations made on glaciers will 

 thereby be explained in the most natural manner, viz. the rclatively rapid 

 progress of the ice masses in summertime and the birth of the glacier rem- 

 nants, wliich, according to eye-witnesses, takes place most frequently in hot ^ 

 weather. Petzholdt in the work already alluded to in this chapter cites 

 some observations concerning this subject by Hugi, Agasslz, Vogt a. O. 

 In every case the breaking np of a rift in the glacier was observed on some 

 very hot suminer day. The rift was at first only a few inches deep, but 

 gradually incrcased, and the bursting of the ice was accompanied by a sharp 

 crack, like a peleton fire. From the behavior of the ice represented on plato 

 23 we might just cxpoct such eftects to take place. The upper laycr of the ice 

 at first expands very regulaiiy under the infiuence of the sunnncr heat, but, 

 as its temperature approaches the melting point, suddenly l)cgins to contract 

 its volume enormously. Tliis causes a härd strain upon its surface, which 

 ends with the breaking up of fissures and remnants. 



19 



