Dr Martens on the Glaciers of 'Spitsbergen. 171 



collects in them in the form of little reservoirs or streamlets. 

 By reason of its maximum of density, which is found to be be- 

 tween + 4° and + 4°. 5, the water, which contact with the ice 

 has brought to zero, reascends to the surface, to descend 

 again after it has acquired, under the influence of the air, a 

 temperature near to this maximum of density. Finally, the 

 freezing of this water, and the enlargement of volume accom- 

 panying its change of condition, which causes it to act as a 

 wedge, and thereby increase the fissure. I shall refer the na- 

 tural philosophers, who may raise objections to this theory, to 

 the profound discussion preserved by Gilbert in his Annalen.* 

 Nature of the Ice. — The ice of which the glaciers of Spitz- 

 bergen are composed resembles, in every respect, that of the 

 superior glaciers {Fime) of the Alps. On the surface it is 

 rough and unequal, never slippery. The crevices are often 

 tinted with the most beautiful azure blue, which is likewise 

 the case with the cavities in the floating pieces of ice proceed- 

 ing from them. The latter sometimes appear striated. When 

 examined closely, we perceive that this appearance is owing 

 to a multitude of small bubbles of air of an elongated and py- 

 riform shape. In some pieces these bubbles are rounded. Saus - 

 suret and Scoresbyj have recorded observations analogous to 

 these. This ice is not formed by the union of crystals or in- 

 timately united fragments, like that of the inferior glaciers of 

 Switzerland. MM. Hugi§ and Kaemtz,|| who have described 

 them very accurately, both insist on this peculiarity, that the 

 crystals diminish in size in proportion as we ascend a glacier. 

 It need not, therefore, excite surprise, if they do not exist in 

 the ice of Spitzbergen. At the same time, I have observed, both 

 at Bell Sound and Magdalena Bay, crystalline formations of a 

 very singular character. At Bell Sound, this was on the margin 

 of a small rivulet ; at Magdalena Bay, near a feeble stream- 

 let, between the glacier Pointe aux Tombeaux and the sea. 

 These two places were perfectly flat, and covered with a not 

 very thick layer of ice and snow. Here and there certain 

 places were to be seen, of several square metres in extent, in 



* Aunalen der Chyeik, t. l.xiv. p. I!!3. 1820. f L- c. § 525. % L. c. t. 1. p. 341. 

 § L. c. p. 341. II Lehrbiich der Meteorologie, t. ii. p. 160. 1832. 



