Beport of the Researches of M. Ayassiz. 151 



ed an account of the placing of the thermometers, in Jame- 

 son's Journal, vol. xxxiii. p. 277. 



Made wise by previous experience,* we determined to be 

 more careful in our use of means for withdrawing the second 

 thermometrograph, which was twelve feet beneath the first. 

 Having noticed the quantity of ice which daily disappeared 

 from the surface, M. Agassiz determined to postpone any at- 

 tempt, until near the termination of our stay. In fact, when 

 we again thought of the undertaking, three weeks later, the 

 surface of the glacier had experienced a sinking of many feet of 

 ice, and our thermometrograph was so much nearer the sur- 

 face. We now set to work, using the most cautious and sure 

 means, and carefully superintending the workmen who were 

 engaged in the service. M. F. Keller of Zurich, who at the 

 time was of the party, imagined, that by heating iron bars, 

 which might be introduced into the tin case, we might 

 possibly easily melt the cylinder of ice. Two iron bars, 

 therefore, an inch in diameter, were at the same time in- 

 troduced into the tube, and around them was constructed, by 

 means of large slabs, a kind of fire-place, in which a fire was 

 lighted. The two bars were thus in the midst of a brazier ; 

 but although they were thereby heated to a very considerable 

 temperature, it was found that they sank exceedingly slowly, 

 so that at the close of the day, they had not melted even 

 two inches of ice. This proceeding, therefore, was much too 

 tedious, as well as very costly, from the quantity of wood which 

 was required. We were forced, then, to return to the use of 

 boiling water; and with this, it required no less than five 

 days to melt the ice in the ice-cylinder, until we reached 

 the thermometrograph. At last, however, we succeeded in 

 detaching it, and it was with the keenest impatience that 

 we waited for its being brought up. Our satisfaction was 

 great when it reached us unhurt. The index was at — 0°.3. C. 

 (31°. 46 F.) We immediately set to work to verify the zero 

 point of our instrument, which we found perfectly correct. 

 It is worthy of remark, that this indication — 0°.3, corre- 

 sponds to the lowest which was given during the observa- 



* The experiment with the first thermometrograph failed, as stated in this 

 Journal, vol. xxxiii. p. 402. 



