﻿Prof. R. Bunsen's Cakrimetric Researches. 171 



order to cause by its heating to 0° C. the formation of ice found 

 by the observations in the first seven hours. 



Since it can be concluded from the determination of specific 

 heats given below that a time of seven hours is more than suffi- 

 cient to equalize a difference of temperature of 7° C. in the instru- 

 ment^ the formation of ice, which (as is seen from Table I.) has 

 lasted more than 100 hours, must be ascribed to some other 

 cause than that considered. Without entering into the question 

 whether (as C. Schultz* assumes) this cause be due to the air con- 

 tained in the snow-water, or whether the transformation of snow 

 into Firneis play a part in it, let it suffice at present to consider 

 only some of the causes which produce a lowering of the melting- 

 point, which deserve particular consideration in using the ice- 

 calorimeter. If the pure snow surrounding the instrument be 

 soaked with just so much boiled water at 0° C, or distilled water 

 at the same temperature which has been first shaken up with air, 

 as will remain in it, no freezing will take place, at all events, 

 during the first twelve hours (the observations were continued no 

 longer) ; but such a melting of the ice will take place as to ren- 

 der the instrument under these circumstances totally useless for 

 observations. On the other hand, the slightest impurity in the 

 snow produces such a constant deposition of ice on the ice cy- 

 linder, as to cause the mercury-thread to move many divisions of 

 the scale in a minute. Snow which contains only traces of salts 

 from the ground, or which has absorbed animal or vegetable im- 

 purities from the street-pavement, shows this lowering of the 

 melting-point in the most striking manner. River-ice, so pure 

 that water melted from it produced scarcely a faint turbidity in 

 solutions of barium and silver, produced in the instrument in 

 three days 2 grms. of ice. After these observations it is clear 

 that only the purest snow can be used in these experiments. 

 It is also advantageous to experiment in a room whose tempera- 

 ture is not much above 0° C, and not to begin the observations 

 until the formation of ice on the ice cylinder does not cause an 

 alteration of more than a few divisions of the scale in an hour. 

 But, above all, it is necessary to see that, before the instrument 

 has assumed a constant temperature in the snow, a small layer 

 of water is formed by melting between the glass sides and the 

 adjacent cylinder of ice, in order to avoid unequal tension and 

 consequent elastic rebound (elastische Nachwirkung), 



The accuracy of the results, however, depends most essentially 

 on the care with which absorbed air is excluded from the water and 

 mercury when the instrument is first prepared. This is managed 

 in the following way. The instrument half-filled with boiled 

 water is inverted, mouth downwards, and secured with a clamp. 

 * PoggendorfF's Annalen, vol. cxxxvii. p. 253. 



