﻿172 Prof. R. Bunsen's Calorimetric Researches. 



The mouth of the tube c, fig. 1, which is not yet provided with the 

 iron collar d, is then plunged in a beaker containing water kept 

 boiling ; and the water in the instrument, reaching to the level /3, 

 is boiled down to one-third of its original bulk. On removing the 

 lamp used in heating the apparatus, it fills itself with water free 

 from air. It is then allowed to cool, placed upright as in the 

 drawing, and filled up to the level ft with so much freshly boiled 

 mercury that the level of mercury in the vessel b and in the 

 tube c is nearly the same height. The water is now for the 

 most part removed from the tube c by means of a siphon, and 

 the tube freed from moisture by means of a dry current of air 

 produced by the water-pump; and now, for the first time, the 

 iron collar d is fastened on with the finest sealing-wax so as to 

 leave a small part of the tube c projecting above the inner 

 bottom of d, so that the cork of the scale may afterwards be 

 placed in the tube c itself and not in the iron collar. The final 

 filling-in of boiled mercury to the level y is done with a hollow 

 capillary glass tube, so as to avoid any air-bubbles remaining on 

 the sides of the tube. 



In order to place the mercury-thread at the same starting- 

 point on the scale in each experiment, it suffices to press the 

 cork of the scale- tube a little deeper into the mouth of the 

 mercury-tube c, fig. 1. If in doing this the mercury-thread 

 goes beyond the desired mark, a brass weight tied to a thread is 

 warmed in the hand or under the tongue and dropped into the 

 fluid a, fig. 1. Suppose the brass weight weigh g grms., and 

 suppose its temperature is denoted by t, its specific heat by s mj 

 the latent heat of water by /, and if we denote by p the weight 

 of melted ice (found by equation 2) which the indication of one 

 scale- division represents, then the variation caused by the brass 

 weight will be 



Smtg 

 Ip 



divisions of the scale. 

 If we take 



t=S7° C, 



^=0-0939, 

 /= 80-03, 

 ^=0-000853, 



and for g the series 0*1, 0*2, 0*4, 0*6 .... grm., then we ob- 

 tain for these weights in round numbers the following retro- 

 gression of the mercury-thread : — 



