282 R Bunsen — Calonmetric Investigations. 



the capillary resistance is overcome, and tlie mercury thread 

 does not retreat by further rapping. 



Table 1 shows already that the mercury thread of the instru- 

 ment is generally not perfectly stationary. The displacement, 

 which, in a positive as in a negative sense may amount to 1 to 

 3 divisions in the hour, is, as the observer may easily perceive 

 by the employment of the calorimeter, nearly proportional to 

 the time. The slight error thus introduced may be eliminated 

 in the following manner : As soon as it is seen that the instru- 

 ment has become sufficiently stationary, the height of the 

 mercury thread is noted from 30 to 30 minutes. If the dis- 

 placement of the latter amounts in m minutes to Tq scale divi- 

 sions, then the displacement to be ascribed to foreign influ- 

 ences is for one minute >. 



The hour Mo and the height of the mercury thread Qo are 

 now observed at the moment, when the substance to be mvesti- 

 gated is allowed to fall from the heating vessel / fig. 4, (see 

 page 173) into the calorimeter vessel a, fig. 1, and both obsem- 

 tions are repeated one hour later, whereby further Mi and (ji 

 result ; finally the movement of the mercury thread independeut 

 of the heat to be measured, is once more determined, as in the 

 beginning of the experiment — . 



The mean displacement of the mercury thread, independent 

 of the experiment, amounts therefore in one minute to 



and throughout the entire duration of the experiment to 



(M,-M„)i(-^ + ^) 

 scale divisions. This value is to be added to the oscillation o 

 the mercury thread Qo-Qi observed during the experiment a. 

 a correction, and especially, with the negative sign pren^ei 

 when the displacement independent of the experiment toc^^ 

 place in the way of ice melting, with the positive sign m | 

 opposite case. For the oscillation T, corresponding to tbe qu ' 

 tit^y of heat to be measured, the equation 



T=(Qo-QO+ (M -Mo) i(^ + -^) 

 is therefore obtained, wherein it is hardly necessary to mention- 

 that, for the direct readings, their values according to 

 table of calibrations are to be substituted. 



