﻿and Expansions of some Liquids. 139 



I do not think the error of these determinations can exceed 

 0-002, including that which may depend on the preparation of 

 the solutions, to which, however, I devoted the greatest care. 



Having given my reasons for thinking my method very accu- 

 rate as well as very simple, I am bound to mention a grave in- 

 convenience attending it, which I do not regard as insurmount- 

 able, but which I have not yet succeeded in removing, and have 

 consequently been for some time hindered in the application of 

 the process. 



The construction of thermometers with reservoir in the form 

 of a very close helix is attended with great difficulties; and 

 the clever maker who supplied me with them appears to have 

 been obliged to take for the purpose tubes of rather thick glass. 

 It results from this thickness, and the torsion they have under- 

 gone, that they are extremely apt to break when subjected to 

 the abrupt changes of temperature produced by the sudden im- 

 mersion, although these changes have never exceeded 45°. None 

 of them has been capable of withstanding a very long series of 

 experiments. Yet I think it would be possible to make them of 

 thinner glass, so that they could support these abrupt transitions. 



Wishing to determine the specific heats of some solutions in 

 sulphide of carbon, I was obliged to abandon the preceding me- 

 thod, which was unsuitable for so volatile a liquid. In this case 

 I operated in the following manner: — 



The solution is contained in a glass balloon of 40 cubic centims. 

 capacity. A very sensitive thermometer indicates the tempera- 

 ture. The neck of the balloon is sufficiently narrow to be nearly 

 filled by the stem of the thermometer ; a thin tube of caoutchouc, 

 squeezed between the two, closes it exactly. 



The liquid having been heated to a temperature a little above 

 that at which the experiment is to commence, is cooled by con- 

 tinual agitation ; and the moment the thermometer marks the 

 degree determined, the whole is immersed in the calorimeter, 

 which is arranged as in the preceding experiments and contains 

 from 140 to 150 grms. of water. Of this a second thermometer 

 indicates the initial temperature (which should be quite station- 

 ary under the influence of exterior causes), and its rise in conse- 

 quence of the heat communicated by the heated solution. Care 

 is also taken to agitate the balloon in the calorimeter continually 

 till the maximum is reached. 



Further, all correction is dispensed with, as before, by com- 

 parison with other experiments, under the same conditions, made 

 with a liquid of known specific heat. 



This process does not yield such accurate results as the pre- 

 ceding, on account of the greater duration of the experiments ; 

 for the maximum, instead of being attained in 30 seconds, is 



