140 ANNUAL OF SCIENTIFIC DISCOVERY. 



** With this view, fresh liquid must continually be brought into 

 contact with the sides by uniform agitation. For this purpose, a 

 basket, of metal gauze, formed of 2 concentric tubes, is im- 

 mersed in the calorimeter. A small machine raises it and lowers 

 it at equal intervals ; a thermometer marking the hundredth of a 

 degree is immersed in the central tube ; it is fixed, and is read 

 with a telescope. When the specific heat of solids is to be 

 measured, they are placed in the basket in the water. 



" This constitutes the entire apparatus ; the operation is one of 

 extreme simplicity. After pouring into the calorimeter the weight 

 of liquid which is to be investigated, and agitating it for some 

 time, the variation (if any) of the thermometer is observed for 

 5 minutes. Generally it does not vary. A current of measured 

 intensity is then made to pass during 1, 2, etc., minutes, until an 

 elevation of 3 or 4 degrees is produced ; this is noted, after which 

 the cooling of the thermometer is observed during 5 minutes. 

 The quantity of heat given off is known, the effect it has produced 

 is calculated, and from the known formulas of calorimetry the 

 desired capacity is deduced. Thus, suppose two experiments to 

 be made with the same current during the same time, with the 

 weights P and P' of water and of the liquid to be studied. The 

 quantities of heat are the- same ; they have heated the liquids 

 and 0' degrees. Denoting the weight of the calorimeter reduced 

 to water equivalents by /r, and the capacity sought by x, we have 

 (/> -}- TT) = (P f x -f- ;:) (i f , from which x can be calculated. 



"The old method required two operations, which were, the 

 first, to heat in a stove for a long time the body to be studied, and 

 to pour it, with minute precautions, into the calorimeter; the 

 second, to observe the thermometer immersed in the calorimeter. 

 In the method which I propose the first operation is omitted, and 

 the second suffices such as it was before. The corrections remain 

 the same, but are simplified. 



" They are simplified, because a lower temperature is sufficient, 

 and because the heat given off being proportional to the time, 

 the method known as Rumford's is applicable. We may even 

 dispense with all correction, as I shall show. 



"I provided the external envelope of the apparatus with a 

 spiral 20 times as long as the first, and immersed the whole in a 

 vessel containing 20 times as much liquid as the calorimeter, and 

 forming a medium in which the latter is immersed. The current 

 passes simultaneously into the 2 spirals ; it produces there heats 

 proportioned to the quantities of liquids, and consequently equal 

 heatings. At each moment the temperatures of the calorimeter 

 and its surroundings are in equilibrium, and the first, neither los- 

 ing nor gaining anything by radiation, is subject only to the action 

 of the current. It is impossible to maintain this equilibrium 

 strictly during the whole time of the experiments if they are pro- 

 longed ; but it is very easy to establish it within a few tenths ; and 

 this is sufficient to obviate all necessity for correction. Thus we 

 can measure for each degree the specific heat of a liquid (water or 

 alcohol, for example) from the lowest temperatures to its boiling- 

 point. 



