88 DR ANDREWS ON THE HEAT DEVELOPED IN THE 



heating and cooling action of the surrounding medium were determined with 

 great care. The mechanical process of adding the acid to the alkaline solution 

 produced no change of temperature, and as the heat disengaged in the com- 

 bination raised the liquid almost instantly to the maximum temperature, the 

 whole correction required was for cooling. The first temperature was read one 

 minute after the addition of the acid to the alkaline solution, the mixture being 

 stirred during the whole of that time. If 8 represents the correction, and e the 

 excess of temperature above the air in centigrade degrees, the value of 8 will be 

 given by the following expression : — 



8 = e x 0°012 . 



As a proof of the accuracy of the method of mixture adopted in this inquiry, 

 I may mention that, being desirous to know whether the dilute acids em- 

 ployed in these experiments produced any change of temperature when mixed 

 with water, I made the experiment with nitric acid by the method just described, 

 substituting water for the alkaline solution, with the unexpected result of a fall 

 of 0°01. On varying the conditions of the observation, so as to obtain a larger 

 effect, it was ascertained not only that a diminution of temperature had actually 

 occurred, but that the observed fall represented approximately its true amount. 

 When hydrochloric acid of equivalent strength was diluted to the same extent, 

 an elevation of temperature of 0°05 was produced. 



The accuracy of experiments of this kind, where the whole thermal effect 

 observed amounts only to 2° or 3°, depends greatly on the thermometer employed. 

 Unless its indications are perfectly trustworthy in every part of the scale, the 

 labour of the inquirer will only end in disappointment. I have therefore taken 

 every precaution to secure this important object. The tube of the thermometer 

 was calibrated and divided with care, according to an arbitrary scale, by means 

 of a dividing instrument contrived for the purpose, and provided with a short 

 screw of great accuracy made by Troughton & Simms. The divisions, etched 

 finely on the glass, corresponded to about o, 05 C, and the readings could be 

 made with certainty to less than 0°01. The division of the scale, corresponding 

 to 0°, was determined from time to time in the usual way ; and another point, 

 about 30° C, was fixed by comparison with four other thermometers similarly 

 constructed, whose scales extended from the freezing to the boiling point of water. 

 The readings of these four instruments, when reduced to degrees, rarely differed 

 from each other within the limits to which they could be read, or o, 02. The 

 reservoir of the thermometer used in these experiments was 75 millimetres 

 long, and, when immersed in the liquid, occupied nearly its entire depth. 



As some uncertainty always exists with regard to the thermal equivalent of 

 glass vessels, I made two sets of comparative experiments — one with a thickly 

 varnished copper vessel, and the other with a vessel of platinum. The mean 



