1 16 Prof. Regnault on the Specific Heat of some Simple Bodies. 



determinations (Phil. Mag. vol. xv. p. 1 15, and vol. xviii. p. 273), 

 gives cobalt a higher atomic weight. My determinations of the 

 specific heats favour the latter conclusion; but the differences in 

 question are so small that it is impossible to decide the question 

 by a determination of the specific heats, until the experiments 

 can be made with absolutely pure metals. 



Tungsten. 



The specific heat of tungsten was given in my first memoir. 

 The metal was obtained by calcining, at furnace-heat in a lined 

 crucible, oxide of tungsten previously reduced by hydrogen. It 

 w r as to be feared, however, that the metal had by this process 

 absorbed carbon and silicon. I found, in fact, the specific heat to 

 be 0-03636, which, multiplied by the atomic weight 1150, now 

 usually admitted for tungsten, gives 41*81, a number distinctly 

 too high for the law of specific heats. 



M. Rousseau placed at my disposal a large quantity of tungsten. 

 He had prepared it by reducing tungstic acid at a high and 

 steady long sustained temperature by means of hydrogen. This 

 tungsten is crystalline, but pulverulent. To determine its 

 specific heat, it was pressed in a circular brass vessel weighing 

 35& r *60, and the calorific equivalent of which was 3& r *3428. The 

 following are the results of the two experiments which have been 

 made upon this substance : — 



M . . . . 420& r '80 4258*05 



p . \ . . 3s r -3428 3* r *3428 



T . . . . 98°- 15 98°-16 



& .... 12°-32 12°-09 



A<9' . . . . 3°-5512 3°*5625 



A . . . . 422^30 422^30 



C . . . . 0-03358 003326 



Mean .... 003342. 



This specific heat, multiplied by 1150 0, the atomic weight of 

 tungsten according to M. Dumas' s last experiments, gives the 

 product 38*43, which is quite within the limits of the law of 

 specific heats. 



Silicon. 



Chemists are at present quite uncertain as to the formula 

 which should be assigned to silicic acid, and therefore as to the 

 true equivalent of silicon. The majority write the formula SiO 3 , 

 and the equivalent is then 266*7; others write it SiO 2 , which 

 puts the equivalent at 1778; lastly, the formula SiO has been 

 proposed, which makes the equivalent 88*9. 



It has hitherto been possible to form only a few definite com- 



