40 On the Relation of the Specific [July 



of the constituent number, according to the law as above 

 stated by theconstant number, and then divide by the weight 

 ofthe compound atom; then 1-225 x -1875— '0525 =0-40846. 

 = the specific heat of water in the solid state or ice/ The 

 specific heat of ice, according to Clement and Desormes, is 

 •72, according to Avogrado, -92. We arrive at a mean of 

 these numbers, by supposing that the atom of the vapour of 

 water is divided into four parts in passing to the solid state ; 

 for then the constituent number will be four times less, and 

 its square root the half of the preceding, while the com- 

 pound atom which enters as the divisor into the expression of 

 the specific heat, will become four times less. Hence the cal- 

 culated specific heat will be 0*8168 or a mean of the two 

 experiments mentioned ; for -92 + 72 -r 2 = -82. If how- 

 ever we calculate the specific heat, according to the view 

 taken of the composition of water in this country, we have 

 then, 2 as the constituent number, by dividing the square 

 root of which by -5625 half the atom of water, we obtain 

 •94 for the specific heat, a close approximation to the de- 

 termination of Crawford and Avogrado. If in place of * 1 875, 

 we had adopted the number of Dulong and Petit, viz., -375; 

 the calculated specific heat of ice would have been -8168 ; 

 but the necessity of admitting a difference in the constitu- 

 tion of the atom of water in the three states in which it is 

 found, is sufficiently indicated independently of all system, 

 by comparing the specific heat of equal weights of water, 

 vapour and ice. For the specific heat of vapour ought to be 

 1*225, taking that of an equal and constant volume of air 

 as unity ; and calculating from this, and taking as unity 

 that of an equal weight of water, we have -3753 according 

 to Delaroche and Berard, while the specific heat is 1, and 

 that of ice is *92. 



2. Oxides — (1). In this class, according to the theoreti- 

 cal considerations which we have been discussing, it is 

 necessary to view those oxides consisting of one atom base 

 and one atom oxygen, as being formed of J atom oxygen 

 and 1 atom of base ; but in order to correspond with the 

 law already announced, the atom in the solid state will 

 consist of \ atom base and J atom oxygen. Thus, in the case 

 ofthe protoxide of lead, the weight ofthe atom of lead being 

 6*5 or half of that commonly adopted ; taking oxygen as unity, 

 the atom ofthe protoxide will be 3*25 -t- 0*25 = 3*5 ; its con- 



