242 PEOGEESS IN CHEMISTEY. 



this case of 2 atoms, hydrogen chloride consists of 1 atom of each of 

 its constituent elements. The amount of that element therefore which 

 combines with 35.5 grains of chlorine ma}^ give the numerical value of 

 the atomic weight of the element if the compound contains 1 atom of 

 each element. In that case the formula of the above compound would 

 be ZnCl and the atomic weight of zinc 32.7; but if the formula is 

 ZnClg the atomic weight of zinc would be 32.7x2; if 7mC\, 32.7x3, 

 and so on. The .specific heat of metallic zinc enables this question to 

 be solved. For it has been found experimentall}^ to be about 0.01>5, and 

 6.2-^0.095 = 65.2, a close approximation to 32.7x2=65.4. The con- 

 clusion is therefore drawn that zinc chloride is composed of 1 atom of 

 zinc in combination with 2 atoms of chlorine, that the atomic weight 

 of zinc is 65.4, and that the molecular weight of zinc chloride is 

 65.4+ (35. 5 X 2) = 136.4. Inasmuch as the relative weight of a molecule 

 of hydrogen is 2 (that of an atom being 1) zinc chloride in the gaseous 

 state should be 136.4-^2 = 68.2 times that of hydrogen, measured at 

 the same temperature and pressure. This has been found experi- 

 mentally to be the case. 



The methods of determining the vapor densities, or relative weights 

 of vapors, are three in number. The tirst method, due to Dumas (1.S27), 

 consists in vaporizing the substance in question in a bulb of glass or 

 of porcelain, at a known temperature, closing the bulb while still hot 

 and weighing it after it is cold. Knowing the capacity of the ))ul)), 

 the weight of hydrogen necessary to fill it at the desired temperature 

 can be calculated and the density of the \'apor thus arrived at. A 

 second method was devised by Gay-Lussac and perfected l)y A. W. 

 Hofmann (1868), and a third, prefera])le for its simplicity and ease of 

 execution, is due to Victor Meyer (1881). 



In 1858, as already remarked, Cannizzarro showed the connection 

 between these known facts, and for the first time attention was called 

 to the true atomic weights, which were up to that time confused with 

 equivalents, or weights of elements required to replace one unit 

 weight of hydrogen. These were generally regarded as atomic 

 weights 1)}' Dalton and his contemporaries. 



EXCEPTIONS TO THE LAW. 



Some exceptions had been observed to the law of Dulong and Petit, 

 viz, ber3"llium or glucinium, an element occurring in emeralds; boron, 

 of which borax is a compound; silicon, the component of quartz and 

 flint, and carbon. It was found b}^ Weber that at high temperatui'cs 

 the specific heats of these elements are higher, and the atomic heats 

 approximate to the number of 6.2; but this behavior is not peculiar to 

 these elements, for it appears that the specific heat of all elements 

 increases with rise of temperature. 



A certain number of exceptions have also been noticed to the law of 

 Gaj'-Lussac, which may be fornndated: The molecular weight of a 



