32 Recent Chemical Researches . [January, 
whose absorption and radiation spedtra are the same, and 
non-metals as substances in which these two spedtra differ. 
Furthermore, these researches introduce a new definition of 
the “ atom ” as being that mass of matter which gives us a 
line spedtrum, while the mass of matter (non-metallic) 
which gives us a channelled space spedtrum is called a 
“ sub-atom.” 
The plasticity of non-metallic molecules is a subjedt to 
which we shall have occasion to make further reference. 
II. Atomic Weights of the Elements. 
Of the various methods which may be employed for as- 
certaining the atomic weight of any element, that which is 
based upon the determination of the density of the element 
when in a state of vapour and the densities of the greatest 
possible number of gaseous compounds of that element 
holds the first place. By checking the results of such 
vapour density determinations by means of gravimetric 
analyses we are able to fix on the maximum atomic weight 
of the element. As a means of determining the atomic 
weight of those elements which do not assume the gaseous 
condition at any attainable temperature, the method of spe- 
cific heat has generally been adopted. This method is based 
upon the fadt that the average produdt obtained by multi- 
plying atomic weight into specific heat (called atomic heat) is, 
in the case of the solid elements, represented by the number 
6*3. The fadt that the atomic heats of the three solid 
elements, silicon, boron, and carbon, are represented by the 
numbers 4*8, 2*7, and i'8 respectively, has rendered the 
application of the method of specific heats to atomic weight 
determinations somewhat unreliable. If these three ele- 
ments deviate so widely from the general rule, what guarantee 
have we for believing that a similar deviation does not occur 
in the case of other elements the determination of the 
atomic weight of which is as yet an unsolved problem ? 
Prof. Weber, of Hohenheim, has made a number of ex- 
ceedingly accurate determinations of the specific heats of 
carbon, boron, and silicon, from which he arrives at the 
result that these elements obey the law of Dulong and Petit. 
On comparing the numbers given by different observers for 
the specific heats of the three bodies in question, it was 
ascertained that these numbers represented the specific 
heats for different intervals of temperature, and that the 
greater the interval of temperature for which the determi- 
nation was made the greater was the number representing 
the specific heat. 
