1879.] Dr. Carnelley. On Melting and Boiling Points. 191 



of their atomic weights, whilst, for compounds, the following con- 

 clusions are drawn : — 



I. That the melting and boiling points and heats of formation of the 

 normal halogen compounds of the elements are a periodic function of 

 the atomic weights of the constituent elements. 



II. That the influence of the halogen on these same physical pro- 

 perties increases with the number of its atoms in the compound. 



III. That in any normal halogen compound the influence of either 

 of the elements on the melting or boiling point increases with its own 

 atomic weight, and decreases with the atomic weight of the other 

 element. 



IV. The melting or boiling point or heat of formation of a bromide is 

 always nearer to that of the corresponding chloride than to that of the 

 corresponding iodide ; and that the melting or boiling points of the 

 halogen compounds of the middle member of three consecutive ele- 

 ments of the same group are always nearer to those of the first member 

 (i.e., the one with least atomic weight) than to those of the last 

 member. 



The former of these phenomena probably depends on the fact that 

 the atomic weight of Br is nearer to that of CI than to that of I, i.e., 

 less than the mean of the two : and the latter on the fact that the 

 atomic weight of the middle member of three consecutive elements of 

 the same group is always less than the mean of those of the other two 

 elements. 



V. That the melting and boiling points of the halogen compounds 

 of the elements belonging to the first and second groups of Mendele- 

 jeff's classification are widely separated from those of the other groups, 

 being in fact considerably higher. Different relations too often 

 appear to exist between the melting points of the even members of 

 these two groups to those which exist between groups (3 — 7) inclu- 

 sive ; while the compounds of the elements, which are often placed in 

 the odd divisions of the first and second groups are generally altogether 

 irregular. In the case of the odd members of the first group this may 

 be explained by the fact that it is very uncertain whether Cu, Ag, and 

 Au really belong to the same group as Na or not, as pointed out by 

 Mendelejeff in his memoir on the " Periodic Law," in which he places 

 these elements not in the first group along with Na, but in the eighth 

 with Fe, Pd, Pt, &c. 



As far as existing data allow us to judge, the compounds of the 

 elements with monatomic organic radicals also obey the same laws as 

 those of the halogen compounds. 



A process for calculating unknown melting and boiling points (in 

 the case of halogen compounds) by a method of limits is next described, 

 which gives very good results when applied to known melting and 

 boiling points, the average error for all known boiling points being 



