586 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



borne out by the facts concerning either chlorides, bromides, or iodides 

 of these three metals. In every case the value of the fraction 



change of volume 

 heat of formation 



grows smaller as the compressibility of the metal is less. In other 

 words, a given energetic effect is shown to produce a smaller change of 

 volume whenever the compressibility is smaller. The facts are given 

 in the tables below. The first table compares the different metals 

 combined with a given halogen. 



As has been said, the prediction is fully verified. The quotient of 

 the contraction divided by the heat of formation is always smaller when 

 the compressibility is smaller. Exactly the same phenomenon is to be 

 observed if the several halogens be varied, in combination with a given 

 metal. 



rine has been found by Knietsch to be 18.6 kilojoules (Z. elektrochemie, 9, 847, 1903), 

 a number not far from that (20.5) calculated in the record of the present series of 

 papers, and tliis amount is subtracted from all the commonly accepted heats of 

 formation of chlorides. In the case of single gram atoms of chlorides only half 

 of this quantity is subtracted. 



It is worthy of note that these thermal quantities also contain the " heat of co- 

 hesion " or latent heat of liquefaction and solidification of the products. If the 

 compounds did not condense, the heats of formation would be less. This CQf- 



