1904] Mills — Molecular Attraction. 99 



constant with rise in temperature in the case of chlor-ben- 

 zene, brom-benzene and iodo-bezene, may easily be without 

 significance, for in the first mentioned the tendency is not 

 marked, and in the last two the final values are not far from 

 the probable true values as shown by the results at 30° and 

 100° C. On the other hand, they may indicate progressive 

 changes commencing- within the molecule, and the remarks of 

 Dr. Young- above given are sig-nificant. 



Stannic chloride may be said to be in agreement with the 

 theory from 0° C to 170° C, but the values from 100° to 280° 

 continually decrease, and there is little doubt that the cause 

 of the variation is operative from the first. I would call 

 attention to the hig-h specific heat per atom that we find in 

 stannic chloride and the conclusion based on grounds having 

 nothing to do with the present theory, drawn therefrom by 

 the author, 1 viz: "A high specific heat per atom indicates 

 that the potential energy of the atoms is being rapidly 

 increased and that the molecule is approaching the point of 

 dissociation." 1 



Prof. Young writes: "Stannic chloride spoils the surface 

 of mercury even at low temperatures and special methods had 

 to be used throughout. The accuracy is certainly not so 

 g^reat as with most of the other substances, but this will not, 

 I think, explain the regular fall in the value of the constant." 



Water shows divergence 0° to 30° C and at 270° C. Both 

 may be explained by the multiplication of error through the 



8P 



«-=- at the end points of the Biot formula curve. Ramsay 



and Young's observations of the vapor pressures at 230° to 

 270° are in themselves rather conclusive evidence that the 

 trouble at 270° C rests with the Regnault-Biot formula. 2 

 See also remarks on page 94. The measurements of Grif- 



1 Jour. Elisha Mitchell Scientific Soc. Vol. 18. 

 » See Phil. Trans. 1892A, p. 112. 



