PHILOSOPHICAL SOCIETY OF WASHINGTON. 105 



ing temperatures would probably be at equal distances from the 

 meltiug points, although this is by no means certain. Under 

 ordinary circumstances, in any given series of similar compounds 

 the molecular heat increases slowly with the atomic weight. The 

 two subjoined series, series which are usually taken to illustrate 

 the supposed equality, will exemplify this fact. Each molecular 

 heat is here deduced from the mean of all the reliable determina- 

 tions of specific heat extant for each substance. The first series 

 is that of the alkaline chlorides. 



No determinations have yet been made for Cs CI. 

 For the carbonates of the calcium groups. 



Substance. Molecular Heats. 



CaCOj 21.09. 



Sr CO., 21.34. 



Ba CO; 21.49. 



PbCOj 21.91. 



Comparing some chlorides, bromides, and iodides, we have — 



Substance. Molecular Heats. 



K CI 12.98. 



K Br 13 4Y. 



K I 13.60. 



Pb Cl,.^ 18.85. 



Pb Br, 19.55. 



Pbl, 19.67. 



Among liquids the increase is yet more striking : one series will 

 serve to illustrate this. 



Substance. Molecular Heats. 



C CI, 31.91. 



Si CI, 32.42. 



Si CI, 35.25. 



Sn CI, 3t.l5. 



I have traced out relations of this kind in over twenty different 

 series of compounds, and find only one or two apparent excep- 

 tions. These exceptions are doubtless due to experimental error, 

 since they occur only where there are scanty materials. 



Mr. J. E. HiLGARD read a memoir communicated by Prof. 

 Stephen Alexander, of Princeton, 



ON THE zodiacal LIGHT. 



(See the Appendix to this Bulletin.) 



