Prof. T. Carnelley on the Periodic Law. 11 



to the fact that, though the method to be presently described 

 of calculating melting-points and boiling-points does not 

 give definite points, but only limits between which the true 

 melting-point or boiling-point will lie, yet in the construction 

 of the tables the mean only of these has been used in order to 

 avoid complication. 



(5) No less than 73 out of the above 180 exceptions are 

 due to the melting-point or boiling-point of the following- 

 eight substances being too high or too low : — CBr 4 and A1C1 3 , 

 the melting-points of which are too high ; All 3 *, SrBr 2 , Srl 2 , 

 and PI 3 , the melting-points of which are too low ; Asl 3 , the 

 boiling-point of which is too high ; and Snl 4 , A1I 3 *, and PI 3 , 

 the boiling-points of which are too low. It is intended to 

 re-determine at the earliest opportunity the melting-points and 

 boiling-points of these compounds, and especially to observe 

 whether they melt or boil with decomposition or not. 



General Conclusions to be drawn from the above tivelve Relations 

 as to the Laws which govern the Melting- and Boiling- 

 Points and Heats of Formation of the Normal Halogen 

 Compounds of the Elements. 



I. If in a series of binary normal compounds one of the 

 elements be common to all, then the melting-points, boiling- 

 points, and heats of formation are periodic functions of the 

 atomic weight of the other element. 



See Table II. and Plate I. 



II. The influence of the halogen on these same three physical 

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



Thus :— 



Melting-point. 



PBr 3 = 252 

 PC1 3 = J200 



52 



Melting-point. 



SiBr 4 = 260 

 SiCl 4 = _205 



55 



Here the difference between the melting-points of the com- 

 pounds containing four atoms of halogen is greater than 

 between those of the compounds containing only three. 

 (Compare also Relation 6.) 



III. In any normal halogen compound the influence of either 



* The numbers used for A1I 3 in the construction of the tables of 

 relations were those obtained by Deville and Troost, viz. melting-point 

 = 125°, boiling-point = 350°. Since writing the above I have found that 

 Weber obtained the following : — melting-point = above 185°, boiling- 

 point = above 359°, thus confirming the supposition that Deville and 

 T roost's numbers are too low. 



