10 Prof. T. Carnelley on the Periodic Law. 



Table VII. — Illustrating Relations 10/11, and 12. 



I. 



Difference 

 of atomic 



weights of 



elements 



in I. 



Chlorides. 



Bromides. 



Iodides. 





•s ft 



'-3 "o 



^1 



^H O 



•3 ^ 

 pq 



is '3 



^i o 



•s ft 



PQ 



P-I 



-96 

 -52 

 -49 



- 4-4 



- 4-5 



- 4-8 

 4-39-5 



42-0 



86-5 



83-0 



1300 



1745 



-98 



-54 



-29 



+ 2 



47 



15 



46 



116 



147 



205 



274 



305 



-22 

 +32 

 65 

 111 

 123 

 119 

 165 

 210 

 256 

 330 

 417 

 463 



-57 

 -14 



+ 1 



23 



54 



44 



66 



112 



134 



171 



229 



251 



54 

 103 

 113 



158 

 159 

 162 

 207 

 218 

 263 

 359 

 418 

 463 



-55 



+36 



19 



30 



55 



110 



121 



129 



140 



171 



245 



256 



102 

 204 

 185 

 202 

 220 

 287 

 304 

 303 

 320 

 391 

 474 

 491 



As-I 



P-Br 



P-Cl 



Sb-I 



As-Br 



As-Ol 



Sb-Br 



Sb-Cl 



Bi-I 



Bi-Br 



Bi-Cl 





tions and do not obey the rules. This is a very small number 

 of exceptions, and especially so if we bear in mind the fol- 

 lowing considerations : — 



(1) The numbers obtained by different observers for the 

 melting-point or boiling-point of one and the same substance 

 vary somewhat in many cases, and considerably in some few. 

 Consequently it is probable that many of the above exceptions 

 are due to errors in the experimental melting-points and 

 boiling-points. 



(2) Several of the experimental values are somewhat un- 

 certain, as is the case with the boiling-point of SC1 2 as found 

 by Dumas and the boiling-point of NC1 3 as found by Pelouze 

 and Fremy ; whilst it is probable that some of the com- 

 pounds, the boiling-points of which have been determined, 

 boil under ordinary circumstances with slight decomposition, 

 which would of course produce a corresponding error in the 

 boiling-point. This is the case with CBr 4 and some others. 



(3) The boiling-points of several of the compounds have 

 only been determined within certain limits, and yet the means 

 of these limits have been employed in the construction of the 

 tables, and these would of course produce corresponding, 

 though small, errors. 



(4) Most of the above exceptions, however, may be ascribed 



* The differences in the second column at these points do not follow 

 exactly the order of size. It is easily seen, however, that they are so very 

 close together that slight errors in the atomic weights used (viz. P = 31, 

 Sb = 122, As=75,Bi=210, 01=35-5, Br =80, 1=127) might easily affect 

 the arrangement, as the following atomic-weight determinations show : — 

 Bi=208 (Schneider), 210 (Dumas) ; Sb=120 (Schneider), 122-5 (Dexter), 

 122 (Dumas), 122-3 (Kessler), 120 (Unger), 120 (Cooke). 



