a If these equations represented approximations, no matter how 
ose, they would not be worthy of consideration. They are 
etieetly exact. The quasi-connection of oxygen with this se- 
will be pointed out further on. : 
: I 10 Cl-19 F Cl = skate 
. ; — 5Br-F yattue . 
3 a 
12 Br—2 Cl Br—7F 
Do ie ae hick alc otanceaa 
| : ee 
ae pe Rann Pov acsee eae 
Br=6C]—”7 F " 10 Cl—I : 
| SSE AS Pe 
: 2 r 6 Cl—Br 
' Pec. 3I-+F 7 
| Gch gre | 
‘galt has been remarked that the atomic weights of the members 
oxygen series proper differ from each other by whole 
: but the extent to 
of cn differences of this kind exist amongst the atomic weights 
: eg bodies has perhaps not been pointed out. 
i 1 aving the highest atomic weights are bismuth 
: zu, commencing with the’ first of these substances, and 
- Guuishing its atomic weight successively by whole multiples (in 
_ me case by a sub-multiple) we obtain the following results :— 
