322 Scientific Intelligence. 
solid particles. After ignition at a bright yellow heat for an 
hour, inside of a second crucible, it was cooled and moistened with 
a strong solution of ammonium carbonate, and re-ignited. The 
crucible containing now pure Al,O,, i 
specially Rage dan glass bottle and weighed. For the second 
method, aluminum bromide was prepared directly by the action 
of ectine upon aluminum and carefully purified by repeated 
fractional distillations until a was perfectly white and boiled 
steadily at 263°3° under 74 . pressure, the last distillation 
being Jrected | in a current of sitet ogen. Pure silver was prepared 
and dissolved in Bue nitric acid, the solution being used to pre- 
si Lebiehe the me um bromide e. The amount of silver used was 
tion of sodium hydrate prepare from metallic sodium. The 
calcium-chloride ees a sulphuric ac cid and pumice tube and one 
containing phosphoric oxide, used successively. For the details 
supe arine ge ponte the minute precautions used we must 
1; for S, 31°996; and for N, 14°010. As to 
bromine and silver, correction was made for the oxygen occluded 
by the silver in Stas’s research, and thus the atomic weight of 
silver became 107°649 and of bromine 79°754. The results as cal- 
culated were as follows: 1st method, series A gave 27°040+-'0073 
as a mean of five experiments; series B, 27°096--" 0054, also a 
m 
mean "bees experiments. Series A of the third method the 
warhied thinks entitled to most weight and series B, first method, 
Ma the least. The mean of all thirty experiments gives Al= 
703240045. If 1, B be excluded, Al=27:019-4-0030. Hence 
ine atomic weight of aluminum is 27:02, or when integers are 
used for O, a, etc., 27. The paper closes by calling 
attention hie ‘the fact that of the eighteen elements whose atomic 
weig 
un 
saeretioa —Phil. Trans., 1003, 1880. 
¥y aa 
