318 Royal Institution* 



To evade this difficulty, some chemists have proposed to accord 

 to the molecule of aluminic chloride the formula Al 2 CI 6 , whereby an 

 indivisible proportion of metal would be habitually represented by a 

 divisible symbol ; for it is agreed on all hands that the proportion of 

 aluminium contained in the molecule of aluminic chloride is the 

 smallest proportion of aluminium found in any aluminic compound 

 whatsoever ; that it is incapable of experimental division by any pro- 

 cess whatsoever ; and consequently that, so far as our present know- 

 ledge goes, it is an indivisible or atomic proportion. 



Now there are undoubtedly certain bodies, elementary and com- 

 pound, of which the ascertained vapour-densities, and consequent 

 volumes, no matter how accounted for, are, as a mere matter of ex- 

 periment, discordant with the chemical analogies of the respective 

 bodies ; but in most instances these anomalous results are rendered 

 unimportant by other determinations of vapour- density, either of the 

 same bodies raised to higher temperatures, or of associated bodies 

 having a more decided volatility. Hence arises the question whether 

 the ascertained volume of aluminic chloride, which is discordant with 

 the specific heat of aluminium, may not be anomalous in a similar 

 manner, and whether the anomaly may not be corrected by an ex- 

 amination of other more volatile aluminic compounds. 



The methide and ethide of aluminium recently obtained by Mr. 

 Buckton and the speaker are, so to speak, varieties of aluminic 

 chloride in which the chlorine has been replaced by methyle and 

 ethyle, and are at the same time far more volatile and manageable 

 than the typical chloride. Now it has been found that two gaseous 

 volumes of the methide and ethide of aluminium contain only 27'5 

 parts of aluminium, united with three atomic proportions of methyle 

 and ethyle ; and accordingly their molecules have to be expressed 

 by the formula? AlMe 3 and AlEt 3 respectively. In other words, the 

 normal results obtained with the methide and ethide correct the 

 anomalous result obtained with the chloride, and confirm the atomic 

 weight and molecular formulae deducible from the specific heat of 

 aluminium. 



That the ascertained vapour-density of aluminic chloride is really 

 anomalous receives a further corroboration from the behaviour of 

 aluminic methide itself. At 220° and all superior temperatures the 

 vapour-density of this compound shows that two volumes of its vapour 

 contain 27*5 parts of aluminium and three times 15 parts of methyle ; 

 but at 130° its vapour-density, corrected for alteration of tempera- 

 ture, becomes very nearly doubled, or, in other words, two volumes 

 of its vapour contain very nearly 55 parts of aluminium and six times 

 15 parts of methyle. According, however, to the well-known rule, 

 based on the separate researches of Cahours and Deville, the mole- 

 cular formula of a body must be calculated from its permanent or 

 ultimate, and not from' its variable or initial vapour-density ; whence 

 the high vapour-density of alummium-methide at 130° does not at 

 all interfere with our attributing to its molecule the formula AlMe 3 , 

 deducible from its vapour-density at 220° and upwards, and harmo- 

 nizing with the specific heat of metallic aluminium. 



