of Ferricum and A luminium . 117 



ment from considerations based on their composition and beha- 

 viour. Cahours represents an ethyl-iodide of aluminium by the 

 formula Al 2 Et 3 I 3 , just as Bunsen represents an iodide of nitrogen 

 by the formula N 2 H 3 1 3 . A more complete examination, how- 

 ever, of aluminium-ethyl and -methyl would doubtless go far to 

 settle the question of hexatomicity regarded merely as a ques- 

 tion of fact. 



What I conceive to be the arguments opposed to the hexato- 

 micity of aluminium and ferricum are drawn from the specific 

 heat of aluminium (a), from the non-existence of any inferior 

 chloride of aluminium ((3), from the improbability of an incre- 

 ment of chlorine combining two molecules of ferrous chloride 

 into one molecule of ferric chloride (y), and from the series of 

 iron chlorides requiring a trichloride for its completion (8). 



a. The proportion of aluminium which exists in the molecule 

 of aluminic chloride is the smallest proportion of aluminium that 

 is known to exist in any combination whatever, and accordingly 

 ought to be regarded as an atomic proportion. But if this pro- 

 portion be taken at 55 parts, the atom of aluminium will have a 

 specific heat twice as great as that of the atom of any other ele- 

 ment. Thus the atomic heats of certainly all the metals will be 

 expressed by a number approaching more or less nearly to 6*2, 

 except indeed the atomic heat of aluminium, which will be ex- 

 pressed by the number 11*7. Now there does not appear to be 

 any peculiarity in the general behaviour of aluminium which 

 would warrant us in according to it such an anomalous atomic 

 heat. -If, however, the 55 parts of aluminium contained in the 

 hexatomic molecule of aluminic chloride be regarded as two atoms, 

 then aluminium will present the yet more striking and inexpli- 

 cable peculiarity of never entering into combination save in the 

 proportion of two inseparable atoms. 



/3. There are several elements of which the lowest known 

 chloride is a trichloride, but none of which it is a hexachloride. 

 The hexachlorides of iridium, Ir CI 6 , of osmium, Os CI 6 , and of 

 dicarbon, C 2 CI 6 , for instance, proceed each from a tetrachloride 

 and dichloride, and so constitute each the third term of a series 

 of chlorides ; whereas hexachloride of aluminium, Al 2 CI 6 , or 

 All CI 6 , would be the first or lowest chloride of aluminium. Here 

 again there is no peculiarity in the general behaviour of alu- 

 minium, and certainly no extreme electro-negative tendency 

 comparable to that of chromium and molybdenum, whose inferior 

 chlorides were for a long time concealed, which would suggest 

 or accord with its possession of a hexatomic lowest chloride. 



7. It is scarcely conceivable, and certainly contrary to all 

 analogy, that continuous chlorination should lead to the syn- 

 thetic formation of a complex molecule. For the action of chlo- 



