128 On Catalysis , and the Nomenclature of Oxides, 



consisting of a series of decompositions and reformations. It 

 is therefore a catalysis. Hydrated sesquioxide of iron and hy- 

 drate of zinc have no action upon dilute hydrogen peroxide. 

 We may therefore argue from this, as from its other proper- 

 ties, that sesquioxide of iron is not a peroxide in the sense in 

 which the term is used in this paper, namely as denoting an 

 oxide in which one or more of the atoms of oxygen are held 

 in such a way as to have an outward strain in presence of 

 hydroxyl or a similar body. This definition being accepted, 

 the sesquioxide of iron must be regarded as the oxide of iron, 

 the term protoxide being reserved for the lower compound. 

 In the same way we should speak of Ni 3 5 as a peroxide of 

 nickel, of MO as the oxide, and of some lower, at present un- 

 known, compound as the protoxide. Our view of catalysis 

 being as above stated, it is easy to explain the action of nickel 

 or cobalt upon hypochlorites. These are capable of peroxidi- 

 zing nickel ; and this constitutes the difference between the 

 action of the latter with them and with hydroxyl. The cycle 

 of operations can be performed and catalysis established. A 

 further explanation of the difference between the action of 

 copper and of cobalt with hydroxyl suggests itself. The change 

 constituting the cycle in the case of copper is from hydrated 

 oxide to an unknown peroxide and back again ; and since the 

 final moment of catalysis, that in which the last hydroxyl dis- 

 appears, must consist in the reduction of the copper to the 

 lower of the two limits between whicli the action alternates, 

 we find oxide of copper and water as the final products. In 

 the catalysis of hypochlorites by cobalt, the two limits are pro- 

 bably tricobaltic penioxide (Co 3 5 ) on one side, and cobaltic 

 dioxide (Cc0 2 ) on the other ; since (Co 3 5 ) has been shown to 

 be the result of the reaction. 



The decompositions of dilute hydroxyl are very distinct from 

 those which occur to the concentrated compound on contact 

 with finely divided silver, gold, or platinum, with various 

 oxides (as, for instance, with nickel oxide), and with many other 

 bodies. We cannot suppose that finely divided gold or platinum 

 is acted upon by hydroxyl; the action is probably connected 

 w T ith some power possessed by the surface, and especially by 

 edges and by fine particles, of these metals to attract oxygen. 

 In concentrated hydroxyl at ordinary temperatures the oxygen 

 is quivering on the verge of liberation, and such a surface 

 action is sufficient to induce decomposition. 



The phenomena of catalysis thus group themselves into 

 physical and chemical. Physical catalyses are those in which 

 the catalyzing body, often an inert substance chemically, 

 remains unchanged and exerts a surface action purely phy- 



