50 SECTIONAL ADDRESSES. 
and it will evidently depend on the closeness of packing of the molecules 
in that layer whether diffusion is easy or difficult. As a rule, it is probably 
more difficult than in the original solid, and we therefore find on micro- 
scopical examination that crystals of the two reacting substances, whether 
pure metals, solid solutions, or intermetallic compounds, are separated by a 
zone consisting of the product of reaction, which may be very persistent, 
although its breadth gradually diminishes on annealing. This effect is 
well seen, for instance, in alloys of copper with antimony. 
An interesting class of reactions is that which includes the de- 
composition of a crystalline solid, one of the products escaping in the form 
of a gas whilst the other remains solid. From the nature of the curves 
connecting decomposition and time Hiittig and others have concluded 
that the escaping molecules must be able to traverse the crystal freely 
without serious dislocation, but this view is not confirmed by examination 
by means of X-rays or in any other independent manner. On the other 
hand, Hinshelwood has examined a number of such reactions in detail, 
giving special attention to the physical condition of the crystals before 
and after decomposition, and his experiments are not only of a higher 
order of accuracy but they include a study of the physical conditions of 
the reaction. The decomposition of the permanganates by heat has been 
found to be a convenient one for this purpose, since it proceeds at a moderate 
temperature, and the reaction is undoubtedly monomolecular. The initial 
rate of decomposition of silver and potassium permanganates is greatest 
when the solid is finely powdered, but when crystals of appreciable size 
are used the decomposition proceeds at an accelerated rate, as the crystals 
become disintegrated. The results prove that the reaction is confined to 
the surface, and that it can only proceed inwards as the texture is loosened, 
so that diffusion does not play a part in the process, at least when the 
temperature is such that the decomposition is nearly complete in an hour 
or two. When solid solutions of potassium permanganate in potassium 
perchlorate are used, the latter salt being stable under the conditions of 
the experiment, the rate of decomposition is lessened, the observed effect 
corresponding closely with that which is calculated from the heat of forma- 
tion of the solid solution, a quantity which has been directly determined. 
Some similar decompositions are more complex owing to the catalytic 
effect of one or other of the products of reaction. The hindering effect of a 
solid coating, already referred to in connection with reactions in the 
interior of metallic alloys, is seen in the decomposition of ammonium 
dichromate by heat, large crystals becoming coated with an adherent 
layer of chromium oxide, which retards further decomposition. 
Very recently Kurnakoff has studied the gradual change in the state 
of oxidation and hydration of vivianite, an hydrated ferrous phosphate. 
When first produced, these crystals are colourless, but they become 
blue as oxygen is absorbed, a part of the iron passing into the ferric state. 
Moreover, the degree of hydration may vary as water is taken up from 
without. During these changes it is stated that the structure of the 
mineral remains unaltered and the crystals remain homogeneous, the 
optical properties varying continuously, but it does not appear from the 
abstract that any X-ray examination has been made. Such behaviour 
recalls that of the zeolites, the structure of which is probably loose. It is 
unlikely that any closely packed crystal could behave in this way. 
~~. 
