ON DIFFUSION IN SOLIDS. 361 
before any penetration takes place.** Certain precautions have to be 
taken to avoid errors in the interpretation of the results. Thus, an 
experiment may fail from the absence of a satisfactory contact between 
the carbon and iron. This may be ensured by the application of 
pressure, which accordingly favours cementation.** On the other 
hand, in the course of experiments in which negative results were 
obtained from solid carbon and iron in a high vacuum, both materials 
being previously freed from gas, it was observed that a small reduction 
in the degree of the vacuum, caused by the admission of gas, led to 
carburisation.** 
_A careful investigation, in which these conditions are taken into 
account, is due to F. Weyl.8° The materials used were sugar charcoal, 
purified by heating in chlorine and in hydrogen; natural graphite, the 
ash of which had been reduced to 0°22 per cent. by treatment with 
sulphuric and hydrofluoric acids; ‘ kish,’ or graphite obtained by stir- 
ring molten iron saturated with carbon, the ash being reduced to 0°82 
per cent. by successive treatment with nitric acid, chlorine, and hydro- 
fluoric acid; and diamond. Iron prepared in an electric furnace was 
used on account of its freedom from slag. The polished cube of iron 
rested by one of its faces on the surface of the carbon. A good vacuum 
was maintained, the pressure being always below 0°05 mm., and fre- 
quently as low as 0°001-0°002 mm. throughout an experiment. Positive 
results were obtained in all cases, whilst the supposition that gases 
played any important part in the transport of carbon was negatived 
by the observation that a layer of kaolin, 0°25 mm. thick, interposed 
between the carbon and the iron, sufficed to prevent all cementation. 
The influence of contact was shown by the fact that very little cementa- 
tion was obtained with fragments of diamond when the iron rested 
upon them, whereas when the diamond dust was placed on the upper 
surface of the iron cementation occurred quite readily. In the latter 
case the finer particles of diamond came into contact with the iron, 
whilst in the former they were separated. In view of these results 
it is not easy to understand the failure of Charpy and Bonnerot to 
obtain’ carburisation under similar conditions, unless the contact 
between the two substances in their experiments was insufficiently 
good. 319 9 
Graphite which is already present in the interior of a mass of iron 
is undoubtedly absorbed to a considerable extent on heating. Thus, 
when grey iron is heated to different temperatures above 700° and 
quenched, the proportion of combined carbon is increased, the quantity 
increasing rapidly with the temperature.*? A similar observation has 
been made in respect to temper carbon, which is finely divided 
graphite.** 
Mention must now be made of a class of experiments in which the 
8 LL. Guillet, Rev. de Métallurgie, 1906, 3, 227. 
& L, Guillet and C. Griffiths, Compt. rend., 1909, 149, 125. 
% G. Charpy and S. Bonnerot, ibid., 1910, 150, 173. 
8 Métallurgie, 1910, 7, 440. 
8 G. Charpy, Compt. rend., 1907, 145, 1277. 
88 H. Le Chatelier, Rev. de Métallurgie. 
