Attempts were then made to obtain metallic aluminium in the laboratory, by exposing to a very high temperature, in a graphite crucible, mixtures of alumina and cyanide of potassium, with and without carbon. In no case was there a trace of the metal discovered.

There is one fact in connection with the properties of this fume which deserves attention. The silica in all the raw materials of the furnace is almost if not entirely insoluble in water, after treatment by hydrochloric acid; whereas in the fume or sublimate this substance is all soluble in that acid. This proves the fact that chemical action has been at work, and the most probable cause of such action lies in the alkaline constituents of the fume.

It is to be observed that the gases contain no trace of any of these four earths, until we have reached a point about 45 feet above the tuyeres, in a furnace having a height of 80 feet. I apprehend that this is due to the condensation of the alkalies, and their existence in a fused state among the earths. Under such circumstances nothing more likely than that we should have a fused mass of alkalies, with silica, lime, alumina, and magnesia. This mode of action does not necessarily account for the formation and possible evaporation of the metallic bases of earthy substances by the intervention of potassium or sodium, for these metals appear to exist in the fume obtained from the upper parts of the furnace chiefly as potassium and sodium chlorides. The earths at E (45 feet above the tuyeres) bear the following relation to each other:

Looking at the fixed nature of these bodies under the most intense heats we can command it would seem as if the only means of accounting for the sublimation, is by the power which currents of highly heated gases are known to possess of carrying away substances otherwise fixed in their nature.

During the passage of this earthy sublimate, if such it really be, through the colder portions of the materials, a good deal will no doubt be condensed, and carried back in the solid form to the lower regions of the furnace; and it would appear that, while the proportion

of silica is not very sensibly affected in quantity by this cumulative process, the lime certainly, as shown in the table, is considerably diminished in amount; its place being taken by alumina and magnesia.

The following figures exhibit the changes just referred to; and to them is added an example, showing the composition of the slag from a furnace smelting Cleveland stone; by which it will be seen that there is little analogy between the fume and the slag, as to the relations which the earths bear quantitively to each other. Where sulphates and chlorides occur, their equivalent of oxide is taken into account:

The composition of the fume deposited from the gases after they leave the furnace, cannot be relied on as correct; for it is contaminated with mechanically projected dust, as evidenced by the large quantity of peroxide of iron and carbon it contains.

Looking at the small proportion of the earths which leave the furnace in the form of this fume, the composition of the slag necessarily represents the relations they bear to each other in the materials charged. Assuming this, and comparing the composition of the slag with that of the fume, before it is contaminated with the dust mechanically carried over from the ironstone and coke, we have the following numbers:

From these we may infer that the silica is most affected by the causes in operation in the furnace, which conduce to sublimation. Compared with the slag, 37 parts of silica ought to have been accompanied by:

Hence it appears that the lime and alumina are the least affected by the sublimating influences.

The fume, during its passage from the furnaces to the fireplaces, where the gas containing it is burnt, deposits its grosser particles. After combustion however there accumulates in the boiler and stove fireplaces a large quantity of a fine powdery matter, having a dirty yellow colour.

From hence the burnt gas, diluted with the air required for its combustion, and bearing with it a large quantity of the finest portions of the sublimate, is carried forward towards the chimney. At the base of one of the boiler chimneys at the Clarence works, the amount of the fume contained in each cubic metre of the diluted gas was ascertained to amount to 1.07 grammes.

I will add to what has been said on this subject the composition of the deposit gathered in one of the boiler fireplaces, and that of the fume as it enters the chimney.

Although the earthy constituents which form the bulk of this vaporous or finely divided matter appear to take their origin in a part of the furnace where the temperature is by no means very intenseprobably not above 1,093° C. (2,000° F.)-I have remarked that when the furnaces in the Middlesbrough district are "working cold," which happens when they are making white iron, little or no fume is given off. On the other hand its comparative absence is to be observed in the Scotch furnaces smelting rich foundry iron, where it might well be

expected to occur, seeing that the slag much resembles in composition that of the works using Cleveland stone. The slag from one of the Coltness furnaces contains the four earths in the proportions given below, which, it will be seen, closely resemble those at the Clarence works:

At Coltness however the coal is used raw, which gives rise to the distillation of a large quantity of tar in the upper region; and it may be that the presence of this substance may serve to intercept the escape of the sublimate in question.

There are no doubt many circumstances which may conduce to changes in the composition of the fume from blast furnaces. The more prominent of those are alterations in the contents of the materials; this observation is chiefly applicable to the ironstone, which is much more subject to changes in composition than the coke or limestone.

If this remark be appropriate in the case of ironstone, taken from one district, which is nevertheless variable in its composition, it is still more so in the case of furnaces employing ironstone from other districts and of a totally different character.

There is to be found in the present Section a good deal of matter of a more or less speculative character. This observation applies particularly to the alterations in the relative proportions which the carbon, oxygen and nitrogen bear to each other in different levels of the blast furnace. It is also applicable to the formation and behaviour of the cyanogen compounds, and other substances found in the fume given off during the process of smelting. Whatever value the speculations themselves may have, it is hoped that the facts, mentioned in connection with the attempted explanations, may justify the space devoted to their description.

SECTION X.

ON THE EQUIVALENTS OF HEAT EVOLVED BY THE FUEL IN BLAST FURNACES.

SINCE the matter relating to the subject of the present Section was printed in this work, further consideration has induced me to add a few more pages on a question of so much practical importance to the iron smelter. This appeared to me especially desirable because at the present time the use of fire brick stoves is engaging increased attention. I have also been led to give the question further notice by having had a recent opportunity of again examining the performance of furnaces which are exclusively engaged in the manufacture of charcoal iron. Besides this, Mr. Charles Cochrane has published his experience with superheated air in his very capacious furnaces at Ormesby;1 so that I am now able to supplement the data given at p. 107, et seq., in reference to the interesting problems connected with this branch of my work.

The value of every description of fuel, so far as its use in the blast furnace is concerned, I repeat may be primarily regarded as dependent on the fixed carbon it contains.

If the fuel used in smelting iron is uncharred, the gaseous matter, instead of being previously burnt in the coke ovens or charcoal pits, is rendered available for purposes outside the furnace; but it is next to useless inside it. Of course some sacrifice of solid carbon must be made in volatilizing the gases, when using raw fuel; but this does not amount to the waste of the fixed portion of the coal generally incurred during the process of charring.

In practice a coal containing 72 per cent. of coke only yields about

' Proceedings of Institution of Mechanical Engineers, August, 1882.