cooling down to a certain point, the usual signs of decomposition of the water took place, it cannot be doubted that that effect must in reality have been produced, unless we imagine that the experiment in question had deprived the iron of this quality, which there is no reason to suppose, and which has never yet been known to have been accomplished. But a very satisfactory explanation may be found, provided the Franklin experiments above mentioned be correct. If the repulsion between iron and water be perfect when the former is at a red heat, or about 800° of Fahrenheit, as the temperature of thoroughly melted cast iron is estimated at 20577° of the same scale, the iron would have to be cooled down through 19777° before any effect could be produced. This accounts for the quiescent state in which the iron remained under water for several seconds, as Mr. Maugham states. If the repulsion be perfect at 800°, iron cannot possess any remarkable degree of power in decomposing water through more than about 400° of temperature, and so small a mass of metal as the experiment was performed with would pass through this range almost instantaneously, surrounded as it was with water. The trifling effect of decomposition which would be produced, would therefore, most probably, escape observation, particularly as from the depth of water which is stated to have been over it, there would be an additional pressure of nearly three quarters of a pound on the square inch, which would be considerable in comparison of the small weight of metal used, and the large horizontal surface that would necessarily be exposed, in proportion to its bulk, when dropped into water in the manner stated.

Secondly, as to the practical utility of the experiment I entirely differ from Mr. Maugham. The residual product of his experiment being so extremely hard might necessarily have been anticipated, because he used cast iron which contained the smallest quantity of carbon, and therefore approached more nearly to the composition of steel; and also from its having been hardened at the highest possible temperature at which this operation could be performed.

Every time that cast iron is melted it loses a portion of its carbon, and becomes harder; and the smaller the casting into which it is run, the greater its liability to become hard, that is, to lose its carbon. The effect seems to be the same as when iron is exposed to the operation of puddling, which is in fact raking or puddling it about, when in a state of fusion, to allow the escape of the gaseous carbon, and fits it for being made into rolled iron. Now the running of cast

iron into very small moulds has something of the same effect; it divides the mass into more minute portions, and thereby allows greater facility for the gaseous escapement. We consequently find that if two castings be made at the same time, and from the same iron, the one being large and the other small, the large casting will be soft, and the small one probably quite hard. Iron nails are perhaps as hard as almost any article made from cast iron; these when fused would become still harder, and in addition to this the hardening process would be increased to the greatest possible degree by the plan Mr. Maugham adopted, because it is well known that the higher the temperature is raised before the article is plunged into cold water, which is the usual way of hardening,-the greater the effect produced. The fused cast-iron nails in question would therefore be acted upon by the water the instant they were cooled down to that point at which the repulsion between the metal and the water ceased, and the result would of course be the maximum state of hardness.

The operation of tempering may be considered as the reverse operation of hardening: it is, in fact, reducing the state of extreme hardness to that point best suited to the particular object for which the steel is required.

Mr. Maugham's supposition from his experiment that it might be desirable to place the moulds for iron castings under water, would not be found to answer in practice. When the quantity of metal was large, the moment the heat was reduced to that point where the repulsion before mentioned ceased, decomposition of the water would take place; the consequence would be the formation of a quantity of inflammable gas within the mould,-hydrogen, no doubt, mixed with a portion of oxygen,-and an explosion would be the inevitable consequence. In the ordinary manner of making castings, this gas frequently forms in the moulds, and has occasionally been the cause of disastrous accidents: in one instance in particular, a few years ago, at Nottingham, where the greater part of the boiling metal for a large casting weighing several tons was ejected out of the mould by an explosion of this kind of gas, which forced a shower of the boiling metal over several persons who were near. Even in small castings a miniature explosion almost always takes place, arising from the decomposition of the water contained in the sand of which the mould is composed. In addition to this objection, the resulting effect of the intense hardness of the castings would in nineteen cases out of twenty be a most important disadvantage; and last, but not least, the plan

would be an impossibility except in a very few cases where iron moulds are used, because the water would dissolve both the loam and sand of which the moulds for iron castings are usually composed.

It were truly a work of supererogation in your correspondent to attempt proving by arguments that chemists have almost invariably drawn wrong conclusions from their experimental researches into the habitudes of iron: his own letter bears such ample testimony to the truth of the assertion, it may readily be conceded to him that he has fully established that fact. I cannot but observe, however, that scarcely any branch of manufacture offers a wider field, or appears more likely to justify the anticipation of a plentiful harvest for scientific research, than this. If Mr. Maugham will only avoid the idola tribus,-the errors into which almost all chemists have fallen, by drawing conclusions from experiments on a small scale, there are none whose talents and resources better qualify them for following out this subject with success than himself. The best experiments of the laboratory must never be depended on; they only afford leading points to which attention should be directed; for the element of heat is so materially altered when practically operating on a large mass to what it is when the experiment is made on a small scale, that similar results are scarcely ever obtained.

But to return to the subject of Mr. Maugham's letter. There can be no doubt as to the possibility of making steel at once from cast iron, without the previous operation of decarbonising and making it into bars. On this subject there is a paper by Mr. Hawkins in the Report of the British Scientific Association for 1832, p. 598, describing the steel for a suspension bridge on the Danube to have been made in this manner in Germany. But although it might be done on a small scale, the cast iron made in this country is not sufficiently pure to be made into steel without passing through the previous processes of refining, puddling, shingling, and rolling; in all of which it successively becomes more and more pure. The minerals used here in the manufacture of iron are by no means so pure as in many other parts of the world, which is of itself sufficient reason why the various intermediate processes of refining become absolutely necessary with us, while they may occasionally be dispensed with in other places.

Although the query which Mr. Maugham puts, Whether iron and carbon combine in definite proportions or not, has never been positively demonstrated, there is the greatest reason to suppose that they do not. The number of combina

tions appear too great to suppose that they can be multiples of some common number, and the quantity of carbon which iron imbibes increases gradually up to its fusing point, and does not proceed per saltum. But after the iron has reached its fusing point, I think, not only does the increment in the quantity of carbon cease, but it gradually recedes back again the longer the iron is kept in a state of fusion. This perfectly accounts for all the various changes which take place. It will explain all the changes which occur in the manufacture of iron into steel; the tempering process; the change in cast iron by repeated fusions; and also the change which cast iron undergoes when it is refined for rolling into bars it explains also the process of annealing, but will not account for the operation of hardening, unless we imagine that the oxygen of the water unites with a portion of the carbon and forms carbonic acid; and as in the order of simple affinities the union of carbon and oxygen is stronger than that of carbon and iron, it is not unlikely this is what really takes place. That something analogous to it happens is certain, as a practised eye will at once detect the alteration by the colour, which is quite different in hard iron or steel to what it is in soft. Another proof is that the quality of iron smelted in hot weather is always much inferior to that made in cold weather; because in the former the air contains much more hygrometric water than at the latter time; and this being forced into the furnace along with the air from the blowing machine, unites with the carbon, is carried off in the gaseous state, and the iron receives too small a proportion of carbon, and is found to be inferior; for the more carbon it contains the richer is the quality.

The length to which these remarks have already extended prevents me from going further into details on this subject, as I fear I have already trespassed far too largely on your pages.

I am, Sir,

Yours obediently,

London, 11th July, 1836.

C.

OBSERVATIONS ON THE EXPLOSION OF STEAMBOILERS.

BY JACOB PERKINS, ESQ.

[Continued from p. 173.]

THE experiment of Perkins which is more particularly referred to in the query, is that in which an opening having been made in one of the generators, containing intensely heated water in contact with red hot metal, neither steam nor water escaped, and in which having affixed a pipe and stop-cock to the same vessel, no steam issued through the cock when opened. To repeat this with a view to ascertain, as required in the query, the size of opening to which such a result would apply, three apertures were made, th, th, and th of an inch respectively, in the sides of a wrought-iron mercury bottle; these were closed by conical plugs connected with levers, by which the plugs could be withdrawn from the sides of the bottle. The fulcrums of these levers were attached to the wrought-iron cylinder already referred to, within which, its axis coinciding with that of the cylinder, the cylindric bottle was placed. An earthenware furnace was placed below the bottle and surrounding cylinder, the latter resting upon wrought-iron bars, supported by the edges of the furnace, and the former supported by a stone placed upon the grate of the furnace. Besides affording a support for the levers, the wrought-iron cylinder was introduced to protect the experimenters against injury, should the bottle explode in the trials to be made with it. This apparatus having been placed in a quarry pit, adjacent to that in which the cylindric boilers were burst, water was poured into the bottle so as to fill it; the screw plug was next passed into the neck and forced home by lateral blows from a hammer. A fire was now made in the furnace, and the fuel heaped up so as to fill the entire space between the mercurybottle and wrought-iron cylinder, and to be about five inches deep above the stopper of the former. A string was attached to the lever connected with the smallest plug and carried up the bank. The fire soon burned briskly, and it was perceived that a small quantity of steam mingled with the feeble smoke and heated air which rose above the apparatus. About twenty minutes after the beginning of the experiments, the leak appearing to increase, an incautious attempt was made to stop it, but without success; at this time the bottle was seen to