The different varieties of iron and steel will not oxidise (rust) in. dry. air, or when wholly immersed in fresh water free from air, but they all do so when exposed to the action of water or moisture and air alternately. Veiy thin iron oxidises more rapidly than thick iron, owing to the scales of rust on the former being thrown off as soon as formed in consequence of the expansion and contraction from alternations of temperature. Iron plates are more durable when united in masses than when isolated. The oxidation of iron is to a great extent arrested by vibration.

The comparative liability to oxidation of ironand steel in moist air, according to Mallet, is -

Cast iron . . 100

Wrought iron • • • • 129

Steel.... 133

Cast iron does not rust rapidly in air. When immersed in salt water, however, it is gradually" softened, made porous, and converted into a sort of graphite. Mallet found that the rate of corrosion decreased with the thickness of the casting, being during a century 1/10 1/40 in. in depth for castings 1 in. thick. Stevenson found the decay to be more rapid than this.

Wrought iron oxidises in moist air more rapidly than cast iron. The evidence as to its rate of corrosion in salt water is rather contradictory. Rennie found that it corroded less quickly than cast iron, but Mallet's experiments showed that it corroded more quickly.

Steel rusts very rapidly in moist air, more quickly but more uniformly than wrought iron, and far more quickly than cast iron. Low shear steel corrodes more quickly than hard cast steel. Recent experiments show that steel immersed in salt water is at first corroded more quickly than wrought iron, but that its subsequent corrosion is slower, and the total corrosion after a long period of immersion is less than that of wrought iron.

In the course of a paper read by McElroy before the Western Society of Engineers, on the causes of corrosion in cast-iron pipes, the author observed that a prominent cause of corrosion is the class of materials used, and also the method of the manufacture of pipes in ordinary foundries. In the first place, a cheap and easily melted pig is selected - specifications and the inspection ot quality and mixture not being strict - and the castings (for convenience of handling) are generally made in green-sand moulds laid at a slope of about 10° from the horizontal. Impure metal is therefore run in a way that aggravates its defects. The core bars are coated with straw ropes, which may be more or less soft and loose, coated with loam more or less soft and wet, and sprinkled with sand.

If not very carefully wedged, these bars will rise; and they are seldom stiff enough to resist the upward pressure of the molten metal. The usual spring at the centre for the core of an 8-in. pipe is 1/16-1/8in.; or as much as 3/16 in. with a 6-in. pipe. The metal, poured in from the upper end, first fills the lower section of the mould; and as it rises round the core to fill the upper section, its weight springs the bar upward to the extent indicated, making the casting thicker at the lower, and thinner at the upper side. The denser, hotter, and purer metal fills the lower portion; the impurities naturally floating upward to settle in the thinner metal as it cools. Here gather portions of the sand coating of the mould; while the bubbles of the metal, caused by the development of gas from the vegetable matter of the loam, and from its dampness, tend to perpetuate themselves in blisters and air cells.

The usual defects in these cheap castings are, therefore, inequality in thickness, air cells and blisters, sand holes, cold chutes from chilled metal, and mixtures of sand and iron. Such pipes are also frequently out of line, from the effect of unequal contraction. Pipes of this description are peculiarly liable to corrosion; containing as they do mixtures of metal of different densities, together with much graphite. The duration of such pipes in the ground is largely affected by the amount of disturbance they receive. If well laid at a good depth, and thoroughly backed, they may continue serviceable for many years; but their defects are likely to become suddenly prominent upon comparatively slight external interference. In favourable circumstances they may last more than thirty years; but the majority, if tested after less use, will show flaws that would have ensured their rejection if detected when new.

Gruner has lately published the results of a year's researches into the comparative oxidisability of cast iron, steel, and soft iron, under the influences of moist air, sea water, and acidulated water. Having done justice to the earlier labours of Mallet, Phillips and Parker, he explains the arrangements made to secure a perfectly fair trial. The following results were obtained. The experiments with moist air are still proceeding; but so far, it was found that in 20 days the steel plates lost 3-4 grm. for every 2 sq. decimeters of surface. Chrome steel rusted more, and tungstated steel less, than the ordinary carburetted steel. Cast iron lost only about half as much as the steel, and spiegeleisen less than grey iron. Sea water dissolves iron rapidly, and acts upon it more powerfully than on steel, most powerfully of all upon spiegeleisen. In 9 days, the steel plates with 2 sq. decimeters of surface lost 1-2 grm., while Bessemer metal lost 3.5 grm., phosphorised iron 5 grm., and spiegeleisen 7 grm. Tempered steel was less affected than the same steel twice annealed, soft steel less than chrome steel, and tungstated steel less than the ordinary steel with the same proportion of carbon.

It is evident from these experiments that manganese sheets ought not to be used on the hull of a vessel. Acidulated water dissolves cast iron much more rapidly than steel, but not spiegeleisen. (La Metallurgy.)

In the rusting of iron there is formed, together with an evolution of hydrogen, which combines with nitrogen, forming a small quantity of ammonia, ferrous carbonate. This changes very quickly into ferric hydrate, mixed with ferrous oxide, and enclosing some unaltered ferrous carbonate. The presence in rust of ferrous oxide, carbonic acid, and ammonia, is thus explicable. Rust formed under water is, in consequence of the smaller amount of acid present, usually richer in protoxide of iron, and therefore a little magnetic, and of a deeper tint than that formed in air. It is accordingly assumed that the carbonic acid present in the atmosphere and in water acts, in the production of rust, similarly to those acids in which iron dissolves, the only difference being that, in the rusting of iron, the ferrous salt first formed changes, before it is dissolved, into basic ferric salt or ferric hydrate, which change is a natural result of the solution of iron in an insufficient quantity of the acid or water present, or both. The denser the iron, the smoother and more even its surface, the less is the contact between it and the attacking substances, and, under otherwise similar conditions, so much the better, of course, will it withstand rusting.