AbstractThe results of routine observations on the corrosion of the small reference specimens of ingot iron and of zinc exposed at the atmospheric testing stations of the Corrosion Committee all over the world are presented and discussed. The data, which relate to more than 20 sites and to up to 20 separate annual tests at each, provide a good indication of the effect of climatic differences on atmospheric corrosion.In Great Britain, where for most of the year the climate is sufficiently humid to promote corrosion, the determining factor for its severity is the sulphur pollution of the air. Tests made in collaboration with the Fuel Research Station at 16 sites where the atmospheric pollution is measured have shown an almost perfect correlation between the amount of sulphur dioxide taken up from the atmosphere by lead peroxide pollution gauges and the rale of corrosion, both for steel and zinc.In the absence of sulphur pollution, corrosion at the sites overseas is generally slight; in dry climates it is negligible. For example, tests at Delhi indicate that the resistance of the Delhi pillar to corrosion is to be attributed to the mildness of the corrosive conditions there rather than to any superiority of the iron itself.Corrosion on surf beaches in the tropics may, however, be exceptionally severe. At Lighthouse Beach, Lagos, Nigeria and at Kure Beach, N.C., U.S.A., the rates of corrosion for steel are about five times greater than those in the most corrosive industrial atmospheres in Great Britain. The severe corrosion at these places is associated with the salt spray from the sea, and the rate of corrosion falls off rapidly with increasing distance from the surf line.Preliminary tests on specimens exposed at different heights on the Sutton Coldfield television mast suggest that the effect of distance from the ground is secondary to the natural variations in rates of corrosion from one year to another.Some details of experimental technique are discussed. Work, in which American investigators have collaborated, has confirmed that at low copper contents variations of a few hundredths of one per cent, in copper content have a marked effect on the corrosion‐resistance of ingot iron. Because of the increasing difficulty in controlling the copper content of different batches of ingot iron, the small reference specimens used in the Corrosion Committee's most recent tests have been made of coppersteel instead of ingot iron. The copper‐steel now in use contains about 0·3% of copper; minor deviations from this figure should not affect the corrosion rate within the experimental error. It has also been found that, although it is advisable to stit the inhibited hydrochloric acid solution (Clarke's solution) used for derusting iron or steel specimens after exposure, failure to do so does not impair the accuracy of the results, at least for ingot iron or copper‐steel.