Results are given for stress corrosion studies carried out on a number of aluminum alloys. Testing was done at Aruba, Netherlands Antilles, Kure Beach, N. C., and Richmond, Va. These sites provided environments that ranged from tropical marine to mild industrial. Failure times are compared with laboratory and service results. One aim was to study the effect of prolonged aging at ambient temperatures. In aluminum-magnesium alloys this aging may produce a metallurgical structure which makes these alloys subject to stress corrosion cracking. Specimens aged at higher temperatures do show this response in laboratory and atmospheric tests. From these accelerated data, Arrhenius plots predict these alloys should become susceptible within two to five years if aged at ambient temperatures. However, test results do not bear this out. It is possible to draw erroneous conclusions from only one type of test. For at least one alloy, the rate of failure is much more rapid in the atmosphere than in an accelerated laboratory test. This has been attributed to an interaction between specimens tested in a communal solution. In conclusion, either a laboratory or an atmospheric test is, in itself, insufficient to characterize the stress corrosion behavior of an engineering material. These two tests give data that are complimentary instead of supplementary. Both types are necessary to evaluate the potential hazards of a material having some degree of stress corrosion susceptibility.