AbstractThe problem of interpreting lung‐function measurements in industrial workers is examined. The data under discussion pertain to FEV1and FVC measurements in smoking and in nonsmoking groups of grain‐elevator workers in British Columbia and of workers in Vancouver City Hall. Initial observations have now been enriched by longitudinal follow up data on the same groups after three and after six years. It is shown that interesting selection phenomena, favouring “fit” individuals, take place over time, with regard both to lung symptoms and lung functions. Thus cross‐sectional and longitudinal studies refer to somewhat different populations. It also appears that longitudinal studies are considerably more sensitive to identifying cumulative lung damage than are corresponding cross‐sectional studies. The nonlinearity of the effect of age on lung functions is noted in the longitudinal data in a number of cases, lending support to the hypothesis of association between quadratic age effect and cumulative exposure to l