Prediction equations for respiratory function tests were obtained by multiple regression of data from 263 healthy males. The material was evenly distributed in the ages 20-70 years, about one third each of non-smokers, smokers and ex-smokers. Measurements were done of lung volumes (with body plethysmograph), airways resistance, ventilatory capacity including flow-volume registration, gas distribution and closing volume, transfer factor and static elastic recoil pressures of the lung with calculation of static compliance. The parameters age, height, weight, years of tobacco smoking and grams of tobacco smoked each day showed significant correlation with the outcome of the test in most of the respiratory function tests. Therefore a set of basic regression equations including these parameters were calculated. In addition an 'extended' set of equations was calculated for prediction of some tests with inclusion of nonlinear terms and the parameter 'years of abstinence from smoking.' The reduction in variance which followed inclusion of tested parameters was moderate (20-69%). There were significant differences between results of the present study and several previously published regression formulas.