In subjects without a history of childhood respiratory problems, ventilatory function was closely related to total cigarette consumption. After accounting for pack-years of smoking, independent effects of age, sex, duration of smoking, and current smoking status could not be demonstrated. The presence of chronic productive cough, however, proved to be a major correlate of the levelof function, even after accounting for pack-years (1). A history of childhood respiratory trouble (CRT) was shown to be an independent predictor of low lung function in adults, and a model could be created suggestingthat CRTleads to an increased susceptibility to later noxious influences. Admittedly, the possibility of preferential recall of childhood problems by adults with respiratory disease made this only an attractive hypothesis rather than a firm conclusion (2). Ventilatoryimpairment was also shown to be related to allergy skin test reactivity (atopy), blood eosinophilia, and serum immunoglobulin E (IgE) level. The relationship to atopy was not close, but it was thought that this could be a result of loss of allergy skin test reactivity with age (3). Eosinophilia appeared to be a more important correlate, but primarily in nonsmokers; significant additive effects of smoking and eosinophilia could not be demonstrated (4). Serum IgE was also shown to be related to ventilatory impairment, but in this case the relationship was significant only when the functional impairment was associated with symptoms suggesting an asthmatic or at least a bronchitic type of disorder (5, 6). On the basis of these observations, we suggested the hypothesis that different types of obstructive airway disorders in the population might have quite different pathogenetic mechanisms (7). One (chronic asthmatic bronchitis) appeared related to immunologic factors and to features suggesting an asthmatic predisposition, reminiscent of the" Dutch hypothesis" proposed by investigators from the Netherlands many yearsago (8). It appeared to be the predominant type of airway disorder in nonsmokers. The second type of disease, with a relatively insidious development of ventilatory impairment, appeared to be a more direct consequence of smoking and could not be shown to be related to immunologic factors. It appeared to be the overwhelmingly important disorder in male smokers and probably represented the emphysematous type ofchronic obstructive pulmonary disease (COPD). Newer longitudinal analyses of data from our own population sample appeared compatible with this hypothesis. Continuing male smokers with mild obstructive impairment showed a relatively predictable, rapid, subsequent decline in lung function, an effect not demonstrable in other groups (9). This led us to question whether the functional abnormalities in women and in nonsmokers of both sexes might represent a disease with a less predictable and fluctuating (or asthmaticlike) course. Even more recent observations indicate that subjects who had a clinically significant degree of airway obstruction on enrollment can be divided into groups with vastly different subsequent courses. Those with features of chronic asthmatic bronchitis (including a known diagnosis of': asthma) had an almost stable course over the following 10yr, whereas those without any features of asthma showed a decline in lung function of 70 mllyr and a relativelyhigh mortality (10). In viewof these findings, which strongly suggest that there are two quite differ-