The Authors' reply

higher than that for Knudson et al or Cotes et al. By age 60, the Miller et al prediction is 10% or 23% higher. If the prediction equations of Knudson et a18 had been used, probably none of the predicted values for forced expiratory volume in one second (FEV,) or forced vital capacity (FVC) would have been significantly reduced, even without adjustment for smoking habits. If the equations of Cotes et a19 had been used, certainly none would have been significantly reduced. In the present paper, Kilburn et aP compared the pulmonary function of 17 male current smokers with radiographic changes with that of 39 male current smokers without radiographic changes. In their earlier paper, those with radiographic changes were reported to be on average six years older, but this information is not presented again. The bias in the prediction equations for pulmonary function invalidate this comparison.

higher than that for Knudson et al or Cotes et al. By age 60, the Miller et al prediction is 10% or 23% higher. If the prediction equations of Knudson et a18 had been used, probably none of the predicted values for forced expiratory volume in one second (FEV,) or forced vital capacity (FVC) would have been significantly reduced, even without adjustment for smoking habits. If the equations of Cotes et a19 had been used, certainly none would have been significantly reduced.
In the present paper, Kilburn et aP compared the pulmonary function of 17 male current smokers with radiographic changes with that of 39 male current smokers without radiographic changes. In their earlier paper, those with radiographic changes were reported to be on average six years older, but this information is not presented again. The bias in the prediction equations for pulmonary function invalidate this comparison.

Other issues
The text refers to Botham and Holt'0 as showing that "fibreglass also causes peribronchiolar fibrosis by inhalation." That paper does not mention peribronchiolar fibrosis, and indeed it would be unlikely to do so because the study was primarily concerned with inhalation of glass powder for one day, followed up for one month, with some comparison with the effects of exposure to fibrous glass. Kilburn et al2 failed to reference any of the long term inhalation studies of fibreglass, none of which has shown any evidence that fibrosis is caused by fibreglass exposure. I' 20 The participants in this study were 284 volunteers from the "500 workers with 20 years of exposure to fibreglass." It is difficult to understand how the average duration of exposure to fibreglass could have been 19-9 years (table 2 from Kilburn et al).2 Was the selection criterion based on duration of employment rather than on duration of exposure?
The non-smokers had higher prevalences of bronchitis and of asthma than did the smokers, with the ex-smokers having the lowest prevalence. This is so different from other studies that a discussion of this would have been appropriate. The only explanation given is that this "may reflect current and ex-smokers with seniority relocating into jobs with less exposure to fibreglass." This is hardly an adequate discussion. It is also irrelevant if the true selection criterion was based on duration of employment. CHARLES  We would like to reply to Rossiter as follows: No dual publication Rossiter begins, in his first sentence, with the accusation of dual publication, which is both unkind and wrong. The proceedings of the second Califomia thermal insulation conference were compiled and provided to the participants-not published.
No fibrosis from cigarette smoking alone He proceeds by renewing his contention that Weiss, using minifilms read by magnification, found irregular opacities. Weiss did not use Intemational Labour Office (ILO) criteria for film interpretation, and his work was done in 1972 not 1991 as implied by Rossiter's reference5 and has not been replicated. Finally he apparently failed to appreciate that silver halide grains may be magnified to haziness in minifilms. The definite study in several thousand subjects with exposure to asbestos showed that smoking apparently enhances the opacities but does not create them.' Differences between 1989 and 1991 radiographic findings The differences in radiographic findings in Rossiter's table are because the preliminary report was based on my readings of the x ray films only. When Power's readings, which showed more positives and slightly higher average profusions were included, the final percentages were slightly higher.
Pulmonary function: predicted equation The suggestion that the pulmonary function prediction equations from Miller et al of a Michigan population and Morris' Oregon population study are "supranormal" was settled 20 or more years ago. The pulmonary function tests of neither Knudson et al, nor Cotes et al were free enough of technical faults to be representative, let alone being normal. Of course, if one lowers the predicted far enough, the pulmonary function tests of many working populations will exceed the predicted.
There is a bias in pulmonary function comparison if height, age and duration of cigarette smoking are not adjusted. As table 3 of our paper makes clear, these data were adjusted and are presented as % predicted. This, as Rossiter knows, adjusts for the six year difference in age (which was actually four years in the present table 3).
Animal studies I am not privy to exclusive informa-tion implied by Rossiter concerning the Botham and Holt findings but their descriptions are quite clear and even their title (Rossiter Ref`0) mentions comparison of glass fibre. Other animal studies seem to be negative but each of his 10 references would require detailed discussion, which does not seem relevant to the major issues.
Our error The minimal criteria for workers entry into this study .was actually 15 years of fibreglass exposure in the plant, which yielded 19 9 years as the average duration of exposure. The planning committee consists of people from varied backgrounds, including also employee and employer representatives, to ensure a wide ranging scientific programme.
The aim is to bring experience from research to practice and from practice to research. Invited lecturers will present both the scientific state of the art and practical applications and there will be workshops, oral sessions, and posters.
Topics will include infection hazards, dermatoses, chemical health hazards, physical risk factors, occupational accidents, ergonomics, developing countries, job content and skills development, work organisation and development, the client/patient and the health care worker, and gender perspectives.
There will be visits to hospitals and health care institutions in the Stockholm area. Some one-day courses on topics related to the conference will be arranged.