A five year prospective study, documenting respiratory symptoms, measurement of lung function and assessment of airborne dust concentrations, was performed in eight Lancashire textile mills. Six mills processed cotton and two mills processed man-made fibre. Provisional analysis of this cohort (Fletcher et al, 1999) demonstrated overall low levels of lung function decline, but with higher levels of decline witnessed in man-made fibre workers (-17mls/yr) and waste cotton workers (-21mls/yr) compared to blend (+4mls/yr) and medium cotton exposed workers (-5mls/yr). The reason for this could relate to a direct effect of man-made fibre dust, or be due to the fact that some workers appear to move from the cotton textile industry to man-made fibre manufacture. To study this in detail we have performed a subsidiary analysis of this data investigating lung function decline in man-made fibre workers only (n=652). Previous cotton dust exposure and current measures of man-made fibre exposure are included in the data set as separate predictor variables and its effect on the measured lung function decline compared to cotton naive workers. A random effects regression model was fitted to the data, firstly overall, then by ethnic group and age group (<35yrs and >35yrs). The statistical analysis demonstrated no association between lung function decline and measures of both past cotton dust exposure and current man-made fibre exposure. It appears that susceptible workers moving from cotton to man-made are not responsible for the excess in lung function decline demonstrated within this exposure group. Consistent statistically significant predictors of lung function decline were age and sex, which have previously been documented.