Previous studies indicate an increased prevalence of knee disorders in some occupations possibly related to kneeling working positions. The purpose of the present study was to examine the relationship among knee-straining work, self-reported knee-complaints, and physical signs of knee disorders. The duration of knee-straining work was estimated from videotapes of representative work tasks. Floor layers (n = 133), carpenters (n = 506), and compositors (n = 327) aged 26 to 72 years without previous acute knee traumas were examined in a cross-sectional study by questionnaire. A stratified random sample of the questionnaire responders; 67 floor layers, 127 carpenters, and 101 compositors had independent double examinations for physical signs of knee disorders. The videotapes showed that knee-straining work constituted 56% of working time for floor layers, 26% for carpenters, and none for compositors. The prevalences of self-reported knee-complaints were positively associated with the amount of knee-straining work and were significantly different for the three trades. Floor layers and carpenters who were presently working in their trade had a higher prevalence of knee complaints than floor layers and carpenters who had left their trade. Age, seniority, weight, body mass index, smoking, and knee-straining sports activity had no significant effects. The clinical study showed a positive association for knee-straining work, hyperkeratosis, and bursitis. A similar pattern was found for signs of intraarticular knee disorders by one physician but not by another. The reproducibility of these signs was low. More studies are needed to define clinically important knee disorders for epidemiological studies.