The accident rates of piecework and salaried workers from the furniture industry were compared in five job categories. A significant difference in relative risk was found between jobs: labourers 36.3%, upholsterers 17.2%, polishers 10.0%, cabinet-makers and carpenters 8.5% and sawyers 7.4% (chi-square = 288.2, p < 10"^). There was no significant difference between piecework and salaried workers on the average risk: 19.5% and 13.9% respectively (chi-square = .26, p < .65). However, when wage modes were compared within each job category, a significant increase in relative risk was found for the piecework labourers (31 observed, 16 expected; chi-square 21.4, p < IO"'*). By contrast, a significant decrease in relative risk was found for the piecework cabinet-makers and carpenters (9 observed, 21 expected; chi-square 7.3, p < .05). The relative risk was not significantly different between wage modes for the other three job categories. When injury sites were compared, piecework labourers were also found to have an excess of accidents involving lower-back injury (5 observed, 1 expected; chisquare 20.5, p. < 10~4); the wrist (7 observed, 1 expected; chi-square 29.5, p < 10_<*); and the fingers (11 observed, 4 expected; chi-square 14.6, p < 10"^). We conclude that the relationship between wage modes and accident rates interacts with job categories producing (depending on the job) a significant increase or decrease in the relative risk. These results are discussed in terms of the intrinsic market of the incentive. For the labourers they are strictly quantitative and incorporating; in the job content of the cabinet-makers and carpenters, they involve quality controls on the finished product. We, therefore, define the labourers as structural concentrators of risks both through their average relative risks and the negative impact of piecework in their case. We believe that more attention should be paid to the analysis of accident rates by making internal-job comparisons. We also suggest that a careful analysis of the structural concentrators of risks in the division of labour should be approached by task redesign which would tend to eliminate incentives solely based on the quantity of production.