The use of critical effects in the determination of occupational expos
ure limits (OELs) in Sweden is subjected to a statistical study. Many
of the present OELs are high in relation to known noeffect levels and
effect levels, and the degree of protection has a surprisingly weak co
rrelation with the seriousness of the adverse effect. Several proposal
s for improved procedures are put forward. One of these is to suppleme
nt the concept of critical effects with that of dominant effects. A do
minant effect of a substance is a health effect that is at some concen
tration the most serious health effect.