The concept of critical loads has been widely accepted as a basis for desig
ning control strategies to reduce regional air pollution. Data from the "10
00 lake survey" carried out in Norway in 1986 were used to derive a dose-re
sponse function for the probability of damage to fish populations as a func
tion of the critical load exceedance, by means of a logistic regression mod
el for fish damage. A corresponding function based on a fish damage databas
e (brown trout, arctic char, and perch) and the national critical load data
base for Norway compared well with the derived dose-response function. When
the critical load is not exceeded there is only a very small probability t
hat the fish population will be damaged, but when the critical load is exce
eded the chance of damage increases with the amount of exceedance. At criti
cal load the probability of damage to fish populations is about 20. These r
esults confirm that the critical load concept is a strong tool for estimati
ng the extent of biological damage caused by long-range transported air pol
lution, and that prognoses based on this method can give reliable estimates
of the extent of fish damage under future scenarios of acidic deposition.