In an experimental study of adaptation to negative pleiotropic effects of a major fungicide resistance mutation in the filamentous fungus Aspergillus nidulans we have investigated the relative effectiveness of artificial selection vs. natural selection on the rate of compensatory evolution. Using mycelial growth rate as a fitness measure, artificial selection involved the weekly transfer of the fastest growing sector onto a fresh plate. Natural selection was approximated by transferring random samples of all the spores produced by the mycelium. Fungicide resistant and fungicide sensitive haploid and diploid strains were used in an evolution experiment over 10 weekly transfers, which is equivalent to 1200 cell cycles. Two different environmental conditions were applied: a constant fungicide-free environment and a weekly alternation between presence and absence of fungicide. Results show that for all strains and conditions used the transfer of a random sample of all spores leads to more rapid adaptation than the transfer of the visually ‘fittest’ sector. The rates of compensatory evolution in the constant and the alternating environment did not differ. Moreover, haploid strains tend to have a higher rate of adaptation than isogenic diploid strains.