The effect of breeding from the white rot fungus Phanerochaete chrysosporium ME446 on performance for lignin mineralization was examined. This model for informed strain improvement without mutagenesis is based on abundant restriction fragment length polymorphisms (RFLPs). Under optimized conditions for lignin mineralization, extracellular manganese peroxidase (MnP) but not lignin peroxidase (LiP) could be detected, so measurement of LiP activity is not a valid assay for lignin degradation. Mineralization of 14C-labelled synthetic lignin (14C-DHP) was used to compare the performance of the wild-type strain ME446 with those of sets of progeny strains. Meiotic progeny from strain ME446, heterokaryotic progeny of crosses between such strains, and meiotic progeny of one heterokaryotic strain were examined. In each case, a minority of strains performed more efficiently than the parental strain ME446. The greatest range of lignin-mineralization performance (70-fold) was found within the set of initial progeny of ME446 and the narrowest was within the set of secondary homokaryotic strains. This is consistent with the view that a moderate number of determinants contribute to lignin mineralization performance. However, performance did not correlate with the possession of any single allele of those for 38 previously defined RFLP markers. The results show that lignin mineralization performance can be improved by cycles of crosses and fruiting, without mutagenesis.