328 X-linked recessive lethal mutations induced in late spermatids by hycanthone methanesulfonate were tested for coverage by duplications that comprised, in total, about 24% of the euchromatic X chromosome; 78 lethals appeared to be covered. Crossover localization tests of a random sample of 38 non-covered lethals revealed 4 chromosomes carrying a lethal within a duplicated segment. Lethals localized to a particular region were crossed to reference deficiencies and single-locus mutations, and inter se, to ascertain their genetic extent. The proportion of multi-locus deletions among these 78 covered and 4 non-covered lethals was 3/48, 1/10 and 13/24 for the distal, medial and proximal regions, respectively. A storage period of 9 days did not noticeably influence these proportions. In the sample of 38 non-covered lethals, and among 17 of the covered single-site lethals, 4 cases of strong crossover suppression were detected. Comparison of these results with data obtained with other mutagens suggests that induction of multi-locus deletions, and possibly of other types of chromosome rearrangement, could in part depend on other mechanisms than those acting in the formation of translocations and chromosome loss. For the purpose of mutagen testing, these findings imply that, in Drosophila, results in the regular genetic tests for chromosome breakage events do not always accurately predict the capacity of a mutagen to induce multi-locus deletions. This is of importance since transmissible multi-locus deletions have been considered a significant source of genetic damage in man.