DNA alkylation by four acridine-linked 'DNA-targeted' aniline mustard derivatives has been studied by 32P-postlabelling. P1 nuclease digestion proved much more efficient than butanol extraction for enhancing the yield of adducted bases for these somewhat hydrophilic compounds. The yield of adducts was maximal after approximately 4 h digestion with micrococcal nuclease/spleen phosphodiesterase and remained relatively constant after that up to 24 h, suggesting that the adducts formed are stable under these conditions. There was some variation in the rates of phosphorylation of the adducts by T4 polynucleotide kinase, with optimal labelling generally occurring after 1 h. The (CH2)5O-linked half-mustard derivative 1 gave five nucleotide 3'-diphosphate adduct spots with calf thymus DNA. Two of these were identified as the adenine N1 and N3 adducts, corresponding to those previously identified as the main base adducts formed by 1 following acid digestion studies. The corresponding full mustard also gave five adduct spots. In contrast, the (CH2)3-linked half-mustard 3 gave only two adduct spots, the most intense of which was identified as a guanine adduct. The corresponding full mustard 4 gave three adduct spots, two of which were identified as guanine adducts. These results agree well with those obtained for the same compounds by the more tedious methods of acid digestion to base adducts, followed by isolation on HPLC, and show that the technique of 32P-labelling can be usefully applied to the study of alkylation of DNA by this class of 'targeted' mustards.