Micronuclei (MN), used as a biomarker of effect in exposure to genotoxic carcinogens, derive from chromosomes and chromosomal fragments lagging behind in anaphase. The two types of MN are usually distinguished from each other by centromeric fluorescence in situ hybridization (FISH), centromere-positive (C+) MN representing entire chromosomes and centromere-negative (C−) MN chromosomal fragments. The incorporation of various types of chromosomal fragments and chromosomes and chromatids to MN is still poorly understood. We used directly labelled pancentromeric and pantelomeric DNA probes to examine the contents of MN in cultured binucleate lymphocytes of four unexposed, healthy subjects (two men and two women) 35–56 years of age. The presence and number of telomeric and centromeric signals was evaluated in 200 MN (50 MN per subject). These data were used to estimate the proportion of MN harbouring terminal/interstitial fragments, acentric/centric fragments, chromatid-type/chromosome-type fragments and entire chromatids/chromosomes. The majority of the C+ MN (96% in men and 86% in women) found contained telomeric (T+) sequences. Most of the C+ T+ MN had one centromere and two or one telomere signals, suggesting that single chromatids were more frequently involved in MN than both sister chromatids. Among the C− MN, telomere signals were found in 91% (men) and 79% (women), showing that fragments in MN were mostly terminal. Most C− T+ MN had one telomere signal, indicating higher prevalence for chromatid-type than chromosome-type terminal fragments. Combined centromeric and telomeric FISH is expected to increase the sensitivity of detecting exposure-related effects, when the exposure induces specific types of MN and its effect is low. This approach could particularly have use in discriminating between MN harbouring chromatid- and chromosome-type fragments in studies of human exposure to chemical clastogens and ionizing radiation. [ABSTRACT FROM AUTHOR]