About 80% to 90% of females are informative for X- inactivation/methylation analysis with the probe M27 beta, which would therefore seem attractive in assessing clonality in hematologic cell populations. Eighteen acute lymphoid or myeloid leukemias, three chronic lymphocytic leukemias, and three chronic myelogenous leukemias as well as 12 malignant non-Hodgkin's lymphomas mostly showed extremely skewed clonal X inactivation (median allelic cleavage ratio [ACR] of unmethylated/inactive M27 beta alleles was 50, range 1 to 100) or hypermethylation of both alleles. Two lymphomas showed random M27 beta X inactivation but clonal antigen-receptor gene rearrangements. In normal peripheral blood leukocytes from 105 healthy females aged 2 to 96 years, the median ACR was 2 (range 1 to 100). Thus, it was significantly lower than in the leukemias (P = .0001, Mann-Whitney test), but extremely skewed patterns (ACR > 10.8, ie, >80th percentile) were seen not only in the leukemias but also in 21/105 samples (20%) of normal leukocytes, and significantly more frequent in a population of elderly women (aged 75 to 96 years) compared with healthy children (aged 2 to 8 years) and younger women (aged 20 to 58 years) (P =.00125; chi 2 test). We conclude that in a population of cells derived from the hematopoietic system where clonality is uncertain, skewed M27 beta patterns are not reliable indicators for the presence of a clonal neoplastic disorder. The basis for severe X-inactivation skewing is unclear at present, but this finding raises interesting questions regarding the composition of the hematopoietic stem cell pool.