Summary. Spontaneous growth of erythroid progenitor cells is one of the hallmarks of polycythaemia vera (PV).In vitro clonogenic assays can be used to support the diagnosis of PV. However, clonogenic assays are difficult to standardize, and their diagnostic value is accepted as a minor criteria for PV only in the case of positive results. Clonogenic assays need to be validated and normal ranges established carefully before they can be used for diagnostic purposes. In a retrospective study, we analysed the results of 146 consecutive clonogenic assays performed in patients with erythrocytosis between January 1992 and December 2000. Normal ranges were established on the basis of the results from 81 controls. The charts of the 146 patients were reviewed and the patients were allocated to the following diagnostic groups: PV (n ¼ 38) and non-primary erythrocytosis (NPE, n ¼ 65) according to the criteria of Pearson & Messinezy. In total, 43 patients were excluded from the study because of missing data. The clonogenic assay results were correlated with the final clinical diagnosis of the patients. At a cut-off value of 5 erythroid colonyforming unit colonies per well, 29 out of 38 patients with PV showed positive results, and 9 out of 38 patients showed false-negative results. In the NPE group, there were 64 out of 65 real-negative results and 1 out of 65 false-positive result. Thus the positive predictive value of the clonogenic assay for the diagnosis of PV was 97%. We conclude that clonogenic assays, in the case of positive results, and if carefully validated, are very useful and reliable tools for the diagnosis of PV.