Neutrophil alkaline phosphatase activity was estimated in 194 patients; 59 cases of chronic myeloid leukaemia (CML), 42 cases of polycythaemia vera (PV), 24 cases of primary myelofibrosis, 7 cases of idiopathic thrombocythaemia, 6 cases of leukaemoid reaction, 19 cases of secondary polycythaemia (PS) and 37 cases of the primary myelodysplastic syndrome (MDS). According to NAP activities the groups proved to be separate entities (p less than 0.00025). The incidence of decreased NAP score in the CML group was 85% and differed significantly from the other groups as a whole, as well as separately (p less than 0.001). The MDS group, the only group besides CML that showed decreased scores, also differed significantly from the others (p less than 0.001). The PS group, nearly always showing normal scores, differed significantly from the PV group (p less than 0.0052). A method evaluating single cell NAP activity proved superior to the score method in discriminating between the different groups. Thus, the incidence of decreased activity in the CML group was 93% compared with 85% by the score method and the incidences of increased activity in the PV, MP, IT, and LR groups were 79% to 100% compared with 25% to 67% by the score method. The latter difference was statistically significant (p = 0.029).