The accuracy of the Coulter Counter Model S Plus III in determining lymphocyte percentage was assessed. In 'routine' adult blood counts accuracy was satisfactory but in infectious mononucleosis, and in chronic lymphocytic leukaemia and other lymphoproliferative disorders, there was a high rejection rate and an often considerable underestimation of the lymphocyte percentage. The S Plus III lymphocyte percentage was of no use in neonates because of inaccuracy and a high rejection rate. In infants the relationship with lymphocyte percentages based on manual differential counts was non-linear, and there was underestimation of some high percentages. The white cell size histogram was useful in altering staff to abnormalities, and also suggested that lymphocytes of neonates and infants were larger than those of adults, with plots being very similar to those of some adults with lymphoproliferative disorders. The S Plus III lymphocyte percentage was useful in reducing the need for differential counts in adult patients but did not replace examination of a blood film or a manual lymphocyte percentage in patients with chronic lymphocytic leukaemia or a lymphoproliferative disorder.