In this study low leukocyte counts were carried out using several methods. The results obtained by the visual method, the Coulter Model "S" using "whole blood", and the Coulter DN, show close correlation, but discrepancies exist when the Coulter S uses "prediluted blood". This is explained by excessive bubbles. This does not allow the Coulter Model S to be used with confidence for low leukocyte counts if "whole blood" is not available.