Yeast suspensions were analysed by flow cytometry after dye staining for determination of total and viable cell densities. Results were comparable to traditional colony counting and, in addition, provided further information on the percentage of total cells that were viable. The flow cytometric methods provided results within 20 min whereas colony counts were not available until 36 h. We evaluated a number of fluorescent dyes: ChemChrome Y (CY), oxonol (Ox), propidium iodide (PI), Fungolight and rhodamine 123, for accurate determination of viability of industrial yeast cultures and freshly re‐hydrated high activity dried yeast (HADY). PI, Ox and CY gave the most conclusive live/dead discrimination and were the simplest to use. Culturing after dye staining and cell sorting demonstrated that the yeast remained viable after cell sorting and incubation with PI, CY or Ox. The methods, therefore, permit physical selection of individual yeast cells from populations of mixed viability. Sorting demonstrated that PI stained non‐culturable cells whilst CY stained culturable cells. Analysis of yeast stained simultaneously with CY and PI or with Ox and PI demonstrated that PI and CY assays were in mutual agreement with respect to viability assessments. The Ox assay was in agreement with CY and PI for live/heat‐killed mixtures. However, for re‐hydrated HADY, Ox stained a significantly (P⩽0·05) higher proportion of cells than did PI. © 1998 John Wiley & Sons, Ltd.