With the aim of developing a standardized program of clonogenic cultures, a multicenter intercomparative study of human CFU-GM, BFU-E and CFU-GEMM cultures was conducted. Aliquots of fresh mononuclear cord blood cells, as well as uniform culture materials and instructions for cell culture and for colony scoring were distributed to 28 national laboratories involved in hematopoietic research and transplantation. High interlaboratory coefficients of variation (CV) in the reported number of progenitors were found in our first intercomparative study (range 67-231%). To investigate the relevance of colony scoring in variations of the reported colony numbers, participants were invited to attend a meeting where a single culture dish was scored. In this case, the CVs ranged from 31% to 81%. A subsequent intercomparative assay was then conducted, and significant reductions in the inter-laboratory CVs were obtained with respect to the first study (CVs for colonies grown with two different media: CFU-GMs, 48% and 55%; BFU-Es, 70% and 62%; CFU-GEMMs, 70% and 51%; respectively). In most instances CVs were not significantly different from those obtained in the single plate scoring study, suggesting that the scoring process was the most relevant parameter accounting for variations in the reported colony numbers.