Current experimental models are poorly suited to study the early biochemical and molecular events of the lineage determination process in myeloid progenitor cells. Viable lineage-committed precursors cannot be identified until after they have expressed their mature phenotype and these precursors cannot be grown to large number while lineage is committed but still immature. Recently, we have identified stable sublines of the HL-60 human leukemia cell line which differ from each other in that they selectively differentiate to either neutrophils (UR-1-4), monocyte/macrophages (MRI), eosinophils (clones 2 and 15), or mixtures of two (clones 7 and 8) or of all three lineages (UR-1-2) when stimulated to mature with butyric acid under identical conditions. Characterization of these sublines provided evidence that the expression of lineage in HL-60 cells is a multistep process and that the lineage tendencies (lineage direction) the clones exhibited when cultured with butyric acid represent a step in that process earlier than irreversible lineage commitment but later than the multipotential wild type HL-60 cells. First, treatment of these sublines with compounds that induce differentiation of HL-60 cells to specific lineages (dimethylsulfoxide, neutrophil; 1,25-(OH)2 vitamin D3, monocyte), generally induced differentiation to the lineage associated with that inducer rather than the butyrate-associated lineage. Second, culture of neutrophil or monocyte-directed sublines in medium of elevated pH for two months leads to the development of eosinophils. Culturing the sublines first in butyric acid for variable lengths of time and switching to either DMSO or VD3 indicated that irreversible lineage commitment develops on a time course similar to the development of the commitment to mature. Markers of monocytic and eosinophilic differentiation could not be simultaneously demonstrated in single mature cells, consistent with the phenomenon of lineage fidelity. In addition, several assays were validated that could reliably classify mature HL-60 cells to their lineage. The collection of these sublines appears to constitute a model system with well-defined behavior with respect to the early events of lineage determination that can be grown to quantities sufficient for biochemical and molecular analysis. Exploring the differences between these clones may provide a new way to examine the early events of the lineage development process in myeloid cells.