We describe two assays to detect the action of colony-inhibiting cells. In the first assay, we used a simple density separation technique to remove dense neutrophils (PMN) from suspensions of blood and of bone marrow cells prior to culture in semisolid agar. Conditions were arranged to ensure that control suspensions of unseparated cells and test suspensions of buoyant mononuclear cells differed only in their content of neutrophils. The control and test suspensions contained equal numbers of mononuclear cells (and granulocyte precursors). Colony (and cluster) formation was invariably enhanced in neutrophil-depleted cultures of normal cells. In the second assay, dense PMN, treated by an adherence separation procedure, were recovered, and the non-adherent dense PMN were added back to PMN-depleted cultures. A reproducible dose- related decrease in colony (and cluster) formation to basal levels resulted. The inhibitory effect was identical when the PMN were added directly to the culture (overlayer) or to the underlayer. In PMN- depleted cultures obtained from patients with leukemia and other hemopoietic disorders, neither colony nor cluster formation was enhanced, and sometimes it was reduced. When we compared the effect of adding patient and normal non-adherent PMN to target cultures of normal and patient PMN-depleted cells, some leukemic PMN were noninhibitory. Our results suggest that abnormalities of cellular interactions in vitro detected in the first assay may have more than one explanation, as shown when they are subjected to the closer scrutiny possible with the second assay.