Circadian changes in in vitro pharmacodynamic effects of recombinant mouse interleukin-3 (rmIL-3), rm granulocyte-macrophage colony-stimulating factor (rmGM-CSF), and recombinant human G-CSF (rhG-CSF) were investigated in 418 male B6D2F1 mice. Seven distinct experiments were staggered from July to December 1991. All mice were standardized for 3 weeks with a lighting schedule consisting of 12 hours of light and 12 hours of dark (LD12:12). In each experiment, bone marrow was sampled from separate groups of nine to 10 mice each every 4 hours for 24 hours. Data were analyzed with analysis of variance (ANOVA) and Cosinor. This latter method computes the probability of rhythm detection and its parameters. Femoral myeloid progenitors were quantified using the colony-forming units granulocyte/macrophage (CFU-GM) assay in the presence or absence of recombinant CSFs. For each CSF, the number of colonies is a function of circadian time of bone marrow exposure (ANOVA and Cosinor; p < 0.0001) with the values at peak time being double those found at the trough. Peak CSF efficacy occurred at 3 hours after light onset (HALO, early rest span) irrespective of CSF type or dose. Furthermore, in the absence of any added CSF, the number of clusters varied significantly according to sampling time, with a similar peak at 3 HALO (ANOVA and Cosinor; p < 0.001). Further in vivo chronopharmacologic experiments are needed to assess the relevance of these in vitro rhythms in bone marrow responsiveness to hematopoietic growth factors.