Information on the kinetics of bone marrow (BM) myeloid precursors (BMMP) is required for integrating cancer chemotherapy with granulocyte-macrophage colony-stimulating factor (rhGM-CSF), with the aim of reducing neutropenia. Using bivariate flow-cytometric analysis of the in vivo incorporation of bromode-oxyuridine (BUDR) vs DNA content we have studied the kinetics of BMMP in 21 patients with SCLC during the first of six chemotherapy courses (etoposide, epirubicin, and cis-platinum, days 1-3, every 21 days), given alone (eight patients) or followed by rhGM-CSF (10 micrograms/kg/day s.c., days 4-14) as BM rescue (eight patients) or both preceded (days -17 to -7, as BM priming) and followed by rhGM-CSF (five patients). At 11-14 days after the start of these therapies there was an increase in the baseline proliferative activity of proliferating BMMP and a shortening in the time needed by the metamyelocyte to mature and to leave the marrow. Both effects were greater and were maintained to a significantly greater degree a week later in patients who received chemotherapy plus rhGM-CSF rescue than in those who received chemotherapy alone or rhGM-CSF priming alone. At day 11-14 the pretreatment median cell production rate of pBMMP was increased by 340%, 150%, and 183% and the maturation time was reduced by 80%, 45%, and 57%, respectively, in the three groups. A week later, the corresponding figures were 206%, 111%, and 157% and 50%, 18%, and 45%. Hence, an identical rhGM-CSF schedule is more effective in increasing the neutrophil production by BMMP when given following chemotherapy as BM rescue than before it as BM priming. In both the rescue and the priming schedule, the increase in proliferative activity of BMMP just at the end of rhGM-CSF stimulation was linked to both an increase in the labeling index and a reduction in duration of S-phase (TS), while a week later it was linked solely to reduction in TS. This could actually reduce one of the two kinetic targets of subsequently administered cytostatics, i.e., a high LI and a long time spent in S phase. From this study, accurate kinetic data can be obtained with the in vivo BUDR technique that are useful in scheduling rhGM-CSF.