Abstract The human placenta is capable of producing a variety of haematopoietic growth factors in vitro. It is not clear, however, whether the placenta produces such factors in vivo and if so, whether placental production of haematopoietic growth factors has a physiological role in fetal haernatopoietic development. As a step toward making this determination, we assessed whether the onset of placental production of granulocyte colony-stimulating factor (G-CSF), in vivo, coincides with the onset of granulocytopoiesis in the developing fetus. To make this assessment, we obtained human placentae between 10 weeks of gestation and term and studied production of G-CSF in several ways. First, we sought to determine whether the onset of production of G-CSF mRNA in the placenta immediately precedes the appearance of neutrophil development in the fetus. Second, we assessed the effect of gestational age on the capacity of the placenta to generate G-CSF in vitro, by incubating cubes of placenta, with or without including interleukin-la (IL-1α) in the culture media, and quantifying G-CSF in the cell culture supernatants 24h later. Third, we assessed the rate of G-CSF production by the placenta, by perfusing two normal, term placentae using a membrane-oxygenator system, and quantifying G-CSF, at intervals, in the perfusates. We found: (1) no evidence that placental production of G-CSF is involved in regulating granulocytopoiesis in the fetus, (2) that the healthy placenta contains little or no G-CSF mRNA in vivo, (3) the placenta at term has a far greater capacity to produce G-CSF, when stimulated, than does the placenta before term, and (4) that although the placenta does not normally produce G-CSF in vivo, it has the capacity of generating very large quantities of G-CSF continuously over at least several days.