We have previously shown that enriched preparations of oligodendrocytes from mature bovine brain or 30d rat brain contain 4-10% ganglioside-GD3+ glial precursor cells, which differentiate into astrocytes in culture. These findings are in contrast to those described by others in cultures of neonatal rat brain, which contain precursors that differentiate into both astrocytes and oligodendrocytes. We have extended this study to determine whether the properties of glial precursor cells vary during development. Cells isolated by the same technique from 5-, 10- and 20-day-old rats, were placed in culture and double-immunostained at 1,2,3,6 and 10 days in vitro (DIV) for GD3/glial fibrillary acidic protein (GFAP), galactosylceramide (GC)/GFAP, GD3/GC, GD3/antigen O4, GC/O4 and GFAP/O4. After 1 DIV the isolates from 5 day rats contained 30% GD3+ cells, 1% oligodendrocytes (GC+) and 3.5% astrocytes (GFAP+). The corresponding percentages from 10 day rats were: 35% GD3+, 3% GC+ and 3% GFAP+; and from 20 day rats: 28% GD3+, 35% GC+ and 1% GFAP+. Thus the number of oligodendrocytes in the initial isolate increased dramatically between 10 and 20 days. At all 3 ages immature cells were a major component of the total isolate. GFAP+ cells increased rapidly in all cultures. In all cultures the numbers of GFAP+/GD3+ cells reached a maximum and then declined coincident with the increase of GFAP+/GD3- cells, but there were many more of these double-stained cells in cultures from 20 day rats. Moreover, all cultures at 1 DIV contained some GFAP+/GD3- cells. Thus astrocytes appeared to derive both from pre-existing GFAP+/GD3- cells and from GD3+ cells, the latter pathway being more significant in the older rats. GC+ cells increased in cultures prepared from 5 day and 10 day rats, but remained relatively constant in cultures from 20 day rats. The number of GD3+/O4+ cells decreased coincident with an increase of GC+/O4+ cells (all GC+ cells were O4+), but the number of GC+/GD3+ cells was insignificant in any culture from any age animal. These findings in developing animals support the scheme shown by others in neonates that oligodendrocytes derive from GD3+ cells via O4+/GC-intermediate cells. These data show that more GD3+ cells differentiate into oligodendrocytes in 5- and 10-day-old animals than in 20-day-old animals, and a larger percentage differentiate into astrocytes in 20-day-old animals than in the younger animals. The most reasonable explanation of these results is that two committed populations of GD3+ precursor cells exist in the brain, and that the ratio of these populations changes during development. The implication of this conclusion is that bipotential progenitor cells do not persist beyond the neonatal period, but become committed to separate lineages during development.