Current studies support the morphological classification of oligodendrocytes proposed by Del Rio Hortega ( Bol. R. Soc. Esp. Hist. Nat. 10:25-29;  C.R. Soc. Biol. 91:818-820), in which cells either myelinate multiple internodes that are associated with small axons, or they myelinate restricted/single internodes of large-diameter axons. The reasons why an oligodendrocyte myelinates a particular calibre of axon are unknown. Because progenitors are generated in restricted, subventricular zones, an intrinsic program would imply that germinal centres contain a mixture of cells, each committed to myelinate axons of a particular size. Conversely, each cell could have the potential ability to myelinate any size axon. We tested this latter hypothesis that oligodendrocyte progenitors are uncommitted in their ability to myelinate a particular axon size. We introduced oligodendrocyte lineage cells from the optic nerve, which normally encounter only small-diameter axons, to a myelin-deficient environment containing a large range of axon sizes. Dissociated, mixed glial cells from the optic nerve were characterised immunocytochemically and were grafted into the spinal cord ventral column of neonatal, myelin-deficient rat mutants. Examination of the patches of myelin produced by these cells at different times after transplantation revealed that optic nerve oligodendrocytes were capable of producing a widespread, nonselective myelination of axons that were destined to have both small or large calibres. Thus, an axonal or local signal, and not an intrinsic program, is probably responsible for the previously described oligodendrocyte diversity.