Ependymin, a prominent protein of the brain's extracellular fluid (ECF) was previously implicated in the consolidation of memory and in the activity-driven sharpening of the retinotectal projection. Because both these phenomena probably involve the growth and elaboration of appropriate synapses, we have tested whether ependymin can serve as a substrate for the growth of axons from goldfish retinal ganglion cells in a culture assay. Ependymin (Ep), laminin (LAM), polylysine (PL), and Concanavalin A (Con A) were plated on glass coverslips either uniformly or in striped patterns. Ep alone, either soluble or partly polymerized (by dropping calcium concentration and pH), was a good substrate for axonal outgrowth, as good or better than PL and Con A, but not as good as LAM. Neurites grew faster on LAM (71 microns/h) than on Ep (32 microns/h) or on PL (22 microns/h). Fasciculation was low on LAM, intermediate on Ep, and highest on PL. In exclusive side-by-side stripe assays, axons preferred LAM over Ep, but gave weak or no preference for Ep over Con A or PL. With stripes of LAM + Ep alongside pure LAM, the axons preferred the mixture of LAM + Ep. When antibodies to Ep were plated in stripes over continuous Ep substrate, the axons avoided the antibody-blocked stripes and grew on the Ep stripes. Antibodies to Ep did not, however, block growth on laminin substrates, nor did antibodies to LAM block growth on Ep. Dot blots and western blots showed very little cross recognition between the antibodies. Ependymin is a good substrate for neurite outgrowth, which is normally present in ECF, and adhesion to Ep is independent of LAM and possibly additive to it.