Metallic electrodes based on iron, nickel, and/or cobalt have re-emerged as promising cost-effective anodes for the alkaline oxygen evolution reaction (OER) due to their simplicity and their in situ formation of a highly active oxy-hydroxide surface catalyst layer, which exhibits state-of-the-art overpotentials for the OER. However, the effect of alloy composition has not been systematically studied. Herein, using metallic anodes with defined Fe-Ni-Co atomic ratios prepared via arc melting, we report the relationship between the initial alloy composition, the OER performance, and the emergent active catalyst composition. After 50 h operation at 0.5 A cm(-2) the most active initial alloys (having a moderate amount of cobalt <40 at. %, an iron proportion between 30 and 80 at. % and a nickel ratio below 60 at. %) gave average overpotentials for 10 mA cm(-2) ca. 300-320 mV and Tafel slopes of 35-50 mV dec(-1). Iron and nickel-rich alloys performed poorer. The oxyhydroxide OER catalyst formed on the anode surface generally showed an increased concentration of Co and Ni and a depletion of Fe compared to the initial metal composition, giving the most active OER catalyst at a composition of Ni and Co of ca. 40 at. % with Fe at ca. 20 at. %. However, the initial alloy composition of Fe12.5Co12.5Ni75, showed a nearly invariant surface metal composition, indicating this as the most stable composition. Further analysis of the surface identified no correlation of the mass of metals leached from the anode surface, the electrochemically active surface area, or the presence of active Ni2+/3+ redox surface sites to the OER performance suggesting these factors do not influence the results.