This study assessed cell voltage development, electricity recovery, and microbial community composition in response to initial substrate including acetate, xylose, acetate/xylose 1:1 mixture (ace/xyl), and bioethanol effluent (BE) during microbial fuel cell (MFC) operation at 1000 a"broken vertical bar external resistance. The BE mainly contained 20.5 g/L xylose, 1.8 g/L arabinose, and 2.5 g/L propionic acid. The MFCs initially fed with acetate showed shorter initiation time (1 day), higher average cell voltage (634 +/- 9 mV), and higher coulombic efficiency (31.5 +/- 0.5 %) than those initially fed with ace/xyl or xylose. However, BE-initiated MFCs only generated 162 +/- 1 mV. The acetate-initiated MFCs exhibited longer adaptation time (21 h) and lower cell voltage (645 +/- 10 mV) when the substrate was switched to xylose, whereas substrate switching to BE produced the highest voltage (656 mV), maximum power density (362 +/- 27 mW/m(2)), maximum current density (709 +/- 27 mA/m(2)), and coulombic efficiency (25 +/- 0.5 %) in the acetate-initiated MFCs. The microbial community in acetate-initiated MFCs was less diverse and contained more electrogenic bacteria (13.9 +/- 0.4 %) including Geobacter sulfurreducens and Desulfuromonas acetexigen than the MFCs initially fed with ace/xyl, xylose, and BE. After switching the substrate to xylose and subsequently to BE, the microbial community in the acetate-initiated MFCs became more diverse, while no significant changes were observed in ace/xyl-, xylose-, and BE-initiated MFCs. The results showed that initial substrate affected the power generation and the capability to adapt to the substrate alteration in MFCs. Acetate-initiated MFCs showed best performance in utilizing BE.