Numerous studies have demonstrated the benefits of ethanol in increasing the thermal efficiency of gasoline-fueled spark ignition engines via the higher enthalpy of vaporization and higher knock resistance of ethanol compared with gasoline. This study expands on previous work by considering a split fuel injection strategy with a boosted direct injection spark ignition (DISI) engine fueled with E0 (100% by volume reference grade gasoline; with research octane number = 91 and motoring octane number = 83), E100 (100% by volume anhydrous ethanol), and various splash-blends of the two fuels. Experiments were performed using a production 3-cylinder Ford Ecoboost engine where two cylinders were de-activated to create a single-cylinder engine with a displacement of 0.33 L. The engine was operated over a range of loads with boosted intake manifold absolute pressure (MAP) from 1 bar to 1.5 bar absolute. The fuel injection timing of single fuel injection events was varied at MAP = 1 bar intake air pressure using different blend ratios (E0, E30, E50, E85 and E100) to identify the maximum thermal efficiency for each fuel blend. At boosted intake air pressures, E0 became knock limited, whereas none of the ethanol blends were knock-limited at any of the intake air pressures. A split fuel injection strategy with 50% of the fuel mass in each of two injection events was investigated for the range of intake air pressures. The different fuel blends showed little sensitivity to the split injection strategy, which indicated fuel air mixing did not significantly affect combustion at the conditions studied. The highest gross indicated thermal efficiencies (GITE) of 38.4% were achieved with E85 and E100 at 1.1 and 1.2 bar intake pressure for an improvement of 4% compared with baseline gasoline for the same intake pressures. The improvement in GITE scaled linearly with the mass fraction of ethanol in the fuel blend and non-linearly with mole fraction of ethanol in the blend for 1 bar intake pressure. The observed increase in GITE with ethanol fraction in blend is understood to be result of heat of vaporization accounting (where the initial state of fuel is assumed to be liquid) and cooling effects on properties, in roughly equal proportions as indicated by comparison of the experimental data with results of GT Power simulations.