In recent years there has been a renewed interest in the use of ethanol as a future fuel for spark ignition engines. The finite availability of oil coupled with concerns about global warming and CO2 emissions has led to an increase in the use of ethanol which is produced from biomass. Generally ethanol is blended with gasoline to make an automotive fuel with the most common blend being E85 (85% ethanol and 15% gasoline by volume), which is used in Sweden and in the USA. Unlike gasoline, ethanol is a single component fuel which has a low volatility especially at low temperatures which means that cold-starting a spark ignition engine with ethanol-gasoline fuel blends is more difficult. In Europe cold-starting emission testing must be completed at ambient temperature and -7°C. During cold-starting tests with ethanol-gasoline blends high fuelling levels are required to form a combustible mixture and start the engine which results in high hydrocarbon emissions. This study investigates the start performance of a range of ethanol-gasoline blends in a port fuel injected spark ignition engine. Tests are completed at both ambient temperature at -rc and the start performance is characterised by the fuelling levels required to start the engine, the emissions of hydrocarbons and the fuel unaccounted for during the test. The optimum first start fuelling level is highly dependent on the proportion of ethanol in the fuel. The lower volatility of the ethanol fuel blends meant that it is harder to form a combustible mixture without injecting high levels of fuel. The high fuel injection rates result in high levels of liquid fuel being inducted into the cylinder which causes high HC emissions and fuel unaccounted for during the start. Under ambient temperature starts E 10 and E25 offer start performance benefits over gasoline, whilst E50 and in some aspects E75 are comparable to gasoline. E85 fuel results in considerably higher emissions of HCs and fuel unaccounted for than gasoline during ambient temperature starts. At -rc the start performance with ethanol suffers compared to gasoline. When using E10 and E25 the start performance is comparable to gasoline, but with E50 and the higher proportion ethanol fuel blends the start performance becomes far poorer. With E85 the mass of fuel emitted from the engine as un-burnt HCs during a -rc start is seven times the quantity with gasoline. Hardware modifications were made to the engine in order to try to improve the start performance of -rc E85 starts. Of the modifications tested heating the fuel within the fuel rail produced the biggest start improvement followed by heating the intake air. An extended cranking period prior to the start of the test only produced a small improvement in start performance and is not worthwhile compared to the energy needed to crank the engine. With both fuel rail and intake air heating the start fuelling level can be 40.5% below the baseline E85 level and peak HC emissions are reduced by 30.5%. The low volatility of E85 meant that even when incorporating the hardware modifications, the start performance of the engine is still not comparable to the standard engine running on gasoline due to high emissions of HCs. When cold-starting, especially at low temperatures, the use of high proportion ethanol-gasoline fuel blends is not suitable if low emissions of HCs are a priority. Longer duration tests were also completed with different ethanol-gasoline fuel blends in order to investigate their influence on engine warm-up characteristics. Larger proportions of ethanol within the fuel resulted in higher engine work output and lower heat rejection to the coolant and oil. With E 100 it takes 7.2% longer for the thermostat to open compared to gasoline. The warm-up tests also showed that the exhaust gas temperature is lower with ethanol in the fuel, which will result in a longer period before the catalyst lights-off. A higher proportion of ethanol in the fuel results in lower NOx emissions due to reduced peak temperatures within the engine.