In 2008-2009, EPA and DOE tested fifteen 2008 model year Tier 2 vehicles on 27 fuels. The fuels were match-blended to specific fuel parameter targets. The fuel parameter targets were pre-selected to represent the range of fuel properties from fuel survey data from the Alliance of Automobile Manufacturers for 2006. EPA's analysis of the EPAct data showed that higher aromatics, ethanol, and T90 increase particulate matter (PM) emissions. EPA focused their analysis only on the targeted fuel properties and their impacts on emissions, namely RVP, T50, T90, aromatics, and ethanol. However, in the process of fuel blending, at least one non-targeted fuel property, the T70 distillation parameter, significantly exceeded 2006 Alliance survey parameters for two of the E10 test fuels. These two test fuels had very high PM emissions. In this study, we examine the impacts of adding T70 as an explanatory variable to the analysis of fuel effects on PM. We then compare an emissions model using just the EPA variables to our new emissions model using T70. Results indicate that for the EPAct test program, the T70 distillation parameter is a better predictor of cold start PM emissions than the other distillation parameters, and a cold start emissions model that includes T70 does not include an ethanol term for cold start emissions. Further results indicate that if T70 is added to the Bag 1 EPAct model and used in EPA’s MOVES2014 emission inventory model, increased ethanol levels beyond E10 are predicted to reduce PM from on-road motor vehicles in the U.S.