Ethanol is a biofuel commonly used in gasoline blends to displace petroleum consumption; its utilization is on the rise in the United States, spurred by the biofuel utilization mandates put in place by the Energy Independence and Security Act of 2007 (EISA). The United States Environmental Protection Agency (EPA) has the statutory responsibility to implement the EISA mandates through the promulgation of the Renewable Fuel Standard. EPA has historically mandated an emissions certification fuel specification that calls for ethanol-free fuel, except for the certification of flex-fuel vehicles. However, since the U.S. gasoline marketplace is now virtually saturated with E10, some organizations have suggested that inclusion of ethanol in emissions certification fuels would be appropriate. The test methodologies and calculations contained in the Code of Federal Regulations for gasoline-fueled vehicles have been developed with the presumption that the certification fuel does not contain ethanol; thus, a number of technical issues would require resolution before such a change could be accomplished. This paper makes use of the considerable data gathered during the mid-level blends testing program to investigate one such issue: estimation of non-methane organic gas (NMOG) emissions. The data reported in this paper were gathered from over 600 cold-start Federal Test Procedure (FTP) tests conducted on 68 vehicles representing 21 models from model year 2000 to 2009. Most of the vehicles were certified to the Tier-2 emissions standard, but several older Tier-1 and national low emissions vehicle program (NLEV) vehicles were also included in the study. Exhaust speciation shows that ethanol, acetaldehyde, and formaldehyde dominate the oxygenated species emissions when ethanol is blended into the test fuel. A set of correlations were developed that are derived from the measured non-methane hydrocarbon (NMHC) emissions and the ethanol blend level in the fuel. These correlations were applied to the measured NMHC emissions from the mid-level ethanol blends testing program and the results compared against the measured NMOG emissions. The results show that the composite FTP NMOG emissions estimate has an error of 0.0015 ±0.0074 g/mile for 95% of the test results. Estimates for the individual phases of the FTP are also presented with similar error levels. A limited number of tests conducted using the LA92, US06, and highway fuel economy test cycles show that the FTP correlation also holds reasonably well for these cycles, though the error level relative to the measured NMOG value increases for NMOG emissions less than 0.010 g/mile.