The chemistries controlling autoignition of primary reference fuels (n-heptane/isooctane binary mixtures) and binary olefin/paraffin mixtures have been inferred from experimental motored-engine measurements. For all n-heptane/isooctane and olefin/paraffin mixtures, each component of the mixture reacted via parallel intramolecular mechanisms with the only interactions being via small labile radicals. The octane qualities of the neat components appears to be dictated not by the initial reaction rate of the fuel, but by the reaction rate of the subsequent fuel-product reactions. In contrast, the blending octane quality of a component appears to be dictated more by the rate of the initial fuel reactions. The abnormally high blending octane qualities of olefins result from them having high rates of initial fuel reaction combined with slow rates of subsequent fuel-product reactions.The vast majority of past and recent experimental and modeling studies of the chemistry of autoignition has focused on single component fuels. Unfortunately, in the real world, virtually all automotive fuels consist of many components. In light of this fact, more recent experimental studies such as that of Filipe et al. * and modeling studies such as that of Westbrook et al.  are beginning to address the chemistries of binary mixtures. Both of these studies are of particular note as both examined the autoignition behavior of primary reference fuels. Understanding the behavior of the primary reference fuels is of particular importance as these reference fuels define the octane scale against which all other fuels are rated. In the present motored-engine study, the autoignition chemistries of neat n-heptane, neat isooctane, and their primary reference fuel (PRF) blends will be examined in greater detail than in Reference 1, and emphasis will be placed on identifying the major intermediate species resulting from their autoignition chemistries and on examining the behavior of these species in the PRF blends.In addition, the autoignition chemistries of other binary blends will be probed to examine the nonlinear octane blending that has been characterized in the past by blending octane numbers. The concept of blending octane numbers originated in recognition of the fact that the octane quality of a binary or a multi-component mixture was not necessarily a linear combination of the pure component octane qualities. This is particularly true for combinations of olefins with paraffins and aromatics with paraffins. The blending octane number of a compound is determined by mixing twenty volume percent of that compound with eighty volume percent of 60 PRF (60% by volume of isooctane and 40% by volume of n-heptane). The octane quality of that mixture is then measured using the standard Research and Motor Octane Rating Techniques. The blending octane number (BON) of the compound of interest is then determined by extrapolating the experimentally measured octane number (ON) to a hypothetical mixture containing 100% of the compound of interest using the formula: Not surprisingly, both the Research and Motor BON's of n-heptane are zero and both the Research and Motor BON's of isooctane are 100. One problem with accurately determining BON's is that the range of measured ON's is collapsed significantly over that for determining pure compound octane qualities. For instance, the mixture required for determining the BON of isooctane has a measured ON of 68 while the corresponding mixture for n-heptane has a measured ON of 48, yielding a range of only 20 octane numbers. In contrast, the range in octane numbers for the measurement of neat isooctane and n-heptane is by definition 100. Thus, there is a five-to-one reduction in the octane range that leads to greater experimental errors in determining blending octane numbers.The Research blending octane qualities of all paraffins listed in the API Research Project 45 compilation  are plotted against their corresponding neat Research octane qualities in Figure 1. The solid line is the one-to-one correspondence line. While there is some scatter in the data, the scatter is evenly distributed about the one-to-one line implying that all paraffins blend reasonably linearly with respect to the PRF's. The corresponding plot for the Motor octane qualities (not shown) is virtually identical with Figure 1 as would be expected based on the fact that paraffins generally exhibit zero fuel octane sensitivity (identical Research and Motor octane qualities) . The corresponding plots for the 68 olefins and 26 aromatics in this compilation are presented in Figures 2 and 3. The vast majority of olefins and aromatics show significant positive deviations from the one-to-one correspondence line, with many olefins and aromatics showing blending octane numbers more than 40 octane numbers greater than their neat octane qualities. Figure 4 presents the analogous plot for the Motor rating conditions for the olefins. Comparing this figure to the Research octane qualities in Figure 2 shows that the neat Motor octane qualities are shifted to lower values due to the fuel octane sensitivity exhibited by olefins [4}, and the significant positive deviations between the blending and neat octane qualities noted for the Research rating conditions carry over to the Motor rating conditions as well. Similar behavior is also exhibited by the Motor blending octane qualities of the aromatics (not shown). These large, consistent differences between the blending and neat octane qualities imply that the autoignition chemistries of the olefins and aromatics are interacting with that of the PRF paraffins in such a way as to retard the overall reactivity of the olefin/PRF mixture. In this report, the nature of this chemical interaction and retardation will be inferred from stable intermediate species measurements and the general chemistries controlling autoignition of paraffins and olefins. Emphasis will be placed on examining the interactions and chemistries of the binary PRF's and on select binary olefin/paraffin mixtures. This paper will begin with a brief description of the experimental equipment and procedures. This will be followed by the results section that will be divided into subsections discussing the individual results for neat n-heptane, neat isooctane, primary reference fuel blends, binary olefin/paraffin mixtures, and finally these results will be combined to discuss the chemical origins of nonlinear octane blending.