Experimental measurements have been made in a series of non-premixed and slightly premixed benzene-doped methane flames to investigate the effect of fuel-side air on soot concentrations and to determine whether the self-reaction of cyclopentadienyl radical is a significant source of naphthalene under these conditions. The flames were generated by adding increasing amounts of primary air to the benzene-doped methane fuel mixture in a co-flowing axisymmetric non-premixed flame. The molar benzene/methane ratio was 2%, which was high enough for the benzene to dominate the production of aromatic hydrocarbons and soot. Measurements were made on the centerline of gas temperature with thermocouples. of soot volume fraction with laser-induced incandescence, and of major species and C1 to C13 hydrocarbon concentrations with online mass spectrometry. The maximum centerline soot volume fraction in the flames decreased strongly and monotonically as air was added to the fuel, with a reduction of one half from the non-premixed flame to a flame with a primary equivalence ratio of 15. This observation shows that even the small amounts of fuel-side air produced by 'accidental' mechanisms Such as entrainment at the flame base can significantly impact soot concentrations. The hydrocarbon measurements showed that premixing caused a greater amount of benzene in the pyrolysis zone to be consumed through phenoxy formation, such that the concentrations of the phenoxy decomposition product cyclopentadienyl increased. whereas the concentrations of substituted benzenes such as phenylacetylene and toluene decreased. The naphthalene concentrations also decreased monotonically. which indicates that most of the naphthalene was formed from phenylacetylene by the HACA mechanism (or from benzyl by propargyl addition), and that negligible amounts of naphthalene were formed from cyclopentadienyl.