Flame radiation, the dominant heat transfer mechanism in full scale fires, is in turn controlled primarily by the fraction of the flame volume occupied by solid carbon. Particulate volume fractions, fv, are measured in situ in small scale, 0(10 cm), buoyant diffusion flames supported in air by solid polystyrene, PMMA and POM, cellular polystyrene, two cellular polyurethanes and liquid isooctane, acetone and alcohol. Two measurement techniques, based on attenuation of known monochromatic laser radiation by the flame, are described. In the small particle absorption limit, valid in the visible for POM, acetone and alcohol transmittance at a single wavelength suffices to determine fv. For the remaining fuels, scattering becomes significant and multiwavelength transmittance measurements are used to determine an approximate two parameter particle radii distribution, N(r)=No (27r3/2rmax4) exp (−3r/rmax) where rmax is the most probable radius and No is the particle concentration. The resulting fv=18.6 No rmax3 may be used to calculate the infrared emission from solid carbon in the flames considered. Volume fractions rank in the expected order of flame luminosity and smokiness from polystyrene, fv≈5×10−6, to alcohol, fv≈10−7. Within the approximations of flame homogeneity, spherical particles, known optical properties and assumed form for the size distribution, the fv, data are from ±5% to ±15% accurate. Good agreement exists with fv of solid polystyrene and PMMA derived independently from infrared flame transmittance and radiance data and between experimental mass pyrolysing rates and calculated rates obtained using these results in a radiation model.