Simultaneous two-dimensional Rayleigh and fuel Raman images have been collected in air-diluted methane and hydrogen jet diffusion flames. Temperature, fuel mass fraction and mixture fraction images are derived by a two-scalar approach based on one-step chemistry and equal species diffusivities. This enables calculation of two components of the scalar dissipation rate chi. The inherently weak Raman signal has been maximised by intra-cavity measurements, using a flashlamp-pumped dye laser. In addition, the Raman signal-to-noise ratio is drastically improved by a novel contour-aligned smoothing technique which exploits the high correlation between the Rayleigh and Raman signals. Quantitative measurements of scalar dissipation are presented, including probability density functions for components of chi. Profiles of mean and rms mixture fraction show the usual features already documented in other published results for this type of flame. Probability density functions of xi are close to Gaussian on the axis, and tend to bimodal at the edge of the flame. Results for the CH4 flames indicate that the mean of chi shows little change with Reynolds number. In the H-2 flame, mean values for the axial and radial components of the scalar dissipation rate, chi, are nearly the same, indicating a more isotropic structure than in the CH4 flames. For both fuels, the pdf of ln(chi) on the axis is more peaky than a lognormal distribution and somewhat skewed. The profiles of (chi/eta) show a nonlinear dependence on mixture fraction and have no clear resemblance to the skewed, monomodal shapes seen in cold flows. In the H-2 flame there is a strong correlation between instantaneous, local values of scalar dissipation and the departure from equilibrium, as measured by temperature depression.