A wire-mesh reactor, with the capability of virtually eliminating secondary reactions, has been used as base-case in the study of product yields and structures from the pyrolysis and hydropyrolysis of a sample of sugar cane bagasse in a fixed-bed ＇hot-rod＇ reactor. Results from the two reactors have been compared to determine how best to assess bench-scale data which might be used for eventual process development. Experiments have been carried out at 600 degrees C at pressures up to 70 bar. Structural features of the bio-oils have been examined by size exclusion chromatography and FT-infrared spectroscopy. In both reactors the effect of increasing pressure was to reduce the bio-oil and total volatile yields; hydropyrolysis bio-oil yields were marginally higher than pyrolysis yields under equivalent operating conditions. About 5 to 6% bio-oil product is lost in the fixed-bed reactor, compared with the wire-mesh reactor, with consequent increase in recovered chars. Pressure and reactor bed depth appear to affect only the thermally more sensitive components of the bio-oils, and increasing pressure beyond 40 bar, or bed depth beyond the level reported, was found not to affect yields to an appreciable extent. Taken together, these data indicate that about one-third of the original biomass may be converted to oil by direct pyrolysis. Size exclusion chromatograms and FT-infrared spectra of bio-oils from the hydropyrolysis experiments conducted at 1 and 70 bar suggest that oils from the hot-rod reactor experience a greater degree of secondary reactions than those from the wire-mesh reactor. Using tetrahydrofuran as eluent, the highest molecular masses were found to be around 1000 u in terms of polystyrene standards. However, comparison with size exclusion chromatograms using 1-methyl-2-pyrrolidinone indicated partial loss of sample when operating in THF. Our findings strongly suggest that THF is not a suitable eluent for the characterisation of biomass-derived pyrolysis oils by size exclusion chromatography.