This study systematically re-examines the titania-catalysed photo-oxidation of methylene blue (MB) in aqueous solution at 20°C, placing particular emphasis on the effects of TiO 2 crystallite size, TiO 2 polymorph (anatase, brookite, rutile and combinations thereof) and experimental test conditions on the rate of MB photo-oxidation. For all TiO 2 samples tested, the highest rate of MB photo-oxidation was observed at pH 6, slightly above the isoelectric point of TiO 2 (*5.8 for P25 TiO 2 ). Increasing the ionic strength at pH 6 induced MB dimer formation in solution, and lowered the rate of MB photo-oxidation by TiO 2 . For all TiO 2 polymorphs, the surface area normalised rate increased with crystallite size reflecting the corresponding reduction in surface and bulk defects (electronhole pair recombination sites). The optimum crystallite sizes were *20-25 nm for anatase and *50 nm for brookite. The photocatalytic activity of the different TiO 2 powders followed the general order P25 [ anatase [ brookite ) rutile, with the high activity of P25 TiO 2 providing strong evidence that anatase-rutile heterojunctions act as ''hotspots'' for MB photo-oxidation. Mixed phase anatase-rutile or brookite-rutile powders, each containing *5 wt% rutile, demonstrated superior area normalized photocatalytic activities for MB photo-oxidation compared to pure phase anatase or brookite powders of comparable crystallite size. Finally, deposition of Pd, Pt or Au nanoparticles decreased the activity of P25 TiO 2 for MB photo-oxidation. This paper clarifies long-standing confusion in the scientific literature about the photo-oxidation of aqueous MB over TiO 2 and M/TiO 2 (M = Pd, Pt and Au) photocatalysts.