1. A study has been made of pure proteins heated either alone or in contact with sugars, so as to cause a severe fall in their reactive lysine contents, and also of commercial protein concentrations. 2. For unheated materials, and for bovine plasma albumin and fat-extracted, dried chicken muscle severely heated in the absence of sugar, Acid Orange 12 binding values (mmol bound dye/kg crude protein (nitrogen times 6-25)) were close to the sum of total histidine, total arginine and reactive lysine contents (mmol/kg crude protein (N times 6-25)), which we have termed HARL values. The dye-binding values and the HARL values were reduced similarly by heat treatment. 3. For materials in which protein and glucose had reacted under mild conditions (37 degrees), the dye-binding capacity with Acid Orange 12 was unchanged even though the HARL value of these materials was considerably reduced. When protein and glucose or sucrose were heated more severely, the dye-binding capacity was slightly lowered but not to the same extent as the reduction in the basic amino acids. 4. Animal feeding-stuffs, whether unheated, industrially processed or diliberately heated, appeared to react with Acid Orange 12 in the same way as the model systems (selected to represent three types of heat damage: 'advanced' and 'early' Maillard and protein-protein damage). 5. Remazol blue binding values and fluorodinitrobenzene (RDNB)-reactive lysine values were similarly reduced in materials that had been severely heated, either with or without added sugars; however, when protein and glucose had reacted under mild conditions the fall in Remazol blue binding was less than that in FDNB-reactive lysine. 6. For the model materials, binding with cresol red was, in general, higher for heated samples but the results showed no correlation with FDNB-reactive lysine values. For meat and groundnut meals, changes in values after heat treatment were smaller than those that have been reported for soya-bean meals.