Abstract. A critical review of the literature on the staling of bread has been presented with particular emphasis on developments during the last decade. The staling of bread has been discussed from a structural point of view with separate contributions due to starch retrogradation, a modification in the gluten structure and a partial re‐distribution of moisture from the gluten to the starch fraction of the crumb. Both model systems and studies on the natural crumb are considered including the contribution to staling of macroscopic moisture re‐distribution and deterioration in flavour volatiles. The staling of baked goods has also been reviewed and it is proposed that certain analogies from the bread situation might provide a suitable basis for an overall theory of cake staling. It seems possible that a considerable contribution to the staling of cake sponge may be due to changes undergone by the various proteins; the free moisture produced from the transforming proteins may be absorbed by the sugars and gelatinised starch. Retrogradation of starch appears to be less important than changes in the proteins during the staling of cake. Dehydration of the cake crumb is perceived to play an important role in the sensory evaluation of cake staleness.

Starting with a typical baker's dough, a model system was constructed in order to examine whether moisture re-distributed itself bet ween the two principal components, gluten and starch, of crumb during ageing.It was estimated that up to 30 % of the moisture associated with thegluten fraction, migrated to the starch during I20 h of storage of the baked system at 2J"c. The rate of migration decreased upproximately exponentially with time. Moisture from the gluten was also found to migrate to the starch during baking, but at an accelerated rate. The rate of moisture migrution in the model system, both at 100 and 2 5 "~. was consistent with a diffusion-controlled process involving migration of the gluten-released moisture to the gluten-starch interface. Experimental results show that staling of bread involves (in addition to the generally accepted retrogradation of the starch) an irreversible modification in the water structure of the gluten, leading to the formation of labile water, which becomes available for absorption by the incompletely gelatinised starch. Since the gluten forms the continuous matrix of the crumb, the modification results in a rigidification or firming process.Separation of the components of dough was achieved by ultra-centrifugation.

The catalytic decomposition of ammonia at 10-5-10-4 torr and in an ultra-high vacuum was studied on nickel, rhodium, tantalum and tungsten wires. The promoting and poisoning effect of carbon when present in wires in limited and excess amounts, respectively, generally confirmed previously reported observations for platinum. A decay of activity occurred after flashing a wire and was attributed to a depletion of carbon in the sur€ace region of the wire as a result of reaction with ammonia (or with a surface nitride). With rhodium, ammonia was observed mass-spectrometrically to react with a wire poisoned by excess carbon to give hydrogen cyanide. Absolute rate theory calculations indicate that at low temperatures where the activity of a rhodium wire decreases rapidly with decreasing temperature, the rate-limiting step in ammonia decomposition is the desorption of nitrogen from the catalyst surface. Such kinetic behaviour is inconsistent with decomposition rates observed at temperatures in excess of about 650°K. The activation energies on tantalum (14 kcaI mole-') and tungsten (25 kcal mole-') are thought to contain terms for migration of the adsorbed ammonia molecules, formation of active sites and evaporation from the surface.

A new technique for studying the staling of bread is described based on conductance and capacitance measurements on the isolated centre-crumb. It is proposed that the observed changes (up to 17%) in the electrical properties are due to retrogradation of the starch as well as the changes undergone by the transforming protein. Moisture redistribution is thought not to play a significant role during staling of the crumb at temperatures of less than about 1 0 "C. At 3 6 "C the firstorder rate-constant for the protein transformation was estimated to be 0.7 day-' and 0.3 day-' for the rateconstant for retrogradation of the starch fraction of the crumb. At 0 "C, the protein transformation was insignificant and the rate constant for the starch change was 1.03 day-'. The thickness of the gluten membrane surrounding the starch granules in bread was approximated to be less than 1 pm.

The decomposition into nitrogen and hydrogen on heated filaments was studied at pressures of 10-4 torr and less with both conventional and ultra-high vacuum (u.h.v.) techniques. The reaction probability P of an ammonia molecule at a collision with the platinum was determined under various conditions. Particular attention was given to initial values of P after flashing, to steady values of P after some reaction, and to the effect on P of various filament pre-treatments. Exposure of an incandescent filament to hydrogen at 10-2 torr and to residual gases revealed important effects. Complex results were observed in the conventional apparatus, including a negative temperature coefficient above about 900"K, and a variable order of reaction, which could be less than unity, depending on the pretreatment. In the absence of residual gases (u.h.v.), simple kinetics were observed above 700°K. The conventional vacuum results were reproduced by a wire gre-treated with ammonia containing hydrocarbon (2 : 2 : 4-trimethylpentane) but not with a wire pretreated with hydrogen only.It is suggested that ammonia on platinum rapidly forms a surface nitrogen-containing layer which may be a surface nitride. This is removed by reaction with carbon diffusing from the interior. Surface carbon is also removed by reaction with ammonia. Excessive carburization gives an inactive surface. Equations are deduced on this basis. The rate-determining step at high temperatures is postulated to be formation of NHn(ads) from NH3(ads). This is supported by ratetheory calculations and hydrogen retardation studies.