Because wheat comprises such a large component of poultry diets, its nutritional characteristics and, in particular, its metabolizable energy (ME) content are critically important. This paper examines the popularly held view that the ME of wheat is variable by seeking out compositional factors which could account for the variation. The bioassays which have been applied to the derivation of the ME of wheat have also been appraised and it has been concluded that substitution techniques which are used in apparent ME studies may have contributed to the uncertainty which has been attributed to the ME value of wheat. Neither the variety nor the site of its cultivation appeared to affect the ME of wheat, but both density (positively) and α-amylase activity (negatively) were correlated with ME; however, neither effect was large and the commercial significance is doubtful. Attention has recently focused on the non-starch polysaccharides in wheat, specifically water-soluble arabinoxylan, and response experiments after xylanase addition to wheat-based diets have produced evidence that the viscosity of the digesta is a factor resulting in poor performance; the mechanisms involved have not, however, been elucidated.