In this review paper, the term 'availability' is used to describe the supply of nutrients and energy from the digestive tract to the body tissues. The traditional method for the estimation of nutrient and energy availability is to determine the apparent digestibility. This method has been a cornerstone in the development of all existing feed evaluation systems, both in ruminant and in non-ruminant nutrition. The apparent digestibility of a nutrient is an underestimation of the true digestibility to the extent that endogenous secretions of this nutrient are excreted in the faeces. Other factors such as microbial degradation and syntheses in the digestive tract contribute to invalidate the apparent digestibility as a measure of nutrient availability. These problems are particularly related to protein and amino acids. Alternative methods to estimate nutrient availability have been developed. Some of these are invasive and some are non-invasive. In the first category, pigs are cannulated at various sites of the digestive tract or catheterized in an artery and the portal vein. Non-invasive methods include isotopic labelling of the feed, in vitro assays and the use of mechanistic modelling. The virtues and drawbacks of these methods are discussed and it is concluded that progress in the field of predicting nutrient availability and in the development of future, nutrient-based feed evaluation systems is most likely to occur when various experimental and modelling techniques are integrated.