The gross energy (GE) of a feed represents its total energy content and is measured by burning the material in an atmosphere of oxygen in a bomb calorimeter. The GE of many feeds is often taken to be 18.4 MJ/kg dry matter (DM) (MAFF et al, 1984). This assumption is likely to be doubtful for feeds which are for example high in lipid or which contain products of fermentation. The gross energy content of grass silages in particular has recently been reviewed by Edwards (1986).The GE content of grass silage can account for some 50% of the variability in metabolisable energy (ME). It is however ignored in present laboratory methods used for the prediction of ME which largely depend on using fibre fractions (Givens, 1986) which are more closely related to digestibility. This paper presents the results of GE determinations made on 109 grass silages made in clamps and since GE cannot be measured routinely, examines a number of factors which may be helpful in its prediction.