Stems of fodder radish (Rhaphanus sativus L.), young lucerne (Medicago sativa L.), leafy Italian ryegrass (Lolium multiflorum Lam.) and silage of maize (Zea mays L.) were dried in seven different temperature regimes ranging from 3 to 105-degrees-C. Eighteen chemical constituents, frequently used in forage quality analysis, were measured or calculated. Freeze drying at 3-degrees-C was used as the control treatment to which other drying treatments were compared. Content of water-soluble carbohydrate (wsc), in vitro digestibility (D(om)) and %N insoluble in neutral detergent (N(cwc)) were affected most by drying temperature. Storage at - 20-degrees-C before freeze drying had no effect on chemical composition. Drying at 30-degrees-C reduced the content of dry matter (%dm), %wsc and %D(om) but increased %cwc and %N(cwc), compared with freeze drying. Drying at 50 and 70-degrees-C (standard practice) gave similar but less severe losses than at 30-degrees-C. Drying at 105-degrees-C further increased %cwc, %N(cwc) and greatly decreased %D(om). Rate of digestion declined with higher drying temperature. Storage at - 20-degrees-C before drying at high temperature increased %N(cwc) and reduced rate of digestion. Amylase incorporated in the cell wall determination reduced %N(cwc) and %cwc free from protein. %N(cwc) was higher in ryegrass than in the other forages. The effects of cultural practices and of the drying facilities on the biological processes during drying are discussed. It was concluded that an effective proteolytic step must be included in the cell-wall analysis for its proper measurement. However the analysis will probably never be perfect. Significant interactions between product and drying temperature for all constituents prevent general recommendations about the optimum drying procedure other than freeze drying. Quick drying of fresh products at 70-degrees-C is the second best option, because alterations during drying are usually smallest.