Respiration is poorly represented in whole plant or ecosystem models relative to photosynthesis. This paper reviews the principles underlying the development of a more mechanistic approach to modelling plant respiration and the criteria by which model behaviour might be judged. The main conclusions are as follows: (1) Models should separate C substrate From structure so that direct or indirect C substrate dependence of the components of respiration can be represented. (2) Account should be taken of the fact that some of the energy for leaf respiration is drawn from the light reactions of photosynthesis. (3) It is possible to estimate respiration associated with growth, nitrate reduction, symbiotic N-2 fixation, N-uptake, other ion uptake and phloem loading, because reasonable estimates are available of average specific unit respiratory costs and the rates of these processes can be quantified. (4) At present, it is less easy to estimate respiration associated with protein turnover, maintenance of cell ion concentrations and gradients and all forms of respiration involving the alternative pathway and futile cycles. (5) The growth-maintenance paradigm is valuable bur 'maintenance' is an approximate concept and there is no rigorous division between growth and maintenance energy-requiring processes. (6) An alternative 'process-residual' approach would be to estimate explicitly respiratory fluxes associated with the six processes listed in (3) above and treat the remainder as a residual with a phenomenological 'residual maintenance' coefficient. (7) Maintenance or 'residual maintenance' respiration rates are often more closely related to tissue N content than biomass, volume or surface area. (8) Respiratory fluxes associated with different processes vary independently, seasonally and during plant development, and so should be represented separately if possible. (9) An unforced outcome of mechanistic models should be a constrained, but non-constant, ratio between whole plant gross photosynthesis and respiration. (C) 2000 Annals of Botany Company.