Limited information exists on the response of grass and legume silage to enzyme and bacterial inoculant treatments when wilted to drier than desired conditions. This study was undertaken to evaluate the impact of cellulase (from T richoderma longibrachiatum) application rate, when combined with a bacterial inoculant (L actobacillus plantarum and Pediococcus cerevisiae), on the fermentation characteristics of cocksfoot (Dactylis glomerata L) and lucerne (Medicago sativa L) ensiled at high dry-matter concentrations. Forages were wilted to near 600 g dry matter kg~1 and cellulase, combined with inoculant, was applied at 0É30 ml kg~1 herbage and at two, four and eight times this concentration (at least 2500 IU ml~1). Cellulase was also applied alone at 0É60 ml kg~1. Wilted forages were ensiled in laboratory silos for 60 days. E †ect of cellulase application rate on neutral detergent Ðbre concentrations of the silages was small and inconsistent. Averaged across species, only the intermediate cellulase concentrations decreased neutral detergent Ðbre concentration (P \ 0É082). The limited cell-wall degradation was probably related to the high silage dry-matter and lignin concentrations. Cellulase combined with inoculant increased total fermentation, when averaged across species. In cocksfoot, cellulase combined with inoculant decreased pH and concentration but increased the lactic : acetic acid NH 3 -N ratio of control silage, with most of the e †ect caused by the inoculant. Cellulase applied alone to lucerne caused a higher lactic : acetic acid ratio than the control or when combined with the inoculant at the same cellulase rate. Thus, the e †ect of cellulaseÈinoculant mixtures on silage quality varied among plant species, with cocksfoot generally more responsive than lucerne.