Information available on the role of Mg for growth and cellulose degradation by rumen bacteria is both limited and inconsistent. In this study, the Mg requirements for two strains each of the cellulolytic rumen species Fibrobacter succinogenes (A3c and S85), Ruminococcus albus (7 and 8) and Ruminococcus flavefaciens (B34b and C94) were investigated. Maximum growth, rate of growth and lag time were all measured using a complete factorial design, 2(3)×6; factors were: strains (2), within species (3) and Mg concentrations (6). R. flavefaciens was the only species that did not grow when Mg was singly deleted from the media, and both strains exhibited a linear growth response to increasing Mg concentrations (P<0.001). The requirement for R. flavefaciens B34b was estimated as 0.54 mM; whereas the requirement for R. flavefaciens C94 was >0.82 as there was no plateau in growth. Although not an absolute requirement for growth, strains of the two other species of cellulolytic bacteria all responded to increasing Mg concentrations. For F. succinogenes S85, R. albus 7 and R. albus 8, their requirement estimated from maximum growth was 0.56, 0.52 and 0.51, respectively. A requirement for F. succinogenes A3c could not be calculated because there was no solution for contrasts. Whether R. flavefaciens had a Mg requirement for cellulose degradation was determined in NH3-free cellulose media, using a 2×4 factorial design, 2 strains and 4 treatments. Both strains of R. flavefaciens were found to have an absolute Mg requirement for cellulose degradation. Based on reported concentrations of Mg in the rumen, 1.0 to 10.1 mM, it seems unlikely that an in vivo deficiency of this element would occur.