Transgenic tobacco (Nicotiana tabacum L. cv. W38) plants with an antisense gene directed against the mRNA of ribulose-1,5-bisphosphate carboxylase/ oxygenase (Rubisco) activase were used to examine the relationship between CO2-assimilation rate, Rubisco carbamylation and activase content. Plants used were those members of the r1 progeny of a primary transformant with two independent T-DNA inserts that could be grown without CO2 supplementation. These plants had from < 1% to 20% of the activase content of control plants. Severe suppression of activase to amounts below 5% of those present in the controls was required before reductions in CO2-assimilation rate and Rubisco carbamylation were observed, indicating that one activase tetramer is able to service as many as 200 Rubisco hexadecamers and maintain wild-type carbamylation levels in vivo. The reduction in CO2-assimilation rate was correlated with the reduction in Rubisco carbamylation. The anti-activase plants had similar ribulose-1,5-bisphosphate pool sizes but reduced 3-phosphoglycerate pool sizes compared to those of control plants. Stomatal conductance was not affected by reduced activase content or CO2-assimilation rate. A mathematical model of activase action is used to explain the observed hyperbolic dependence of Rubisco carbamylation on activase content.