After two years of tapping in Chachoengsao province, uprooting of 12 trees (clones RRIM 600, PB 235, GT1, 4 trees per clone) provided estimations of the relations existing between the total dry biomass of rubber trees and their trunk girth. These relations were found linear at least inside the studied range (trunk girth varying from 45 cm to 65 cm), but different between clones. A general model, able to estimate in one location the biomass of all clones might perhaps be set up by taking into account not only the trunk girth but also some other variables describing canopy structure, as the canopy biomass cannot be described by the trunk girth only. Nevertheless, finding and parameterizing such a model requires further survey. Thanks to the linear regression models obtained between biomass and trunk girth, biomass of the trees could be estimated at the moment of opening and after two years of tapping. Biomass partition variables could be calculated, allowing estimation of the harvest index and k factor, accounting for rubber production efficiency. Nevertheless, as these models rely on the uprooting of only 4 trees per clone, an extended survey is required to confirm their accuracy. Competition between growth and production was found linear for all three clones, confirming numerous previous studies. Clone GT1 showed a rather bad behavior regarding the competition between growth and rubber production. Clones PB 235 and RRIM 600 showed much better behaviors with lower competition between growth and production. Moreover, RRIM 600 seemed to maintain a significant growth compared to other clones during the tapping stop period (February to April 2000, dry season), sign of a better water use effectiveness. Rubber breeders in rather dry areas should aim at such a characteristic .and systematic survey of clonal girth increments during the dry season could provide a reliable picture of clonal susceptibility to draught and water stress.