[en] Detached flag leaf cultures were not suitable for evaluation of stem-rust resistance in our screening programme. On the basis of yield evaluation it was possible to screen out ten stem-rust ''tolerant'' wheat lines in field experiments. Rusted and protected microplots of each line were paired within a replicate. After artificial inoculation, the protected plants were sprayed with fungicides (benomyl plus dithiocarbamate plus copper salt) at weekly intervals until maturation to keep each protected plot rust-free. The thousand-kernel weights of rusted and protected plots were compared. When the thousand-kernel weight of protected plot increased only slightly and the rust reaction type of plants was susceptible in the rusted plot, the line was screened out as putative ''tolerant''. On the basis of three-year field trial ten ''tolerant'' lines were selected. Nine out of ten lines proved to be resistant to two stem-rust races in greenhouse tests in the seedling stage, when resistance was determined on the basis of reduced spore production instead of infection types. Resistance of these seedlings related partly to the reduced number of pustules and partly to a slow rusting character of plants. It seems possible to screen resistant cultivars in the greenhouse by the method outlined in this paper, when resistance is determined on the basis of a reduced number of infection sites and/or by the slow rusting capacity. (author)