Two hundred and fifty-two open-pollinated faba bean accessions from diverse origins were evaluated for their ability to retard the development of the rust, Uromyces viciae-fabae (Pers.) Schroet. in small adjacent plots during the period 1978–1981. A mixture of rust isolates from Manitoba was used as the inoculum in the first 2 years and cultures of the rust races 1 and 3 were used in subsequent years. Significant differences at the 1% level were found between accessions for the area under the disease progress curve (AUDPC) and the final rust severity in the various trials. Mass selections (MS) and single-plant selections (SPS) were made from accessions for the slow-rusting character. Eight populations and/or families from accessions 2N6 (MS), 2N29 (SPS), 2N43 (MS and SPS), 2N122 (SPS), ILB938 (MS), ILB(332X133)A (MS) and ILB(332X133)B (MS) consistently had low AUDPC values and were considered to be slow rusters. Four additional progenies from accessions 2N291 (MS), 2N319 (SPS), ILB697 (MS) and ILB866 (MS) had intermediate AUDPC values and were considered to be moderate rusters. The other progenies were either fast rusters or showed inconsistent AUDPC values. Rust development and spread were also evaluated in isolated test plots using one slow-, one moderate- and one fast-rusting mass-selected plant population. Populations were found to be similarly ranked in both adjacent and isolated plots, confirming the adequacy of using small adjacent plots in evaluating the slow-rusting characteristics of large numbers of accessions or progenies derived by MS or SPS.