In Brazil, guava is attacked by several pests and diseases that damage mainly the fruits. The rust, caused by Puccinia psidii, it is the most serious disease under our conditions, constituting, in some areas, the limiting factor of the crop due to the injury it causes. Currently the only control measure available is weekly fungicide sprays, which increase the production cost. The use of resistant cultivars is the most effective and economic control of this disease. The objective of this research was to select individual guava plants to obtain rust resistant varieties. Twenty-two thousand nine hundred and fifty (22950) seeds from fruits originated through open pollination of 306 accesses were used, constituting half-sib progenies. Seedlings were grouped into different number of accesses as: A: 35 primary selections of white guava, obtained in the breeding program IAC (identification: White LxPy); B: 64 primary selections of red guava, obtained in that same program (identification: Red LxPy); C: 118 commercial varieties (some with 2 up to 6 accesses) as Supreme, Indiana, Weber, FAO, Australian, Patillo, Paluma, Rica, Ruby Supreme, IAC-4 and others; D: 55 advanced selections of IAC program (with acronym MAS), of Conceicao de Almeida, BA (with acronym EEFT) and others named Sigla (II to XIII) and; E: 34 accesses not clearly identified or without identification. Selection was applied in the initial stages of the seedlings and after artificial inoculation with the fungus. After two years, 105 individual plants were selected with absolutely no symptoms of the disease and will now go under selection for other traits, such as yield, fruit characteristics, color, and flavor. The heritability for rust resistance was estimated in a broad sense being h'=0.275.
