Genotype by environment (G x E) interactions for grain yield were investigated in 13 perennial rice (Oryza sativa L./Oryza longistaminata) derivatives over three sites and 2 years in Lao PDR. Genotype accounted for 29.0% of the total sum of squares, with environment and the G x E interaction responsible for 60.2 and 10.8%, respectively. Cluster analysis identified three environment and six genotype groups, which accounted for 49.7, 98.0 and 42.8% of the E, G and G x E sums of squares, respectively. Principal component axes 1, 2 and 3 accounted for 54.0, 30.6 and 11.7% of the G x E sum of squares, respectively, with PCA1 indicating yield potential, PCA2 timing of cessation of rainfall in the 2011 wet season, and PCA3 environmental stresses affecting regrowth in the 2012 wet season. Genotype groups differed in adaptation to these contrasting conditions. G6 (Line 213, 240 and RD23) was widely adapted to all environments, with G5 (Line 248) being especially adapted to the 2012 environments. G3 and G4 were neutral, though G3 (Line 53) showed some preference for the Na Pok environments. G1 and G2 were poorly adapted everywhere, with the tall and late G1 (Line 11) being especially poor at Na Pok 2011 in a dry finish. While yields were lower in 2012, all derivatives survived the dry season with access to life-saving irrigation. This is promising, as the annual rice RD23 was unable to ratoon under these conditions, and had to be re-sown. Importantly, Line 213, 240 and 248 yielded comparably to RD23 from regrowth in 2012. Development of perennial rice should target rainfed and especially upland environments.