Genotype by environment interactions for performance of perennial rice genotypes (Oryza sativa L./Oryza longistaminata) relative to annual rice genotypes over regrowth cycles and locations in southern China

Abstract Genotype by environment (GxE) interactions for performance of 5 perennial rice genotypes ( Oryza sativa L./ Oryza longistaminata ) were examined relative to 1 mutant and 3 annual rice genotypes over 2–6 growth cycles at 7 locations in southern China between 2014 and 2017. Environment main effects accounted for 25.7% of the total sum of squares (SS), with genotype 33.8% and GxE 37.7%. Cluster analysis identified 6 genotype x 6 environment groups, which accounted for 77.9% of the GxE-SS. Principal component axes 1, 2 and 3 accounted for 54.7%, 25.1% and 9.4% of the GxE-SS, respectively, with PCA1 indicating yield potential, PCA2 performance over ratoon cycles, and PCA3 ratoon percentage. Environment groups differed in yield potential, which related to site favourability and whether it was low or high in the ratoon cycle. Genotype groups differed in yield potential and how well they performed in higher ratoon cycles. The medium-maturity (125 days) seasonally-replanted annual rice check BN21 was highest yielding (6.13 t ha −1 ). Perennial rice PR23 was high yielding and stable (5.25 t ha −1 ), with earlier maturity (119 days) and strong regrowth (82%). Ratooned annual rice RD23 was high yielding in original crops but poor yielding in ratoon crops, with a low ratoon percentage (16.5%). Similarly, perennial Bt71 and ratooned BN21 were high yielding in original crops and low-cycle ratoons under favourable conditions, but yielded poorly in high-cycle ratoons and less favourable conditions, with moderate regrowth (59.6%). Despite strong regrowth (77.5%), perennials 264 and Bt69 had low yield, as did perennial 139A and mutant TZ, and both these groups were late maturity. A combination of high yield potential, strong regrowth and earlier maturity resulted in higher performance of perennial rice over environments and regrowth cycles, with PR23 outstanding, and able to perform similarly to the seasonally-replanted annual check, BN21, over up to six growth cycles. Ratoon performance and trade-offs need to be examined further.