Résumé
Rice crop duration in tropical-arid, irrigated environments, such as the Sahel, varies strongly among seasons and years. For rice double-cropping systems, cultivars are needed that have a stable duration under variable daylengths and temperatures. No efficient selection tools are currently available to screen for comparatively thermo- and photoperiod-insensitive cultivars, and little is known about the genetic diversity of rice in this respect. A previous study presented a model that disaggregates photothermal effects on rice phenology for the different crop development stages and the two main factors, daylength and temperature. The present study characterizes differences in the photothermal response of 18 rice lines for three major phases of their development, namely, germination, the appearance of early leaves, and the induction phase (IP) between the end of the basic vegetative phase (BVP) and panicle initiation (PI). A field experiment was conducted at Ndiaye in Senegal, using 11 staggered sowing dates at 15 day intervals. Three check cultivars were replicated four times per date, and the others were not replicated. Phenological observations included leaf tip appearance rates on a daily basis, and heading and flowering dates. Water and air temperatures were also monitored. The duration from seed soaking to the appearance of the first leaf depended linearly on water temperature, with a base temperature of about 8°C. No differences were observed among lines. The appearance rate of the first to the fourth leaf had an optimum temperature (
T
opt) between 22°C and 25°C, with significant differences between lines. Lines with low
T
opt had high maximal leaf appearance rates during seedling stage. The quantitatively most important component of the seasonal variability of crop duration was IP, which depended on temperature and photoperiod. Photoperiodic and temperature effects were disaggregated using an empirically based model leaf appearance (LAP), which, by way of parametrization, helped quantify genotypic differences. The photothermal differences between lines were associated with adaptation to specific seasons. Lines such as IR13240-108-2-2-3, which is adapted to several seasons, were comparatively insensitive to photoperiod and temperature. That line also had a low
T
opt and a high Leaf Appearance. In conclusion, screening for temperature-insensitive leaf appearance rates may be promising for rice breeding for arid, irrigated environments. Selection for combined low thermal and photoperiod sensitivity, however, would be impractical on an experimental basis, and would therefore require genetic markers.