Some Effects Of Water Deficit On Rainfed Landraces Of Rice (Oryza Sativa L.) Indigenous To Kenya
MetadataShow full item record
Landraces of rice (Oryza sativa L.) have been selected by farmers over a long period of cultivation in regions of Kenya which are subject to drought. The response of some of these varieties to water deficit has been investigated with the aim of identifying the characteristics responsible for their field reputation for drought resistance. Three landraces (Sindano Buhari KR-35, Matama KR-108 and Mdadho KR-135) were compared with drought-susceptible and drought-resistant checks, IR~20 and IR-52 respectively. Experiments were conducted in controlled conditions, in glasshouses and the field. water deficit was imposed either by restricting the watering regime, or by the use of polyethylene glycol 6000. When plants were grown in pots and water deficit imposed by restricted watering, growth was reduced at low mean water potentials. Root growth was reduced more than shoot growth. Leaf water potent1al and stomatal conductance decreased, but the mean length of roots increased. All varieties however, including the checks, showed a similar pattern of response to water deficit suggesting that varietal differences in the control of water loss could not account for their differing performance under conditions of water deficit in the field. When plants were grown in the field, KR-35was the most resistant to water deficit in terms of relative yield, and IR-20 the most sensitive. Visual observation of rooting pattern when the plants were excavated revealed that plants in well-watered soil had a large number of shallow roots. Witholding water resulted in a decrease in water potential in the upper regions of the soil profile and this was accompanied by a reduction in root biomass and a change in the proportion of roots at different depths. KR-35 showed the least reduction in long roots, and IR-20 the greatest. Plants were grown in soil surrounded by a semipermeable membrane and immersed in solutions of controlled water potential. As the water potential was reduced, the number of long roots decreased dramatically in IR-20, but in KR-35 the proportion of longer roots increased. Water uptake by mature roots, measured by micropotometry, was greater per unit length of root in KR-35 than in IR-20. Whole plant water use efficiency increased as the water potential of the nutrient solution was decreased, and KR-35 was about twice as efficient in its use of water as IR-20. The results are discussed in relation to the characteristics which should be selected for rice varieties bred for rainfed cultivation.