© CSIRO 2004 10.1071/EA02068 0816-1089/04/030333 Australian Journal of Experimental Agriculture, 2004, 44, 333–342 CSIRO PUBLISHING www.publish.csiro.au/journals/ajea Characterisation of rice (Oryza sativa) F 3 populations selected for salt resistance. 2. Relationships between yield-related parameters and physiological properties G.Y. Zhu A,B , J-M. Kinet A and S. Lutts A,C A Unité de Biologie Végétale, Institut des Sciences de la Vie, Université Catholique de Louvain, Place Croix du Sud 5 (Bte 13), B-1348 Louvain-la-Neuve, Belgium. B Present address: Centre of Microbial and Plant Genetics, Katholieke Universiteit Leuven, Kasteelpark Arenberg 20, B-3001, Heverlee, Belgium. C Author for correspondence; e-mail: lutts@bota.ucl.ac.be Abstract. One somaclonal family (1-23) regenerated from a salt-resistant callus, its initial cultivar I Kong Pao (IKP, salt-sensitive), the breeding line IR31785 (extremely salt-sensitive), the moderately salt-resistant cultivar Aiwu and F 3 populations derived from several crosses were exposed until maturity to a low sub-lethal (30 mmol NaCl/L) dose of salt. The agronomic performance of this material was compared with its physiological behaviour, as recorded during the vegetative growth phase. Although grain yield of 1-23 was similar to the yield of its initial cultivar IKP in both control and salt-stress conditions, a strong increase in salt resistance in terms of grain yield per plant was observed in their F 3 population derived from IR31785 × 1-23, which performed better than the corresponding parents and better even than the cultivar Aiwu in the presence of 30 mmol/L NaCl. Such improvement was due to a high number of spikelets per plant and a high spikelet fertility. An improved salt resistance in terms of grain yield was also observed in the F 3 population derived from IR31785 × IKP. Among physiological parameters, only K:Na ratios quantified in stress conditions are useful as a predictive criteria for yield potential, and salt resistance to low NaCl does not involve any adaptation to the osmotic component of salt stress. No physiological parameters quantified in the absence of stress were correlated with yield in the presence of NaCl. Results are discussed in relation to the usefulness of somaclonal variation for plant breeding. Additional keywords: in vitro selection, NaCl, salinity, somaclonal variation, yield. Introduction Soil salinity is a major limitation to agricultural productivity in many parts of the world. Accumulation of excess soluble salts, especially sodium salt, in the root-zone results in a loss of soil productivity in relation to (i) the inability of the plant to absorb water, and (ii) the nutritional disorders induced by the absorbed toxic ions (Bohnert and Jensen 1996). Rice (Oryza sativa L.) has been classified as a salt-sensitive species by several workers (Asch et al. 2000; Khatun et al. 1995; Lutts et al. 1995; Thomas and Nambisan 1999; Yeo et al. 1990). In O. sativa, salinity induces a decrease in panicle weight and panicle length, in primary branches/panicle, filled seeds/panicle, 1000-seed weight and total seed weight/plant (Abdullah et al. 2001; Khatun et al. 1995; Zeng and Shannon 2000; Zhu et al. 2000), while secondary water stress may have an impact on remobilisation of nutrients from vegetative tissues to grains (Yang et al. 2001, 2002). Breeding for salt resistance in rice is one approach that may be used to deal with the problem of increasing soil salinity. The success of any breeding programme devoted to the improvement of salinity resistance depends on the existence of an intraspecific variability for parameters controlling salinity resistance and reliable precocious selection criteria susceptible to be quantified at an earlier stage and to be correlated with yield potential. Somaclonal variation is a convenient way to induce variability since it may be usefully associated with in vitro selection (Tal 1994), and direct utilisation of somaclonal variant for improvements of salt resistance in rice has been reported by several workers (Lutts et al. 1999; Sathish et al. 1997; Winicov 1996; Zapata et al. 1991; Zhang et al. 1995). In these studies, valuable plant material was directly identified among regenerates or their progenies when screened for salinity resistance. However, somaclonal variation may induce deleterious modifications; in this view, somaclonal variation is a destructive rather than a constructive process. It was hypothesised that some variants may possess interesting properties in terms of salinity resistance but that these variants would be rejected in a mass screening because they also exhibit several negative traits.