Drought response of modern and old oat lines in greenhouse and long-term field trials

This study compares theresponse of old and modern oat (Avena sutiva L.) lines to pre-anthesis precipitation in long-term field trials, and to pre- and post-anthesis drought in the greenhouse. Long-term field trials were carried out at the Experimental Farms of Hankkija Plant Breeding Institute and the University of Helsinki between 1965 and 1988. Grain yield of 12 oat lines (released since 1959) was compared with that of the check lines. The effect of differences in May-June precipitation on grain yield was established for different lines. Greenhouse experiments included 19 oat lines (re-leased since 1921) and three wild species of oat (A. barbata L., A. sterilis max. L. and A. fatua L.). The data from greenhouse experiments were analysed using discriminant analysis in groups of old (released before 19705), modern, and wild oat types. Ranking of the oat lines according to results from long-term field trials and greenhouse experiments was not consistent; contrary to the field experiments, the old lines tended to be the most drought sensitive when tested in the greenhouse. There-fore, the simple and non-laborious methods used in this study for ranking of drought sensitivity of oat lines are not recommended.


ntroduction
The results from long-term oat cultivar trials in Finland have indicated that grain yield has increased through plant breeding by 30-40% (Rekunen 1988, Peltonen-Sainio andKarjalainen 1991). This can be attributed to a higher grain to straw ratio following selection for short straw, and high weight and filling rate of the panicle (Peltonen-Sainio 1990). This may contribute to drought sensitivity because plant height often correlates with root depth, and most root mass of the short-strawed cultivars is concentrated in the topsoil (Mac Key 1988). Postulated changes in drought sensitivity are emphasized due to early summer drought that is characteristic in southern and south-western Finland (Mukula and Rantanen 1989).
Considerable effort has been made to develop reliable and rapid methods for screening for cultivar differences in drought tolerance. Expression of drought resistance is associated with numerous changes in plant phenology, morpholo-gy, and physiology (Larsson and Gorny 1988), which makes its evaluation very complex and ranking of cultivars often inconsistent. Extrapolation of drought tolerance from long-term field experiments is a rapid method for evaluating cultivar differences, but differences may be masked due to numerous factors other than precipitation which cause variation in grain yield. Comparison of grain yields produced in favourable versus drought-stressed environments, simulated in a controlled environment, possibly represents a more accurate method for evaluating differences in drought tolerance. This study compares the response of old (released before 19705) and modern oat lines to pre-anthesis precipitation in long-term field trials and to pre-and post-anthesis drought simulated in the greenhouse.
Three greenhouse experiments were conducted at the Department of Plant Production, University of Helsinki, Finland. Day-length was 18 hours at light intensity (PAR) of about 200 |imol m' 2 s l . Day and night temperatures were respectively 20°C and 17°C. The same oat lines that were included in the long-term field experiments and seven additional lines [Esa (1922), Jalostettu maatiainen(1921), Kytö (1925), Osmo (1921), Pellervo (1935, and breeding lines Hja 76420 and Hja 77200] were tested in the greenhouse. Wild species of oat Avena sterilis max. and A. abyssinica were included in all experiments and A. fatua ('Laihia') in experiment (exp) 3. Seeds were pre-germinated on moist blotting paper for two days at 20°C before sowing. In total 22 pregerminated seeds were placed in a 7.5 1 plastic pot containing peat-vermiculite mixture (1:1). Following emergence, seedlings were thinned to 20. The pots were fertilized by top-watering every 14 days with a nutrient solution (500 ml 1%, The experiments included two treatments: 1) a control, where plants were continuously wellwatered (200 to 500 ml per pot), and 2) drought stress. When simulating temporary and moderate pre-anthesis drought (exps 1 and 2), the oat stands were subjected to a water deficit for two weeks which was followed by adequate watering until yellow ripening. When severely drought stressed (exp 3), plants were subjected to water deficit for two days close to pollination and thereafter they were slightly stressed throughout grain-fill. The severity of drought was measured using a test-tube (0 2 cm) to remove a soil core from each pot and determine the water content. pF-values were determined based on soil water retention curves. Physiological measurements were carried out in exps 1 and 2 at the four-five leaf stage on the two uppermost fully expanded leaves, and in exp 3 on the flag leaf close to pollination. Photosynthetic parameters were measured on two plants per treatment, three measurements per leaf, with a LI-COR LI-6200 portable photosynthesis system (Licor Inc., NE, USA). The two uppermost fully expanded leaves of six plants in exps 1 and 2 and 15 flag leaves 200 AGRICULTURAL AND FOOD SCIENCE IN FINLAND Vol. 6 (1997): 199-205. in exp 3 were collected and endogenous abscisic acid (ABA) concentration was measured as described by Peltonen-Sainio and Mäkelä (1995). Similarly, accumulation of proline was measured as described by Troll and Lindsley (1955) and Hanson et al. (1977). Relative water content (RWC) was evaluated on three and two uppermost fully expanded leaves in exps 2 and 3, respectively, according to Ritchie et al. (1990). Mature oat plants (14 per treatment, exps 1 and 3) were cut at soil-level, air-dried, and the total weight of grains per main shoot panicle (i.e., panicle weight) of each cultivar was measured.

Statistical analyses
Average grain yields over different field trials, in which May and June precipitation differedby < 5 mm, were calculated. To establish the cultivar response to early summer precipitation in long-term field trials, the grain yield ofeach line was calculated relative to the check cultivar in as many experiments as possible, and the results were interpreted according to Larsson (1982). In greenhouse experiments, response of oat lines to drought was compared by calculating indices. The index of each physiological trait was calculated for each line by dividing the value for the trait measured under water deficit by thatrecorded under adequate watering. The greater the deviation of each physiological index from 1, the greater the response of the specific line to drought stress. The data from greenhouse experiments, arranged in a groups of old (released before 19705), modern, and wild oat types, were analyzed by discriminant analysis (Statgraphics 1992).

Results and discussion
Differences among oat lines in response to early summer precipitation were moderate when extrapolated from long-term field trials (Fig. I).
The regression coefficients for different oat lines ranged from -0.041 to 0.093. Therefore, 100 mm decrease in May-June precipitation resulted in 4.1% higher to 9.3% lower grain yield relative to Puhti (i.e., from ca. 180kg more to 400 kg ha' 1 less grains, respectively). Grain yields of the modern lines Hankkijan Vouti, Karhu, Pol, Ryhti, and Veli, were reduced more than that of the long-strawed old cultivars Kyrö and Hannes due to reduced early summer precipitation. Virma and Sisu were exceptions to this tendency. Hja 76416 is a short-statured breeding line, characterized as one of the most stable in northern growing conditions (Peltonen-Sainio et al. 1993) but showed, however, average responsiveness to early summer precipitation.
Discriminant analysis demonstrated differences among the modern, old, and wild oat types in theirresponse to moderate pre-anthesis (exps 1 and 2) and severe post-anthesis drought (exp 3) simulated in the greenhouse. In exp 1, the response to pre-anthesis drought was recorded immediately after the first signs of loss of turgidity and in exp 2, after prolonging the effects of drought. The response of wild species (A. barbata and A. sterilis max.) differed from those of the cultivated oats (Fig. 2). In exp 1, this was mainly due to fewer drought-induced changes in their relative water content and stomatal conductance of the uppermost fully expanded leaf blades, in exp 2, due to lower accumulation ofABA and proline, and more constant stomatal resistance and leaf temperature, and in exp 3, due to more constant stomatal resistance and better ability to fill grains under inadequate watering (Tables 1   and 2). Old oat lines tended to be least able to maintain high water content during early stages of drought, but according to discriminant analysis there were no differences between modern and old lines subsequently (see exps 1 and 2 in Fig. 2). However, mean values of indices for ABA and proline content, stomatal resistance, and leaf temperature (Table 2) indicated that, in general, the modern lines were slightly more efficient than the old lines underdrought stress also in exp 2. When drought-stressed close to heading (exp 3), weight of grains per panicle was   Larsson (1982) and Larsson and Gorny (1988) concluded that the most reliable ranking of oat lines according to theirresponse to water deficit is established by combining the results from field trials and greenhouse experiments. However, ranking of the oat lines according to long-term field trials and greenhouse experiments differed. For example, Hannes and Kyrö identified as the most drought tolerant oat lines according to field trials, were among the most sensitive lines (exp 1) or were of average performance (exps 2 and 3) when tested in the greenhouse. Probably the main reason for the inability of the old cultivars to cope with pre-anthesis drought when tested in the greenhouse, is that growing them in pots masks the effect of root depth as a tolerance mechanism. However, evidently it is difficult to rank cultivars reliably based solely on the results from field trials, despite some differences in response of oat lines to pre-anthesis precipitation. Numerous factors other than precipitation cause variation in grain yield. This is possibly emphasized when testing modern cultivars solely and lacking such large intrinsic differences in plant stand structure as included in this study. Furthermore, comparing drought response of the most recently released cultivars may be restricted due to too few data points. Therefore, the simple and non-laborious methods represented in this study seem to be biased and are not suggested for use when ranking drought tolerance of oat lines.