Go back to JSRRweb homepage:ホームページへ戻る
Minoru Yamauchi*1 and D.V. Aragones2
1:Japan International Research Center for Agricultural Sciences,
Presently, Chugoku National Agricultural Experiment Station,
Nishifukatu 6-12-1, Fukuyama 721, Japan and
2:International Rice Research Institute (IRRI),
P.O. Box 933, 1099 Manila, Philippines
Grain yield is determined by the interaction between genes and environmental factors including climate, soil, farming practice, and pests. Grain yield is expressed as follows:
Grain yield = Shoot dry matter x Harvest index
or
Grain yield = Spikelet number/panicle x Panicle number/m2 x Percentage filled spikelet x 1000 grain weight
Improvement of these characteristics would increase grain yield, although improving one characteristic often results in deterioration of others.
Hybrid vigor increases grain yield. F1 rice hybrids have larger shoot dry matter accumulation and spikelet number per panicle than the parental lines, producing more grain (Virmani and Edwards, 1983; Yamauchi, 1994). The physiological characteristics of F1 hybrids are expressed by heterosis which is the ratio to the mean of parental lines.
Nutrient uptake is a prerequisite for dry matter accumulation and is determined by the morphological characteristics and metabolic activity of roots. The major morphological characteristics related to grain yield are the dry matter partition to roots and root length. In the field, the root length per unit volume of soil (i.e. root density) would determine the efficiency of nutrient uptake. The number of elongated roots is correlated with spikelet number and grain yield (Harada et al., 1984). Grain yield increased up to 5500-6000kg/ha with increases in superficial roots or root length (Kawata et al. 1978; Morita et al., 1988).
Larger root systems were observed in F1 hybrids (Lin and Yuan, 1980). The root morphological characteristics of F1 hybrids have little been measured in the field. Root activity measured by nitrogen uptake was higher in F1 hybrids than parental lines when expressed per plant but not per unit weight (Suzuki and Morooka, 1986; Hasegawa et al., 1991).
This study was performed to clarify the relationships between the characteristics related to grain yield and roots with various cultivars/lines and F1 hybrids and in different growth seasons. In addition, we analyzed the heteroses in root morphological characteristics related to dry matter accumulation and yield components to understand the relative importance of root heterosis to grain yield heterosis.
Materials and methods
We conducted three experiments each in 1990-wet and 1991-dry seasons at IRRI, Philippines. Each experiment included 2 F1 hybrids, their parental lines, and check cultivar IR72. The design was a randomized complete block with three replications. Two-week old seedlings were transplanted at one per hill at 20 x 20 cm spacing. No fertilizer was applied in the 1990-wet season experiments while 100 kg of N, and 30kg each of K2O and P2O5/ha were applied before transplanting in the 1991-dry season experiments.
A root sampler (20 x 20 x 50 cm inner dimensions), which was a modification of that described by Thangaraj and O'Tool (1986), was driven into the soil to a depth of 40cm with a hammer and pulled out at flowering . The removable side sheet of the sampler was opened. The soil in the sampler was cut at 5, 10, 20, 30, and 40 cm depths. The soil blocks were frozen, and then roots in the soil were collected using running tap water. Root length was measured with a Comair Root Length Scanner. The roots of three hills were sampled per replication.
Results and Discussion
There was a significant correlation between grain yield and root length (correlation coefficient r=0.549**, significant at 1% level) when the data from six experiments were combined. The coefficient was larger than that between grain yield and leaf area index at flowering (r=0.371*, significant at 5% level), suggesting the importance of root length to achieve high grain yield. The correlation between grain yield and root density was higher in the soil layers from 5-10 (r=0.573**), 10-20 (r=0.540**), and 30-40 cm (0.466**) than in those from 0-5 (r=0.392*) and 20-30 cm (r=0.297ns, not significant).
Check cultivar IR72 produced more grain in the dry (6150 kg/ha) than in the wet season (4530 kg/ha), which was due to the increase in harvest index. Yield components of spikelet number/m2, percentage filled spikelets, and 1000-grain weight were increased in the dry season by 15, 11, and 7%, respectively. Root density at the soil layers of 5-10, 10-20, and 30-40cm increased in the dry season, resulting in a 27% increase in total root length. The root density in the 5-10 and 10-20 cm soil layers showed a significant correlation with grain yield with r=0.841* and 0.828*, respectively.
Mean heterosis in grain yield of was 1.18 , which was the result of increased shoot dry matter at maturity (heterosis=1.14) and increased spikelet number/panicle (1.14). Heterosis in leaf area was 1.17, correlated to heterosis in grain yield (r=0.634*). The mean heterosis in root length was 1.09 and was not correlated with grain yield (r=0.138ns), suggesting that heterosis in leaf area is primarily important in achieving high grain yield in F1 hybrids.
The same level of grain yield was obtained with rice plants having varying root length. The rice with longer roots tended to have larger shoots, and higher dry matter partition to roots or finer roots than the rice with shorter roots. The rice with shorter roots produced the same level of grain as the rice with longer roots because it had high harvest index. There was a negative correlation between harvest index and dry matter partition to roots in the data from six combined experiments (r=-0.414*) and in heterosis (r=-0.918**). Thus, grain yield would be increased by using rice plants with high harvest index and a small root system or with large shoot and root size.
References