Assessment of grain yield and quality improvements in rice by modern breeding techniques in Bangladesh and projections of future rice improvements

R.K. Dutta, M.A. Baset Mia, B.P. Lahiri and M.A. Salam
Crop Physiology Division, Bangladesh Institute of Nuclear Agriculture, PO Box 4, Mymensingh, Bangladesh

INTRODUCTION

Increased grain yield and improved rice quality are absolutely necessary to feed the world's galloping population and to maintain its health and nutrition. Thus, the recent approach for rice production includes the improvement of both yield and grain quality to cater for consumer demand and also to increase the nutritional level of the general public. But grain yield and protein content are generally correlated negatively (Govindaswami and Ghosh, 1974); therefore, improving both without breaking the existing correlation is not possible. There are many modern techniques for improving rice, but a suitable method for improving one particular trait has not yet been identified. After the development of IR8, on the basis of the proposed rice ideotype model by morphological defect elimination, other similar miracles in rice improvement could not be achieved because of the unavailability of superior characters in the available germplasm. The first international symposium on hybrid rice, which was held in China in 1986, felt that the productivity of modern semi-dwarf varieties of rice had reached a ceiling in many countries with advanced production technology (Ikehasi, 1990).

In China, a commercial hybrid rice known as F1, which was developed by the exploitation of heterosis in rice by developing a cytoplasmic male sterility gene restoration system with a reliable male sterile line using the cytoplasm of wild rice species, was able to show a breakthrough regarding rice yield, recording a jump in yield of up to 25 percent with a potential 6.5 tonnes/ha maximum. But, unless the days to maturity decrease and quality improves, the standard is not satisfactory for consumer demand.

In Bangladesh, scientists have been working for a long time on improving the yield and quality of rice using different techniques; hybridization, mutation breeding and tissue culture have stimulated the creation of variability in rice in respect of crop and grain quality characteristics. In addition to grain yield and hulling recovery, grain physico-chemical properties have far-reaching significance in attracting consumers. Grain fineness and cooking quality are the most important characteristics for consumers and protein content and amino acid composition are important factors from a nutritional point of view.

The Bangladesh Rice Research Institute at Gazipur has developed 32 rice varieties by hybridization with a yield of up to 6.5 tonnes/ha (Table 2). On the other hand, the Bangladesh Institute of Nuclear Agriculture at Mymensingh has developed some mutants from HYVs and local varieties by gamma irradiation from 60-Co sources, hybrid lines by hybridization of mutants with standard varieties and somaclones by creating somaclonal variations of standard rice varieties using 2,4-D in tissue culture media. Among the rice lines developed, some achieved superior qualities over their parents and were released as varieties.

IRRI scientists have recently proposed a new model for rice improvement with an ambitious yield attainment of 11-15 tonnes/ha. The model emphasized a minimum of three to four tillers per hill, a sturdy root system, a dark green leaf colour and a better vascular system, etc. (That and Tran, 1989). The objectives of this investigation were:

1. to analyse critically the achievements obtained by different techniques;
2. to find out the criteria for further improvements which may be possible using different techniques; and
3. to analyse the grain yield components in the light of the minimum tiller hypothesis and the strategy for further yield improvement.

MATERIALS AND METHODS

Two experiments were carried out with two different objectives: Experiment I aimed at evaluating the developed rice lines along with their parents; and Experiment II aimed at testing the validity of the minimum tiller hypothesis.

Experiment I. Twenty-four rice hybrid lines/mutants/somaclones/varieties were used in this experiment (see Tables 1 and 3). The experiment was conducted from August to December 1990 at the farm of the Bangladesh Institute of Nuclear Agriculture, Mymensingh. The experiment was laid out in a randomized complete block design with three replications. The unit plot size was 5 x 3 m and the plants were spaced at 20 x 15 cm intervals. Fertilizers were applied at the rate of 70 kg N, 30 kg P2O5, 30 kg K2O and 10 kg S in the form of urea, triple superphosphate, muriate of potash and gypsum.

Experiment II. This experiment was conducted from January to June 1994 in field conditions with three varieties (TL, BR 4 and BINA-19) and three spacings (20 x 15 cm, 15 x 10 cm and 10 x 10 cm) designed as a randomized complete block with three replications. The unit plot size was 4 x 3 m. Forty day-old single seedlings were transplanted per hill. Fertilizers were applied as per Experiment I. Fifteen hills were harvested randomly from three locations to collect data on plant height and yield components. Yield data were recorded from a 10 m2 harvested area. The data were analysed statistically and the mean values were evaluated by DMRT.

RESULTS AND DISCUSSION

Grain yield improvements by different techniques

The results presented in Table 1 show that BINA-101 and BINA-115, obtained by crossing BR 4 and IRATOM 38, did not achieve any higher yield than the better parent BR 4 and grandparent IR8. Rather, these lines showed intermediate characteristics regarding yield indicating a quantitative gene effect for yield expression.

The lines obtained from the crosses between IRATOM 24 and BR 4, viz. BINA-6, BINA-14, BINA-17, BINA-28 and BINA-163, also showed less grain yield in respect of any parent and grandparent. The lines obtained from crosses between BINASAIL and BR 4, viz. BINA-2, BINA-10, BINA-16 and BINA-90, also showed reduced yield in respect of both parents. Mutant varieties obtained by gamma irradiation of IR8, viz. IRATOM 38 and IRATOM 24, showed earliness and less photosensitivity with some yield reduction. BINASAIL developed from NIZERSAIL (a strongly photosensitive variety), showed a 10 to 15 percent higher yield and less photosensitivity than its mother (Table 1).

Rice lines developed from crosses between BR 4 and IRATOM 24 showed no improvement in protein content over their parents. Again, hybrid lines developed from crosses between BR 4 and BINASAIL, viz. BINA-2, BINA-10, BINA-16 and BINA-19, showed a reduction of protein content compared with their parents (Table 3). However, in most cases, rice lines showed an increased grain length and decreased breadth compared with their parents (Table 3). Thus, it was evident that grain fineness could be achieved by hybridization. Mutants IRATOM 24, IRATOM 38 and BINASAIL showed a much higher protein content than that of their mothers. These lines also showed grain fineness, e.g. increased grain length in BINASAIL, IRATOM 38 compared with their mother and a decreased grain breadth in IRATOM 38 and IRATOM 24.

Somaclonal lines, viz. P-3, P-7, P-8 and P-9, showed a reduced protein content compared with their mother. However, similar grains were obtained in most of the somaclonal lines, generally reducing the length and breadth and changing the shape to coarseness.

Relationship of grain yield with yield components and population pressure

Table 4 shows that the grain yield of rice, in general, does not show much correlation with panicle length
(r = 0.262), grains/panicle (r = 0.236), panicles/hill (r = 0.309) and grain size. Among the yield components, the number of grains/panicle generally shows dominance in contribution to grain yield/m2. In many cases, these characters are compensatory in nature. As yield/m2 is the function of the number of panicles/hill x grains/panicle x grain size x population/m2. Thus there is scope for increasing the number of panicles that have a decreased number of grains, or increasing the grain size of those with a reduced number of grains, or increasing the number of grains and the grain size where there is a decreased number of panicles/hill under adequate population pressure (Table 4).

TABLE 1

Agronomic characteristics of selected rice mutants/hybrids/somaclones/varieties recorded as autumn rice

Note: W - winter season.

The duration of rice lines (Table 1) shows an apparent relation with grain yield which is not true for all the rice lines. The results presented in Table 4 reveal that higher population pressure reduces the number of tillers to a minimum of three to four per hill. The effect of population pressure was the increase of tillers/m2 from 268 to 408. The number of panicles/hill decreased but the number of panicles/m2 increased in parallel with the number of tillers. This decrease in the number of grains/panicle complied with a proportional increase of panicles/m2. Panicle length, grain size and seed yield/m2 were unaffected. TDM per m2 increased but the harvest index decreased in proportion to the increasing population pressure, suggesting a compensatory mechanism operating in relation to the yield potential of a variety with its yield components.

The experiment results also suggested that, with three to four tillers/hill achieved by population pressure, the rice yield under Bangladesh's spatial and temporal conditions reaches its ceiling of 6.5 tonnes/ha for indica-japonica type in the winter. Manipulating the spacing shows a compensatory mechanism between the number of panicles/m2 and the number of grains/panicle. Why the number of grains per panicle decreases under population pressure is not known. Table 1 also shows that the number of tillers in some varieties, e.g. BR 3 and NIZERSAIL, was higher but that the number of effective tillers or panicles/hill was proportionally higher in NIZERSAIL than in BR 3. However, the higher number of panicles/hill in NIZERSAIL did not guarantee a higher yield. In many varieties where the ratio of panicles/hill is higher, the yield does not increase substantially because either the number of grains/panicle or the grain size decreases while the number of unfilled grains on terminal panicles increases, showing photosynthetic wastage.

TABLE 2

Plant height, growth duration and yield of rice varieties of Bangladesh

It has been reported that the reproductive structure in rice is predetermined at the panicle initiation stage (Matsushima, 1975), controlled by its genetic make-up. Thus, enhancement of the potential by increased photosynthetic reserve before panicle initiation by breeding might contribute to yield increase. In doing so, an attempt should be made to build a proper plant structure before panicle initiation. This may be done on the basis of a plant structure capable of getting the maximum benefit from solar energy. For rice, the duration of grain filling is predetermined but the rate varies according to the size of the grain - a higher rate for larger grains (Mia et al., 1995).

In rice, seed size is predetermined genetically; if the rate is increased, the grain will be coarser. If the duration of grain filling is lengthened, a greater number of seeds may be fed. On the other hand, a long-duration rice crop is not desirable; therefore, increasing the number of grains/panicle only is the most logical means of rice yield enhancement. In doing this, the most important phenomenon to which rice plants can adapt is increasing the photosynthetic reserve before panicle initiation with better translocation. This may be done on the basis of a leaf area index (LAI) at panicle initiation (PI) and a higher harvest index at harvest. For varieties with a higher biomass and harvest index the yield is higher. But there is also a compensatory mechanism with the number of grains/panicle and number of panicles/unit area. Therefore, it is usually impossible to increase the biomass/unit area and the harvest index. In somaclonal lines, the number of grains/panicle increases. Thus, recent photosynthates must be generated at the time of grain filling to support the maximum number of grains/panicle, which is ensured by proper plant structure development before panicle initiation.

Under efficient light absorption conditions during grain filling, rice varieties with a longer growth duration can support larger panicles than those with a shorter growth duration. However, panicle weight can also be increased more satisfactorily by increasing the number of grains rather than the grain size.

The latest IRRI model of rice plant type emphasizes low tillering (3-4)/hill, large panicles (250 grains/panicle), a sturdy stem with a large vascular bundle, a strong root system and higher harvest index, resistance to diseases and pests, dark green, thick, erect leaves, etc. with a yield potential of 11-15 tonnes/ha. This theoretical model has not yet been achieved in the rice plant. However, there is scope for analysis regarding the validity of this ambitious model. First of all, the yield of IR8 worldwide has reached the highest level. IR20, although said to be resistant to many diseases and pests and water stress, is not capable of achieving such results.

TABLE 3

Physico-chemical characteristics of rice genotypes

Research on morphophysiological parameters demonstrates that there is a compensatory mechanism between low tillers, large panicle versus the number of panicles/unit area. If the tillers are low there may be a large panicle; in that case, the population must be kept high otherwise the yield/unit area will not reach a peak. But, in practice, when the population is high, tillers/hill are, of course, low (three to four), but the number of grains/panicle decreases in proportion to the increased population.

On the other hand, if the population is low, the number of tillers will be higher and panicles will again be smaller. If, however, such a type is achieved where tillering is predetermined under any population pressure, there is still a possibility of smaller panicles under adequate population pressure. This is quite paradoxical and makes a dilemma in rice breeding for increasing the yield. However, if the harvest index is increased considerably, large panicles may be achieved, although this is unlikely because there is a tendency for the harvest index to be lower under population pressure (Table 4). In order to achieve further yield improvement in indica-japonica type rice, the preflowering photosynthetic potential must be given maximum emphasis. This will lead to the maximal biomass being attained in the minimum number of days (120 days seed to seed). It has been explained that sink size is a major factor limiting the potential yield, particularly in medium and long growth-duration varieties. Under full sunshine conditions during the grain filling period, rice varieties with a longer growth duration can probably support larger panicles than those with a shorter duration, if the translocation of preheading stored carbohydrate is efficient.

TABLE 4

Effect of population pressure on the architectral parameters of winter rice

It has been observed that in long growth-duration varieties many non-structural carbohydrates remain in the stem and leaf sheath at maturity. The authors noted that a higher leaf area index and a higher crop growth rate after panicle initiation lead to a longer duration of the crop with a decreased harvest index (Mia et al., 1995). It is considered that the IRRI model may help to increase the yield of indica-japonica type of rice further if the model is strictly followed. Several new points may be added with regard to a phenological development pattern as well as some biochemical approaches to yield improvement. The proposal is based on the fact that rice under adequate population pressure shows nutrient and photosynthetic energy limitations under closer canopy development.

These limitations may have four dimensions:

• Nutrient uptake limitations owing to limited rhizosphere and spatial conditions in a closer canopy.
• Lack of nutrient uptake efficiency and photosynthetic assimilate production with regard to plant development vigour.
• Limitations in the distribution of phenomenal utilization throughout ontogenetic development.
• Genetic limitation of the coordination of gene action under stressful conditions.

The proposal suggests that the following manipulation of indica-japonica type of rice in addition to the proposed IRRI model may overcome the limitations of nutrients and photosynthates:

• Efficient and extensive root system. Within the limited rhizosphere where the nutrient supply may be limited owing to a soil availability problem, higher uptake efficiency of nutrients by the root system may enhance the availability of nutrients in the soil solution. The growth of plants and their yield and nutrient uptake are highly influenced by the cation-exchange capacity (CEC) of their roots (Drake, Vengris and Colby, 1951; Rao, Krishnamurthy and Rao, 1967; Singh, Srivastava and Shanker, 1990). In that case, root cation-exchange capacity must be increased. On the other hand, the adventitious root system must be extensive rather than a mass so that roots can build up a network extension system to take up nutrients from a distance rather than competing among themselves in a limited space.
• Erect and narrow top canopy leaves. The top canopy leaves must be erect and narrow but the flag leaf should be broad and erect to avoid mutual shading. Leaves must not be droopy at any position on the plant. The top canopy leaves must be highly efficient for light absorption. For this purpose, light green leaves are preferred as they contain a higher chlorophyll a/b ratio and are more efficient in light absorption. However, the lower leaves may be deep green with higher chlorophyll b for the purpose of shade adaptation.
• Distribution of phenomenal utilization. The production of photosynthates and translocation towards the sink must be synchronized with economic needs. Although a rice variety may have a high growth potential, it must be expressed fully before panicle initiation so that post panicle initiation photosynthesis may be utilized for grain development, thus accelerating the growth cycle and providing maximal economic benefits from a minimal crop duration.
• High genetic potential of varieties/genotypes. The genotypes/varieties must have a higher growth rate, i.e. the hormonal relationship of the genotypes must be positive even under stressful conditions. Sufficient enzymes such as the major enzyme of photosynthesis, RuBP Carboxylase/Oxygenase, and the major enzyme of N-metabolism, Glutamine synthetase, should be synthesized for binding particular hormones as signals for the purpose of growth. Therefore, the genotypes must be high in RuBP Carboxylase/GS. For this purpose, the preflowering level of soluble protein in the leaf should be the determining factor.

REFERENCES

Drake, M., Vengris, J. & Colby, W.G. 1951. Cation-exchange capacity of plant roots. Soil Sci., 72: 139-142.

Govindaswami, S. & Ghosh, A.K. 1974. Breeding for high protein content in rice. Indian J. Genet. Plant Breed., 34(A): 628-641.

Ikehasi, H. 1990. Problems and potential of hybrid rice breeding. Int. Rice Comm. Newsl. p. 8-9.

Matsushima, S. 1966. Crop science in rice. Tokyo, Fuji Publ. Co. 365 pp.

Matsushima, S. 1975. Analysis of the development factor determining yield and yield production in lowland rice. Bull. Natl Inst. Agric. Sci., Ser. A(5) (Japan). p. 1-271.

Mia, M.A.B., Ali, M.A., Khan, M.A.H. & Dutta, R.K. 1995. Crop growth and grain development pattern of some coarse and fine rice varieties. Bangladesh J. Crop Sci.

Mia, M.A.B., Salam, M.A., Chowdhury, S.I., Islam, A.F.M.S. & Dutta, R.K. 1994. Morpho-physiological studies in relation to yield potential of rice. Bangladesh J. Nuclear Agric.

Rao, K.C., Krishnamurthy, T.N. & Rao, J.T. 1967. Cation-exchange capacity of roots and yield potential in sugar cane. Plant Soil, 26: 314-318.

Singh, S. & Ram, L.C. 1978. Studies of relationships between cation-exchange capacity of plant roots and tillering and plant growth of different varieties of paddy and wheat crops. Plant Soil, 49: 661-665.

Sing, B., Srivastava, O.P. & Shanker, H. 1990. Root cation-exchange capacity (CEC) and yield contributing parameters in rice in normal and alkali soils. Int. Rice Res. Newsl., 15: 19-20.

That, T.T. & Tran, D.V. 1989. Prospects for a new generation of high-yielding rice varieties. Int. Rice Comm. Newsl. p. 15-19.

EVALUATION DU RENDEMENT EN GRAIN ET AM�LIORATION DE LA QUALIT� DU RIZ AU MOYEN DE TECHNIQUES MODERNES D'OBTENTION AU BANGLADESH

L'�tude comparative se penche sur la r�ussite relative des diff�rentes techniques d'obtention visant � am�liorer le riz, par exemple l'hybridation, la s�lection par mutations et la culture tissulaire. Elle r�v�le que l'hybridation a r�ussi � am�liorer la finesse du grain mais non le rendement par rapport aux parents. La s�lection par mutations au moyen du rayonnement gamma s'est r�v�l�e une technique susceptible d'introduire la pr�cocit� du riz, d'am�liorer la finesse du grain, d'accro�tre sa teneur en prot�ines et son rendement en grain. La culture tissulaire visant � induire une variation somaclonale a permis d'am�liorer le rendement en grain et sa finesse. Les �l�ments du rendement n'ont gu�re fait appara�tre de corr�lation avec le rendement en grain et ont r�v�l� un m�canisme de compensation r�ciproque. La pression accrue de la population n'a pas amen� d'am�lioration des rendements du fait de la r�duction proportionnelle du nombre de grains par panicule. Il a �t� propos� de manipuler les caract�res morphophysiologiques afin d'am�liorer encore le riz de type indica-japonica.

EVALUACI�N DE MEJORAS EN EL RENDIMIENTO Y LA CALIDAD DEL GRANO DEL ARROZ MEDIANTE T�CNICAS GEN�TICAS MODERNAS EN BANGLADESH

Se llev� a cabo un estudio comparativo sobre diferentes t�cnicas gen�ticas, como por ejemplo hibridaci�n, mejoramiento de mutaciones y cultivo de tejidos, para mejorar variedades de arroz.
El estudio revel� que la hibridaci�n hab�a conseguido mejorar la finura del grano, pero no el rendimiento con respecto a las l�neas parentales. El mejoramiento de mutaciones mediante radiaci�n gamma demostr� ser una t�cnica con posibilidades de inducir precocidad, mejorar la finura del grano e incrementar el contenido de prote�na y el rendimiento en grano. El cultivo de tejidos para inducir variaciones som�ticas clonales ofreci� resultados positivos en cuanto al rendimiento y finura del grano. Los componentes del rendimiento no mostraron una estrecha correlaci�n con el rendimiento en grano y pusieron de manifiesto un mecanismo de compensaci�n entre ellos. El aumento de la presi�n de la poblaci�n no contribuy� a la mejora del rendimiento a causa de la reducci�n proporcional del n�mero de granos por pan�cula. Se ha presentado una propuesta de manipulaci�n de los caracteres morfofisiol�gicos para seguir mejorando tipos de arroz Indica y Japonica.