[ホーム＞発表論文＞本文]

吉田総合案内へ戻る．
本頁は PPS 11に掲載．図省略．

Yield Stability of Aromatic Upland Rice with High Yielding Ability in Indonesia

Totok Agung Dwi Haryanto1), Suwarto1) and Tomohiko Yoshida2)
(1)Faculty of Agriculture, Jenderal Soedirman University, Purwokerto, Indonesia; 2)Faculty of Agriculture, Utsunomiya University, Japan)
Abstract : Aromatic rice variety, Mentikwangi, was crossed with high-yielding upland rice variety, Poso, and the pedigree was selected to obtain lines with high yielding and aromatic characters. The objectives of the research were to study the yield stability of aromatic upland genotypes across different locations and to select aromatic upland rice genotypes having wide adaptability, and or specific location adaptability. Yield stability of genotypes was estimated by using regression lines proposed by Finlay and Wilkinson. Some genotypes showed high yield stability and wide adaptability in different locations, and others showed good adaptability to a specific location. The lines having high yield stability and wide adaptability were G10 (405 g m-2), G19 (400 g m-2), G39 (418 g m-2), and G136 (411 g m-2), which may be considered as candidates of new aromatic upland rice cultivars. Situpatenggang had specific adaptability at the fertile locations; and Poso and G13 at the infertile locations. Genotype x location interactions for the yield and its components performance were observed.
Key words: Adaptability, Aromatic upland rice, Yield stability.

Rice is one of the most important food crops in the world and the second largest cereal crop, rice is the staple food of nearly one-half of the world’s population. It contributes over 20% of the total calorie intake of the human population (Chaudhary and Tran, 2001). In general, rice is boiled and eaten as a main dish; however, with the development of the processing industry and increased prosperity, it is processed to create a variety of rice products, as well as forming a constituent of a wide range of snack foods, baked products and beverages. Different kinds of preparation of rice demand different grain qualities; both physical and chemical. There are rice varieties which produce grain of a different kind in terms of physical appearance, chemical composition or aroma. A number of aromatic rice varieties are known in various countries. In India, Pakistan, Thailand, Bangladesh, Nepal, Iran, Afghanistan, Myanmar, and also Indonesia these rice varieties are the most prized.

There are many studies on aromatic rice in Asia. Breeding, production and future prospects of aromatic rice have been studied in China (Tang and Wang, 2001), India (Rani and Krishnaiah, 2001; Singh et al., 2001; Bentur and Krishnaiah, 2001; Modgal and Gupta, 2001), Cambodia (Sarom, 2001), Myanmar (Nwe et al., 2001), Pakistan (Mann and Ashraf, 2001), Thailand (Narula and Chaudhary, 2001) and Viet Nam (Nghia et al., 2001a; 2001b).

In Indonesia, aromatic rice which have a perfumed, nutty flavour, aroma, and have a light, fluffy texture when cooked was already cultivated for a long time. Farmers are usually growing local aromatic rice cultivars in an irrigated land, such as Pandanwangi, and Rojolele. However, the local cultivars are late maturity (more than 120 days) and sensitive to pest and disease. Several new aromatic rice cultivars already released in recent years such as Bengawan Solo (1993), Sintanur (2001) and Batang Gadis (2002) which have high yielding, early maturing, and tolerant to pest and disease relatively.

Kato et al. (2006a) mentioned that developing of new water-saving rice production systems, besides enhancing and stabilizing of current rainfed rice production systems, was one of the major options for the increase of rice production using the limited water resource. Studies on the cultivar x environment interactions for yield of upland rice (Lafitte and Courtois, 2002), dry matter production of upland rice (Kato et al., 2006a), and grain yield of upland rice (Kato et al., 2006b) have been reported. However, information about aromatic rice specifically for cultivation in the upland is still limited. An upland aromatic rice cultivar, Situpatenggang, already released in 2002.

In Indonesia, development of aromatic upland rice is of importance to improve upland rice quality and upland productivity. The upland area in Indonesia is around 11.6 million ha, and is not used for crop production optimally yet. The area for upland rice production is only 1.2 million ha, and it produces 2.6 million tons per year with productivity of 2.27 t ha-1 (Central Bureau of Statistic, 2004). 　

We have studied the development of the aromatic upland rice in Indonesia. It started with crossing upland rice tolerant to drought with lowland aromatic rice. Poso cultivar (high-yielding upland rice, tolerant to drought, Indica type) and Mentikwangi (low-land aromatic rice, Javanica type) has been conducted in 2001. Genetic studies and the maternal effect of aromatic character have been reported (Totok et al., 2005). The pedigree has been selected from the F2 to F5 generation for aromatic upland rice lines (Totok, 2004). In 2004, 50 selected lines of F5 were examined for their growth and yield performance. Among them, 25 lines were aromatic, and among those 25 aromatic lines, 19 lines were high yielding in terms of grain yield per plant. Preliminary yield trial was conducted for 19 lines, and 9 lines were selected as aromatic and high yielding upland lines (Totok and Utari, 2005).

It is necessary to examine the yield stability of the aromatic upland rice, and to obtain lines having high yield stability across different locations and/or having specific locational adaptability. The yield stability across different locations varies with the genotype (Fehr, 1987). Finlay and Wilkinson (1963) used regression analysis for estimating the stability. Relationship between the yield at each location and the location index was repressed to a straight line (regression line) for each genotype and for mean yield of all genotypes. Then, the regression lines for the yield of each genotype were compared with that for the mean yield of all genotypes over all locations to estimate the yield stability of each genotype. Yield and yield components of a crop are influenced by genotype (G), environment (E), and their (GxE) interaction. Every factor of the environment has a potential to cause differential performance, associated with GxE interaction (Fehr, 1987). GxE interaction was measured by the analysis of variance. Materials and Methods

1.Materials
Nine rice lines obtained by crossing between Poso (Ps) cultivar (high yielding upland rice, tolerant to drought, Indica type) and Mentikwangi (Mw) cultivar (lowland aromatic rice, Javanica type) were used in this experiment. Four other rice cultivars, namely Ps and Mw (as parents), Silugonggo (Slg), and Situ Patenggang (Stp) were also used as reference genotypes. Slg and Stp were upland rice cultivars released in 2002 by Ministry of Agriculture, Indonesia. Thus, 13 genotypes in total were used in this field experiment.

2.Methods and design
The 13 genotypes were planted in the upland at eight different locations in Java, Indonesia in January 2006 and harvested in May 2006. The locations were Purworejo (Pwj), Banyumas (Bms), Kudus (Kds), Tegal (Tgl), Batang (Btg), Kebumen (Kbm), Cirebon (Crb) and Banjarnegara (Bjn). At each location, seeds of each genotype were sown directly in a 5 x 2.5 m plot in the rate of 3-4 seeds per hole. Seedlings were thinned to 2 seedlings per genotype. Planting distance was 25 cm between rows and between plants. No irrigation was applied. Water supply was depended on the rainfall. The experiment was a randomized complete block design with three replications. Fertilizer rate was 200 kg ha-1 N, 100 kg ha-1 P2O5 and 100 kg ha-1 K2O. Data for plant height (cm), the number of productive tillers per hill, panicle length (cm), the number of grains per panicles, 1000-grain weight (g), and grain weight (g) per hill were collected from 5 randomly chosen plants from each plot. Yield (g) was measured from 5 m2 effective plots. Data were analyzed combined with variance. Means were separated by Duncan’s Multiple Range Test (DMRT) when the variance analysis revealed significant differences (Steel and Torrie, 1980). Regression linear analysis proposed by Finlay and Wilkinson (1963) was used to analyze the yield stability of genotype, as follows.

Yij = μ + gi Ij + σij
where:
Yij: yield mean of a genotype i at the j location
μ : Population mean
gi : Regression coefficient of the i genotype
Ij : Environmental index of the j location
σij : Regression deviation of the i genotype at the j location

The genotype having the regression line above that for mean yield of all genotypes over all locations is considered to have high yield stability and capable of adapting to all the locations. Such genotype would increase the yield as the productivity of the location improves. The genotype having a regression line crossing that for the mean yield of all genotypes over all locations is considered to have adapted well to a specific location. The genotype having a regression line below that for the mean yield of all genotypes over all locations is considered to have low yield adaptability across locations (Finlay and Wilkinson, 1963).

Results and Discussion
All of the nine upland rice lines and four reference cultivars sown at eight different locations grew well. Plants headed between 59 and 92 days after planting (DAP). The grains were harvested between 94 and 115 DAP.

Analysis of variance showed a significant effect of location, genotype and their interaction on yield and all yield components (data not shown). Table 1 shows the plant height of each genotype at each location. Plant height varied significantly with the genotype of each location. The lowest plant was Slg in Batang (67 cm) and the highest was G35 in Kebumen (156 cm). They also had the lowest and the highest mean values as the average at all locations.

Tables 2, 3, 4, 5, 6 and 7 show the number of productive tillers per hill, panicle length, the number of grains per panicle, 1,000-grain weight, grain weight per hill, and yield, respectively, of each genotype at each location. These values significantly varied with the genotype at each location in plant height, indicating the genetic variability. This is because the genotypes used in this study originated from F2 population of Mw x Ps crossing.

The results indicated that each line responded differently to the location conditions, showing that the genotype’s ranking differed with the sowing location. Different responses of upland rice cultivars to water conditions have been reported (Kato et al., 2006b). These results suggested that yield and yield components were influenced by genotype x location interaction as shown in our experiment. Lafitte and Courtois (2002) also reported that the anthesis date, leaf fresh weight, root pressure, leaf area and rooting depth of upland rice were changed by cultivar x environmental interaction, The number of productive tillers per hill significantly varied with the genotype at each location. The number of productive tillers per hill was lowest in Stp (8 tillers) at Bjn and highest in Slg at Kds (26 tillers). It was not different among G10, G12, G136 and Mw at Kds. Stp and Slg had the smallest and the largest number of productive tillers per hill, 12 and 22 tillers, respectively, on the average.

Panicle length was significantly different among genotypes at each location. Slg at Bjn had the shortest panicle (18 cm), which was not different from that at Btg. G9 at Kbm had the longest panicle (29 cm), which was not different from that of G12, G13, G35, G39 and Ps. Slg had the shortest panicle as the average at all locations, 21 cm. On the other hand, G12 had the longest panicle as the average at all locations (26 cm).

The number of grains per panicle was significantly different among genotypes at each location. The lowest number of grains per panicle was observed in Slg at Bjn and the highest in Stp at Kbm (265), which was not different from that in G39. Slg and Stp had the smallest (96) and largest (175) mean number of grains per panicle as the average at all locations.

The 1,000-grain weight was significantly different among genotypes at each location. Slg at Kds had the lightest 1,000-grain weight (20 g) and G13 at Pwj the heaviest (30 g), which was not different from that in G10. Slg and G10 also had the lightest and the heaviest mean 1,000-grain weight as the average at all locations, 22 and 28 g, respectively. La fitte and Courtois (2002) reported that the mean 1,000-grain weight of 48 upland rice cultivars grown under nine different environments varied from 16.4 to 24.7 g. However, Fukushima et al. (2006) showed that the yearly variation of 1,000-grain weight was from 20.3 to 22.5 g in Akisayaka and from 22.0 to 22.9 g in Yumehikari.

Grain weight per hill was significantly different among genotypes at each location. Slg at Btg had the lightest (12 g) and G13 at Pwj the heaviest (31 g) grains. Slg and G10 had the lightest and the heaviest grain weight per hill as the average at all locations, 22 and 28 g, respectively.

Yield (g m-2) at each location was significantly different among genotypes. The yield was lowest in Stp at Bjn (65.0 g m-2) and highest in G10 at Btg (666.6 g m-2) which was not different from that in G12 at Bjn. Slg and G39 were genotypes having the lowest and the highest yield as the average at all locations, 304.8 and 418.0 g m-2, respectively (Table 7).

The lines having higher mean values of yield components than the reference cultivars had a higher yield than the reference cultivars. For instance, G10 had larger number of tillers per hill than Ps, Mw and Stp; longer panicle than Mw, Slg and Stp; heavier 1,000-grain weight than any other cultivar, resulting in high average yield (405 g m-2 ). G19 had a larger grain number per panicle than either Mw or Slg, heavier 1,000-grain weight than either Stp or Slg, heavier grain weight per hill than any other cultivar, resulting in a high average yield (400 g m-2). G39 had a larger number of tillers per hill than Ps and Stp, longer panicles than any other cultivar, higher grain number per panicle than Ps, Mw and Slg, heavier 1,000-grain weight than Stp and Slg; and heavier grain weight per hill than any other cultivar, resulting in a high average yield (418 g m-2). G136 had a larger number of tillers per hill than either Ps or Stp; longer panicle than Mw, Slg and Stp; a larger number of grains perpanicle than either Mw or Slg; and heavier grain weight per hill than any other cultivar, resulting in high average yield (411 g m-2). Thus, G10, G19, G39 and G136 had higher yield components than the others. On the other hand, Fukushima et al. (2006) indicated that a larger number of spikelets per unit area and having sink and source abilities during the late ripening stage were the two characters required for high yielding ability of Akisayaka rice cultivar. Our results confirmed the high yield component on the average contributed to high yielding ability.

Table 7 shows that yield of G39 was higher than or not significantly different from that of at least 3 reference cultivars at all locations. Yields of G136 and G19 were higher than those of 3 reference cultivars at all locations, except for G136 at Tgl and G19 at Btg. However, the yield of G10 was higher than that of reference cultivars, at 5 locations; Kbm, Btg, Tgl, Kds and Bms. Our study suggested that G10, G19, G39 and G136 are aromatic rice lines having high yielding ability across different locations. The yield of G10, G19, G39, and G136 are 405 g m-2, 400 g m-2, 418 g m-2, and 411 g m-2, respectively.

The genotype having the regression line above that for the mean yield of all genotypes over all locations is estimated to have high yield stability. Our study showed that G10, G12, G19, G39, G136 and Mw had regression lines above that for the mean yield of all genotypes over all locations. It shows that these genotypes may have high stability and good adaptability across 8 locations (Fig. 1).

The genotype having the regression line crossing that for the mean yield of all genotypes is estimated to have specific adaptability. In this study, Stp is considered to have specific adaptability to the fertile (high productive) locations, whereas Ps and G13 have specific adaptability to the infertile (low productive) location (Fig. 2).

The regression lines crossing each other revealed the genotype x location interaction. The locational difference in yield of the genotype relative to that of other genotypes shows the locational change of the ranking of the genotype. The genotype having the regression line below that for the mean yield of all genotypes over all locatioins is estimated to have low yield stability. G9, G34, G35 and Slg are considered to have low stability (Fig. 3).

Based on the high yielding (Table 7), high yield components (Table 2-6) and high yield stability (Fig. 1), G136, G39, G19 and G10 are considered as prospective aromatic upland rice lines having high yielding ability, high yield stability and wide adaptability.

In conclusion, yield stability across different locations varied with the genotype. Some genotypes had high yield stability and wide adaptability to all locations, and some others had high adaptability to specific location. The lines having high yield stability and wide adaptability were G10 (405 g m-2), G19 (400 g m-2), G39 (418 g m-2), and G136 (411 g m-2), which are considered as candidates for new aromatic upland rice cultivars. The genotypes having specific adaptability were Stp at the fertile locations, and Ps and G13 at the infertile locations. Genotype x location interaction for the yield and its components were observed. Yield components contributing to higher yielding ability was confirmed.

Acknowledgements
This work was supported in part by a grant from the Technological and Professional Skill Development Project (TPSDP) ADB Loan and by a grant (Beasiswa Unggulan) from Directorate General of Higher Education, Department of National Education, Indonesia. 
References

Bentur, J.S. and Krishnaiah, K. 2001. Insect pest of Basmati and other quality rice 
　and their management in India. In R. Duffy,  R.C. Chaudhary and D.V. Tran eds. 
　Speciality rices of the world. Breeding, production and marketing. FAO of the UN Rome. 
　Science Publishers, Inc. Plymouth, UK. 89-100.
Central Bureau of Statistic, 2004. Indonesian statistic. Indonesian Central Bureau of 
　Statistic,  Jakarta. 35.
Chaudhary, R.C. and Tran, D.V. 2001. Speciality rice of the world: a prologue. In R. 
　Duffy, R.C. Chaudhary and D.V. Tran eds. Speciality rices of the world. Breeding, 
　production and marketing. FAO of the UN Rome. Science Publishers, Inc. Plymouth, UK. 3-14.
Fehr, W.R. 1987. Principles of Cultivar Development. Vol.1. Theory and Technique. Collier 
　Macmillan Publishers, London. 249-259.
Finlay, K.W. and Wilkinson, G.N. 1963. The analysis of adaptation in a plant breeding 
　programme. Aust. J. Agron. Res. 114:742-754.
Fukushima, A., Kusuda, O., Nakano, H. and Morita, S. 2006. Analysis of high yielding 
　ability in a rice cultivar Akisayaka. Plant Prod. Sci. 9:369-372.
Hayashi, S., Kamoshita, A. and Yamagishi, J. 2006. Effect of planting density on grain 
　yield and water productivity of rice (Oryza sativa L.) grown in flooded and non flooded 
　field in Japan. Plant Prod. Sci. 9:298-311.
Kato, Y., Kamoshita, A., Yamagishi, J. and Abe, J. 2006a. Growth of three rice (Oryza
　sativa L.) cultivars under upland conditions with different levels of water supply. 
　1. Nitrogen content and dry matter production. Plant Prod. Sci. 4:422-434. 
Kato, Y., Kamoshita, A. and Yamagishi, J. 2006b. Growth of three rice (Oryza sativa L.) 
　cultivars under upland conditions with different levels of water supply. 2. Grain yield. 
　Plant Prod. Sci. 9:435-445. 
Lafitte, H.R. and Courtois, B. 2002. Interpreting cultivar x environment interactions 
　for yield in upland rice: assigning value to drought-adaptive traits. Crop Sci.42:1409-1420.
Mann, R.A. and Ashraf, M. 2001. Improvement of Basmati and its production practices in 
　Pakistan. In R. Duffy, R.C. Chaudhary and D.V. Tran Eds. Speciality rices of the world. 
　Breeding, production and marketing. FAO of the UN Rome. Science Publishers, Inc. Plymouth, 
　UK. 129-148.
Modgal, S.C. and Gupta, P.C. 2001. Package of practices to cultivate Basmati type aromatic 
　rices in Uttar Pradesh, India. In  R. Duffy, R.C. Chaudhary and D.V. Tran eds. Speciality 
　rices of the world. Breeding, production and marketing. FAO of the UN Rome. Science 
　Publishers, Inc. Plymouth, UK. 101-110.
Narula, A. and Chaudhary, R.C. 2001. Current status and future of the famous aromatic rice 
　variety Khao Dawk Mali in Thailand. In R. Duffy, R.C. Chaudhary and D.V. Tran Eds). 
　Speciality rices of the world. Breeding, production and marketing. FAO of the UN Rome. 
　Science Publishers, Inc. Plymouth, UK. 163-174.
Nghia, N.H., Buu, B.C., Trinh, L.N. and Thao, L.V. 2001a. Speciality rice in Viet Nam: 
　breeding, production and marketing. In R. Duffy, R.C. Chaudhary and D.V. Tran Eds). 
　Speciality rices of the world. Breeding, production and marketing. FAO of the UN Rome. 
　Science Publishers, Inc. Plymouth, UK. 175-190.
Nghia, N.H., Buu, B.C., Trinh, L.N. and Thao, L.V. 2001b. Improvement of aromatic rice in 
　Viet Nam. In R. Duffy, R.C. Chaudhary and D.V. Tran Eds. Speciality rices of the world. 
　Breeding, production and marketing. FAO of the UN Rome. Science Publishers, Inc. Plymouth, 
　UK. 191-200.
Nwe, K.T., Myint, T.T.and Garcia, A.G. 2001. Breeding and cultivation of superior quality 
　rices in Myanmar. In R. Duffy, R.C. Chaudhary and D.V. Tran Eds. Speciality rices of the 
　world. Breeding, production and marketing. FAO of the UN Rome. Science Publishers, Inc. 
　Plymouth, UK.115-128.
Rani, N.S. and Krishnaiah, K. 2001. Current status and future prospects for improvement 
　of aromatic rices in India. In R. Duffy,  R.C. Chaudhary and D.V. Tran eds. Speciality 
　rices of the world. Breeding, production and marketing. FAO of the UN Rome. Science 
　Publishers, Inc. Plymouth, UK. 49-78.
Sarom, M. 2001. Current status of aromatic and glutinous rice varietas in Cambodia: their 
　breeding, production and future. In R. Duffy,  R.C. Chaudhary and D.V. Tran eds. Speciality 
　rices of the world. Breeding, production and marketing. FAO of the UN Rome. Science 
　Publishers, Inc. Plymouth, UK. 19-34.
Singh, R.K., Sahai, V.N., Sharma, R.N., Singh, U.S., Singh, S. and Singh, O.N. 2001. Current 
　status and future prospects for improving traditional aromatic rices varieties in India. 
　In R. Duffy, R.C. Chaudhary and D.V. Tran eds. Speciality rices of the world. Breeding, 
　production and marketing. FAO of the UN Rome. Science Publishers, Inc. Plymouth, UK. 79-88.
Steel, R.G.D. and Torrie, J.H.1980. Principles and Procedures of Statistic. A Biometrical Approach. 
　2nd.ed. McGrow-Hill, New York. 401-437.
Tang, S. and Wang, Z. 2001. Breeding for superior quality aromatic rice varieties in China. 
　In R. Duffy,  R.C. Chaudhary and D.V. Tran eds. Speciality rices of the world. Breeding, 
　production and marketing. FAO of the UN Rome. Science Publishers, Inc. Plymouth, UK. 35-44.
Totok A.D.H. 2004. Growth, yield and rice quality of F5 genotypes from crossing of Mentikwangi 
　X Poso for developing of aromatic upland rice. J. of Rur.  Dev. Jend Soedirman Univ. 2:122-128.
Totok, A.D.H., Suwarto, Daryanto and Soesanto, L. 2005. Constructing of high yielding and aromatic 
　upland rice variety for improving of its production and economic value. J. Agroland. Fac. 
　Agric. Tadulako Univ.3:93-99.
Totok A.D.H. and Utari, R.S. 2005. Yield trial of F6 pure lines from crossing of Mentikwangi x 
　Poso in comparison with their parents. Res. Report. Fac.  of Agric. Jend. Soedirman Univ. 22-30.

Table 1. Plant height (cm) in each genotype at eight locations.
────────────────────────────────────────
GenotypesPwjBmsKdsTglBtgKbmCrbBjnAvg.
────────────────────────────────────────
G9121 de133 ce107 bd135 de116 bc143 eg 111 cd109 de122
G10117 cd123 c112 bf125 cd95 ab136 de111 cd101 cd115
G12132 f127 cd116 df 144 e 105 b148 g125 ef121 f127
G13132 f133 de117 ef143 e111 bc146 fg116 de 111 de126
G19113 bc111 b104 b122 bd 83 ab127 bc103 bc 95 bc107
G34108 b125 cd104 b125 cd 99 ab136 de109 cd106 de114
G35136 f128 cd116 df 125 cd100 ab156 h127 f109 de125
G39116 cd127 cd120 f126 d115 bc139 ef117 de 112 e121
G136121 de138 e109 be108 b 92 ab124 bc 110 cd104 ce113
Ps121 de128 cd114 cf 133 de106 b146 fg121 ef121 f124
Mw124 e127 cd106 bc121 bd139 c130 cd111 cd106 de 121
Slg 85 a 83 a 79 a85 a67 a90 a81 a 69 a80
Stp111 bc104 b107 bd 109 bc 97 ab121 b 97 b89 b104
────────────────────────────────────────
Values with the same letter in a column do not differ significantly at p=0.05.
(Pwj, Bms, Kds, Tgl, Btg, Kbm, Crb, Bjn were Regency of Purworejo, Banyumas, Kudus, Tegal, Batang, Kebumen, Cirebon, Banjarnegara, respectively).

Table 2. The number of productive tillers per hill in each genotype at eight locations.
────────────────────────────────────────
GenotypesPwjBmsKdsTglBtgKbmCrbBjnAvg.
────────────────────────────────────────
G921.2 de18.5a16.5 ac16.0 be21.9 cd19.1 bc18.1bd15.1bc18.3
G1023.3 e17.8 a22.6 de19.3 de15.8 ab20.3 c19.5 cd13.6 ac19.0
G1218.0 bc12.9 bd22.2 ce12.9 ab15.0 ab19.7 bc17.5 bd13.7 ac16.5
G1317.9 bc12.0 cd20.5 bd13.0 ab17.6 bc19.8 bc18.7 bd11.8 ac16.4
G1920.3 ce15.7 ac19.5 bd15.0 bd11.8 a21.2 c15.2 ab10.7 ac16.2
G3420.4 c15.2 ac12.3 a15.9 be13.3 ab20.4 c15.4 ab10.9 ac15.5
G3523.1 e14.3 ac16.3 ab13.9 bc12.0 a20.3 c20.7 de11.2 ac16.5
G3919.5 bd11.9 cd19.5 bd16.3 be2l.9 d16.0 ab16.7 ac10.0 ab16.9
G13618.7 bd14.7 ac21.4 be15.3 bd11.7 a19.5 bc20.2 cd12.2 ac16.7
Ps16.4 b13.9 ac20.3 bd18.0 ce14.4 ab18.3 ac15.5 ab13.4 ac16.3
Mw19.7cd15.9 ac20.8 be16.9 be17.7 bc22.0 c23.9 ef12.3 ac18.7
Slg19.5 bd16.9 ab26.4 e20.3 e 24.5 d22.2 c25.9 f16.9 c21.6
Stp13.3 a89 d17.7 ad 9.2 a11.4 a14.9 a13.1 a 7.9 a12.1
────────────────────────────────────────
Values with the same letter in a column do not differ significantly at p=0.05. (For　Pwj, Bms, Kds, Tgl, Btg, Kbm, Crb and Bjn see Table 1).

Table 3. Panicle length (cm) in each genotype at eight locations.
────────────────────────────────────────
GenotypesPwjBmsKdsTglBtgKbmCrbBjnAvg.
────────────────────────────────────────
G923.2 d25.0 ac24.9 bc24.4 ac23.8 bc29.3 g24.9 ce22.3 bc24.7
G1023.6 d25.3 ab25.0 bc25.8 be23.2 bc27.8 df24.1 bd23.6 bd24.8
G1223.8 d 24.3 ac24.7 bc26.3 ce28.0 d29.0 fg25.4 de24.6 bd25.8
G1323.8 d23.4 be25.7 bc27.1 de24.1 d28.4 fg 23.2 ac24.2 bd25.0
G1921.8 b23.0 ce24.8 bc24.1 ac18.9 a26.7 cd22.8 ab23.0 bd23.1
G3422.0 bc22.6 e22.7 a24.8 ad23.7 bc26.2bc23.4 ad23.5 bd23.6
G3525.5 e22.5 e25.1 bc27.5 e22.8 bc28.7 fg25.0 ce23.9 bd25.1
G3923.5 d24.9 ad26.3 c25.3 be22.2 bc28.2 eg26.2 e23.5 bd25.0
G13623.0 cd25.7 a23.9 ab25.1 be23.2 bc27.0 ce23.9 bd24.8 cd24.6
Ps23.5 d24.1 ae26.1 c26.0 ce22.0 bc28.1 eg23.9 bd25.5 d24.9
Mw21.6 b22.9 de24.1 ab25.0 be21.6 b27.2 ce23.2 ac22.7 bc23.6
Slg19.9 a20.7 f22.3 a22.3 a17.7 a22.6 a21.7 a17.7 a20.6
Stp21.2 b24.6 ad25.0 bc23.3 ab21.7 b25.4 b22.4 ab22.0 b23.2
────────────────────────────────────────
Values with the same letter in a column do not differ significantly at p=0.05. (For Pwj, Bms, Kds, Tgl, Btg, Kbm, Crb and Bjn see Table 1).

Table 4. The number of grains per panicle in each genotype at eight locations.
────────────────────────────────────────
GenotypesPwjBmsKdsTglBtgKbmCrbBjnAvg.
────────────────────────────────────────
G9114 a100 bd100 a104 a 99 bc161 bd105 a82 b108
G10121 ab113 bd122 ac115 ab160 de149 bc120 ac84 b123
G12147 de127 b145 c152 cd201 e233 e149 cd119 cd159
G13141 ce116 bc137 bc146 c143 cd185 d148 cd103 bc140
G19135 be107 bd174 de136 bc80 a187 d132 ad114 cd133
G34130 ad116 bc115 ab140 bc157 cd165 bd135 ad122 cd134
G35140 ce112 bd115 ab128 ac117 bc139 ab126 ac112 cd124
G39149 ef124 b182 e137 bc107 bc243 ef184 e134 de158
G136127 ac111 bd137 bc141 bc111 bd175 cd138 bd125 cd133
Ps141 ce123 b151 cd138 bc112 bc175 cd144 cd147 e141
Mw122 ab 91 cd124 ac130 ac136 bd164 bd126 ac104 bc125
Slg121 ab83 d110 ab101 a 73 a117 a111 ab 56 a96
Stp165 f184,9 a175 de176 d152 de265 f163 de115 cd175
────────────────────────────────────────
Values with the same letter in a column do not differ significantly at p=0.05. (For Pwj, Bms, Kds, Tgl, Btg, Kbm, Crb and Bjn see Table 1).

Table 5. One-thousand-grain weight (g) in each genotype at eight locations.
────────────────────────────────────────
GenotypesPwjBmsKdsTglBtgKbmCrbBjnAvg.
────────────────────────────────────────
G925.4 bc27.1 bd20.9 ab23.2 ab24.4 d25.4 bc22.7 ab23.4 ab24.1
G1029.7 g29.1 a24.2 e29.2 e27.1 g27.7 d25.4 cd29.2 ef27.7
G1227.6 e26.8 bd20.3 a25.9 be26.0 g26.3 bd22.3 a26.5 cd25.2
G1329.9 g27.7 ac23.7 de27.2 ce26.7 g27.6 d24.4 bd29.6 f27.1
G1926.9 d26.9 bd21.2 ab28.6 de19.2 a25.4 bc24.0 ad23.6 ab24.5
G3425.5 bc25.8 ce20.0 a22.7 ab25.9 fg25.1 b23.8 ad23.2 ab24.0
G3525.6 c26.9 bd21.5 ac23.2 ab24.3 d23.3 a23.8 ad21.6 a23.8
G3927.4 de25.2 de21.3 ab24.9 bd24.9 ef25.7 bc23.5 ac24.9 bc24.7
G13624.9 b24.8 e20.6 ab19.7 a22.3 b26.8 cd22.8 ab24.7bc23.3
Ps28.3 f26.9 bd22.2 bd24.0 bc24.8 de26.0 bd24.3 bd25.1 bc25.2
Mw28.2 f27.9 ab23.1 ce25.3 be22.8 bc26.3 bd25.2 cd27.2 de25.8
Slg23.0 a21.1 f19.7 a21.9 ab22.1 b23.5 a 22.9 ab23.3 ab22.2
Stp27.3 de24.7 e20.9 ab23.6 ac23.6 cd25.6 bc25.6 d23.9 b24.4
────────────────────────────────────────
Values with the same letter in a column do not differ significantly at p=0.05. (For Pwj, Bms, Kds, Tgl, Btg, Kbm, Crb and Bjn see Table 1).

Table 6. Grain weight per hill (g) in each genotype at eight locations.
────────────────────────────────────────
GenotypesPwjBmsKdsTglBtgKbmCrbBjnAvg.
────────────────────────────────────────
G924.9 a22.7 ab17.1 a17.2 a30.825.2 a23.9 ab18.2 ac22.5
G1025.3 a21.1 ab34.0 bc27.5 ab42.327.7 ab23.7 ab16.9 ac27.3
G1229.9 bc14.0 bc27.7 ac18 .1 ab37.926.6 ab23.1 a23.6 ac25.1
G1330.8 ce19.2 ab32.4 bc21.7 ab26.725.2 a26.0 ab18.0 ac25.0
G1933.2 df23.4 ab32.4 bc27.4 ab12.936.4 d26.2 ab17.8 ac26.2
G3426.3 ab22.7 ab19.1 a27.3 ab16.327.4 ab25.6 ab18.7 ac22.9
G3527.7 ac25.8 a18.3 a19.2 ab20.825.0 a24.4 ab13.4 a21.8
G3931.6 ce20.9 ab37.9 c23.7 ab30.436.0 d33.2 c20.3 ac29.3
G13635.9 f18.9 ab36.3 c18.9 ab31.933.5 bd26.4 ab26.6 bc28.6
Ps29.8 be26.5 a38.1 c29.0 b22.127.3 ab24.3 ab29.6 c28.3
Mw28.7 ac22.8 ab31.3 bc28.0 ab32.928.2 ac29.5 bc20.9 ac27.8
Slg29.0 ad18.9 c22.8 ab22.1 ab11.727.2 ab25.3 ab14.0 ab20.1
Stp33.5 ef20.8 ab27.7 ac16.9 a30.234.5 cd28.0 ac10.8 a25.3
────────────────────────────────────────
Values with the same letter in a column do not differ significantly at p=0.05. (For Pwj, Bms, Kds, Tgl, Btg, Kbm, Crb and Bjn see Table 1).

Table 7. Yield (g m-2) in each genotype at eight locations
────────────────────────────────────────
GenotypesPwjBmsKdsTglBtgKbmCrbBjnAvg.
────────────────────────────────────────
G9375.4 a324.0 bd113.4 a197.4 a493.2 ef372.0 ab401.8 ac109.0 ac325.4
G10381.4 a344.8 bd190.0 cd440.0 de666.6 g434.8 bd379.0 ab101.2 ac405.2
G12466.6 bc401.8 cd156.6 ac245.4 ab606.6 fg360.0 ab370.2 a141.6 ac372.4
G13492.6 cf293.0 bc230.0 de296.6 ad426.6 de317.6 a416.4 ab108.0 ac353.4
G19524.0 fg400.2 cd186.6 cd463.0 e206.6 ab575.2 f445.6 bc107.0 ac400.2
G34426.0 b 326.6 bd106.6 a401.0 ce260.0 ac467.2 cd436.2 ac112.0 ac346.2
G35431.4 bc366.4 bd126.6 ab253.2 ac333.2 bd315.8 a389.8 ac80.2 a316.6
G39479.4 de281.8 b273.4 e387.0 be486.6 ef486.4 de531.6 d121.6 ac418.0
G136546.6 g308.6 bd243.4 e284.0 ac510.0 ef560.4 f421.8 ac159.6 bc410.6
Ps494.0 ef275.6 b180.0 bd389.8 be353.2 cd418.4 bd389.4 ac177.6 c357.2
Mw442.6 bd407.0 d183.4 cd356.4 be526.6 ef393.6 bc458.0 c125.2 ac395.4
Slg470.6 ce148.8 a233.4 de293.8 ac186.6 a396.6 bc404.2 ac84.2 ab304.8
Stp535.4 fg349.6 bd186.6 cd258.6 ac483.2 e549.0 ef448.6 bc65.0 a401.6
────────────────────────────────────────
Values with the same letter in a column do not differ significantly at p=0.05. (For Pwj, Bms, Kds, Tgl, Btg, Kbm, Crb, and Bjn see Table 1). 
Fig. 1 Regression lines for genotypes above that for the mean yield (×1000 g m-2) 
of all genotypes at different locations.
(L1 - L8 were: Pwj, Bms, Kds, Tgl, Btg, Kbm, Crb and Bjn, respectively).   


Fig. 2. Regression lines for genotypes above that for the mean yield (×1000 g m-2) 
of all genotypes at different locations. (For L1 - L8 see Fig. 1).

Fig. 3. Regression lines for genotypes below that for the mean yield (×1000 g m-2) 
of all genotypes at different locations. (For L1 - L8 see Fig. 1).
以上