Agro-Morphological Diversity of some Improved and Local Rice (Oryza sativa) Varieties in Irrigated Lowland in Guinea Savannah of Côte D'ivoire

 +225 49 02 56 65 / 01 34 92 40 This article is published under the terms of the Creative Commons Attribution License 4.0 Author(s) retain the copyright of this article. Publication rights with Alkhaer Publications. Published at: http://www.ijsciences.com/pub/issue/2019-08/ DOI: 10.18483/ijSci.2154; Online ISSN: 2305-3925; Print ISSN: 2410-4477 Agro-Morphological Diversity of some Improved and Local Rice (Oryza sativa) Varieties in Irrigated Lowland in Guinea Savannah of Côte D'ivoire Kouassi Abou Bakari, Aka Stéphane Yves, Kone Brahima, Sorho Fatogoma


I. Introduction
Rice is a major staple food for peoples in both urban and rural zones worldwide (Ojo et al., 2009) as occurring in Côte d'Ivoire since the year 1980 (ONDR, 2016) thanks to its culinary qualities and relatively accessible prices. However, only half of local demand is covered by domestic rice production because of poor seed quality and the high consumption quantity. To mitigate this deficit, Côte d'Ivoire uses massive imports of milled rice estimated in 2016 at 1,498,102 tons for about 383 billion FCFA (ONDR, 2017) in coast.
Unfortunately, only 5% of world rice production, meaning 31 million tonnes out of 650 million tonnes of annual production is available for international trade. In addition, rice production in West Africa covers only 2/3 of local demand and requires the import of 3 million tonnes of milled rice (Bahan et al., 2012). Thus, the population of Côte d'Ivoire, estimated to 22.67 million persons with an annual growth rate around 2.6% and a national poverty rate of 46.3%, is exposed to a risk of food insecurity (INS, 2016).
To address this threat of food insecurity, Côte d'Ivoire has undertaken various agricultural policies since 1960 with the establishment of institutional organisms such as the Rice Development Company (SODERIZ) and the National Office of Rice Development (ONDR) targeting the increase of national production. The creation of these organisms for Ivorian rice production is based on the exploitation of several assets of the country such as large areas of land suitable for rice cultivation (lowlands and plains); a very favourable climate with abundant rainfall; a high level of know-how of the producers and the existence of improved varieties with relatively high potential production and good organoleptic qualities (ONDR, 2016). However, the average yields are still very low (1 to 2 t / ha) because the available improved varieties are hybrids whose high production potential is due to heterosis that is gradually lost after several production cycles since rice mainly reproduce by self-pollination (Yoshida, 1981). It is therefore necessary to ensure the permanent production of good quality F1 hybrid seeds to maintain high yields across growing seasons.
The present study is based on the principle that any genetic improvement program or conservation of genetic resources of a given plant species first requires the knowledge of the available genetic diversity (Tendro, 2010). The overall objective is to analyse the phenotypic diversity of 20 rice accessions including 10 improved varieties produced by AFRICARICE and 10 local varieties of Côte d'Ivoire in order to determine the most discriminating traits to characterize the most yielding irrigated lowland rice varieties.

II.1. Site of the study
The study was undertaken in the valley of M'be around the city of Bouaké in central Côte d'Ivoire. The locality belongs to the district of Bandama Valley and is situated about 8°06 North Latitude and 6°00 West Longitude. It is a guinea savannah zone characterized by an average temperature of 26.2°C and 1119 mm as annual rainfall. The studied lowland is classified in third order with alluvial deposits in the middle part against colluvial sand in the fringe.

II.3.1 Experimental design
The experimentation was performed in three Fischer blocks separated by 3 m. Within each block, local and improved rice varieties were arranged separately in two sub-blocks 3 m apart. In each sub-block, seedlings of the different rice varieties were arranged separately in micro-plots of 3 m in wide and 5 m in long. The micro-plots were separated by 1 m and arranged randomly.

II.3.2. Transplantation of rice seedlings
After plowing with a tiller and successive flooding and drainage of the experimental plot, picketing was done according to the experimental design. The NPK fertilizer (15% -15% -15%) was applied as basal manure at the rate of 200 kg/ha or 200g / 15m 2 . After 21 days in the nursery, one plantlet of each rice varieties was transplanted per hole spaced by 20 cm between plants and between lines.

II.3.3. Monitoring of Experimental Plots and Data Collection
Maloriz ® , a post-emergence herbicide, was applied at the rate of 200 ml per 15l of water, 21 days after transplantation of rice seedlings. Two weeks after application of the post-emergence herbicide, the plots were irrigated to reach and maintain a maximum water height of 5 cm. The insecticide Decis ® was applied at the rate of 2.5ml in 5l of water per 50m 2 in cases of insect attacks. Chemical weeding was performed with the herbicide Garize ® at the rate of 200ml per 15l of water prior to maximum tillering and heading. This chemical weeding was supplemented by manual weeding to complete weed control. Urea was applied at the rate of 14 g per 15 m 2 at tillering stage and at the rate 19 g per 15 m 2 at heading stage.
Nine qualitative traits (Table I) and 20 quantitative traits (Table II) from the list of rice, Oryza spp, descriptors (IRRI 1980, Bioversity International, Africarice, 2007) were scored on each of the 20 rice varieties. At the maturity stage of the grains, plants of each rice variety were cut at the base of the tillers, then dried and beaten. The grains were then collected and the straw fragments and pebbles were removed. All the data were collected by leaving two border lines on each side of the micro-plots.

II.4. Statistical analyses:
Agro-morphological diversity of the 20 rice varieties related to qualitative traits was assessed by Multiple Correspondence Analysis (MCA). Descriptive statistics (mean values, standard deviations, and coefficients of variation) of each of the quantitative traits were calculated. One-way analysis of variance was performed to test the variety effect on values of quantitative traits. For each quantitative trait, the Newman-Keuls test was used to compare average values of rice varieties to identify the ones with the best agro-morphological characteristics. Agromorphological diversity of rice varieties related to quantitative traits was evaluated by a Hierarchical Ascending Classification (HAC) based on the Unweighted Pair-Group Method with Arithmetic Average. A Discriminant Factorial Analysis (DFA) was also performed using as categorical variables the groups provide by the HAC. All statistical analyses were performed with the software Statistica 7.1 (2005) and SPSS 24 (2017)

III.1.1 Diversity of the rice varieties due to qualitative traits
The projection of rice varieties in the space determined by the factorial axes 1 and 2 of the Multiple Correspondence Analysis distinguishes 3 agro-morphological groups. The Group I is represented on the negative side of the Axis 1. It includes the local varieties Soungrouba and WITA9 plus the new improved varieties AR624H, AR593H, AR034H, AR630H, AR043H, AR601H, AR051H, AR638H and AR629H. These varieties are characterized by straw-yellow grains, medium pubescent grain and easy ginning. The Group II consists of the local varieties Danané and GT11 plus the improved variety AR597H. This group is represented on the positive side of Axis 2 and is characterized by non-pubescent and tawny-red grains, medium apicoid and brown-red caryopsis. The Group III gathers the local varieties Demamba, Marigbé, Kpaté, Palawan, Djoukeme and Kouikloné represented on the negative side of axis 2. This group is characterized by very pubescent grains, falling flag leaves and semi-compact branches of panicles.

III.1.2. Structuration of the diversity of rice varieties based on quantitative traits III.1.2.1. Analysis of variances
Analysis of variance (ANOVA) showed highly significant effect of studied varieties (p <0.0001) with significant differences of quantitative traits (Tables  III and IV). The significance levels of the differences between rice varieties is high for seed thickness (p = 0.0043 <0.01) and flag leaf length (P = 0.0159 <0.05).

III.1.2.2. Analysis of the Diversity by Hierarchical Ascending Classification (HAC)
The dendrogram obtained by the Hierarchical Ascending classification (FIG. 2) shows, at the point of truncation 20, a clear separation of rice varieties into six (6) groups. The Group I consists of 8 new improved varieties AR624H, AR593H, AR601H, AR034H, AR051H, AR629H, AR630H and AR638H. The Group II includes local varieties WITA9 (control) and Djoukeme. The improved varieties AR597H and AR043H are in the group IV. The local varieties Kpaté and GT11 form the group V. The local varieties Marigbè and Danané constitute the group VI.

III.1.2.3. Discriminant Factorial Analysis
The classification matrix of groups of rice varieties provided by the discriminant factorial analysis confirms all the groups determined the hierarchical ascending classification (Table V). The λ Wilk test reveals that 9 of the 20 variables used for this analysis (Table VI) allow a significant discrimination of the groups (p <0.05). These are weight of 1000 grains, grain shape, grain yield, grain length, leaf width, number of tillers per plant, number of panicles per plant, number of grains per panicle and plant height. According to the centered-reduced coefficients of the canonical discriminant function, components 1 and 2 account for 99.04% of the total variance (Table VII). The first canonical component explains 95.3% of the total variance and makes it possible to distinguish groups II, III, IV, V and VI. The Group I is located on the second canonical component (Table VII).
The projection of rice varieties in the space determined by components 1 and 2 of the discriminant canonical analysis ( Figure 3) shows that the group I is negatively correlated to the component 2 and groups II, III, IV and VI are represented on the negative side of the axis 2. The Group V is positively correlated with the component 2. The Group I consists of eight (8) improved varieties (AR624H, AR593H, AR034H, AR051H, AR630H, AR601H, AR638H and AR629H) already grouped by the HAC. They are high-yielding varieties with high tillering and long and round grains. Among these varieties, 5 (AR624H, AR593H, AR034H, AR051H and AR629H) have higher average yields than the variety WITA9 (control). These varieties are also characterized by medium plant heights and very few empty spikelets per panicle. The Group II contains the same varieties as those of the group II determined by the HAC. This group is characterized by high rates of fertile spikelets per panicle and high weights of 1000 grains, medium grain lengths, low grain yields and narrow leaves. These are semi-round grain varieties. The variety WITA9 belongs to this group but has a grain yield close to those of the varieties in the group I. The Group III consists of local varieties, Palawan, Demamba, Kouiklonlé and Soungrouba which have tall plants. These varieties have low tillering and small numbers of panicles per plant so that the yield is also low except to the variety Soungrouba which has an average yield of 4.6 t / ha. On the other hand, the Group III is characterized by high weights of 1000 grains. Improved rice varieties AR043H and AR597H form the group IV. They have the same characteristics as the varieties in the Group I but they have highest grain yields. The Group V consists of local rice varieties GT11 and Kpaté. They are characterized by large stems, semi-round and medium-sized seeds and excellent tillering. These local varieties, on the other hand, have very small numbers of panicles per plant and low grain yields. The Group VI consists of the local varieties Marigbé and Danané as provided by the HAC. These varieties have medium plant heights, long and semi-round grains. This group is also characterized by low grain yields, very low tillering and very low numbers of panicles per plant.

III.2. Discussion
Agro-morphological characterization is a very important preliminary study in breeding programs, plant conservation and even agricultural policy planning. Thus, many analyses of agromorphological diversity have been performed for different crops such as rice (Sanni et  The results obtained in the current study are consistent with these earlier observations. Indeed, the analysis of the agro-morphological diversity of 20 rice varieties in agro-ecological conditions of Bouaké using multiple correspondence analyse (MCA) distinguished three (3) groups. The group I contains 9 out of the 10 new improved varieties (AR624H, AR593H, AR034H, AR051H, AR630H, AR601H, AR638H, AR629H and AR043H, plus the local varieties WITA9 and Soungrouba). This group is likely specific to sativa improved rice variety knowing that WITA9 was released by AfricaRice in 2000 and Soungourou is also an improved variety adopted as local landrace. The group I is characterized by yellow-straw coloured grains, moderately pubescent seeds and easy ginning. The group II, which includes the local varieties Danané and GT11 plus the improved variety AR597H, is characterized by non-pubescent seeds, medium apicles, brown-red caryopsis and tawny-red grains. Finally, the group III consisting of the 6 local rice varieties Djoukeme, Kpaté, Marigbé, Palawan, Demamba, Kouikloné is characterised by highly pubescent grains, drooping flag leaves and semicompact branches of the panicles. Different authors used two postulates to explain the relatively weak usefulness of qualitative traits for the agromorphological discrimination of crop plant varieties. According to N'Da et al (2014), the poor contribution of qualitative traits to the structuration of agromorphological diversity could be explained by the fact that local varieties are selected by peasants whose perception of the different modalities of qualitative characteristics is very subjective. For Sarla and Swamy (2005); Bezançon and Diallo (2006) and Akakpo, (2011), the geographical and / or specific origin of varieties are factors determining the structuration of agro-morphological diversity based on qualitative characteristics. Our results support these assertions. Indeed, all these varieties belong to the species Oryza sativa which has an asian origin. In addition, local and improved varieties tend to form two distinct groups.
One of the specific objectives of this study was to identify irrigated rice varieties with the best yielding potential. The highest grain yield was 9.64 t / ha for the improved variety AR043H. This result, although much higher than average yields of local rice varieties and that of the popular WITA9, is far below the 15t / ha which is the potential yield of improved varieties according to AfricaRice. Nevertheless, the yield was increased by 90% comparing to the national average yield of 1 t/ha. Such level of yield increasing underlines the possibility to resorb the gap between need and national rice production. Therefore, our study can recommend the use of hybrid rice seeds for balancing rice production trends in Côte d'Ivoire. Furthermore, there is still an opportunity to improve the observed relative low yield of the improved varieties by assessing genotype x environment interaction effects. Indeed, the mineralogical composition of the soil of the experimental plot has not been preliminary analysed. So the nitrogen fertilizer was not applied according to the characteristics of the soil in a way of rationalization of fertilization in order to reach a targeted yield. Taking these parameters into account in oncoming crop cycles should lead to better yields. On the other hand, the grains have been harvested at an advanced maturity stage which certainly caused losses. In fact, the improved rice varieties provided by AfricaRice are characterised by easy ginning, whereas the later the harvest is performed the greater the losses of grain yield (Lacharme, 2001).
The factorial analysis of quantitative traits revealed significant differences between minimum and maximum values for each rice variety. This shows a strong phenotypic heterogeneity between these 20 varieties. This morphological diversity has identified six groups of rice varieties that differ in plant height, number of tillers per plant, number of panicle per plant, width of plant leaf, grain shape, grain yield and weight of 1000 grains.
Group I contains the 8 improved varieties AR624H, AR593H, AR034H, AR051H, AR630H, AR601H, AR638H and AR629H characterized by medium height, high tillering and very high grain yield with long round grains. Group IV closes the other two improved varieties AR043H and AR597H which are distinguished from those of Group I by the highest numbers of tillers and grain yields. These varieties can thus be of interest in the sense that the high yield is a key feature for rice farmers, but also because the high tillering plays a very important role in the fight against weeds and in general against grassing. (2012) classified millet varieties into two groups: a medium-sized and early millet group and a large and late millet group. In the current study, the sowingflowering duration was not considered in the analysis of variance as well as in the Discriminant Factor Analysis. Indeed the data collected for this variable presented outliers that suggest a systematic scoring error. Corrective action for future ratings of this variable is expected to yield more reliable data. The height of the plants, however, discriminates between Group III and V rice varieties with respect to those belonging to groups I, II, IV and VI. Groups III and V, for example, consist of local rice varieties with high plant height collected in Daloa and Danané, two cities located respectively in west-central and northwestern Côte d'Ivoire. These two regions are characterized by agro-ecological zones where rainfall is very high. Groups I and IV, on the other hand, consist of improved varieties of AfricaRice, one parent of which has been collected in the Sahel, and groups II and VI, on the other hand, consist of varieties collected in Bouaké and Ouangolodougou, respectively in central and northern Côte d'Ivoire. These agro-ecological zones are characterized by a relatively low rainfall which determines a medium height of the plants. Like the observations made on different cereals such as millet (Akanvou et al., 2012) and maize (N'da et al., 2014), our results show a structuration of the morphological diversity of rice varieties according to agro-ecological zones of origin and therefore according to the availability of water.

IV. Conclusion
The agro-morphological characterization of the 20 varieties of rice made it possible to show the importance of the qualitative and quantitative descriptors in the analysis of the diversity expressed. Qualitative variables showed relatively low variability by distinguishing overall improved varieties from local varieties. On the other hand, quantitative descriptors such as plant height, number of tillers per plant, number of panicles per plant, length of single leaf, grain shape, grain yield and weight of 1000 grains are the most relevant and have distinguished 6 groups of rice varieties.

Acknowlegments
We thank Madame Ly-Ramata, former Minister of High education and research for supporting the study and for her personal interaction with scientists in a way to tackle food insecurity in Cote d'Ivoire. All the members of GIPS (Groupe d'Initiatiave pour la Production de Sémence) are also concerned by this paper because of their supports. We also thank Dr El Namaky A. for providing the seeds of new improved rive varieties from AfricaRice Senegal 3098-3106.