Two species of cultivated rice are known: Oryza sativa L. and Oryza
glaberrima Steud. Oryza sativa or Asian rice was domesticated more
than 10,000 years ago from O. rufipogon (Kovach
et al., 2007) and has been adapted to a wide range of conditions
including different water regimes and a wide range of soils (Takahashi,
1984). Oryza sativa includes two ecologically different subspecies.
Oryza sativa japonica is cultivated in the upland ecology, while the
other ecologies are occupied by O. sativa indica. Oryza glaberrima
(or African Rice) was domesticated some 2,000-3,000 years ago (Porteres,
1970) from the wild ancestor O. barthii in the floodplains of the
bend of the Niger River, which is the primary center of origin (Sarla
and Swamy, 2005). Oryza glaberrima is increasingly replaced in Africa
rice growing environment by O. sativa which has been introduced by the
Portuguese (Linares, 2002). Many African farmers are
shifting from O. glaberrima to O. sativa because of its low yielding
caused by grain shattering and lodging susceptibility (Linares,
2002). Oryza glaberrima accessions are considered as a rich reservoir
of genes and alleles suitable for the genetic improvement of rice targeting
the African continents rice growing constraints (Jones
et al., 1997). Previous research work in interspecific crossing has
helped to combine weed competitiveness of O. glaberrima and high yield
potential of O. sativa japonica (Jones et al.,
1997). Moreover, O. glaberrima is considered to have early growth
vigor, which is the most important character in weed competitiveness (Dingkuhn
et al., 1998, 1999; Saito
et al., 2010). Oryza glaberrima was ecologically divided into
two groups (Oka, 1974): upland and deep water types.
However, other studies pinpointed that there are no sub-groups in O. glaberrima
(Morishima et al., 1962). In support of this
opinion, some upland O. glaberrima accessions have been shown to have
resistance to lowland constraints (Futakuchi and Sie, 2009),
while so-called upland types and deep-water types showed the same response to
submergence (Futakuchi and Jones, 2005). On the other
hand, Semon et al. (2005) demonstrated that population
can be divided into five different groups using genetic diversity analysis mean.
Knowledge of ecological adaptation to variable water regimes is limited, especially
for O. glaberrima. This study further considers the morphological and
phenological response of O. glaberrima across the hydrological toposequence
from upland to irrigated lowland.
MATERIALS AND METHODS
Site characteristics: The study was conducted at the experimental farm of the International Institute of Tropical Agriculture (IITA) Ibadan station, Nigeria. The station is located at latitude 7°3N and longitude 3°45E in the forest-savanna agro-ecology. Rainfall during the wet seasons (from May to December) was 872.25 mm in 2009 and 1680.25 mm in 2010 (Fig. 1).
Plant materials: Two hundred and seventeen accessions of O. glaberrima were selected with the objective of covering the geographic origins of African rice. Table 1 shows the frequencies distribution of the different accessions of O. glaberrima involved in this experiment across countries. Five O. sativa cultivars were included in the experiment as checks: IDSA 6, Morobérékan and WAB56-104 adapted to upland conditions and IR 64 and WITA 4 adapted to lowland conditions. The seeds were obtained from the Africa Rice Center germplasm collection.
Experimental design and crop management: Seeds of all accessions were sown at three levels in the toposequence (upland, non-flooded hydromorphic and irrigated lowland) at the IITA station experimental field in Ibadan, Nigeria. In the upland and hydromorphic plots, dry seeds were dibbled, whereas pre-germinated seeds were sown in a nearby seedbed nursery and transplanted into the main plots in lowland at 20 Days After Sowing (DAS). An Alpha Lattice design with three replications was used at each level. Before sowing or transplanting, the field was tilled manually using hoes. Plots were fertilized with 200 kg ha-1 of NPK as a basal (15-15-15) dressing at 25 DAS and with 50 kg ha-1 of urea at 45 DAS. Weeding was done by hand to avoid possible herbicide effects on crop phenology. Plots were kept weed-free by regular hand weeding.
In the lowland field, groundwater level was maintained at 5 to 10 cm depth.
In upland and hydromorphic conditions, water level below the soil surface was
determined using 6 cm diameter Poly Vinyl Chloride (PVC) tubes.
|| Rainfall in 2009 and 2010 (source IITA Ibadan)
|| Distribution of plant material across geographic origins
Piezometers were inserted to 80 cm soil depth at three randomly selected positions
in each replicate. Each 100 cm piezometer was made with 80 cm of perforations
from the bottom up; 20 cm was above ground and covered by plastic sheet. The
experiment was performed during the 2009 and 2010 wet seasons but differed in
the sowing date, which was May 15 in 2009 and July 2 in 2010.
Measurements of agro-morphological characters: Water level in each PVC tube was determined daily using a long ruler. Data were collected from four plants randomly identified and labeled, selected from the middle plants in each experimental plot. Method of data measurement is shown in Table 2. Vegetative vigor was scored in each plot at 30 DAS. Plant height and tiller number were measured at maturity. Phenological response including days to 50% heading and maturity (crop duration) were visually assessed on a daily basis starting from booting stage on all plots throughout the experiment. Soil samples at different depths (0-20; 20-40 cm) were collected separately at each ecosystem before crop establishment to determine major chemical components and soil physical characteristics.
|| List of characters and their respective classes used for
the evaluation of diversity
|| Fluctuation of water levels in 2009 and 2010
Statistical data analysis: The phenotypic frequencies of the five characters
were analyzed by the Shannon-weaver index (H) using the formula of Jain
et al. (1975):
where, k is the number of phenotypic classes for the character and Pi
is the proportion of the total number of entries in that class. H was
estimated for each character by ecology. Each value of H was standardized
by dividing it by its maximum value (log2k), in order to keep the
values in the range of 0-1 (Abdi et al., 2002).
Analysis of variance was performed for non-transformed Shannon-weaver index.
The genetic correlation rgxy between two ecologies (x and y) was
calculated according to Burdon (1977):
where, Covgxy is the covariance of both ecologies x and y, σgx
is the standard deviation for ecology x and σgy is the standard
deviation for ecology.
Groundwater fluctuations: Piezometric characterization of the experimental sites (upland and hydromorphic) was important to knowing the water availability during plant growth, with the soil surface being considered as zero level. The rainfall-dependent fluctuations in groundwater along the toposequence were between 0 and 20 cm in the hydromorphic zone in 2009 and 2010. In the upland zone, groundwater fluctuated between 37 and 78 cm in 2009 and between 14 and 68 cm in 2010. Rainfall was greater in 2010 (1680.25 mm) than in 2009 (872.25 mm). In upland 2009, drought occurred when groundwater level was closed to -80 cm. Figure 2 shows the fluctuations in groundwater in upland and hydromorphic zones in both seasons. Globally, groundwater level was obviously higher in hydromorphic condition than in upland ecology irrespective of seasons. However, this water level was similar in hydromorphic condition across seasons while it was higher in upland in 2010 than it was in 2009.
Soil characteristics: Soil samples at different depths were analyzed
to determine the chemical components and the soil physical characteristics of
each ecosystem (Table 3). Using Jamagne
(1967) triangle of soil texture, the soils from upland and lowland were
found to be sandy and similar.
|| Physico-chemical characteristics of soil samples collected
from different ecologies
|ECEC: Effective Cation exchange capacity; C: Carbon; N: Nitrogen;
pH: Hydrogen potential, Meh P: Mehlich phosphorus
|| Phenotypic frequencies (%) of five variables across ecologies
in 2009 and 2010
Results showed that the Hydrogen potential (pH) of the soil was slightly acidic
and ranged between 6.0 and 6.4. The ecosystems displayed a similar pattern for
Carbon nitrogen ratio (C/N), which was between 8.1 and 9.1 with a low Effective
Cation Exchange Capacity (ECEC) and a high concentration in phosphorus.
Morphological and phenological variability across hydrological conditions: Table 4 shows the phenotypic frequencies of five variables-vegetative vigor, plant height and number of tillers at maturity, days to 50% heading and maturity-for the 2009 and 2010 seasons. Frequencies were calculated as a percentage of the number of accessions in each class. The individuals differed from one class to another.
Monomorphic phenotypic frequency distribution was exhibited by vegetative vigor in the upland ecosystem whereas intermediate frequency was observed in hydromorphic and lowland ecosystems in 2009. In upland, 49.3% (classes 1 and 3 together) of O. glaberrima accessions used were vigorous. Nearly 90% of entries in hydromorphic ecosystem were either extra vigorous or vigorous equaling to 61.37% for class 1 and 28.4% for class 2. The lowland condition displayed 71.6% of O. glaberrima for these combined classes. Normal vigor was recorded in frequency of 45% in upland, 9.8% in the hydromorphic zone and 26.4% in lowland. The weak class bore very few individual across ecologies equaling to 5.6, 0.5 and 2%, respectively in upland, hydromorphic and lowland.
In 2010, vegetative vigor was of intermediate type in upland and lowland ecosystems. Monomorphic frequency was observed in the hydromorphic ecosystem. Forty two percent of accessions were classified as extra vigorous and 26% as vigorous in upland conditions. In hydromorphic extra vigorous and vigorous class bore 47.2 and 51% of O. glaberrima, respectively. In lowland ecosystem, extra vigorous and vigorous class indicated 81.7 and 18.3% of tested entries, respectively. The normal vigor recorded 24% in upland and only 1.8% in hydromorphic condition. The weak covered 8% of the tested O. glaberrima in the upland.
Plant height at maturity showed monomorphic phenotypic frequency in upland conditions but intermediate frequency was found in hydromorphic and lowland conditions in 2009. Most O. glaberrima accessions used were tall, except in upland conditions: 82.2% in lowland, 77.5% in hydromorphic and 32.5% upland. Under upland conditions, 46.6% of O. glaberrima tested were of intermediate height.
In 2010, intermediate phenotypic frequency was observed in all ecosystems. The majority of plants in the different ecosystems were found to be tall: 84.8% in lowland, 87.5% in hydromorphic and 54.4% in upland conditions.
Number of tillers at maturity was monomorphic in both years in all ecosystems. In 2009 the frequency of O. glaberrima having an intermediate number of tillers was: 54.5% in upland, 70.9% in hydromorphic and 45.3% in lowland. High tillering ability was observed in frequency of 37.2% of accessions in lowland, 14.4% in hydromorphic and 31.8% in upland.
In 2010, the intermediate class accounted for 63.2% in upland, 79.1% in hydromorphic and 82.1% in lowland. High tillering ability was recorded in low frequency in upland (10.3%), hydromorphic (10.7%) and lowland (8.3%) but the upland condition further displayed 24.6% of tested entries as low tillering.
In 2009, polymorphic phenotypic frequency was observed in the hydromorphic condition for days to 50% heading, whereas frequency was shown to be intermediate value in the upland and lowland ecosystems. In the upland ecosystem, the late and very late headings individuals combined at 84.5% while these of very early, early and intermediate headings were low (15.5%). In hydromorphic conditions, very early and early headings individuals were recorded for 39.2% of accessions, followed by late and very late headings (32.2%) and intermediate heading (28.6%) of O. glaberrima. Late and very late headings individuals in lowland conditions accounted for 90.1% of all accessions.
Monomorphic phenotypic frequency was given by days to 50% heading in all ecosystems in 2010. In the upland condition, the very early and early headings O. glaberrima represented were 33.7%, intermediate heading O. glaberrima accounted for 49.3% and the late and very late heading accounted for 17%. The intermediate heading entries represented 45.6% followed by the very early and early headings (43.8%) and the late heading (10.6%) in hydromorphic condition. Results obtained in lowland condition were 43.1, 41.1 and 11%, respectively for early heading, intermediate heading and late heading.
Intermediate frequency was shown for maturity in upland and lowland conditions and frequency was monomorphic in the hydromorphic condition in 2009. Late and very late maturing accessions represented together 89.9 and 91.4% in upland and lowland, respectively. In the hydromorphic condition, accessions of both class (very early and early) accounted for 38.1% followed by late and very late maturing (31%) and intermediate maturing (30.9%).
In 2010, frequency was monomorphic in all ecologies. About fifty five per cent (55.2%) of O. glaberrima were intermediate maturing in upland followed by early maturing (32.2%) and late and very late maturing (10.6%). In the hydromorphic ecosystem, intermediate maturity was recorded for 48.3% of tested entries followed by early maturing (38.7%) and late maturing (8.3%). In lowland 56.8% of O. glaberrima tested were found as intermediate maturing, 32.4% as early maturing, 5.3% as very early maturing and 5% as late maturing.
Table 5 and 6 showed the estimated Shannon-weaver
diversity index (H) values of five variables for each ecological condition
in 2009 and 2010. In 2009, the analysis of variance indicated significant difference
(p<0.05) across ecologies for all variables while only vegetative vigor and
plant height at maturity showed significance (p<0.01) in 2010. In 2009, vegetative
vigor showed H value of 0.73 in upland, 0.55 in hydromorphic and 0.68
in lowland conditions; plant height 0.94 in upland, 0.61 in hydromorphic and
0.52 in lowland.
|| Estimates of Shannon-weaver diversity index in 2009
|*Significance at p<5% level, **Significance at p<0.01
|| Estimates of Shannon-weaver diversity index in 2010
|**Significance at p<0.01 level, ***Significance at p<0.001
level, ns: Non-significant
Number of tillers at maturity has H value of 0.64 in upland, 0.55 in
hydromorphic and 0.67 in lowland; while days to 50% heading showed 0.69 in upland,
0.94 in hydromorphic and 0.58 in lowland. For crop duration (maturity), H
indicated 0.60 in upland, 0.94 in hydromorphic and 0.52 in lowland conditions.
In the repeated experiment in 2010, H values were significantly different
only for vegetative vigor and plant height at maturity. Vegetative vigor indicated
0.67 in upland, 0.43 in hydromorphic and 0.29 in lowland. For plant height,
Shannon-weaver index were 0.84, 0.40 and 0.43 in upland, hydromorphic and lowland,
respectively. The other variables (number of tillers at maturity, days to 50%
heading and crop duration) did not show significant difference from upland to
Genetic correlations rgxy between ecologies were used to estimate significant genotypexenvironment interaction (Table 7). Low values of genetic correlations were observed for plant height and tiller number at maturity in both years. When combining ecosystems, upland/hydromorphic gave 0.177, upland/lowland showed 0.227 and hydromorphic/lowland indicated 0.214 for plant height at maturity in 2009. Number of tillers at maturity revealed 0.288, 0.20 and 0.244 for the same combinations. Days to 50% heading showed 0.55 for the upland/hydromorphic combination, 0.65 for the upland/lowland combination and 0.46 for hydromorphic/lowland. Crop duration indicated 0.49, 0.61 and 0.38 for the three combinations. Positives and significant correlations were obtained for plant height, tiller number, days to 50% heading and crop duration in both seasons. The second season showed globally higher genetic correlations for all variables but little difference was observed.
|| Genetic correlation coefficients for variables between ecosystem
|***Significance at p<0.001 level
Rice is able to grow at pH levels between 4 and 8. From upland to lowland in
this study, pH values were between 6.0 and 6.4, which cannot hamper rice production.
Soil physico-chemical characteristics were similar from upland to lowland. Variability
of morphological and phenological response from upland to lowland was caused
by the hydrological condition. Variability between ecologies was confirmed by
low values of genetic correlation coefficients. Each ecosystem is considered
as an environment and the genetic correlation between all pairs of ecosystem
combinations can be used to access genotype by environment interaction (Hammami
et al., 2009). In the absence of genotype by environment interaction,
the expected genetic correlation across environments is one (Hammami
et al., 2009). Previous studies by Cooper and
DeLacy (1994) reported that when the genetic correlation among environments
is less than one, genotype by environment is significant. In this study genetic
correlation coefficients were less than one for all variables and combinations.
Phenotypic frequency of O. glaberrima was analyzed along the toposequence.
This study observed that most of the accessions showed best vigor from upland
to lowland; vegetative vigor scores ranged between 1 (extra vigorous) and 5
(normal vigor); with other classifications being very low. Previous studies
indicated that O. glaberrima cultivars achieve superior vegetative growth
vigor (Asch et al., 1999; Dingkuhn
et al., 1998, 1999; Johnson
et al., 1998). This study also pinpointed that most of the O.
glaberrima accessions were tall in hydromorphic and lowland conditions,
which can be explained by the time spent by plants in water. Long time spending
of plants in water increased plant height (Sie, 1991).
Number of tillers at maturity indicated monomorphic frequency in 2009 and in
2010 from upland to lowland. This result confirms Sanni
et al. (2008) studies that revealed monomorphic distribution of tillers
number at maturity using plant material from different ecological zones. Most
accessions tested were late or very late maturing in 2009 in upland (89.9%)
and lowland (91.4%). Hydromorphic showed early maturity (38.1%) and intermediate
maturity (30.9%). In 2009 long crop durations are probably due, in lowland to
transplanting shock received by plants and to drought in upland (Dingkuhn
and Asch, 1999; Dingkuhn et al., 1990).
The variation observed among O. glaberrima tested was probably related
to ecosystem. Manzelli et al. (2005) reported
that local environment conditions influence genotypic constitution. They suggested
a close relationship between agro-ecological conditions and morphological variation
of germplasm. Diversity analysis of germplasm collections of several crop species
and their wild relatives revealed considerable variability (Yang
et al., 1991). High variability was obtained by Sanni
et al. (2008) using O. sativa landraces from different ecological
zones. The variability observed in O. glaberrima accessions along the
toposequence was due to the ecological heterogeneity as reported by Sanni
et al. (2008). Drought in August 2009 delayed flowering and crop
duration period in the upland ecosystem with high percentage of very late and
late maturing accessions. In 2010, most accessions were early and intermediate
maturing and the percentage of late and very late maturing accessions were very
low. This result can be explained by two factors: rainfall and late planting
in 2010. Rainfall was 872.25 mm in 2009 and 1680.25 mm in 2010. No drought incidence
was observed in 2010. Crops planted in May and July flowered at almost the same
time. The high percentage of late maturing accessions was due to the early planting,
which increased crop duration in 2009 (Dingkuhn and Asch,
1999). Crop duration determines the ability of the crop to avoid abiotic
stress and related to the genotype and the environment Dingkuhn
and Asch (1999). The maximal crop duration was observed for the May planting.
This study shows morphological and phenological variability of O. glaberrima
along the toposequence. High variability is observed from upland to lowland
ecologies and confirmed by variable phenotypic frequencies, high Shannon-Weaver
index and low genetic correlation coefficients. Rainfall contributes to variability
between ecosystems. Transplanting in lowland and drought in upland contribute
to delayed crop phenology. Crop duration depends on the date of sowing; plants
sown in May and July flowered at the same time. These findings will support
efforts to breed rice varieties with short duration or phenological plasticity
in different rice growing environments.
This study was financially supported by the Stress-tolerant Rice for Africa and South Asia Project funded by the Bill and Melinda Gates Foundation and the Ministry of Foreign Affair, Japan. The authors would like to acknowledge this support with deep thanks.