Cowpea is the primary source of protein to vast majority of the populations
in sub-Saharan Africa. It is an important component of the cropping system in
most of the drought prone areas where it serves as a vital source of income
to farmers and seed traders (Fatokun et al., 2000).
FAO data of 2010 showed that it is grown on about 10.5 million ha, with an annual
grain production of approximately 5.5 million tonnes worldwide an estimated
5300 hg ha-1 (http://faostat.fao.org).
The grain yield of cowpea is usually low due to a complex of biotic and abiotic
stresses (Van Ek et al., 1997; Ajeigbe
et al., 2008). These result in extremely low productivity with a
mean yield ranging between 100 to 400 kg ha-1 in Africa as compared
to above 800 kg ha-1 in other areas of the world (Singh,
2002). In sub-Saharan Africa, cowpea is mainly grown under rain fed system
with a minimum annual rainfall of about 600 mm (Valenzuela
and Smith, 2002). Thus drought is a key abiotic constraint of cowpea yield
in this production area (Singh et al., 1997). The
drought situation is compounded by the erratic pattern of rainfall in the cowpea
belt. Cowpea is mainly exposed to drought at the onset and the end of the rainy
season (Singh and Matsui, 2002). Therefore, it is necessary
to identify drought tolerant genotypes that can withstand this condition. Farmers
in Nigeria, especially in the moist savanna grow cowpea towards the end of raining
season in form of relay with cereals like maize and sorghum (Singh
and Tarawali, 1997). This system therefore exposes the cowpea even more
Drought is a condition caused by unavailability of rainfall (Acquaah,
2007). It is said to occur whenever the distribution, amount or duration
of precipitation deviates from the normal. In some instances, the amount may
be normal but drought may occur when long dry spells occur in the season at
critical stages of crop growth and development. According to Acquaah
2007, The effects of drought depend on the stage of plant growth and
development. The most sensitive stages as per yield are pod setting and
pod filling (Turk et al., 1980).
Crops grown under rainfed agriculture in the dry areas of tropical Africa are
subject to dry hot conditions which are synonymous to drought (Hall,
2004). There have been several reports of drought occurrence in the drier
part of these areas (Hall, 2004). The most recent history
of occurrence of drought in Nigeria however was that of the year 1972 to 1974.
The consequences of drought could be dare and painful. According to Derrick
(1977), the great drought of 1973 caused devastating economic loss for millions
of people especially those whose livelihood depended on crop animal husbandry.
These negative effects of drought can be avoided through planting of drought
resistant crop varieties, short season crops as well as desert ephemerals (Ayoade,
Singh and Matsui (2002) listed the methods used to estimate
drought tolerances which they noted are expensive and time consuming. These
involved measurement of water potential, relative turgidity, diffusion pressure
deficit, chlorophyll stability index and carbon isotope discrimination. They
reported the identification of several drought tolerant lines using simple,
relatively cheap and reliable screening methods. Some of these have been incorporated
into improved lines. However, there is the need to identify lines based on their
performance in both stressed and non stressed environment so as to enhance productivity
across the gradient of moisture stress. There is also need identify simple methods
for the classification of these line based on their performance in these environments.
Drought also tend to compound other form of stresses in cowpea, for instance,
Striga (causes considerable yield reduction in cowpea) infection was reported
to be more devastating in areas with sandy soils, low fertility and low rainfall
Therefore, this study was conducted to assess the performance of genotypes
under the stresses of drought and Striga with a view to identify lines
that are tolerant as well as to identify selection criteria for easy identification
of drought tolerance under moisture stressed and optimum field conditions.
MATERIALS AND METHODS
Plant material, location and experimental design: Twenty two Cowpea
genotypes differing in seed and plant characteristics were used in the studies
(Table 1). The genotypes were obtained from the cowpea breeding
programme of the Institute for Agricultural Research (IAR), Ahmadu Bello University
Zaria, Nigeria and from the International Institute of Tropical Agriculture
(IITA), Ibadan, Nigeria. These genotypes were evaluated under 2 different dates
of planting, viz., 7th August (optimum moisture condition) and 7th September
(moisture stressed condition), for two years (2008 and 2009) at the IAR station,
Minjibir (12°10 52N 08°39 22E), Kano State,
Nigeria. The trials were laid in a randomized complete block design with four
|| Origin and description of the genotypes used
Each plot consisted of four ridges 4 m long spaced 75 cm between rows and 20
cm within rows with two stands per hill. Recommended management practices for
cowpea were observed. Data were taken on days taken for 50% of plants in plots
to flower, days to pod maturity as days taken for 90% of the plant in the net
plot to reach 90% physiological maturity, seed per plant and grain yield, i.e.,
the total grain harvested from the net plot which is the two middle rows. Striga
score was recorded as the number of Striga emerged (two middle rows).
The record for rainfall for the period between planting and maturity for the
entire period of this study was obtained from the Metrological unit, Agricultural
Research Station of the IAR, Minjibir Kano State, Nigeria.
Selection indices: Arithmetic Means (AM) and Geometric Means (GM) of
seed yield in stress and non stress environments, response of seed yield to
drought across stress and non stress environments (RD) and percent reduction
in seed yield across stress and non stress environments (PR) were calculated
using the method of White and Singh (1991). Drought response
index was calculated using the regression model of Bidinger
et al. (1987a, b) and Drought Susceptibility
Index (DSI) was calculated as described by Fischer and Maurer
Pearson correlation was employed to determine relationship of genotype yields
of date 1 and 2 across 2008 and 2009. The analysis was also run to determine
the level of association between the genotypes yields and the different selection
indices across the two years.
The relationship between yield across dates of planting and rainfall over the
two years was determined. For the purpose of the analysis, the corresponding
rainfall for any given genotype in a particular planting date was estimated
as the total rainfall received from planting to maturity. For example, if a
given genotype matured in 60 days from planting, total rainfall from planting
to the 60th day is recorded as the rainfall associated with the yield of that
Data analysis: The data was subjected to statistical analyses to compare
performance of the genotypes in the two dates over two years using the general
linear model procedure of SAS 9.0 (SAS, 2002). Least Significant
Difference (LSD) was computed (p≤0.05) to compare performance among genotypes.
Result of the field evaluation of the 22 genotypes is presented in Table
2. There was a highly significant difference (p≤0.01) between dates for
days to 50% flowering, maturity, seed per plant, seed yield, Striga score
and total rainfall. However, there was no significant difference (p>0.05)
between years for days to 50% flowering, maturity, seed per plant, yield Striga
score. Rainfall however, differed highly significantly (p≤0.01) between the
two years. There was more rainfall in 2009 with a mean rainfall of 206 mm as
compared to 169 mm in 2008. Similarly, the genotypes differed highly significantly
(p≤0.01) in terms of days to 50% flowering, maturity, seed per plant, yield
and Striga score. The amount of rainfall received by each genotype did
not differ. The date by genotype interaction was highly significant (p≤0.01)
for days to 50% flowering, maturity, seed per plant, yield and Striga
score. Date by genotype interaction was not significant for rainfall.
Means performance of the genotypes for days to 50% flowering, maturity, seed
per plant, yield, Striga score and rainfall for the drought stressed
and optimum conditions are presented in Table 3. The genotypes
generally flower earlier under the optimum condition. The mean for days to 50%
flowering was 41.05 days and 43.82 days for optimum and stressed conditions,
respectively. IT98K-311-8-2 had the earliest flowering with 34.67 days while
Biu local had the latest with 47.67 days for the optimum condition. However,
the earliest flowering days of 40.17 days was recorded for IT97K-499-35 for
the stressed condition. While 48.17 days was the longest recorded in IT97K-1069-6
In terms of days taken to maturity, the genotypes mature earlier in the stressed
condition with a mean maturity period of 69.48 days. The mean maturity period
for the optimum condition was 74.7 days. Biu local had the latest maturity days
of 82.67 days for optimum condition whereas the earliest was 70.67 days in IT89KD-288.
The earliest period for the stressed condition was 65 days in IT99K-1122 and
the latest was 72.67 days in IT98K-128-3 (Table 3).
There were also greater numbers of seeds per plant in the optimum condition.
The mean number of seeds per plant was 18.45 for the optimum condition. IT99K-216-24-2
recorded the highest number 45.47 seeds per plant while the lowest number of
7 seeds per plant was recorded in Kanannado. The mean for the stressed condition
was 3.5 seeds per plant with IT98K-412-13 recording the highest number of 8.75
seeds per plant and the lowest were 1.07 seeds per plant in IT00K-1263 (Table
||Mean squares for combine analysis of variance of 22 genotypes
of cowpea grown in two dates (early and late sowing) for two years at IAR
station Minjibir, Kano State Nigeria
|**Significant p≤0.01, *Significant p≤0.05
||Mean performance of some putative drought tolerant cowpea
varieties evaluated in field under optimum moisture (date 1) and moisture
stressed condition (date 2) over two years in the Sudan Savanna of the North-West
Similarly, the genotype yield was by far greater in the optimum condition .The
mean grain yield was 940.04 kg ha-1 for the optimum condition as
compared to 198.12 kg ha-1 for the stressed condition. IT98K-412-13
had the greatest yield for both optimum and stressed conditions with yields
of 1792.47 and 526.83 kg ha-1 for optimum and stressed conditions
respectively. However, the least yield for the optimum condition was 395.5 kg
ha-1 recorded in Biu local and the least yield for the stressed condition
was 71.53 kg ha-1 recorded in IT97K-1069-6 (Table 3).
On the other hand, the Striga infestation was greater in the optimum
moisture than in the stressed condition. The mean Striga score was 6.35
and 5.53 for the optimum and stressed conditions respectively. The genotypes,
IAR-07-1050, IT97K-499-35 and IT99K-1122 had no Striga emergence under
the optimum moisture while Sa Babba Sata had the highest score of 14.33 for
the optimum condition. In the moisture stressed condition, IAR-07-1050, IT97K-499-35
and IT99K-1122 recorded 0, 1 and 2 Striga scores respectively and IT97K-1069-6
had the highest score of 14.33 (Table 3).
Generally, the genotypes received more rainfall under the optimum condition
across the two years irrespective of the mean annual rainfall. The mean rainfall
received by the genotypes was 278.22 mm and 96.45 mm for the optimum and the
stressed condition respectively. The relationship between yield across dates
of planting and rainfall over the two years was linear and positive (Fig.
1). The magnitude of the relationship shows a high coefficient of determination
(R2 = 0.95).
Mean genotypes grain yield and the corresponding drought selection indices
are presented in Table 4. The mean AM across dates for yield
was 569.1 kg ha-1. IT98K-412-13 had the highest AM of 1159.69 kg
ha-1 but the lowest was 287.01 kg ha-1 in Biu local. Similarly,
the highest GM of 971 kg ha-1 was found in IT98K-412-13 but the lowest
was 215.1 kg ha-1 in IT97K-1069-6. The mean RD was 7.69 with highest
and lowest of 2.25 and 13.15 for Biu local and IT98K-628, respectively. Similarly
Biu local had the lowest PR of 54.86 whereas the highest was in IT98K-128-3
and the across dates was 78.18. In terms of DRI, only 2 genotypes exceeded the
threshold of 1.3, i.e. IT98K-412-13 and IT93K-452-1 with 1.79 and 1.88, respectively.
The DRI ranged between -1.88 and 1.88. The mean DSI was .099 and Biu local had
the least index of 0.7 whereas the highest was 1.17 in IT98K-412-13.
The correlation coefficients between genotype yields of date 1 and 2 across
2008 and 2009 and between yields and the different selection indices are presented
in Table 5. Mean yield of date 1 was highly correlated (p≤0.01)
with yield of date 2 and also with AM, GM and RD but not significant with PR,
DRI and DSI. Yield of date 1 has the highest correlation with AM (r = 0.9824**).
|| Relationship between seed yield over genotypes and rainfall
||Mean performance in grain yield kg ha-1 and its
corresponding drought selection indices for some cowpea genotypes evaluated
in field under optimum moisture (date 1) and moisture stressed condition
(date 2) over two years in the Sudan Savanna of the North-West Nigeria
|AM: Arithmetic means, GM: Geometric, RD: Response to drought,
PR: Percent reduction in seed yield, DRI: Drought response index and DSI:
Drought susceptibility index
||Correlation between grain yield and drought selection indices
for 22 cowpea genotypes evaluated in field under optimum moisture (date
1) and moisture stressed condition (date 2) over two years in the Sudan
Savanna of the North-West Nigeria
|AM: Arithmetic means, GM: Geometric, RD: Response to drought,
PR: Percent reduction in seed yield, DRI: Drought response index, DSI: Drought
susceptibility index, y1: Grain yield under optimum, y2: Grain yield under
stressed condition, **Significant at p≤0.01, *Significant at p≤0.05
Yield of date 2 was however highly correlated (p≤0.01) with yield of date
1 and with all the selection indices. AM was also highly correlated (p≤0.01)
with RD and DRI but not with PR and DSI. Significant correlation (p≤0.05)
was obtained between RD, PR and DSI but not DSI. PR was also highly correlated
(p≤0.01) with DRI and DSI. DRI and DSI were also highly correlated.
The stressed condition received just about one third of the total rainfall
of the optimum planting date. This signifies the extent of drought exposure
by the experimental materials. This explains why cowpea farmers do not plant
cowpea in these areas in September as such date is considered too off season
and also suggests the suitability of the date used in this experiment to illicit
response to drought condition.
The highly significant difference of dates for the parameters measured showed
that genotype performed better when sown early in August when rainfall is at
its peak than in September and this is relatively consistent over the years
despite the difference in rainfall pattern in the two years as indicated by
the significant differences between the two years in terms of rainfall.
Similarly, the significant difference of genotypes indicates a wide range of
drought responses by the genotypes.
The relationship between seed yield across date over the years and rainfall
indicated the dependence of seed yield to great extent on the amount of rainfall.
The seed yield increases by 3.9 kg ha-1 with every mm increase in
rainfall. This further stresses the advantage of the earlier sowing date. This
result was in agreement with findings of Abebe et al.
(1998) on relationship between seasonal rainfall and dry bean yield in the
The significant date by genotype interaction was an indication of lack of consistency
in the performance of the genotypes in the two dates since they generally performed
better when sown early in August. The mean seed yield for date 1 was 940 kg
ha-1 and the varieties differ for that date. The highest yield for
date 1 was 1792.47 kg ha-1 in IT98K-412-13 and the lowest was 395.5
kg ha-1 in Biu local. Similarly, the highest seed yield in the date
2 was 526.83 kg ha-1 in IT98K-412-13. However, the lowest was 76.82
kg ha-1 in IT98K-128-3 not Biu local, whereas the mean yield for
this date was 198.12 kg ha-1.
Based on Am, IT98K-412-13, IT99K-216-24-2, IT99K-7-21-2-2 and IT98K-628 were
the best genotypes while Biu Local and IT00K-1263 and Kanannado were the least.
In terms of GM, IT98K-412-13, IT99K-216-24-2, IT99K-7-21-2-2, IT93K-452-1 were
the highest and IT00K-1263, IT97K-1069-6 and IT89KD-288, lowest. Whereas AM
is influenced by seed yield performance in date 1, GM is influenced by the yield
performance in date 2. This is consistent with the result obtained by Abebe
et al. (1998) that these indices appear to be suitable for selection
of lines that perform well across stress and nonstress environments. However,
RD has much less correlation to yield of date 2 (r = 0.44, p<0.05) when compared
to AM and GM and so its less suitable for use in drought prone environment.
PR and DSI are similarly as indicated by the perfect correlation (r = 1) between
them. IT98K-128-3, IT98K-628, IT97K-1069-6 and IT97K-390-2 were ranked as the
most susceptible and Biu Local, Kanannado, IT93K-452-1, Sa Babba Sata and IT98K-412-13
as drought tolerant by PR and DSI. This is based on amount of decline in yield
between date 1 and date 2 and the model presupposes that the more decline in
yield between the dates the less tolerant the genotypes. This is because in
estimating these indices, emphases were on the change in genotype performance
rather than overall performance. Thus, Abebe et al.
(1998) suggested that, PR and S should only be used in combination with
yield data to identify productive cultivars in soil-moisture-stress environments.,
DRI showed highly significant positive correlation (r = 0.78, p<0.01) with
date 2 yield, which indicates its suitability for assessing yield in stressed
environments. This also indicates drought tolerance. This conforms to Acevedo
et al. (1999) who reported that high correlation between DRI and
yield under salinity stress indicates that salinity resistance (the yield not
accounted for by yield potential and escape) has a strong influence on the yield
under stress. However, Richards (1978) suggested that
selection in stressed condition is more efficient for yield improvement than
selection in optimum environment, or selection based on a drought response index.
Based on DRI, IT93K-452-1 and IT98K-412-13 were drought tolerant while IT99K-529-1,
Biu Local, Kanannado, IT99K-216-24-2, Sa Babba Sata, IT96D-610 and IT97K-819-118
have some level of tolerance and the other genotypes are susceptible. These
results highlighted the significance of combining screening for drought tolerance
in areas of production using both yields and drought selection indices.
Muranaka et al. (2011) suggested a close association
between drought stress and Striga resistance because there is probably
restriction of water uptake by Striga that magnifies the effect of drought
stress. IAR-07-1050 was completely resistant to Striga which highlighted
its potential in improving yield performance in drought stressed environment.
Other genotypes such as IT98K-412-13, IT99K-216-24-2 etc also performed well
in presence of Striga. This indicates some level of tolerance to Striga
in addition to drought tolerance observed. It is also clear that their seed
yield may be much greater if improved for Striga resistance; a potential
which can be derived from other genotypes such as IAR-07-1050. In the same vein,
several workers have reported the identification of many other genotypes with
resistance to various strains of Striga (Singh, 2005).
This study assessed the performance of 22 genotypes of cowpea under the stresses
of drought and Striga. The performances of the genotypes varied with
date of planting. Those planted early in August generally performed better.
This reaction is generally consistent over years. One genotype, IAR-07-1050
was found to be resistant to Striga while others were infected to different
extent as indicated by their seed yield. IT93K-452-1 and IT98K-412-13 were drought
tolerant based on DRI and other indices while IT99K-529-1, Biu Local, Kanannado,
IT99K-216-24-2, Sa Babba Sata, IT96D-610 and IT97K-819-118 showed some level
This work was funded through the Tropical Legume II Project of the IITA.