Evaluation of cowpea genotypes for variations in their contribution of N and

P to subsequent maize crop in three agro-ecological zones of West Africa

R.C. Abaidoo1∗ , J.A. Okogun1 , G.O. Kolawole1 , J. Diels1 , P. Randall2 , and N. Sanginga1,3
1
International Institute of Tropical Agriculture (IITA), Oyo Road, PMB 5320, Ibadan, Nigeria, Mailing address:
c/o L.W. Lambourn & Co., Carolyn House, 26 Dingwall Road, Croydon CR9 3EE, England; 2 Commonwealth
Scientific and Industrial Research Organization (CSIRO) Plant Industry, GPO Box 1600, Canberra, ACT, 2601,
Australia; 3 present address, Tropical Soil Biology and Fertility, Institute of CIAT, P.O. Box 30677, Nairobi, Kenya;
∗
Corresponding author E-mail: r.abaidoo@cgiar.org, Tel: +234-(0)2-241 2626 ext 2300

Abstract

Cowpea is an important source of food, cash, and fodder in West Africa. It is perceived to be tolerant to low available
soil phosphorus (P) conditions and nodulates promiscuously contributing to soil fertility through its high nitrogen
(N)-fixing capacity. Cowpea can contribute substantial amounts of N to a subsequent cereal crop but little is known
of the effect of cowpea on P nutrition of a cereal crop grown in rotation. This study, therefore evaluated eight cowpea
genotypes for N2 fixation, tolerance to low soil P (0P), and response to P fertilizer application (90 kg P ha−1 as
Rock Phosphate (RP) or 30 kg P ha−1 as Triple Superphosphate (TSP)), and their potential contribution of N and
P to a subsequent maize crop in three agroecological zones (at Shika, northern Guinea savanna; at Fashola, derived
savanna; and Davié, coastal savanna). Grain yield and P uptake of cowpea genotypes were influenced by location.
Relative response of cowpea genotypes to P application was highest at Shika where soil resin P content was lowest.
The responses to RP application across locations ranged from –14 to 94%; and 194 to 358% for TSP application.
Variation between genotypes was minimal within the 0P and RP treatments but more pronounced within the TSP
treatment at all the locations. The results showed that N2 fixation, N exported in grain, and N balances of cowpea
genotypes were different depending on P nutrition conditions. Grain yield of maize following cowpea was influ-
enced largely by the previous cowpea genotype through its N and P dynamics, location, and VAM colonization of
roots. There was, however, no consistent evidence of a significant effect of P applied to the legume, on the residual
value to the subsequent maize crop.

Key words: Cowpea, Maize, Nitrogen fixation, Phosphorus efficiency, Rotation

Introduction need for N fertilizers through biological nitrogen fix-

Cowpea is widely cultivated in the drier regions of
West Africa and is found in rotation, strip cropping,
relay cropping, and intercropping systems with other
legumes and non-legumes, or as a mono-crop. It plays
an important role in the livelihood of millions of peo-
ple as a source of food, cash, and fodder for livestock.
Cowpea also plays an important role in the nutrient
economies of these cropping systems by reducing the
A. Bationo (eds.), Advances in Integrated Soil Fertility Management in Sub-Saharan Africa: Challenges and Opportunities, 401–412.
© 2007 Springer.
ation (BNF). Estimates of the benefits of cowpea to
soil N supply in the savanna through BNF range from
60 to 80 kg N ha−1 , depending on how the residues
are managed (Dakora et al. 1987; Horst and Härdter
1994). Lower, and even negative values have also been
reported (Awonaike et al. 1990; Sanginga et al. 2000).
Cowpea is a common component of the intensified
cropping systems being adopted in the derived and the
Guinea savanna where the crop is relied upon to fix N
402
and improve soil fertility. However, the BNF process
has a high requirement for P (Vance 2001), yet many
soils in this region have a bicarbonate-extractable P
content of less than 8 ppm (Mokwunye et al. 1986).
Consequently, unless the P nutrition is improved, the
full potential of cowpea to the nitrogen economy of
soil may not be realized, and there is a risk that cowpea
based cropping systems will fall short of expectations
as integrated soil fertility management options in many
parts of the savanna.
One approach to improving crop P nutrition is
through better P acquisition efficiency, and there is
evidence that legume genotypes may have developed
various strategies to take up P from sparingly solu-
ble sources (Marschner 1998; Bagayoko et al. 2000).
The potential therefore exists for the identification
of cowpea genotypes that are productive under rela-
tively low soil P fertility conditions or are better able
to extract residual P or to use applied P fertilizers
more efficiently, particularly sparingly soluble fertiliz-
ers. Successful identification and use of such genotypes
in prevalent cropping systems will improve P nutri-
tion of the legume and possibly also the non-legume
components.
Unlike the situation with N, the contribution of cow-
pea to soil P availability, and its subsequent use by
cereals has not been extensively studied in the West
African moist and derived savanna. Sanginga et al.
(2000) observed large differences among cowpea geno-
types in their P requirements for growth and N2 fixation
at a site in derived savanna in southern western Nigeria,
but suggested further testing to examine the effects on
the soil P status and P nutrition of subsequent crops. The
objectives of the study were (i) to evaluate eight cow-
pea genotypes in three savanna environments (northern
Guinea, derived, and coastal) for inter-genotypic vari-
ations in N2 fixation, tolerance to low P, and response
to P fertilizer application (90 kg P ha−1 as Rock Phos-
phate and 30 kg P ha−1 as TSP), and (ii) to assess their
contribution of N and P to the subsequent maize crop.
Materials and methods
Study sites
The field experiments were conducted at Fashola and
Shika in Nigeria, and Davié in Togo during 2001, 2002
and 2003. Fashola (7º50 N, 3º55 E; soil type, Ferric
Luvisol) is located in the derived savanna of southern
western Nigeria and has a bimodal rainfall pattern with
Evaluation of cowpea genotypes
about 1200 mm rainfall per annum. Shika (11◦ 13 N, 7◦
12 E; soil type, Haplic Lixisol) is located in the north-
ern Guinea savanna (northern Kaduna State, Nigeria)
and has a unimodal rainfall pattern with about 1100
mm rainfall per annum. Davié (6◦ 23 N, 0◦ 50 E; soil
type, Rhodic Ferralsol) is located in the coastal savanna
of Togo with a mean annual rainfall of 1000 mm.
Soil characteristics
Soil chemical characteristics determined at the begin-
ning of the trials included pH (1:1, soil:H2 O), available
P and exchangeable bases (in Mehlich-3 solution at
pH 2.5), using standard laboratory procedures (IITA
1989). In addition, sequential P fractionation was done
using the modified Hedley’s procedure, as described by
Tiessen and Moir (1993), and the P in the solution was
measured by the method of Murphy and Riley (1962).
Available P (Olsen P) in previous legume plots was
measured at maize planting.
Experimental layout, planting and crop management
Fields were cleared at the three locations, ploughed
and ridges spaced 75 cm apart were constructed with a
tractor-mounted ridger. The plot sizes were 4 m x 4 m
with 1 m gap between plots, 1.5 m between treatments,
and 2 m between replications.
The experiment was laid out as a split plot design
with four replications. The treatments consisted of
three P treatments (nil (0P), 30 kg P ha−1 as triple
superphosphate (TSP) and 90 kg P ha−1 as rock
phosphate (RP) applied on the main plots; and eight
cowpea genotypes (Dan-ila, IT82D-716, IT82D-849,
IT86D-715, IT89KD-374, IT89KD-349, IT89KD-391,
IT90K-59), and one maize genotype (Oba Super 1)
grown on the sub-plots. Both the cowpea and maize
genotypes were obtained from the germplasm collec-
tion of IITA, Ibadan, Nigeria. Both species were sown
at four seeds per hole at 25 cm spacing and thinned to
one plant at 2 weeks after planting (WAP). The plots
were weeded manually using hoes and plants were
sprayed periodically, with Karate insecticide from 4
WAP until late podding stage, to control pod borers
(Maruca testularis), aphids, (Aphis craccivora), and
thrips (Megalurothrips sjostedti).
An early harvest was made at 50% podding stage
(R3.5) for cowpea (Fehr et al. 1971) to estimate shoot
R.C. Abaidoo et al.
biomass, vesicular arbuscular mycorrhiza (VAM) colo-
nization of roots, percentage N derived from N2 fixation
(%Ndfa), and total N and P uptake. The colonization
of roots by VAM was estimated using the method of
Giovannetti and Mosse (1980) after staining and clear-
ing following the procedure of Philips and Hayman
(1970). The %Ndfa was estimated by determining the
concentrations of ureides in xylem sap at cowpea pod-
ding stage (R3.5 ), as described by Peoples et al. (1989).
The amount of N fixed (Nf ) was estimated from %Ndfa
and the amount of N in the crop at the same growth
stage. The net contribution of N2 fixation to the N bal-
ance in the soil was estimated using the equation pro-
posed by Peoples and Craswell (1992): Net N balance
= Nf – Ng , where Ng is the total N in the grain.
Harvesting was done at physiological maturity to
estimate grain yield and total N and P exported in
the grain. Harvested plants were weighed and sub-
samples of shoots and grain were oven dried at 65◦ C
for seven days for dry weight determination, and were
then ground and analysed for total N and P using an
autoanalyser following wet acid digestion (IITA 1989).
The effects of the cowpea genotypes and P applied
in 2001 and 2002 on the subsequent maize crop were
determined in 2002 and 2003. In 2003, the previous
0P plots were split into two; one half received 15 kg
P ha−1 as a fresh addition of TSP while the other half
did not. Weeds that had grown in the previous crop
cycle or during the fallow period were cut and cleared
by hand ensuring that remaining legume residues were
not dislodged from their original plots. Soil samples
(0–10 cm) from the individual cowpea and maize plots
were analysed for changes in pH, Olsen P content, and
P fractions, before maize was planted. Four seeds of
maize (Oba Super I) were sown at 75 cm x 25 cm spac-
ing in the previous cowpea and maize plots and the
seedlings were later thinned to one plant per stand at 2
WAP. The maize plots received the same management
as the cowpea crop cycle plots except for the use of
insecticide. Five plants were randomly harvested from
the two middle rows at 8 WAP for estimation of VAM
infection, and total tissue N and P, while 7 m2 of the
harvest (two inner) rows were harvested at physiolog-
ical maturity for grain yield and determination of total
N and P exported in grain.
Statistical analysis
Data collected were combined over two years for each
location and analysed using the mixed model ANOVA
403
procedure of SAS (Littel et al. 1996). In this model, P
source and genotype were considered as fixed effects
and replications as random effects. Data collected for
maize in rotation were analysed separately for each
year since treatments for the years were not the same.
Treatment means were separated using standard error
of the mean.
Results
Soil physico-chemical characteristics
The pH of the field sites ranged from 5.5 (Davié) to 6.50
(Shika) while the exchangeable cations were lowest
at Fashola and highest at Davié (Table 1). The resin-
extractable P (mg P kg−1 , mean of two years) was 1.6 at
Shika, 2.4 at Fashola and 5.4 at Davié. The comparable
values for readily available P (sum of the labile frac-
tions) ranged from 10 mg P kg−1 at Fashola to 15 mg
P kg−1 at Davié. The NaOH-Po fraction ranged from
22.6 mg P kg−1 at Fashola to 37.7 mg P kg−1 at Davié
and was the largest fraction in the labile plus moder-
ately labile pool, comprising about 60% of this pool in
each soil.
The first crop phase
Grain yield of cowpea and response to P
The grain yield of cowpea was considerably lower
at Shika, the driest site, than at Fashola or Davié,
probably reflecting the effect of differences in rain-
fall. Inter-genotypic variation was slight within the 0P
and RP treatments but this was more pronounced with
the TSP application in all locations. Grain yield differ-
ences among genotypes within the 0P treatments were
not significant at Shika and Fashola, but at Davié, the
grain yield of IT86D-715 was significantly higher than
of Dan-ila, IT82D-716, IT89KD-374, and IT89KD-
391, while IT82D-716 produced significantly lower
yields than IT89KD-849, IT89KD-349, and IT90K-59
(Table 2).
Relative grain yield response of cowpea genotypes
to P application was highest at Shika where soil resin
P content was lowest (Table 1). At this location, grain
yield responses to RP application ranged from 14 to
94%, and from 194 to 358% in response to TSP appli-
cation. Responses to RP and TSP applications at the
two other locations were lower and ranged from 0 to
15% for RP and from 20 to 53% for TSP. Cowpea
404 Evaluation of cowpea genotypes
Table 1. Physico-chemical characteristics of top soil (0–15 cm) from the three locations, sam-
pled in 2001 prior to sowing of the first (cowpea) crop phase. At each location, the sites used
in 2001 and 2002 were adjacent

Soil characteristics Fashola Shika Davié

2001 2002 2001 2002 2001 2002

pH (H2O) 6.3 6.0 6.3 6.5 5.5 6.0

Soil texture (g kg−1 )
Sand 670 730 790 790 430 450
Silt 220 100 100 100 400 380
Clay 110 170 110 110 170 170
Exchangeable cations
(c molc kg−1 )
Ca 2.0 1.7 2.1 1.9 3.2 3.1
Mg 0.8 1.0 1.5 1.3 2.2 2.0
K 0.2 0.3 0.2 0.2 0.8 0.4
P fractions (mg kg−1 )
Labile
Resin-P 3.2 1.5 0.7 2.5 6.9 3.8
HCO3-Pi 3.2 0.8 2.6 3.8 4.6 1.5
HCO3-Po 4.1 7.1 4.3 9.2 5.0 8.2
Moderately labile
NaOH-Pi 7.3 3.6 7.2 10.5 10.4 10.0
NaOH-Po 27.7 17.5 27.1 34.3 38.8 36.6
Ca bound
1M HCl 2.7 3.7 2.0 4.6 5.4 13.5
Stable
Conc. HCl-Pi 27.4 6.0 26.2 8.9 23.8 17.3
Conc. HCl-Po 1.0 5.4 22.2 5.9 6.9 11.1
Residual P 15.3 10.0 31.5 48.7 31.5 32.9
Total P 91.6 55.6 123.6 128.4 133.2 134.8

Table 2. Grain yield (kg ha−1 ) of cowpea genotypes grown at three locations without P or with P applied as RP or TSP1

Genotype Location
Shika Fashola Davié
0P RP TSP Mean 0P RP TSP Mean 0P RP TSP Mean

Dan-ila 167 224 510 301 791 880 1010 894 668 655 968 763
IT82D-716 159 193 468 273 652 721 998 791 498 410 600 503
IT89KD-849 163 186 528 292 792 868 1039 900 880 840 1145 955
IT86D-715 213 355 768 445 684 693 968 782 1013 905 1090 1002
IT89KD-349 172 270 607 350 879 914 1054 949 868 720 675 754
IT89KD-374 150 207 530 296 800 797 1129 909 760 863 873 832
IT89KD-391 198 314 745 419 784 835 959 860 593 570 595 586
IT90K-59 166 322 761 416 860 990 1244 1031 940 915 1230 1028
Mean 173 259 614 780 838 1050 777 744 897
SEM2
P source 15 38 28
Genotype 25 65 45
P×Genotype 43 107 78
1 Data represent means across 2001 and 2002, and 2 SEM, the standard error of the mean.
R.C. Abaidoo et al. 405

genotypes IT90K-59 and IT86D-715 had the highest
grain yield response to both RP and TSP at Shika and
Fashola while IT89KD-374 and IT90K-59 gave posi-
tive responses to RP at Davié. The cowpea genotype
IT82D-716 produced the lowest grain yield response
to RP and TSP at Shika and was among the genotypes
with lowest response to P application at Fashola and
Davié.
Total P exported in cowpea grain
The application of TSP significantly increased grain P
content in all cowpea genotypes at all locations except
for a few genotypes in Davié. The percentage increases
in grain P content ranged from 234 to 386 at Shika, from
23 to 66 at Fashola, and from 11 to 69 at Davié. The
increases in grain P content due to PR application were
not significant at Fashola and Davié, but a few geno-
types (e.g., IT86D-715, IT89KD-391 and IT90K-59)
accumulated higher P from RP than the corresponding
0P treatments (Table 3).
N fixation and N balance
Phosphorus fertilizer application significantly enhanced
N2 fixation at Shika but not at the other two loca-
tions. The mean percentage N derived from the atmo-
sphere (%Ndfa) within the 0P treatment was increased
by 18% with RP application and by 26% with TSP
application (Figure 1). Variations between genotypes
were, however, more pronounced. The %Ndfa by
Dan-ila, IT89KD-349, and IT89KD-374 at Shika was
significantly greater than that of IT82D-716 in the
0P treatment. Within the TSP treatments, %Ndfa by
IT82D-716 and IT89KD-849 was significantly lower
than that measured for Dan-ila, IT89KD-349, and
IT89KD-374.
Application of TSP increased total N accumulation
in shoot biomass at mid podding stage (R3.5). Increases
over 0P treatments were 32% at Davié, 69% at Fashola,
and 294% at Shika (Table 4). Variations among geno-
types within the 0P and RP treatments at all locations
were not significant but the differences among some
genotypes within TSP treatments at Shika and Davié
were significant. At Shika, for example, the N accu-
mulation of 145 kg ha−1 for IT82D-715 was signifi-
cantly higher than the 77 kg N ha−1 accumulated in
IT89D-374, while at Davié, N accumulation of 133–
138 kg ha−1 measured for IT82D-716, IT82D-715, and
IT89D-391 was significantly higher than the 85 kg N
ha−1 of IT89D-849.
The N contribution to the subsequent maize crop
(N balance) was significantly improved with TSP but
not with RP application in all the locations. At Shika,
mean N balance across genotypes within the TSP treat-
ment was 31.7 kg ha−1 compared to the 1.8 kg ha−1
and 6.1 kg ha−1 for the 0P and RP treatments, respec-
tively (data not shown). Similarly, TSP application at
Fashola resulted in a mean N balance of 27.9 kg ha−1

Table 3. Total P exported in grain (kg ha−1 ) of cowpea genotypes grown at three locations under low and high P conditions1

Genotype Location
Shika Fashola Davié

0P RP TSP Mean 0P RP TSP Mean 0P RP TSP Mean

Dan-ila 0.58 0.83 1.94 1.12 2.66 2.94 4.17 3.26 2.31 2.43 3.91 2.88
IT82D-716 0.61 0.75 2.19 1.18 2.54 2.98 4.73 3.41 2.09 1.81 3.12 2.34
IT89KD-849 0.46 0.56 1.73 0.92 3.45 3.92 4.87 4.08 3.33 2.63 5.20 3.72
IT86D-715 0.66 1.19 2.62 1.49 3.17 3.37 4.88 3.80 3.53 3.14 4.66 3.78
IT89KD-349 0.57 0.98 2.24 1.27 3.32 3.21 4.10 3.54 3.18 2.40 2.82 2.80
IT89KD-374 0.45 0.70 1.98 1.04 3.48 3.39 4.82 3.90 2.77 2.95 3.35 3.02
IT89KD-391 0.72 1.10 2.82 1.55 2.58 2.99 4.27 3.28 2.54 2.40 3.26 2.74
IT90K-59 0.56 1.14 2.72 1.48 3.18 3.73 4.74 3.88 3.45 3.15 4.71 3.77
Mean 0.58 0.91 2.28 3.05 3.32 4.57 2.90 2.62 3.89
SEM2
P source 0.06 0.20 0.15
Genotype 0.09 0.33 0.25
P × Genotype 0.16 0.58 0.43
1 Data represent means across 2001 and 2002, and 2 SEM, the standard error of the mean.
406 Evaluation of cowpea genotypes

Shika
100

80

60

40

20

–20

–40

Fashola
80

60

40

20

–20

–40

–60
Davie
110

90

70

50

30

10 N export
N fixed
–10 N balance

–30

–50
Dan-ila IT82D- IT89K IT86D- IT89K IT89K IT89K IT90K-
716 D-849 715 D-349 D-374 D-391 59

Figure 1. Nitrogen fixation, N exported in grain, and N balance of cowpea genotypes grown at three locations with no added P or with 30 kg
P.ha−1 (as TSP). For each pair of bars, the left represents low-P and the right, the high-P (30 kg P ha−1 ) condition.
R.C. Abaidoo et al. 407

Table 4. Total N (kg ha−1 ) in shoot biomass at R3.5 growth stage of cowpea genotypes grown at three locations without P or with P applied as
TSP or RP

Genotype Location

Shika Fashola Davié

0P RP TSP Mean 0P RP TSP Mean 0P RP TSP Mean

Dan-ila 28.3 49.0 106.6 61.3 70.7 94.5 126.9 97.4 97.4 103.4 125.1 108.6
IT82D-716 34.3 45.8 120.2 66.8 56.5 74.9 94.5 75.3 86.5 115.8 133.1 111.8
IT89KD-849 28.3 41.4 107.9 59.2 59.5 61.8 151.3 90.9 72.8 90.6 85.4 82.9
IT86D-715 34.2 36.1 144.8 71.7 70.3 64.7 99.0 78.0 108.6 107.9 133.4 116.6
IT89KD-349 15.6 36.8 89.9 47.4 76.9 69.4 101.9 82.7 76.8 98.4 98.6 91.3
IT89KD-374 31.2 36.7 76.6 48.2 77.8 59.4 133.8 90.3 88.9 108.1 106.0 101.0
IT89KD-391 20.9 47.6 104.3 56.6 73.3 83.7 108.1 88.4 94.5 121.6 137.6 117.9
IT90K-59 18.9 26.7 84.6 43.4 66.7 88.4 118.0 91.0 86.1 91.9 119.9 99.3
Mean 26.5 40.0 104.4 69.0 74.6 116.7 88.9 104.7 117.4
SEM
P source 6.92 8.44 5.43
Genotype 11.31 13.79 8.78
P × Genotype 19.58 23.88 15.15
1 Data represent means across 2001 and 2002, and 2 SEM, the standard error of the mean.

compared to 7.2 kg ha−1 for 0P and 7.9 kg ha−1 for
the RP treatment. Variation among genotypes was also
location-specific. For example, Dan-ila and IT82D-715
had significantly higher N balances than IT90K-59
at Shika within the TSP treatments (Figure 1), but
the N balance estimated for IT89D-849 was signifi-
cantly higher than those of IT82-715 and IT90K-59 at
Fashola.
RP (4.4), and TSP (3.7) were not significantly different.
Differences between previous cowpea genotype plots
within the same P treatment were not significant at any
of the locations.
The second crop phase

Grain yield of maize grown after cowpea

Olsen P concentration in soil of plots previously sown
to cowpea genotypes
Olsen P contents of previous (2001) P treatment plots
measured in 2003 (at maize planting) varied signifi-
cantly at Davié, but not at Fashola (the 2001 Shika
plots were lost in 2002, and therefore unavailable for
sampling in 2003). However, differences among pre-
vious cowpea genotype plots within the P treatments
were not significant. The ranges of Olsen-P values (mg
P kg−1 soil) measured from the previous P treatment
plots at Davié were: 0P, 2.3–4.0; RP, 3.1–5.0; and TSP,
4.0–8.8; and at Fashola were: 0P, 2.9–5.3; RP, 3.3–5.5;
and TSP 2.6–5.3. Similar trends were observed for the
previous 2002 legume x P treatment plots sampled in
2003. At Shika and Davié, mean Olsen P contents (mg
kg−1 soil) of previous TSP plots (Shika, 7.1; Davié,
7.30) were significantly higher than that of the previ-
ous RP (Shika, 4.2; Davié, 4.5), and 0P (Shika, 3.3;
Davié, 3.6) plots. At Fashola, mean values for 0P (3.8),
In 2002, mean maize grain yields across plots previ-
ously planted to cowpea genotypes were generally low
and ranged from 1086 to 2069 kg ha−1 but the mean
maize grain yield after pooling data for all genotypes
in response to previous TSP application was signifi-
cantly higher than the mean yields from the previous
RP treatment. Grain yields from previous 0P treatments
were also significantly higher than those from the RP
treatment but were comparable to yields from previ-
ous TSP plots at Fashola (data not shown). In Fashola,
mean maize grain yield was influenced by the previous
cowpea genotype treatment but only the differences
between maize grain yields from previous Dan-ila, and
IT89KD-391, and the previous maize plot were signif-
icant within the previous TSP plots. Mean total N and
P exported in the grain of maize grown in previous 0P
and TSP plots were significantly higher than N and P
exported from previous RP plots at Fashola but not at
Davie (data not shown).
408 Evaluation of cowpea genotypes

Table 5. Grain yield (kg ha−1 ) of maize grown in 2003 following cowpea genotypes or maize grown in 2002. P treatments were 0P or a current
application of 15 kg P ha−1 or residual P from RP or TSP applied in 2002

Previous crop Location

Shika Fashola Davié

0P 15 kg P RP TSP 0P 15 kg P RP TSP 0P 15 kg P RP TSP

Cowpea
Dan-ila 1669 3004 2110 2115 700 1110 1260 1018 3120 3152 2839 3721
IT82D-716 1025 2503 1795 1820 1219 1807 1593 1522 2547 3051 3067 3082
IT89KD-849 1063 2501 2271 2265 1116 1204 972 1539 1978 2310 2419 2885
IT86D-715 909 2649 1668 2275 1100 1261 1406 1372 2906 3166 2527 3315
IT89KD-349 587 1820 1850 2057 1217 1533 1347 1222 2515 3396 2850 3348
IT89KD-374 717 2041 1778 1807 808 1318 1189 1210 2428 3167 2684 2873
IT89KD-391 975 2171 1744 2298 875 1116 1146 1249 2507 3800 2502 2525
IT90K-59 756 1982 2171 2013 1253 1683 1410 934 2394 3332 2478 3125
Maize 449 1544 1632 1422 729 940 1063 1136 2646 3068 3267 2637
Mean 906 2246 1891 2008 1002 1330 1265 1245 2560 3160 2737 3057
SEM1
P source 97 81 115
P × Genotype 292 245 344
1 SEM: standard error of the mean.

Maize grain yields (from previous 2001 cowpea and
maize plots) from the 0P and TSP plots were generally
higher than yields from previous RP plots at Fashola
but not at Davié (data not shown). Also, differences
between maize grain yield from previous cowpea geno-
types and maize plots were significant (P < 0.05) only
for the Dan-ila–RP plot and IT89KD-391-TSP plot at
Fashola. The fresh addition of 15 kg P ha−1 at planting
did also significantly improve maize grain yields over
the yields from the previous P treatments at Fashola.
The application of 15 kg P ha−1 to the 2002 cow-
pea plots increased maize grain yields over yields from
previous 0P plots within the range of 18–95% at Shika,
18–92% at Fashola, and 3–24% at Davié (Table 5). At
Shika, mean maize yields from previous 0P plots were
significantly lower than those from the previous RP and
TSP plots. At Fashola, mean maize yields from previ-
ous RP and TSP plots were significantly higher than
yields from the previous 0P plots but the addition of
15 kg P ha−1 made no significant contribution to yield
increases in previous 0P plots, as observed at Shika. The
effect of previous cowpea genotypes and the fresh addi-
tion of 15 kg P ha−1 on grain yield of maize followed
no consistent pattern across the locations. For example,
at Shika, the addition of 15 kg P ha−1 resulted in a sig-
nificantly higher grain yield from previous 0P plots for
Dan-ila, IT82KD-716, IT89KD-849, and IT86KD-715
compared with yields from the previous maize plot.
Total N and P in grain of maize grown after cowpea
Total grain N and P from previous (2002) TSP plots and
the plots that received the 15 kg P ha−1 were signifi-
cantly higher than the 0P yields at Shika and Fashola but
not at Davié (Data not shown). Percentage increases in
total N in maize grain following cowpea genotypes over
that of maize following maize ranged from –12 to 341
at Shika, –22 to 134 at Fashola and –22 to 49 at Davié
(Figure 2). With respect to residual P benefits, percent-
age increases over that of maize after maize ranged
from –13 to 226 at Shika, –41 to 212 at Fashola, and
–10 to 76 at Davié. Despite these gains, there were only
a few isolated instances where total maize grain N and
P contents from previous cowpea genotype plots were
significantly higher than from previous maize plots, but
this was not consistent across the locations and within
P treatments.
Mycorrhizal infection of maize roots following
cowpea
Percentage VAM infection of roots of maize grown in
previous 0P and TSP treatment plots was generally
higher than those from the previous RP plots while
the fresh application of 15 kg P ha−1 resulted in sig-
nificant suppression of VAM infection of maize roots
(Table 6). The effect of previous cowpea genotype on
VAM infection of the root of the subsequent maize crop
R.C. Abaidoo et al. 409

Shika TSP (19.6)

Previous P source and location

Shika 0P (6.3)

Fashola TSP (13.9)

Fashola 0P (8.3)

Davie TSP (34.6)

Davie 0P (43.4)

–100 –50 0 50 100 150 200 250 300 350 400 450
Percent change in N exported in gran

Davie 0P (8.9)
Previous P source and location

Davie TSP (8.3)

Fashola 0P (1.54)

Fashola TSP (4.7)

Shika 0P (1.1)

Shika TSP (2.7)

–50 0 50 100 150 200 250 300

Percent change in P exported

Figure 2. Gains in total N and P in grains of maize grown after cowpea genotypes over maize after maize. Figures in parenthesis represent
absolute N and P contents of grain of maize grown after maize.

was also inconsistent but higher infection rates were by the cow pea crop and amounts from the expected
more commonly associated with previous 0P plots in release and accumulation of P from the RP were not
all the locations. substantial due to the low solubility of RP.
Despite its reputation for tolerance to low P, cow-
pea has shown good response to external P application
Discussion (Sanginga et al. 2000; Ahloowalia et al. 1995). This has
been attributed to inherent low availability of soil P in
The available P data indicated that the sites selected the locations used and the high demand by legumes
were low in soil available P. The NaOH-Po fraction for P for N2 fixation. Cowpea breeding programmes in
was, however, large indicating that it may potentially Africa advocate genotypes that are capable of produc-
serve as an available P source upon mineralization ing high yields under low soil P conditions. The variety
(Kolawole et al. 2003). However, this depends on other IT90K-59, relative to the others in this study, invari-
soil factors and the capability of the legume plants to ably meets this requirement and would fit well in other
use this sparingly available source readily. The Olsen cropping systems such as, in rotation with previously
P contents of previous cowpea treatment plots mea- fertilized maize.
sured during the maize crop cycle suggest that much of The differential responses of cowpea genotypes to
the TSP applied during the legume cycle was utilized applied P were also reflected in the amount of P
410 Evaluation of cowpea genotypes
Table 6. VAM colonization (%) of roots of maize in 2003 following cowpea genotypes or maize grown in 2002. P treatments were 0P or a
current application of 15 kg P.ha−1 or residual P from RP or TSP applied in 2002

Previous crop Location

Shika Fashola Davié

0P 15 kg P RP TSP 0P 15 kg P RP TSP 0P 15 kg P RP TSP

Cowpea
Dan-ila 41.2 4.1 15.9 32.2 41.8 3.5 5.7 21.8 24.4 5.0 13.2 32.8
IT82D-716 41.8 1.7 12.8 38.4 42.7 7.0 12.7 12.9 51.5 4.6 5.5 26.2
IT89KD-849 47.6 3.1 34.6 43.0 41.4 5.5 3.4 32.3 23.3 5.8 8.5 31.8
IT86D-715 48.4 2.5 17.2 41.0 20.9 6.1 12.6 31.1 30.3 5.7 1.9 22.1
IT89KD-349 36.8 3.1 25.1 39.9 42.6 8.0 11.5 27.4 23.3 3.7 5.3 27.5
IT89KD-374 49.5 3.5 15.9 24.9 42.6 7.3 1.0 23.4 33.9 1.3 7.8 32.0
IT89KD-391 33.2 4.4 19.2 31.9 28.2 8.1 4.2 29.5 41.0 2.5 5.5 34.7
IT90K-59 43.9 3.2 37.3 41.5 37.1 1.7 1.1 25.2 30.1 1.8 6.4 19.2
Maize 44.5 4.4 28.7 46.4 44.5 6.0 5.9 28.1 31.5 5.8 17.7 29.6
Mean 43.0 3.3 23.1 37.7 38.0 5.9 6.6 25.9 31.4 4.0 11.0 28.4
SEM1
P source 0.45 0.34 0.45
P x Genotype 1.35 1.03 1.50
1 SEM, the standard error of the mean

exported in grain. The total P exported in the grain of the
responsive cowpea genotype IT90K-59 was higher than
the uptake in the inefficient non-responsive IT82D-716
in almost all the P treatments (data not shown). This
suggests that IT90K-59 may not be a suitable candidate
when considering cowpea genotypes with the potential
of enhancing the P nutrition of the cereal crop in rota-
tion, due to its high P export in grain and which may
suggest a high P harvest index (PHI). Blair (1993) sug-
gested that an ideal P efficient plant should have low
PHI so as to retain more P in straw for subsequent use
by less efficient plants in the cropping system. A suc-
cessful improvement of P availability in cowpea–maize
cropping systems in the savannas will require cowpea
genotypes efficient in P utilization for high grain and
stover productivity for sustained benefits to farmers.
The relatively wide range of %Ndfa recorded for the
cowpea genotypes under the various P environments
compares with levels reported elsewhere (Eaglesham
et al. 1982; Awonaike et al. 1990; Sanginga et al. 2000)
but was lower than the maximum reported for cow-
pea (Peoples and Craswell 1992). BNF was generally
increased by P application but this was most pro-
nounced at Shika where initial soil P availability levels
were lowest which confirmed the high P requirements
for the N2 fixation process (Vance 2001). However, the
moderate increases in %Ndfa resulting from RP and
TSP applications in the other locations may suggest
that cowpea genotypes can maintain a high rate of N2
fixation under moderately low available P conditions.
The relatively low response may also be attributed to
other nutrient limitations. An interesting observation
was that the cowpea genotype IT82D-716 recorded a
relatively low N2 fixation rate at all the locations but
its total N accumulation and N balance remained rela-
tively high. This may be attributed to an efficient uptake
and use of available soil N and possibly accounts for
the low response to P application, as soil N uptake is
less dependent on P availability than N2 fixation.
Nitrogen benefits to cereals through legume cultiva-
tion have been attributed to N sparing for the cereal’s
use, or from mineralization of the N2 fixed in the
legume stover and belowground plant parts (Sanginga
et al. 2000). The N balances estimated in this study
due to previous cowpea cultivation ranging from –2.6
to 52 kg N ha−1 were lower than values reported else-
where (Dakora et al. 1987; Horst and Härdter 1994), but
higher than some of the estimates reported by Awon-
aike et al. (1990) and Sanginga et al. (2000). These
differences could have been due to the variations among
genotypes used in the various studies and, more impor-
tantly, to environmental influences such as the popula-
tion sizes and effectiveness of indigenous rhizobia and
soil physico-chemical characteristics.
R.C. Abaidoo et al.
The significant maize grain yield response to the
fresh application of 15 kg P as TSP in some of the loca-
tions may corroborate the suggestion by Linquist et al.
(1996) that smaller but regular applications of P fertil-
izer would act as an insurance for maintaining moderate
yields and also contribute to the gradual build-up of the
soil P capital of tropical soils.
The anticipated benefits of previous cowpea cultiva-
tion were characterised by wide variations that could
be attributed to cowpea genotype and soil P availability
(Figure 1). The ranges of total N and P accumulated in
the grain of maize following cowpea genotypes were
relatively wide but higher percentage increases over
values from maize-after-maize plots were common
within 0P plots, especially at the Shika location where
inherent soil P availability was low. Even though these
increases recorded in the three locations are similar to
those earlier reported (Dakora et al. 1987), the high
incidence of more and even negative gains in maize
grown after certain cowpea genotypes calls for thor-
ough evaluations of cowpea genotypes before they are
used in rotations with maize in anticipation of positive
and significant N and residual P benefits to the maize
component. The relatively low grain yield and P accu-
mulation values recorded for maize after the P efficient
genotype (IT90K-59) suggest that the use of efficient
genotypes may not necessarily lead to P sparing for use
by the following cereal crop.
VAM colonization of maize roots was considerably
higher in the 0P plots than in plots where P had been
applied to the maize at sowing and consistent with
observations made by Borie et al. (2002) of increased
VAM infection in plants under P stress. VAM infection
was also low in RP and high in TSP, a surprising result
considering that the soil Olsen P values were higher
following TSP than following RP, at least at Shika and
Davié, and the reason for this is unclear from our data.
There is the need for a holistic analysis of the factors
that have the potential to interact with N and P availabil-
ity to limit the full expression of the potential of grain
legumes to improve soil fertility and yield stability in
low input cowpea–maize cropping systems.
Conclusions
The cowpea genotypes used in the study varied in
their efficiency at maintaining high yields under low
P conditions, their response to P application, and to
maintain N balance for potential use by the subsequent
maize. Phosphorus application to cowpea improved the
411
N contribution to subsequent maize through enhanced
N2 fixation. Unlike N contribution, growth of P effi-
cient cowpea genotypes may not necessarily lead to P
sparing effect on subsequent maize growth due to the
high P levels exported in grain from system. The fresh
application of moderate levels of P rather than depen-
dence on residual P from previous P application would,
therefore, be necessary for maintaining yield stability
in cowpea–maize rotation systems in sites extremely
deficient in P, such as those used at the Shika location.
References
Ahloowalia B.S., Kumarasinghe K.S., Sigurbjoernsson B. and
Maluszynski M. 1995. Genotype selection for improved phos-
phorus utilization in crop plants. In: Johansen C. Lee K.K.,
Sharma K.K., Subbarao G.V. and Kueneman E.A. (eds.), Genetic
manipulation of crop plants to enhance integrated nutrient man-
agement in cropping systems-1. Phosphorus: Proceedings of an
FAO/ICRISAT Expert Consultancy Workshop, 15–18 Mar. 1994.
ICRISAT Asia Center, Pantancheru 502 324, Andhra Pradesh,
India: International Crops Research Institute for the Semi-Arid
Tropics, pp. 49–54.
Awonaike K.O., Kumarasinghe K.S. and Danso S.K.A. 1990. Nitro-
gen fixation and yield of cowpea (Vigna unguiculata) as influ-
enced by cultivar and Bradyrhizobium strain. Field Crops Res. 24:
163–171.
Bagayoko M., Buerkert A., Lung G., Bationo A. and Römheld V.
2000. Cereal/legume rotation effects on cereal growth in Sudano-
Sahelian West Africa: soil mineral nitrogen, mycorrhizae and
nematodes. Plant and Soil 218: 103–116.
Blair G. 1993. Nutrient efficiency – what do we really mean? In: P.J.
Randall (ed), Genetic Aspects of Plant Mineral Nutrition. 1993.
Kluwer Academic Publishers, The Netherlands. pp. 205–213.
Borie F., Redel Y., Rubio R., Rouanet J.L. and Barea J.M. 2002.
Interactions between crop residues application and mycorrhizal
developments and some soil-root interphase properties and min-
eral acquisition by plants in an acidic soil. Biol. Fert. Soils 36:
151–160.
Dakora F.D., Aboyinga R.A., Mahama Y. and Apaseku J. 1987.
Assessment of N fixation in groundnut (Arachis hypogea L.) and
cowpea (Vigna unguiculata L. Walp) and their relative N contri-
bution to a succeeding maize crop in northern Ghana. MIRCERN.
J. 3: 389–399.
Eaglesham A.R.J., Ayanaba A., Rao V.R. and Eskew D.L. 1982. Min-
eral N effects on cowpea and soybean crops in a Nigerian soil. Plant
Soil 68: 171–181.
Fehr W.R., Caviness C.E., Burmood D.T. and Pennington J.S. 1971.
Stages of development descriptions for soybeans Glycine max (L.)
Merill. Crop Sci. 11: 929–931.
Giovannetti M. and Mosse B. 1980. An evaluation of techniques
for measuring vesicular-arbuscular mycorrhizal infection in roots.
New Phytol. 84: 489–500.
Horst W.J. and Härdter R. 1994. Rotation of maize with cowpea
improves yield and nutrient use of maize compared to maize
monocropping in an Alfisol in the northern Guinea savanna of
Ghana. Plant Soil. 160: 171–183.
412
International Institute of Tropical Agriculture (IITA) 1989.
Automated and semi-automated methods for soil and
plant analysis. Manual series No. 7. IITA, Ibadan,
Nigeria.
Kolawole G.O., Tian, G. and Tijani-Eniola H. 2003. Dynamics of
phosphorus fractions during fallow with natural vegetation and
planted Pueraria phaseoloides in south western Nigeria. Plant Soil
257: 63–70.
Linquist B.A., Singleton P.W., Cassman K.G. and Keane K. 1996.
Inorganic and organic phosphorus and long-term management
strategies for an Ultisol. Plant Soil 184: 47–55.
Littel R.C., Milliken G.A., Stroup W.N. and Wolfinger R.C. 1996.
SAS system for mixed models. Statistical Analysis Systems Inc.
Cary, NC., USA. 633pp.
Marschner H. 1998. Role of root growth, arbuscular mycorrhiza, and
root exudates for the efficiency in nutrient acquisition. Field Crops
Res. 56: 203–207.
Mokwunye A.U., Chien S.H. and Rhodes E. 1986. Phosphorus
reaction in tropical African soils. In: Mukwunye, A.U. and Vlex
P.L.G. (eds), Management of nitrogen and phosphorus fertilizers
in sub-Saharan Africa. Martinus Nijhoff Publishers, Dordrecht,
The Netherlands. pp 253–281.
Murphy J. and Riley J.P. 1962. A modified single solution method
for the determination of phosphate in natural waters. Anal Chim.
Acta 27: 31–36.
Evaluation of cowpea genotypes
Peoples M.B. and Craswell E.T. 1992. Biological nitrogen fix-
ation: Investments, expectations, and actual contributions to
agriculture. In: Ladha, J.K., George, T. and Bohlool B.B.
(eds), Biological Nitrogen Fixation for Sustainable Agricul-
ture. Kluwer Academic Publishers, Dordrecht, The Netherlands.
pp 13–40.
Peoples M.B., Faizah A.W., Rerkasem B. and Herridge D.F. 1989.
Methods for Evaluating Nitrogen Fixation by Nodulated Legumes
in the Field. ACIAR Monograph No 11, ACIAR, Canberra,
Australia. 76 pp.
Philips J.M. and Hayman D.S. 1970. Improved procedure for clearing
roots and staining parasitic and vesicular-arbuscular mycorrhizal
fungi for rapid assessment of infection. Trans. Br. Mycolo. Soc.
55: 159–161.
Sanginga N., Lyasse O. and Singh B.B. 2000. Phosphorus use effi-
ciency and nitrogen balance of cowpea breeding lines in a low P
soil of the derived savanna zone in West Africa. Plant Soil. 220:
119–128.
Tiessen H. and Moir J.O. 1993. Characterization of available P by
sequential extraction. In: Carter, M.R. (ed), Soil Sampling and
Methods of Analysis (Chapter 10). Lewis Publishers, Boca Raton,
FL. pp. 75–86.
Vance C.P. 2001. Symbiotic nitrogen fixation and phosphorus acqui-
sition, plant nutrition in a world of declining renewable resources.
Plant Physiol. 127: 390–397.