FieldCropsResearch, 33 (1993) 21’7-233a 217 ...
FieldCropsResearch, 33 (1993) 21’7-233a
217
Elsevier Science Publishers B.V., Amsterd m
Varietal intercropping
the yields and stability
of cowpea
semiarid Senegal
and D.R. Parkerb
ABSTRACT
l
Thiaw, S., Hall, A.E. and Parker, D.R.,
Varietal intercropping and the yields and stability of
cowpea production in semiarid
17-233.
Crop productivity is low and variable under rainfed conditions in semiarid zones. This problem is
particularly acute in northem Senegal where p oductivity also is limited by infertile soils. Studies were
conducted to determine whether varietal intercrops of cowpea are better adapted to semiarid climates
and infertile soils than sole crops of cowpea. The same experiment was conducted over two years in
three locations which have contrastmg rainfall and soi1 fertility, and are representative of the major
cowpea-production zones of northem Senega . Two varietal intercrops and six cultivars in sole crops
were grown in a randomized complete block split-plot design with four replications. The split plots
consisted of a starter fertilizer treatment ( 150 kg/ha of 6:9:8, N:P:K) and a control where no fertilizer
was applied. The varietal intercrops consiste< of altemating rows of a medium-cycle spreading culti-
var (58-57) and an early erect cultivar (Bambey 21 or Califomia Blackeye No. 5). The six cultivars
grown as sole crops included the three cultivars used in the intercrops, and three others with spreading
(N’diambour and Mougne) or semierect (TV 3236) growth habits, that are well adapted to northem
Senegal.
Grain and hay yield responded significantly to fertilizer application at the two drier locations, Thil-
makha and Louga, but not at the wetter location Bambey, which had higher levels of soi1 nitrogen
and phosphorus. There were no genotype x fertilizer interactions. The varietal intercrops produced
higher yields of grain and hay than the sole crops at Thilmakha and Louga and had high land equiva-
lency ratios ( 1.42 for grain and 1.50 for hay), and above average stability. In the wetter and more
fertile location (Bambey), the performance of the varietal intercrops was intermediate compared to
the sole crops, and the dense canopy made it diffrcult ta harvest the grain of the two cultivars in the
intercrops separately. These studies demonarated that intercrops of early erect and medium-cycle
spreading cowpeas cari have higher and more stable yields of grain and hay than sole crops, in condi-
tions where drought and infertile soi1 limit trop production.
Correspondence
to: Dr. A.E. Hall, Department of Botany and Plant Sciences, University of Cal-
ifomia, Riverside, CA 9252 1, USA.
0378-4290/93/$06.00 0 1993 Elsevier S4rente Publishers B.V. Al1 rights reserved.
218
S . T H I A W E T A L .
INTRODUCTION
In semiarid zones trop yields are generally low and variable as a conse-
quence of limited and variable rainfall and other constraints. In the semiarid
region of Senegal, the average rainfall provides sufficient moisture for a 60-
to 1 OO-day cropping season which is followed by a long dry season (Dancette
and Hall, 1979). From year to year, there is substantial variation in useful
length of the rainy season (Khalfaoui, 199 1 ), and drought cari occur at dif-
ferent times during the cropping season, substantially reducing trop
production.
Cowpea ( Signa unguicuYata (L.) Walp.) is a warm-season trop that has
been grown in Senegal for thousands of years (Ng and Marechal, 1985 ) to
produce grain and hay. Formerly, late-klowering spreading varieties were
planted in home gardens or intercropped with cereals or other crops. During
the past thirty years, new cultivars have been developed with early or medium
maturity, and erect or spreading growth habit. More than 90% of the area
planted to cowpea at this time is sole-cropped.
Since the droughts began in 1968, cowpea has become an important trop
in the northern peanut basin of Senegal due to its ability to produce food
under conditions of drought and infertile soils, where the other staple crops
(Pearl millet, sorghum and peanut) have produced virtually no food in some
years. For subsistence farmers growing crops under conditions of .variable
drought and other soi1 and biotic constraints, yield stability may be more im-
portant than high yield. Cropping systems of semiarid West Africa were re-
viewed by Fusse1 and Serafïni ( 1985) who concluded that Pearl millet/cow-
pea intercrops could enhance and stabilize yields compared with sole crops.
Pearl millet/cowpea intercrops have been compared with sole crops in the
semiarid zone of Senegal over a 5-year period (C. Dancette, unpublished
studies). These studies indicated that sole crops were more effective than in-
tercrops in the drier parts of the zone. Reddy et al. ( 1988) reached similar
conclusions from studies conducted in a similar ecological zone in Niger. Mil-
let/cowpea intercrops may have an advantage over sole crops of cowpea where
flower thrips (A4egaZurothrip.s sjostedii) are abundant and pesticides are not
used (Singh et al., 1990). The levels of flower thrips, however, are usually
low in northern Senegal.
Varietal intercrops, combining morphologically and phenologically con-
trasting cultivars of cowpea, could enhance yield stability and contribute to
food requirements (Hall, 1988). In northem Senegal, farmers usually expe-
rience food shortages just before the Pearl millet and late- or medium-cycle
cowpea cultivars are ready to harvest. Cowpea cultivars with early maturity
cari provide food and cash during the traditional period of hunger, as well as
greater grain yield than later-maturing cultivars when the rainy season is very
short (Hall and Patel, 1985 ) . In contras& medium-cycle, spreading cultivars
VARIETAL INTERCROPPING OF COWPEA IN SEN$GAL
219
cari produce more grain and hay t
the early erect cultivais when the rainy
season is of adequate duration or
ought occurs in the middle of the crop-
ping season. On-faim studies in S
a1 have demonstrated that farmers pre-
fer to grow both types of cowpea
e early erect cultivars and the medium-
cycle spreading ones (M. N ‘diaye
d A. Bal, unpublished studies). Possible
advantages from varietal intercr
consisting of alternating rows of these
two types of cultivars may be
he following examples. Where drought
occurs early in the season, th
ivar may senesce, but if rains resume,
a late-maturing cultivar grown
rietal intercrop could compensate by
spreading into the space made
by the senescence of the early culti-
var. In contrast, when the rainy
continuous but short, only the early
cultivar would produce grain but
ater, spreading cultivar would produce
substantial hay.
Preliminary studies in Se
iouf, and M. Diagne, unpublished
studies) demonstrated that
crops of cowpea cari be more effec-
tive than sole crops. In the studies
. Diagne varietal intercrops, consisting
of alternating rows of an early er
cultivar and a medium-cycle spreading
cultivai-, were compared with sole rops of the same cultivars. He evaluated
land equivalency ratio (LER), whi h is the relative land area under sole crops
required to produce the same yiel as an intercrop. In the first year, the LER
for grain yield was only slightly gre ter than unity ( 1.08 ) . In the second year,
the varietal intercrop made much
ore eflkient use of the land than the sole
crops with an LER for grain of 1.11. In addition, the varietal intercrop pro-
duced 20% more grain than the hi est,-yielding sole trop, and 58% more grain
and 40% more hay than the aver ge yields of the sole crops. These studies
were a11 conducted at Bambey in t e wetter part of the semiarid zone of Sen-
egal. It is particularly important t evaluate varietal intercrops over a range
of environments, especially ones 1
w’th more extreme and more variable levels
of drought.
and when leguminous crops were
in soi1 with low supplies of nitrogen.
These conclusions would apply
a cowpea varietal intercrop of early and
the infertile soils in the drier part of the
studies at Louga demonstrated
that in the northern
2 2 0
S. THIAW ET AL.
Yield response to variation in environment due to years and locations is
also an important criterion for assessing the adaptation of cropping systems
or cultivars to semiarid environments. Long-term average yield and yield sta-
bility are important indices of adaptation (Hall et al., 1979). Several meth-
ods are available to evaluate yield stability (Blum, 1988 ) , but they cari make
different predictions concelming the relal.ive stability of the same set of culti-
vars (Hall et al., 1979). A widely used method to evaluate stability consists
of the regression of the yield of individual cultivars against an environmental
index consisting of the mean yield of a11 cultivars in the tria1 using data from
different locations and years (Finlay and Wilkinson, 1963 ) . Linear regres-
sion, as used by Eberhart and Russell ( 1966) would, however, appear to be
more appropriate than the logarithmic regression used by Finlay and Wilk-
inson ( 1963 ), because it provides equal weighting to the different environ-
ments. The final step consists of plotting the regression coefficient against the
overall mean yield of the cultivars (Finlay and Wilkinson, 1963). Cultivars
with high mean yields and regression coefficients less than or close to unity
would be considered to be broadly adapted to a11 environments experienced
in the trials, whereas cultivars with very low regression coefficients would
tend to be adapted to the test environments with low yield potential (Blum,
1988).
The main objective of tlnis study was to compare the average yield, yield
stability and efflciency of land use of varietal intercrops and sole crops of
cowpea grown in several locations with contrasting levels of drought and soi1
fertility.
MATERIALS AND METHODS
During the summers of 1988 and 1989, the same experiment was con-
ducted in three semiarid locations: Bambey ( 14”42’N 16”28’W ), Thil-
makha (15”2’N 16” lS’W), and Louga (15”36’N 16”13’W). These loca-
tions have contrasting rainfall and soi1 fertility, and are representative of the
major cowpea-production zones of Senegal. At Bambey the experiment was
conducted on a deep, slightly leached, tropical ferruginous soil, called a “Dior”
soil. Based on the International Soi1 Science Classification, this soi1 has 7%
clay, 3% silt, 68% fine sandl, 22% coarse Sand, and little variation in texture
in the first two meters. The volumetric moisture content at field capacity is
16 + 2%. Using the United States Department of Agriculture soi1 taxonomy,
this soi1 is classifïed as an Wstipsamment (Hall and Dancette, 1978). Thil-
makha and Louga are located in the northern part of the peanut basin and the
soils are sandy with a low percentage of clay (3%) and a low field capacity
(8%). The pHs of the soils are about 5.5 at Louga and Thilmakha, and 7 at
Bambey.
VARIETAL INTERCROPPING OF COWPEA IN SENbAL
221
Cultural practices
Fields were chosen where pearl lmillet had been grown the previous year
and they were plowed during the dry season in May. Al1 trials were hand
planted when the soi1 was suffrcien ly wet from rainfall to insure germination
and plant establishment, which oc urred in early August in 1988 and early
July in 1989. Two seeds of the six wpea cultivars were planted per hill, and
hills with zero emergence were rep anted 1 week later. After seedling emerg-
ence, plants were thinned to one pe hi11 in a11 locations. Weeds were removed
about 15 days after emergence, w en the cowpea seedlings were well estab-
1
lished. The plots were hand weede three times during the cropping season.
Chemical control of insects was maintained in a11 three locations. Hairy
caterpillar (Amsacta moloneyi), a d cowpea aphid (Aphis craccivora) were
controlled by Thimul 35 applied a the rate of 2.5 1 per ha (the active ingre-
dient is endosulfan at 800 g/ha), a d flower thrip (Megalurothrips sjostedti)
was controlled with Decis applied t the rate of 25 g per ha (the active ingre-
dient is deltamethrine at 15 g/ha) :
.
Experimental design and treatmqbzts
Six cowpea cultivars (Table 1) nd two varietal intercrops were grown as
d
eight genotype treatments in a ran omized complete block split-plot design
with four replications. The genotyp treatments were the main plots, and there
\\
were two fertilizer levels as sub-plo s . Each main plot consisted of 12 rows, 5
m long, and each sub-plot had 6 rows. The distance between rows was 50 cm
for a11 cultivars. Within the rows, t e distance between hills was 25 cm for the
erect cultivars, and 50 cm for the s rea.ding and semi-erect cultivars. The two
intercrop treatments consisted of
early erect cultivar, Bambey 2 1 or Cali-
fornia Blackeye No. 5 (CB5 ), and medium-maturity spreading cultivar, 58-
57, sown in alternate rows. The
r central rows of each sub-plot were har-
vested at the end of the rainy
to determine dry weight yields of grain
and hay.
TABLE 1
~
Origin and characteristics of the cowpea cultijars used
Cultivars
Period from
Growth habit
sowing to 50%
flowering
(days)
58-57
44
spreading
N’diambour
44
spreading
Mougne
46
spreading
Tvx 3236
45
semi-erect
Bambey 2 1
41
erect
Califomia Blackeye 5
38
erect
222
S. THIAW ET AL.
Prior to applying fertilizer, soi1 samples were taken in the O-20 cm depth at
Bambey, Louga, and Thilmakha. The concentrations of total N, P, and K were
determined. Total N was determined by the Kjeldahl method. Soils were di-
gested in a mixture of HF, lHC104, HC1 and HN03, and P concentration was
determined using a molybdovanadate colorimetric method, and K concentra-
tion was determined by atomic absorption spectrophotometry (AAS). One
of the split plots received ferrtilizer, which was broadcast and incorporated at
the rate of 150 kg/ha of 6:9:8 (N:P:K), which is a low level presently recom-
mended as a general practice to farmers in Senegal by the extension service.
This fertilizer also contains a small amount of sulfur in that the nitrogen is
provided by ammonium sulfate.
Fully expanded mature leaves were sampled at mid-bloom (between 35
and 45 days after sowing) from the cultivars 58-57, Mougne, Bambey 2 1 and
CB5 in both control and fertilized plots. Leaf blades were ground and passed
through a 0.85-mm sieve and digested in a mixture of nitric and perchloric
acids. The K and micronutrient (Fe, CU, Zn and Mn) concentrations in the
leaves were determined by AAS. The content of P in the leaf blades was mea-
sured using a modilïcation of the vanadate yellow method of Berg and Gard-
ner ( 1978). Total N concentration was determined using a carbon-nitrogen
analyzer (NA- 1500 Carlo-Erba, Milan, Italy ) . The number of peduncles, pods
per plant and pods per peduncle were determined for the cultivars 58-57,
Bambey 2 1 and CB5. Weather stations at Bambey and Louga provided data
on rainfall, daily pan evaporation (US Weather Bureau Class A), and the
daily maximum and minimum shelter air temperatures. At Thilmakha, only
rainfall was measured.
RESULTS AND DISCUSSION
Weather conditions
In 1988 and 1989, annual rains were 6 39 and 805 mm at Bambey, 409 and
550 mm at Thilmakha, and 442 and 470 :mm at Louga, respectively. Compar-
ison of rainfall and pan evaporation data indicated that the rainfall in 1988
would have supported a 70,.day growing season at Bambey, but only a 60-day
growing season at Louga (Fig. 1). A drought occurred at Louga during the
fourth 1 0-day interval whem the cowpeas were in the early flowering stage. In
1989 there was suftïcient ralin to support a 90-day growing season at Bambey,
but only a 70-day growing season at Louga where the sandier soi1 has limited
ability to store water in the root zone (Fig. 2). An extreme drought occurred
at Louga during the second. and third 1 0-day intervals after sowing when the
cowpeas were in the late vegetative stage. The rainfall at Thilmakha (Thiaw,
1990) was more similar to the rainfall at Louga than at Bambey during both
years. The daily mean pan evaporation was high and averaged about 6 mm/
day at Bambey and Louga in both years. Daily maximum and minimum air
VARIETAL INTERCROPPING OF COWPEA IN SEI’
223
1988
25
BAMEEY
_J
0 R A I N
I E T P
!
LOUGA
dJl.ll
I
i
1 2 3 1
ilb 4 5 6 7 8 9
TEN-DAY INTERVALS FROM SOWING
Fig. 1. Mean rainfall and pan evaporatio for IO-day intervals after sowing at Bambey and
Louga in 1988.
temperatures also were similar at Yambey and Louga in both years with av-
erage values of 34 and 24”C, respe :ively.
Response to fertilizer
Soi1 analyses conducted in 198!
prior to applying the fertilizer, showed
that total N and P were lower at Tl lmakha and Louga than at Bambey (Ta-
ble 2). Other studies at these same ocations gave available soi1 P ( NaHC03
extract) as being lower at Thilma’ la than at Bambey. The level of total K
was similar in the three locations. {a~ signifkant differences were observed
between the cultivars, and the fer-t: zed and control treatments in leaf blade
minera1 content, and average value! u-e presented (Table 3 ) . Differences were
observed between locations in a11 rr nerals in leaf blades except P. The K level
was lower at Thilmakha than at t : two other locations and probably was
defïcient. Among the micro-nutri us, Mn was higher at Louga and Thil-
makha, possibly reflecting the lowc soi1 pH as compared to Bambey. There
was a trend towards higher N levi 5 in leaves from Bambey than the other
locations, which is consistent with he higher total N in the soi1 at Bambey.
Cowpea has been shown to tïx con! clerrable atmospheric nitrogen in a11 three
locations (M. N’Diaye, unpublishc I data). The levels of P in the leaf blades
-.--,*
224
S. THIAW ET AL.
2 5 -
BAMBEY
2
2
z- 2 0 -
r 19690 RAINI ETP
2’ 6
zv
* E’5 -
CE
w
l
1 0 -
-
5 - I
I
3 0 -
LOUGA
2 5 -
9
2
Z-20 -
6’ z
z- 1 5 -
*E
EE
w
10 -
5-
O -
1
2
3
4
5
6
7
6
9
TEN-DAY INTERVALS FROM SOWING
Fig. 2. Mean rainfall and pan evaporation for ICI-day intervals after sowing at Bambey and
Louga in 1989.
TABLE 2
Soi1 minera1 nutrient concentrations in the O-20 cm depth at three locations in Senegal
Minera1
Location
content
Bambey
Thilmakha
Louga
total N @/kg)
0.33
0.18
0.19
cv C%l
8
25
10
total P (g/kg)
0.14
0.10
0.08
cv (%)
4
19
16
Total K (meq/lOO g)
3.17
3.06
3.11
cv (%)
6
16
9
CV =Coefficient of variation.
were intermediate and not significantly dlifferent between the three locations
despite the higher P content in the soi1 at Bambey.
The fertilizer treatment resulted in significantly higher yields of grain and
hay than in control plots at Thilmakha and Louga but not at Bambey (Table
4). The yield responses at Thilmakha antd Louga may have been due to the
small amount of N (9 kgjha) in the starter fertilizer. Studies of Agboola
VARIETAL INTERCROPPING OF COWPEA 1N SEN/EGAL
225
TABLE 3
Leaf blade minera1 nutrient concentrations (dry weight basis) of cowpea in Senegal in 1989, averaged
across four cultivars and the control and fertijized treatments
Location
M n
F e
Zni
C U
N
K
P
WW
WW
(QW
WMg)
WW
ww
WW
Bambey
281
145
28~
12
55.1
,13.1
3.2
Louga
488
260
30 ;
9
41.1
12.0
3.3
Thiimakha
417
134
24 ~
31.1
2.7
LSD (0.05)
70
31
21
8 f6.7
*p:i
i;
signif.
ft
Q
*, Y3ignifïcant at Pc 0.08 and < 0.01, respectively.
TABLE 4
Grain and hay yields of cowpea averaged over genotypic treatments for control and fertilized plots at
Bambey, Thilmakha and Louga
Bambey
l
Thilmakba
Louga
-i
1988
19891
1988
1989
1988
1989
Grain (kg/ha)
Control
1394
1334
671
599
616
501
Fertilized
1438
1399~
930
933
875
575
LSD (0.05)
ns
ns ~
67
53
60
54
Hay (Wha)
l
Control
2424
4766
530
843
461
847
Fertilized
2437
4642
184
1147
158
956
LSD (0.05)
ns
n s
67
12
78
101
( 1978) in Nigeria demonstrated ihat yield of cowpea cari respond to appli-
cation of 10 kg/ha of N when the #erc:entage of soi1 organic matter is less than
1 .O%, but not with 2% or greater organic matter. The soi1 organic matter was
estimated to be less than 1% at Bambey, Louga and Thilmakha. The genotype
treatment xfertilizer interaction was not significant in either year at any lo-
cation for either grain or hay yield. This indicates that trials to evaluate cul-
tivars or varietal intercrops should give similar genotypic or intercrop rank-
ings either with or without starter;fertilizer. In the subsequent analyses yields
are examined which represent thk average across the control and fertilized
treatments.
Overall grain and hay yields we’ le much higher at Bambey during 1988 and
1989 than at Thilmakha and Lo ga (Table 4). Shoot biomass (grain plus
hay) was positively correlated ith seasonal rainfall (r2 ~0.93, P=O.O02)
indicating that the higher yields t Bambey were associated with wetter con-
ditions. The higher soi1 fertility
Bambey was an additional factor contrib-
226
S. THIAW ET AL.
uting to yield. Despite the higher rainfall in 1989, average grain yîeld was
higher in 1988 in a11 locations, possibly due to a more uniform distribution of
the rain in 1988 (compare Figs. 1 and 2 ). Disease and insect pressures also
differed among the experiments. Plots at Bambey suffered a severe aphid in-
festation coupled wîth mosaic virus infection of cultivars 58-57 and
N ‘diambour in 1988, while at Louga and Thîlmakha a11 genotype treatments
suffered from hairy caterpillar in 1989.
Genotypic yield response
The genotypexyear interaction was highly signilïcant at Bambey for both
grain and hay but not at Thilmakha and Louga. Also, there were no significant
genotype x location interactions for Thilmakha and Louga, and mean values
across these two locations are presented (Table 5 ) . The genotypic treatments
differed significantly in grain production except at Thilmakha and Louga in
1989 (Thiaw, 1990). In both years, highest mean grain yields at Bambey were
achieved by sole crops of three cultivars with different growth habits, semi-
erect Tvx 3236, erect Bambey 2 1 and spreading N’diambour, and the varietal
intercrops had întermediate: yields (Table 5 ). In contrast, the varietal înter-
crops had the hîghest grain yields at Thilmakha and Louga (Table 5 ) .
The genotypic treatments differed signilïcantly in hay production except at
Louga in 1988 (Thiaw, 1990). At Bambey highest mean hay yield was
achieved by a sole trop of the spreading cultivar N’diambour, which also had
a high grain yield (Table 5 ) . The hay yields of the varietal intercrops were
TABLE 5
Grain and hay yields (kg/ha) avera8ed across fertilizer treatments and years for cowpea at Bambey
and the combined averages for Thilmakha plus Louga
Genotypic treatment
Bambey
Thilmakha and Louga
-
Grain
Hay
Grain
Hv
58-57
1036
3674
758
766
N’diambour
1468
4388
738
908
Mougne
1349
3982
639
752
Tvx 3236
1828
3998
510
495
Bambey 2 1
1600
2842
596
740
CB5
1230
2276
640
615
58-57/Bambey
21
12159
3500
896
1100
SS-57/CB5
1387
3872
917
952
Mean
1391
3567
712
787
LSD (0.05)
233
818
130
122
gen. X year
f*
tt
ns
ns
gen. X locat.
-
ns
ns
LSD tests the differences between the different treatments. ** = signitïcant at Pc 0.0 1.
VARIETAL INTERCROPPING OF COWPEA IN SENIGAL
227
intermediate at Bambey (Table 5 ) . ~Mean hay yields for Thilmakha and Louga
combined (Table 5) were highest for the varietal intercrops and low for the
erect cultivars (Bambey 2 1 and CB5 ) . These data demonstrated that under
water-limiting conditions and infertile soils, such as at Thilmakha and Louga,
varietal intercrops cari produce higher yields of grain and hay than sole crops
of either erect or spreading cultivars. A,t Bambey, with higher rainfall and soi1
fertility, intercrops were not as productive as the best sole crops and would be
difficult to manage due to the dense canopies that developed.
Yield component data for three ~contrasting genotypes, 58-57, Bambey 2 1
and CB5, in the three locations demonstrated that the higher yields at Bam-
bey were associated with a greater number of peduncles and pods (Thiaw,
1990). There was little difference~ in number of pods per peduncle among
locations. Other studies with cowpea under different irrigation treatments have
also shown a strong positive associltion between grain yield and the number
of pods per plant (Ziska and Hall, ~198#3).
Few studies were discovered in the literature on the performance of varietal
intercrops. Rattunde et al. ( 1988 ) evaluated morphologically and phenolog-
ically différent cultivars of peanut as mixtures or as intercrops with pairs of
varieties sown in altemating rows. They found that the highest pod yields of
the mixtures and row intercrops failed to exceed the yield of the best cultivars.
They concluded that future experimentation with peanut mixtures in the
tropics should focus on stability rather than maximization of yield. Nages-
wara Rao et al. ( 1990) evaluated the performance of intercropping short- and
long-duration peanut genotypes in environments subjected to end-of-season
droughts. The intercrop treatments resulted in LERs of 0.93 to 1.25 with a
mean of 1.1 for pod yield and 0.99 to 1.15 with a mean of 1.1 for total biomass
at the end of the season. They concluded that intercropping peanut genotypes
with different season length reqmrements is a better solution to variable sea-
son length than simply spreading the risk by growing a range of genotypes as
sole crops. However, the LER values were close to 1 .O indicating only small
improvements in land-use efIiciency, and mechanizing the harvest of peanuts
would be difficult with varietal intercrops.
Land-use and biological efficiencies
Mean LER values showed that the varietal intercrops of cowpea had aver-
age land-use efficiencies for grain and hay that were 55% and 57% higher,
respectively, than the sole crops at Thilmakha, and 30% and 43% higher, re-
spectively, for grain and hay at Louga (Table 6). The average advantage of
the varietal intercrops was smaller t Bambey with + 15% for grain and + 22%
for hay. The extent to which the g!eater droughts and lower soi1 fertility con-
tributed to the greater LERs at Thilmakha and Louga, compared with Bam-
bey, is not known. There was no difference in the LERs of fertilized compared
with control plots at Thilmakha or Louga (data not shown). In the studies
---.
--A-
-t.
228
S. THIAW ET AL.
TABLE 6
Values for LER and ATER averaged across fertilizer treatments and years for grain and hay yield of
cowpea at Bambey, Thilmakha and Louga
Location
Intercrop
LER
ATER
Grain
Hw
Grain
Hw
Bambey
58-57/B 21
1.05
1.07
0.97
0.98
SS-57/CB5
1.25
1.37
1.10
1.15
Thilmakha
58-57/B21
1.64
1.67
1.47
1.57
SS-57/CB5
1.38
1.50
1.14
1.32
LWP
58-57/B21
1.12
1.47
1.06
1.32
SS-57/CB5
1.48
1.40
1.21
1.24
LER= land equivalency ratio. ATER = area X time equivalency ratio.
reviewed by Hiebsch and McCollum ( 1987 ) there was a general tendency for
lower LERs under more fertile soi1 conditions, but more fertilizer was applied
than in the present studies. They also showed that intercrops grown under low
soi1 N had ATER values greater than unity, whereas intercrops under high
soi1 N had average ATER values close to unity (Hiebsch and McCollum,
1987). In this study, the ATER values showed that varietal intercrops cari
have higher biological effici.ency than sole crops of the same cultivars (Table
6). At Thilmakha and Louga, the biological effkiencies of the intercrops for
grain (average ATER of 1.22 ) and hay (average ATER of 1.36) were sub-
stantially higher than the sole crops. In contrast, average ATER values at
Bambey for grain and hay were close to 1 .O indicating little difference in bio-
logical effkiency between the sole crops and intercrops. The magnitude of the
partial LERs show which cultivar made the major contribution to the total
LERs. In 1988, the partial LERs for grain (Table 7 ) were higher for the early
erect cultivars (Bambey 2 1 and CB5 ) at Bambey and Thilmakha, whereas in
1989, the medium-cycle cultivar had the highest partial LERs in ail locations.
In 19 out of 24 cases, the partial LERs were greater than 0.5, indicating con-
tributions from both cultivars to the higher LERs of the varietal intercrops.
Partial LERs for hay (Table 8) indicated that in most cases both cultivars
contributed to the high LERs of the varietal intercrops at Thilmakha and
Louga in both 1988 and 1989. The low partial LERs for grain and hay during
1988 for the medium-cycle spreading cultivar at Bambey were partially due
to the high incidence of cowpea aphids amd mosaic virus. The data did not
show a general suppression of partial LER in the early erect cultivars due to
competition with the spreading cultivar.
The data in parentheses in Table 7 demonstrated that the medium-cycle
spreading cultivar provided the major contribution to total grain yield in 1989
in all locations, whereas in 1988 the early erect cultivars contributed more to
VARIETAL INTERCROPPING OF COWPEA IN SEiNbGAL
229
TABLE 7
Partial LERs for grain, averaged across fert
er treatments, for each component cowpea cultivar in
the varietal intercrops during two years in e;
location
-
Location
Intercrop treatme:
58-57/Bambey 21
58-57/CB5
-
58-57
Bambey 2 1
58-57
CB5
Bambey
1988
0.45 (39)*
0.59
0.48 (35)
0.66
1989
0.71 (54)
0.30
0.73 (59)
0.54
Thilmakha
1988
0.66 (46)
0.82
0.60 (31)
0.87
1989
0.82 (57)
0.77
0.66 (63)
0.43
I--wP
1988
0.54 (58)
0.55
0.61 (48)
0.60
1989
0.62 (67)
0.48
0.75 (81)
0.58
‘Number in parentheses indicates the perce
contribution of the medium cycle spreading cultivar
(58-57) to the total yield of the intercrop.
TABLE 8
Partial LERs for hay, averaged across fertil
r treatments, for each component cowpea cultivar in
tbe varietal intercrops during two years in et
location
-
Location
Intercrop treatmet
58-57/Bambey 21
58-57/CB5
-
58-57
Bambey 2 1
58-57
CB5
-
Bambey
1988
0.34 (38)”
0.55
0.39 (35)
0.78
1989
0.66 (68)
0.50
0.76 (74)
0.80
Thilmakha
1988
0.93 (43)
0.87
0.74 (46)
0.80
1989
0.77 (57)
0.75
0.77 (73)
0.53
LQUga
1988
0.55 (42)
10.64
0.68 (38)
0.89
1989
0.69 (52)
0.84
0.85 (86)
0.22
-
*Number in parentheses indicates the perce
contribution of the medium cycle spreading cultivar
(58-57) to the total yield of the intercrop.
total grain yield in virtually ail ci
s. The data for hay (Table 8) exhibited
similar responses. The drought du og the vegetative stage at Louga (Fig. 2)
and Thilmakha in 1989 stressed tl early erect cultivars and could have been
responsible for the yield advantag lfthe later cultivar (58-57) in this year.
--
-.a-
-.
---
230
S . T H I A W E T A L .
Presumably, the early erect cultivars had completed podding by the time the
spreading cultivars encroached on their ,space. In contrast, the low yields of
58-57 in 1988 could have been due to mosaic virus infection. The erect culti-
vars have genetic resistance to this virus. This year-to-year variation in the
type of cultivar making the major contribution to total yield and in partial
LERs illustrates a mechanism whereby varietal intercrops could have en-
hanced yield stability compared with sole crops.
Stability and adaptation
Regression analyses were conducted on the mean yield of each genotype or
intercrop in a given environment against an environmental index consisting
of the mean yield of a11 treatments in thai environment. High r2 values (0.65
to 0.95 for grain, and 0.92 to 0.99 for hay) demonstrated that linear regres-
sion was appropriate. With respect to gratin yield, 58-57 and the two varietal
intercrops had the lowest regression coeflicients (Fig. 3 ) indicating that they
had the highest stability. In addition, the two varietal intercrops had the high-
est mean yields (Fig. 3 ) which indicates that they are well adapted to a11 en-
vironments but are most effective in the dry infertile locations. In contrast,
58-57 had low mean yield indicating ada.ptation only to the poorer environ-
ments. N ‘diambour had high mean grain yield and a regression coefficient of
1.1 indicating reasonable adaptation to a11 environments. In contras& TVX
3236 had a regression coefficient of 1.7’7 indicating adaptation only to the
more favorable environments.
The varietal intercrops and two spreading varieties, N’Diambour and
2 . 0
-l
T V X 3 2 3 6
A
6 2 1
A
AN”‘IAMBDUR
5%57/CB5
J
800
850
9 0 0
9 5 0
1 0 0 0
1 0 5 0
1 1 0 0
G E N O T Y P I C MIEAN G R A I N Y I E L D , KG/HA
Fig. 3. Coeffkient for the regressian of genotypic treatment grain yield with mean grain yield of
a11 treatments in one environment versus mean grain yield in a11 environments.
VARIETAL INTERCROPPING OF COWPEA IN SENkGAI.
231
2 . 0
1 .a
1 .6
2
; 1 . 4
T V X $ 2 3 6
N’DIANEOUF
:
k 1 . 2
A
MOUGNE
A
A
A
A58-57/cB5
0
CJ 1 . 0
58-57
%a-57/e 2 1
6
;tO.E
w
E 0 . 6
k?
0 . 4
0 . 2
0.0 l-
L
1100
1300
1 5 0 0
1700
1 9 0 0
21 10
G E N O T Y P I C MEAN H A Y Y I E L O . KG/HA
Fig. 4. Coeffkient for the regression of genotypic treatment hay yield with mean hay yield of a11
treatments in one environment versus mean hay yield in a11 environments.
Mougne had high mean hay yields and regression coefficients close to 1 .O (Fig.
4 ), indicating that they are well-adapted for hay production in a11 of the test
environments. The two early-erect ‘cultivars (Bambey 2 1 and CB5 ) generally
had low hay production, due to their small, erect, and determinate growth
habit.
No studies were discovered on the stability of varietal intercrops, but some
data are available for varietal mixtures. Stability of multiline mixtures of soy-
bean has been compared with pure litres. In one study, certain mixtures had
greater stability and higher yields than the pure lines (Schutz and Brim, 197 1).
The four cultivars used by Shutz and Brim ( 197 1) had substantial differences
in maturity date (29-day range), but only modest differences in morphology;
they were upright and determinate. In another study, there were no signitï-
tant differences in yield and regression coefficients between mixtures and pure
lines (Walker and Fehr, 1978). Walker and Fehr ( 1978) used many pure
lines but the maturity dates of the lines were within a lO-day range and the
morphologies of the lines were similar. For mixtures or intercrops to have
greater stability than pure lines gro,wn as sole crops, the component lines may
have to exhibit substantial differences in morphology and phenology.
CONCLUSIONS
Intercrops of early-erect and me 4mm-cycle spreading cowpea cultivars were
shown to be more effective than sole crops of these cultivars in dry locations
with infertile soils in Senegal. In these locations, fax-mers growing both types
---
-
-u
232
S . T H I A W E T A L .
of cultivars would have to plant 42 or 50% more land area of the sole crops to
obtain the same grain or hay yields as the varietal intercrops. Farmers in these
locations who are seeking the highest yielding or most stable cropping system
could also benefit from varietal intercrops. The intercrops had higher mean
grain and hay yields than a.ny of the cultivars tested as sole crops in the dry
locations with infertile soils, and above average yield stability across a11 loca-
tions. The varietal intercrop is compatible with mechanical sowing and cul-
tivation, and, in the drier zone where plants are small, it is possible to sepa-
rately harvest the individual cultivars, an important consideration for seed
quality and marketing. Further studies are needed to determine the impact of
varietal intercrops on problems due to insect pests, and to guide the breeding
of early erect and medium-cycle spreading cultivars that are most compatible
in varietal intercropping systems.
ACKNOWLEDGMENTS
This research was partially supported ‘by the Bean/Cowpea Collaborative
Reseach Support Program of the United States Agency for International De-
velopment Grant No. DAN- 13 l O-SS-6008-00. The opinions and recommen-
dations are those of the authors and not necessarilly those of USAID.
REFERENCES
Agboola, A.A., 1978. Influence of soi1 organic matter on cowpea’s response to N fertilizer. Agron.
J., 70: 25-28.
Berg, M.G. and Gardner, H., 1978. Methods of soi1 analysis in the laboratory at Oregon State
University. Special Report 32 1, Revised Sept. 1978. Agric. Exp. Station, Oregon State Uni-
versity, Corvallis.
Blum, A., 1988. Plant Breeding for Stress Environments. CRC Press, Boca Raton, FL.
Dancette, C. and Hall, A.E., 1979. Agroclimatology applied to water management in the Sudan-
ian and Sahelian zones of Africa. In: A.E. Hall, G.H. Canne11 and H.W. Lawton (Editors),
Agriculture in Semi-Arid Environments. Ecological Studies 34. Springer, Heidelberg, pp.
98-l 18.
Eberhart, S.A. and Russell, W.A., 1966. Stability analysis in plant breeding. Plant Breeding,
101: l-23.
Finlay, K.W. and Wilkinson, G.N., 1963. The analysis of adaptation in a plant-breeding pro-
gramme. Aust. J. Agric. Res., 14: 742-754.
Fussell, L.K. and Serafini, P.G., 1985. Crop association in the semiarid West Africa. Research
strategies, past and future. In: W. Ohm and J.C. Nagy (Editors), Appropriate Technology
for Farmers in the Semiarid West Africa. TJniversity of Purdue, 359 pp.
Hall, A.E., 1988. Stabilizing dryland production of cowpeas by using varietal intercrops. In:
P.W. Unger, T.V. Sneed, W.R. Jordan and R. Jensen (Editors), Challenges in Dryland Ag-
riculture: A Global Perspective. Texas Agricultu:ral Experiment Station, College Station, TX,
pp. 744-746.
Hall, A.E. and Dancette, C., 1978.. Analysis of fallow-farming systems in semiarid Africa using
a mode1 to simulate the hydrologie budget. Agraln. J., 70: 8 16-823.
VARIETAL INTERCROPPING OF COWPEA IN SENEGAL
233
Hall, A.E. and Patel, P.N., 1985. Breeding for resistance to drought and heat. In: S.R. Singh and
K.O. Rachie (Editors), Cowpea Research, Production and Utilization. John Wiley, New
York, NY, pp. 137-151.
Hall, A.E., Foster, KW. and Waines, J.G,, 1979. Crop adaptation to semi-arid environments.
In: A.E. Hall, G.H. Canne11 and H.W. Lawton (Editors), Agriculture in Semi-Arid Environ-
ments. Ecological Studies 34. Springer, Heidelberg, pp. 148-179.
Hiebsch, C.K. and McCollum, R.E., 1987, AreaxTime Equivalency Ratio: A method for eval-
uating the productivity of intercrops. Agron. J., 79: 15-22.
Khalfaoui, J.-L.B., 199 1. Determination ofpot.ential lengths of the trop growing period in semi-
arid regions of Senegal. Agric. Forest Meteorol., 55: 25 l-263.
Nageswara Rao, R.C., Wadia, K.D.R. and Williams, J.H., 1990. Intercropping short and long
duration groundnut (Arachis hypogeu) genotypes to increase productivity in environments
prone to end-of-season droughts. Exp. Agric., 26: 63-72.
Ng, N.Q. and Marechal, R., 1985. Cowpea taxonomy, origin and germplasm. In: S.R. Singh and
K.O. Rachie (Editors), Cowpea Research, Production and Utilization. John Wiley, New
York, NY, pp. 11-21.
Nicou, R. and Poulain, J.F., 1968. La fumure minerale du niebe au Senegal. In: Colloque Sur
La Fertilité des Sols Tropicaux, Tananarive, Madagascar, 19-25 November 1967. Agron-
omie Tropicale, Supplement to Nos. 2 and 5: 73 l-754.
Rattunde H.F., Ramraj, V.M., Williams, J.H. and Gibbons, R.W., 1988. Cultivar mixtures: a
means of exploiting morpho-developmenta.1 differences among cultivated groundnuts. Field
Crops Res., 19: 201-210.
Reddy, K.C., Mahamadou, 1. and Van der Ploeg, J., 1988. Appropriate cropping systems for the
Nigerian semiarid tropics. In: P.W. Unger, T.V. Sneed, W.R. Jordan and R. Jensen (Edi-
tors), Challenges in Dryland Agriculture - A Global Perspective. Texas Agricultural Exper-
iment Station, College Station, TX, pp. 767-769.
Shutz, W.M. and Brim, C.A., 1971. Intergenotypic competition in soybean. III. An evaluation
of stability in multiline mixtures. Crop Sci., 11: 684-689.
Singh, S.R., Jackai, L.E.N., dos Santos, J.H.R. and Abdalla, C.B., 1990. Insect pests of cowpea.
In: S.R. Sir@ (Editor), Insect Pests of Food Legumes. John Wiley, New York, NY, pp. 45-
89.
Thiaw, S.. 1990. Stability of dryland production of cowpeas ( Signa unguicuhtu (L. ) Walp. )
with varietal intercrops. MS. Thesis, University of California, Riverside, 57 pp.
Walker, A.K. and Fehr, W.R., 1978. Yield stability of soybean mixtures and multiple pure stands.
Crop Sci., 18: 7 19-723.
Ziska, L.H. and Hall, A.E., 1983. Seed yield and water use of cowpea subjected to planned-
water-deficit irrigation. Irrig. Sci., 3: 237-245.
217
Elsevier Science Publishers B.V., Amsterd m
Varietal intercropping
the yields and stability
of cowpea
semiarid Senegal
and D.R. Parkerb
ABSTRACT
l
Thiaw, S., Hall, A.E. and Parker, D.R.,
Varietal intercropping and the yields and stability of
cowpea production in semiarid
17-233.
Crop productivity is low and variable under rainfed conditions in semiarid zones. This problem is
particularly acute in northem Senegal where p oductivity also is limited by infertile soils. Studies were
conducted to determine whether varietal intercrops of cowpea are better adapted to semiarid climates
and infertile soils than sole crops of cowpea. The same experiment was conducted over two years in
three locations which have contrastmg rainfall and soi1 fertility, and are representative of the major
cowpea-production zones of northem Senega . Two varietal intercrops and six cultivars in sole crops
were grown in a randomized complete block split-plot design with four replications. The split plots
consisted of a starter fertilizer treatment ( 150 kg/ha of 6:9:8, N:P:K) and a control where no fertilizer
was applied. The varietal intercrops consiste< of altemating rows of a medium-cycle spreading culti-
var (58-57) and an early erect cultivar (Bambey 21 or Califomia Blackeye No. 5). The six cultivars
grown as sole crops included the three cultivars used in the intercrops, and three others with spreading
(N’diambour and Mougne) or semierect (TV 3236) growth habits, that are well adapted to northem
Senegal.
Grain and hay yield responded significantly to fertilizer application at the two drier locations, Thil-
makha and Louga, but not at the wetter location Bambey, which had higher levels of soi1 nitrogen
and phosphorus. There were no genotype x fertilizer interactions. The varietal intercrops produced
higher yields of grain and hay than the sole crops at Thilmakha and Louga and had high land equiva-
lency ratios ( 1.42 for grain and 1.50 for hay), and above average stability. In the wetter and more
fertile location (Bambey), the performance of the varietal intercrops was intermediate compared to
the sole crops, and the dense canopy made it diffrcult ta harvest the grain of the two cultivars in the
intercrops separately. These studies demonarated that intercrops of early erect and medium-cycle
spreading cowpeas cari have higher and more stable yields of grain and hay than sole crops, in condi-
tions where drought and infertile soi1 limit trop production.
Correspondence
to: Dr. A.E. Hall, Department of Botany and Plant Sciences, University of Cal-
ifomia, Riverside, CA 9252 1, USA.
0378-4290/93/$06.00 0 1993 Elsevier S4rente Publishers B.V. Al1 rights reserved.
218
S . T H I A W E T A L .
INTRODUCTION
In semiarid zones trop yields are generally low and variable as a conse-
quence of limited and variable rainfall and other constraints. In the semiarid
region of Senegal, the average rainfall provides sufficient moisture for a 60-
to 1 OO-day cropping season which is followed by a long dry season (Dancette
and Hall, 1979). From year to year, there is substantial variation in useful
length of the rainy season (Khalfaoui, 199 1 ), and drought cari occur at dif-
ferent times during the cropping season, substantially reducing trop
production.
Cowpea ( Signa unguicuYata (L.) Walp.) is a warm-season trop that has
been grown in Senegal for thousands of years (Ng and Marechal, 1985 ) to
produce grain and hay. Formerly, late-klowering spreading varieties were
planted in home gardens or intercropped with cereals or other crops. During
the past thirty years, new cultivars have been developed with early or medium
maturity, and erect or spreading growth habit. More than 90% of the area
planted to cowpea at this time is sole-cropped.
Since the droughts began in 1968, cowpea has become an important trop
in the northern peanut basin of Senegal due to its ability to produce food
under conditions of drought and infertile soils, where the other staple crops
(Pearl millet, sorghum and peanut) have produced virtually no food in some
years. For subsistence farmers growing crops under conditions of .variable
drought and other soi1 and biotic constraints, yield stability may be more im-
portant than high yield. Cropping systems of semiarid West Africa were re-
viewed by Fusse1 and Serafïni ( 1985) who concluded that Pearl millet/cow-
pea intercrops could enhance and stabilize yields compared with sole crops.
Pearl millet/cowpea intercrops have been compared with sole crops in the
semiarid zone of Senegal over a 5-year period (C. Dancette, unpublished
studies). These studies indicated that sole crops were more effective than in-
tercrops in the drier parts of the zone. Reddy et al. ( 1988) reached similar
conclusions from studies conducted in a similar ecological zone in Niger. Mil-
let/cowpea intercrops may have an advantage over sole crops of cowpea where
flower thrips (A4egaZurothrip.s sjostedii) are abundant and pesticides are not
used (Singh et al., 1990). The levels of flower thrips, however, are usually
low in northern Senegal.
Varietal intercrops, combining morphologically and phenologically con-
trasting cultivars of cowpea, could enhance yield stability and contribute to
food requirements (Hall, 1988). In northem Senegal, farmers usually expe-
rience food shortages just before the Pearl millet and late- or medium-cycle
cowpea cultivars are ready to harvest. Cowpea cultivars with early maturity
cari provide food and cash during the traditional period of hunger, as well as
greater grain yield than later-maturing cultivars when the rainy season is very
short (Hall and Patel, 1985 ) . In contras& medium-cycle, spreading cultivars
VARIETAL INTERCROPPING OF COWPEA IN SEN$GAL
219
cari produce more grain and hay t
the early erect cultivais when the rainy
season is of adequate duration or
ought occurs in the middle of the crop-
ping season. On-faim studies in S
a1 have demonstrated that farmers pre-
fer to grow both types of cowpea
e early erect cultivars and the medium-
cycle spreading ones (M. N ‘diaye
d A. Bal, unpublished studies). Possible
advantages from varietal intercr
consisting of alternating rows of these
two types of cultivars may be
he following examples. Where drought
occurs early in the season, th
ivar may senesce, but if rains resume,
a late-maturing cultivar grown
rietal intercrop could compensate by
spreading into the space made
by the senescence of the early culti-
var. In contrast, when the rainy
continuous but short, only the early
cultivar would produce grain but
ater, spreading cultivar would produce
substantial hay.
Preliminary studies in Se
iouf, and M. Diagne, unpublished
studies) demonstrated that
crops of cowpea cari be more effec-
tive than sole crops. In the studies
. Diagne varietal intercrops, consisting
of alternating rows of an early er
cultivar and a medium-cycle spreading
cultivai-, were compared with sole rops of the same cultivars. He evaluated
land equivalency ratio (LER), whi h is the relative land area under sole crops
required to produce the same yiel as an intercrop. In the first year, the LER
for grain yield was only slightly gre ter than unity ( 1.08 ) . In the second year,
the varietal intercrop made much
ore eflkient use of the land than the sole
crops with an LER for grain of 1.11. In addition, the varietal intercrop pro-
duced 20% more grain than the hi est,-yielding sole trop, and 58% more grain
and 40% more hay than the aver ge yields of the sole crops. These studies
were a11 conducted at Bambey in t e wetter part of the semiarid zone of Sen-
egal. It is particularly important t evaluate varietal intercrops over a range
of environments, especially ones 1
w’th more extreme and more variable levels
of drought.
and when leguminous crops were
in soi1 with low supplies of nitrogen.
These conclusions would apply
a cowpea varietal intercrop of early and
the infertile soils in the drier part of the
studies at Louga demonstrated
that in the northern
2 2 0
S. THIAW ET AL.
Yield response to variation in environment due to years and locations is
also an important criterion for assessing the adaptation of cropping systems
or cultivars to semiarid environments. Long-term average yield and yield sta-
bility are important indices of adaptation (Hall et al., 1979). Several meth-
ods are available to evaluate yield stability (Blum, 1988 ) , but they cari make
different predictions concelming the relal.ive stability of the same set of culti-
vars (Hall et al., 1979). A widely used method to evaluate stability consists
of the regression of the yield of individual cultivars against an environmental
index consisting of the mean yield of a11 cultivars in the tria1 using data from
different locations and years (Finlay and Wilkinson, 1963 ) . Linear regres-
sion, as used by Eberhart and Russell ( 1966) would, however, appear to be
more appropriate than the logarithmic regression used by Finlay and Wilk-
inson ( 1963 ), because it provides equal weighting to the different environ-
ments. The final step consists of plotting the regression coefficient against the
overall mean yield of the cultivars (Finlay and Wilkinson, 1963). Cultivars
with high mean yields and regression coefficients less than or close to unity
would be considered to be broadly adapted to a11 environments experienced
in the trials, whereas cultivars with very low regression coefficients would
tend to be adapted to the test environments with low yield potential (Blum,
1988).
The main objective of tlnis study was to compare the average yield, yield
stability and efflciency of land use of varietal intercrops and sole crops of
cowpea grown in several locations with contrasting levels of drought and soi1
fertility.
MATERIALS AND METHODS
During the summers of 1988 and 1989, the same experiment was con-
ducted in three semiarid locations: Bambey ( 14”42’N 16”28’W ), Thil-
makha (15”2’N 16” lS’W), and Louga (15”36’N 16”13’W). These loca-
tions have contrasting rainfall and soi1 fertility, and are representative of the
major cowpea-production zones of Senegal. At Bambey the experiment was
conducted on a deep, slightly leached, tropical ferruginous soil, called a “Dior”
soil. Based on the International Soi1 Science Classification, this soi1 has 7%
clay, 3% silt, 68% fine sandl, 22% coarse Sand, and little variation in texture
in the first two meters. The volumetric moisture content at field capacity is
16 + 2%. Using the United States Department of Agriculture soi1 taxonomy,
this soi1 is classifïed as an Wstipsamment (Hall and Dancette, 1978). Thil-
makha and Louga are located in the northern part of the peanut basin and the
soils are sandy with a low percentage of clay (3%) and a low field capacity
(8%). The pHs of the soils are about 5.5 at Louga and Thilmakha, and 7 at
Bambey.
VARIETAL INTERCROPPING OF COWPEA IN SENbAL
221
Cultural practices
Fields were chosen where pearl lmillet had been grown the previous year
and they were plowed during the dry season in May. Al1 trials were hand
planted when the soi1 was suffrcien ly wet from rainfall to insure germination
and plant establishment, which oc urred in early August in 1988 and early
July in 1989. Two seeds of the six wpea cultivars were planted per hill, and
hills with zero emergence were rep anted 1 week later. After seedling emerg-
ence, plants were thinned to one pe hi11 in a11 locations. Weeds were removed
about 15 days after emergence, w en the cowpea seedlings were well estab-
1
lished. The plots were hand weede three times during the cropping season.
Chemical control of insects was maintained in a11 three locations. Hairy
caterpillar (Amsacta moloneyi), a d cowpea aphid (Aphis craccivora) were
controlled by Thimul 35 applied a the rate of 2.5 1 per ha (the active ingre-
dient is endosulfan at 800 g/ha), a d flower thrip (Megalurothrips sjostedti)
was controlled with Decis applied t the rate of 25 g per ha (the active ingre-
dient is deltamethrine at 15 g/ha) :
.
Experimental design and treatmqbzts
Six cowpea cultivars (Table 1) nd two varietal intercrops were grown as
d
eight genotype treatments in a ran omized complete block split-plot design
with four replications. The genotyp treatments were the main plots, and there
\\
were two fertilizer levels as sub-plo s . Each main plot consisted of 12 rows, 5
m long, and each sub-plot had 6 rows. The distance between rows was 50 cm
for a11 cultivars. Within the rows, t e distance between hills was 25 cm for the
erect cultivars, and 50 cm for the s rea.ding and semi-erect cultivars. The two
intercrop treatments consisted of
early erect cultivar, Bambey 2 1 or Cali-
fornia Blackeye No. 5 (CB5 ), and medium-maturity spreading cultivar, 58-
57, sown in alternate rows. The
r central rows of each sub-plot were har-
vested at the end of the rainy
to determine dry weight yields of grain
and hay.
TABLE 1
~
Origin and characteristics of the cowpea cultijars used
Cultivars
Period from
Growth habit
sowing to 50%
flowering
(days)
58-57
44
spreading
N’diambour
44
spreading
Mougne
46
spreading
Tvx 3236
45
semi-erect
Bambey 2 1
41
erect
Califomia Blackeye 5
38
erect
222
S. THIAW ET AL.
Prior to applying fertilizer, soi1 samples were taken in the O-20 cm depth at
Bambey, Louga, and Thilmakha. The concentrations of total N, P, and K were
determined. Total N was determined by the Kjeldahl method. Soils were di-
gested in a mixture of HF, lHC104, HC1 and HN03, and P concentration was
determined using a molybdovanadate colorimetric method, and K concentra-
tion was determined by atomic absorption spectrophotometry (AAS). One
of the split plots received ferrtilizer, which was broadcast and incorporated at
the rate of 150 kg/ha of 6:9:8 (N:P:K), which is a low level presently recom-
mended as a general practice to farmers in Senegal by the extension service.
This fertilizer also contains a small amount of sulfur in that the nitrogen is
provided by ammonium sulfate.
Fully expanded mature leaves were sampled at mid-bloom (between 35
and 45 days after sowing) from the cultivars 58-57, Mougne, Bambey 2 1 and
CB5 in both control and fertilized plots. Leaf blades were ground and passed
through a 0.85-mm sieve and digested in a mixture of nitric and perchloric
acids. The K and micronutrient (Fe, CU, Zn and Mn) concentrations in the
leaves were determined by AAS. The content of P in the leaf blades was mea-
sured using a modilïcation of the vanadate yellow method of Berg and Gard-
ner ( 1978). Total N concentration was determined using a carbon-nitrogen
analyzer (NA- 1500 Carlo-Erba, Milan, Italy ) . The number of peduncles, pods
per plant and pods per peduncle were determined for the cultivars 58-57,
Bambey 2 1 and CB5. Weather stations at Bambey and Louga provided data
on rainfall, daily pan evaporation (US Weather Bureau Class A), and the
daily maximum and minimum shelter air temperatures. At Thilmakha, only
rainfall was measured.
RESULTS AND DISCUSSION
Weather conditions
In 1988 and 1989, annual rains were 6 39 and 805 mm at Bambey, 409 and
550 mm at Thilmakha, and 442 and 470 :mm at Louga, respectively. Compar-
ison of rainfall and pan evaporation data indicated that the rainfall in 1988
would have supported a 70,.day growing season at Bambey, but only a 60-day
growing season at Louga (Fig. 1). A drought occurred at Louga during the
fourth 1 0-day interval whem the cowpeas were in the early flowering stage. In
1989 there was suftïcient ralin to support a 90-day growing season at Bambey,
but only a 70-day growing season at Louga where the sandier soi1 has limited
ability to store water in the root zone (Fig. 2). An extreme drought occurred
at Louga during the second. and third 1 0-day intervals after sowing when the
cowpeas were in the late vegetative stage. The rainfall at Thilmakha (Thiaw,
1990) was more similar to the rainfall at Louga than at Bambey during both
years. The daily mean pan evaporation was high and averaged about 6 mm/
day at Bambey and Louga in both years. Daily maximum and minimum air
VARIETAL INTERCROPPING OF COWPEA IN SEI’
223
1988
25
BAMEEY
_J
0 R A I N
I E T P
!
LOUGA
dJl.ll
I
i
1 2 3 1
ilb 4 5 6 7 8 9
TEN-DAY INTERVALS FROM SOWING
Fig. 1. Mean rainfall and pan evaporatio for IO-day intervals after sowing at Bambey and
Louga in 1988.
temperatures also were similar at Yambey and Louga in both years with av-
erage values of 34 and 24”C, respe :ively.
Response to fertilizer
Soi1 analyses conducted in 198!
prior to applying the fertilizer, showed
that total N and P were lower at Tl lmakha and Louga than at Bambey (Ta-
ble 2). Other studies at these same ocations gave available soi1 P ( NaHC03
extract) as being lower at Thilma’ la than at Bambey. The level of total K
was similar in the three locations. {a~ signifkant differences were observed
between the cultivars, and the fer-t: zed and control treatments in leaf blade
minera1 content, and average value! u-e presented (Table 3 ) . Differences were
observed between locations in a11 rr nerals in leaf blades except P. The K level
was lower at Thilmakha than at t : two other locations and probably was
defïcient. Among the micro-nutri us, Mn was higher at Louga and Thil-
makha, possibly reflecting the lowc soi1 pH as compared to Bambey. There
was a trend towards higher N levi 5 in leaves from Bambey than the other
locations, which is consistent with he higher total N in the soi1 at Bambey.
Cowpea has been shown to tïx con! clerrable atmospheric nitrogen in a11 three
locations (M. N’Diaye, unpublishc I data). The levels of P in the leaf blades
-.--,*
224
S. THIAW ET AL.
2 5 -
BAMBEY
2
2
z- 2 0 -
r 19690 RAINI ETP
2’ 6
zv
* E’5 -
CE
w
l
1 0 -
-
5 - I
I
3 0 -
LOUGA
2 5 -
9
2
Z-20 -
6’ z
z- 1 5 -
*E
EE
w
10 -
5-
O -
1
2
3
4
5
6
7
6
9
TEN-DAY INTERVALS FROM SOWING
Fig. 2. Mean rainfall and pan evaporation for ICI-day intervals after sowing at Bambey and
Louga in 1989.
TABLE 2
Soi1 minera1 nutrient concentrations in the O-20 cm depth at three locations in Senegal
Minera1
Location
content
Bambey
Thilmakha
Louga
total N @/kg)
0.33
0.18
0.19
cv C%l
8
25
10
total P (g/kg)
0.14
0.10
0.08
cv (%)
4
19
16
Total K (meq/lOO g)
3.17
3.06
3.11
cv (%)
6
16
9
CV =Coefficient of variation.
were intermediate and not significantly dlifferent between the three locations
despite the higher P content in the soi1 at Bambey.
The fertilizer treatment resulted in significantly higher yields of grain and
hay than in control plots at Thilmakha and Louga but not at Bambey (Table
4). The yield responses at Thilmakha antd Louga may have been due to the
small amount of N (9 kgjha) in the starter fertilizer. Studies of Agboola
VARIETAL INTERCROPPING OF COWPEA 1N SEN/EGAL
225
TABLE 3
Leaf blade minera1 nutrient concentrations (dry weight basis) of cowpea in Senegal in 1989, averaged
across four cultivars and the control and fertijized treatments
Location
M n
F e
Zni
C U
N
K
P
WW
WW
(QW
WMg)
WW
ww
WW
Bambey
281
145
28~
12
55.1
,13.1
3.2
Louga
488
260
30 ;
9
41.1
12.0
3.3
Thiimakha
417
134
24 ~
31.1
2.7
LSD (0.05)
70
31
21
8 f6.7
*p:i
i;
signif.
ft
Q
*, Y3ignifïcant at Pc 0.08 and < 0.01, respectively.
TABLE 4
Grain and hay yields of cowpea averaged over genotypic treatments for control and fertilized plots at
Bambey, Thilmakha and Louga
Bambey
l
Thilmakba
Louga
-i
1988
19891
1988
1989
1988
1989
Grain (kg/ha)
Control
1394
1334
671
599
616
501
Fertilized
1438
1399~
930
933
875
575
LSD (0.05)
ns
ns ~
67
53
60
54
Hay (Wha)
l
Control
2424
4766
530
843
461
847
Fertilized
2437
4642
184
1147
158
956
LSD (0.05)
ns
n s
67
12
78
101
( 1978) in Nigeria demonstrated ihat yield of cowpea cari respond to appli-
cation of 10 kg/ha of N when the #erc:entage of soi1 organic matter is less than
1 .O%, but not with 2% or greater organic matter. The soi1 organic matter was
estimated to be less than 1% at Bambey, Louga and Thilmakha. The genotype
treatment xfertilizer interaction was not significant in either year at any lo-
cation for either grain or hay yield. This indicates that trials to evaluate cul-
tivars or varietal intercrops should give similar genotypic or intercrop rank-
ings either with or without starter;fertilizer. In the subsequent analyses yields
are examined which represent thk average across the control and fertilized
treatments.
Overall grain and hay yields we’ le much higher at Bambey during 1988 and
1989 than at Thilmakha and Lo ga (Table 4). Shoot biomass (grain plus
hay) was positively correlated ith seasonal rainfall (r2 ~0.93, P=O.O02)
indicating that the higher yields t Bambey were associated with wetter con-
ditions. The higher soi1 fertility
Bambey was an additional factor contrib-
226
S. THIAW ET AL.
uting to yield. Despite the higher rainfall in 1989, average grain yîeld was
higher in 1988 in a11 locations, possibly due to a more uniform distribution of
the rain in 1988 (compare Figs. 1 and 2 ). Disease and insect pressures also
differed among the experiments. Plots at Bambey suffered a severe aphid in-
festation coupled wîth mosaic virus infection of cultivars 58-57 and
N ‘diambour in 1988, while at Louga and Thîlmakha a11 genotype treatments
suffered from hairy caterpillar in 1989.
Genotypic yield response
The genotypexyear interaction was highly signilïcant at Bambey for both
grain and hay but not at Thilmakha and Louga. Also, there were no significant
genotype x location interactions for Thilmakha and Louga, and mean values
across these two locations are presented (Table 5 ) . The genotypic treatments
differed significantly in grain production except at Thilmakha and Louga in
1989 (Thiaw, 1990). In both years, highest mean grain yields at Bambey were
achieved by sole crops of three cultivars with different growth habits, semi-
erect Tvx 3236, erect Bambey 2 1 and spreading N’diambour, and the varietal
intercrops had întermediate: yields (Table 5 ). In contrast, the varietal înter-
crops had the hîghest grain yields at Thilmakha and Louga (Table 5 ) .
The genotypic treatments differed signilïcantly in hay production except at
Louga in 1988 (Thiaw, 1990). At Bambey highest mean hay yield was
achieved by a sole trop of the spreading cultivar N’diambour, which also had
a high grain yield (Table 5 ) . The hay yields of the varietal intercrops were
TABLE 5
Grain and hay yields (kg/ha) avera8ed across fertilizer treatments and years for cowpea at Bambey
and the combined averages for Thilmakha plus Louga
Genotypic treatment
Bambey
Thilmakha and Louga
-
Grain
Hay
Grain
Hv
58-57
1036
3674
758
766
N’diambour
1468
4388
738
908
Mougne
1349
3982
639
752
Tvx 3236
1828
3998
510
495
Bambey 2 1
1600
2842
596
740
CB5
1230
2276
640
615
58-57/Bambey
21
12159
3500
896
1100
SS-57/CB5
1387
3872
917
952
Mean
1391
3567
712
787
LSD (0.05)
233
818
130
122
gen. X year
f*
tt
ns
ns
gen. X locat.
-
ns
ns
LSD tests the differences between the different treatments. ** = signitïcant at Pc 0.0 1.
VARIETAL INTERCROPPING OF COWPEA IN SENIGAL
227
intermediate at Bambey (Table 5 ) . ~Mean hay yields for Thilmakha and Louga
combined (Table 5) were highest for the varietal intercrops and low for the
erect cultivars (Bambey 2 1 and CB5 ) . These data demonstrated that under
water-limiting conditions and infertile soils, such as at Thilmakha and Louga,
varietal intercrops cari produce higher yields of grain and hay than sole crops
of either erect or spreading cultivars. A,t Bambey, with higher rainfall and soi1
fertility, intercrops were not as productive as the best sole crops and would be
difficult to manage due to the dense canopies that developed.
Yield component data for three ~contrasting genotypes, 58-57, Bambey 2 1
and CB5, in the three locations demonstrated that the higher yields at Bam-
bey were associated with a greater number of peduncles and pods (Thiaw,
1990). There was little difference~ in number of pods per peduncle among
locations. Other studies with cowpea under different irrigation treatments have
also shown a strong positive associltion between grain yield and the number
of pods per plant (Ziska and Hall, ~198#3).
Few studies were discovered in the literature on the performance of varietal
intercrops. Rattunde et al. ( 1988 ) evaluated morphologically and phenolog-
ically différent cultivars of peanut as mixtures or as intercrops with pairs of
varieties sown in altemating rows. They found that the highest pod yields of
the mixtures and row intercrops failed to exceed the yield of the best cultivars.
They concluded that future experimentation with peanut mixtures in the
tropics should focus on stability rather than maximization of yield. Nages-
wara Rao et al. ( 1990) evaluated the performance of intercropping short- and
long-duration peanut genotypes in environments subjected to end-of-season
droughts. The intercrop treatments resulted in LERs of 0.93 to 1.25 with a
mean of 1.1 for pod yield and 0.99 to 1.15 with a mean of 1.1 for total biomass
at the end of the season. They concluded that intercropping peanut genotypes
with different season length reqmrements is a better solution to variable sea-
son length than simply spreading the risk by growing a range of genotypes as
sole crops. However, the LER values were close to 1 .O indicating only small
improvements in land-use efIiciency, and mechanizing the harvest of peanuts
would be difficult with varietal intercrops.
Land-use and biological efficiencies
Mean LER values showed that the varietal intercrops of cowpea had aver-
age land-use efficiencies for grain and hay that were 55% and 57% higher,
respectively, than the sole crops at Thilmakha, and 30% and 43% higher, re-
spectively, for grain and hay at Louga (Table 6). The average advantage of
the varietal intercrops was smaller t Bambey with + 15% for grain and + 22%
for hay. The extent to which the g!eater droughts and lower soi1 fertility con-
tributed to the greater LERs at Thilmakha and Louga, compared with Bam-
bey, is not known. There was no difference in the LERs of fertilized compared
with control plots at Thilmakha or Louga (data not shown). In the studies
---.
--A-
-t.
228
S. THIAW ET AL.
TABLE 6
Values for LER and ATER averaged across fertilizer treatments and years for grain and hay yield of
cowpea at Bambey, Thilmakha and Louga
Location
Intercrop
LER
ATER
Grain
Hw
Grain
Hw
Bambey
58-57/B 21
1.05
1.07
0.97
0.98
SS-57/CB5
1.25
1.37
1.10
1.15
Thilmakha
58-57/B21
1.64
1.67
1.47
1.57
SS-57/CB5
1.38
1.50
1.14
1.32
LWP
58-57/B21
1.12
1.47
1.06
1.32
SS-57/CB5
1.48
1.40
1.21
1.24
LER= land equivalency ratio. ATER = area X time equivalency ratio.
reviewed by Hiebsch and McCollum ( 1987 ) there was a general tendency for
lower LERs under more fertile soi1 conditions, but more fertilizer was applied
than in the present studies. They also showed that intercrops grown under low
soi1 N had ATER values greater than unity, whereas intercrops under high
soi1 N had average ATER values close to unity (Hiebsch and McCollum,
1987). In this study, the ATER values showed that varietal intercrops cari
have higher biological effici.ency than sole crops of the same cultivars (Table
6). At Thilmakha and Louga, the biological effkiencies of the intercrops for
grain (average ATER of 1.22 ) and hay (average ATER of 1.36) were sub-
stantially higher than the sole crops. In contrast, average ATER values at
Bambey for grain and hay were close to 1 .O indicating little difference in bio-
logical effkiency between the sole crops and intercrops. The magnitude of the
partial LERs show which cultivar made the major contribution to the total
LERs. In 1988, the partial LERs for grain (Table 7 ) were higher for the early
erect cultivars (Bambey 2 1 and CB5 ) at Bambey and Thilmakha, whereas in
1989, the medium-cycle cultivar had the highest partial LERs in ail locations.
In 19 out of 24 cases, the partial LERs were greater than 0.5, indicating con-
tributions from both cultivars to the higher LERs of the varietal intercrops.
Partial LERs for hay (Table 8) indicated that in most cases both cultivars
contributed to the high LERs of the varietal intercrops at Thilmakha and
Louga in both 1988 and 1989. The low partial LERs for grain and hay during
1988 for the medium-cycle spreading cultivar at Bambey were partially due
to the high incidence of cowpea aphids amd mosaic virus. The data did not
show a general suppression of partial LER in the early erect cultivars due to
competition with the spreading cultivar.
The data in parentheses in Table 7 demonstrated that the medium-cycle
spreading cultivar provided the major contribution to total grain yield in 1989
in all locations, whereas in 1988 the early erect cultivars contributed more to
VARIETAL INTERCROPPING OF COWPEA IN SEiNbGAL
229
TABLE 7
Partial LERs for grain, averaged across fert
er treatments, for each component cowpea cultivar in
the varietal intercrops during two years in e;
location
-
Location
Intercrop treatme:
58-57/Bambey 21
58-57/CB5
-
58-57
Bambey 2 1
58-57
CB5
Bambey
1988
0.45 (39)*
0.59
0.48 (35)
0.66
1989
0.71 (54)
0.30
0.73 (59)
0.54
Thilmakha
1988
0.66 (46)
0.82
0.60 (31)
0.87
1989
0.82 (57)
0.77
0.66 (63)
0.43
I--wP
1988
0.54 (58)
0.55
0.61 (48)
0.60
1989
0.62 (67)
0.48
0.75 (81)
0.58
‘Number in parentheses indicates the perce
contribution of the medium cycle spreading cultivar
(58-57) to the total yield of the intercrop.
TABLE 8
Partial LERs for hay, averaged across fertil
r treatments, for each component cowpea cultivar in
tbe varietal intercrops during two years in et
location
-
Location
Intercrop treatmet
58-57/Bambey 21
58-57/CB5
-
58-57
Bambey 2 1
58-57
CB5
-
Bambey
1988
0.34 (38)”
0.55
0.39 (35)
0.78
1989
0.66 (68)
0.50
0.76 (74)
0.80
Thilmakha
1988
0.93 (43)
0.87
0.74 (46)
0.80
1989
0.77 (57)
0.75
0.77 (73)
0.53
LQUga
1988
0.55 (42)
10.64
0.68 (38)
0.89
1989
0.69 (52)
0.84
0.85 (86)
0.22
-
*Number in parentheses indicates the perce
contribution of the medium cycle spreading cultivar
(58-57) to the total yield of the intercrop.
total grain yield in virtually ail ci
s. The data for hay (Table 8) exhibited
similar responses. The drought du og the vegetative stage at Louga (Fig. 2)
and Thilmakha in 1989 stressed tl early erect cultivars and could have been
responsible for the yield advantag lfthe later cultivar (58-57) in this year.
--
-.a-
-.
---
230
S . T H I A W E T A L .
Presumably, the early erect cultivars had completed podding by the time the
spreading cultivars encroached on their ,space. In contrast, the low yields of
58-57 in 1988 could have been due to mosaic virus infection. The erect culti-
vars have genetic resistance to this virus. This year-to-year variation in the
type of cultivar making the major contribution to total yield and in partial
LERs illustrates a mechanism whereby varietal intercrops could have en-
hanced yield stability compared with sole crops.
Stability and adaptation
Regression analyses were conducted on the mean yield of each genotype or
intercrop in a given environment against an environmental index consisting
of the mean yield of a11 treatments in thai environment. High r2 values (0.65
to 0.95 for grain, and 0.92 to 0.99 for hay) demonstrated that linear regres-
sion was appropriate. With respect to gratin yield, 58-57 and the two varietal
intercrops had the lowest regression coeflicients (Fig. 3 ) indicating that they
had the highest stability. In addition, the two varietal intercrops had the high-
est mean yields (Fig. 3 ) which indicates that they are well adapted to a11 en-
vironments but are most effective in the dry infertile locations. In contrast,
58-57 had low mean yield indicating ada.ptation only to the poorer environ-
ments. N ‘diambour had high mean grain yield and a regression coefficient of
1.1 indicating reasonable adaptation to a11 environments. In contras& TVX
3236 had a regression coefficient of 1.7’7 indicating adaptation only to the
more favorable environments.
The varietal intercrops and two spreading varieties, N’Diambour and
2 . 0
-l
T V X 3 2 3 6
A
6 2 1
A
AN”‘IAMBDUR
5%57/CB5
J
800
850
9 0 0
9 5 0
1 0 0 0
1 0 5 0
1 1 0 0
G E N O T Y P I C MIEAN G R A I N Y I E L D , KG/HA
Fig. 3. Coeffkient for the regressian of genotypic treatment grain yield with mean grain yield of
a11 treatments in one environment versus mean grain yield in a11 environments.
VARIETAL INTERCROPPING OF COWPEA IN SENkGAI.
231
2 . 0
1 .a
1 .6
2
; 1 . 4
T V X $ 2 3 6
N’DIANEOUF
:
k 1 . 2
A
MOUGNE
A
A
A
A58-57/cB5
0
CJ 1 . 0
58-57
%a-57/e 2 1
6
;tO.E
w
E 0 . 6
k?
0 . 4
0 . 2
0.0 l-
L
1100
1300
1 5 0 0
1700
1 9 0 0
21 10
G E N O T Y P I C MEAN H A Y Y I E L O . KG/HA
Fig. 4. Coeffkient for the regression of genotypic treatment hay yield with mean hay yield of a11
treatments in one environment versus mean hay yield in a11 environments.
Mougne had high mean hay yields and regression coefficients close to 1 .O (Fig.
4 ), indicating that they are well-adapted for hay production in a11 of the test
environments. The two early-erect ‘cultivars (Bambey 2 1 and CB5 ) generally
had low hay production, due to their small, erect, and determinate growth
habit.
No studies were discovered on the stability of varietal intercrops, but some
data are available for varietal mixtures. Stability of multiline mixtures of soy-
bean has been compared with pure litres. In one study, certain mixtures had
greater stability and higher yields than the pure lines (Schutz and Brim, 197 1).
The four cultivars used by Shutz and Brim ( 197 1) had substantial differences
in maturity date (29-day range), but only modest differences in morphology;
they were upright and determinate. In another study, there were no signitï-
tant differences in yield and regression coefficients between mixtures and pure
lines (Walker and Fehr, 1978). Walker and Fehr ( 1978) used many pure
lines but the maturity dates of the lines were within a lO-day range and the
morphologies of the lines were similar. For mixtures or intercrops to have
greater stability than pure lines gro,wn as sole crops, the component lines may
have to exhibit substantial differences in morphology and phenology.
CONCLUSIONS
Intercrops of early-erect and me 4mm-cycle spreading cowpea cultivars were
shown to be more effective than sole crops of these cultivars in dry locations
with infertile soils in Senegal. In these locations, fax-mers growing both types
---
-
-u
232
S . T H I A W E T A L .
of cultivars would have to plant 42 or 50% more land area of the sole crops to
obtain the same grain or hay yields as the varietal intercrops. Farmers in these
locations who are seeking the highest yielding or most stable cropping system
could also benefit from varietal intercrops. The intercrops had higher mean
grain and hay yields than a.ny of the cultivars tested as sole crops in the dry
locations with infertile soils, and above average yield stability across a11 loca-
tions. The varietal intercrop is compatible with mechanical sowing and cul-
tivation, and, in the drier zone where plants are small, it is possible to sepa-
rately harvest the individual cultivars, an important consideration for seed
quality and marketing. Further studies are needed to determine the impact of
varietal intercrops on problems due to insect pests, and to guide the breeding
of early erect and medium-cycle spreading cultivars that are most compatible
in varietal intercropping systems.
ACKNOWLEDGMENTS
This research was partially supported ‘by the Bean/Cowpea Collaborative
Reseach Support Program of the United States Agency for International De-
velopment Grant No. DAN- 13 l O-SS-6008-00. The opinions and recommen-
dations are those of the authors and not necessarilly those of USAID.
REFERENCES
Agboola, A.A., 1978. Influence of soi1 organic matter on cowpea’s response to N fertilizer. Agron.
J., 70: 25-28.
Berg, M.G. and Gardner, H., 1978. Methods of soi1 analysis in the laboratory at Oregon State
University. Special Report 32 1, Revised Sept. 1978. Agric. Exp. Station, Oregon State Uni-
versity, Corvallis.
Blum, A., 1988. Plant Breeding for Stress Environments. CRC Press, Boca Raton, FL.
Dancette, C. and Hall, A.E., 1979. Agroclimatology applied to water management in the Sudan-
ian and Sahelian zones of Africa. In: A.E. Hall, G.H. Canne11 and H.W. Lawton (Editors),
Agriculture in Semi-Arid Environments. Ecological Studies 34. Springer, Heidelberg, pp.
98-l 18.
Eberhart, S.A. and Russell, W.A., 1966. Stability analysis in plant breeding. Plant Breeding,
101: l-23.
Finlay, K.W. and Wilkinson, G.N., 1963. The analysis of adaptation in a plant-breeding pro-
gramme. Aust. J. Agric. Res., 14: 742-754.
Fussell, L.K. and Serafini, P.G., 1985. Crop association in the semiarid West Africa. Research
strategies, past and future. In: W. Ohm and J.C. Nagy (Editors), Appropriate Technology
for Farmers in the Semiarid West Africa. TJniversity of Purdue, 359 pp.
Hall, A.E., 1988. Stabilizing dryland production of cowpeas by using varietal intercrops. In:
P.W. Unger, T.V. Sneed, W.R. Jordan and R. Jensen (Editors), Challenges in Dryland Ag-
riculture: A Global Perspective. Texas Agricultu:ral Experiment Station, College Station, TX,
pp. 744-746.
Hall, A.E. and Dancette, C., 1978.. Analysis of fallow-farming systems in semiarid Africa using
a mode1 to simulate the hydrologie budget. Agraln. J., 70: 8 16-823.
VARIETAL INTERCROPPING OF COWPEA IN SENEGAL
233
Hall, A.E. and Patel, P.N., 1985. Breeding for resistance to drought and heat. In: S.R. Singh and
K.O. Rachie (Editors), Cowpea Research, Production and Utilization. John Wiley, New
York, NY, pp. 137-151.
Hall, A.E., Foster, KW. and Waines, J.G,, 1979. Crop adaptation to semi-arid environments.
In: A.E. Hall, G.H. Canne11 and H.W. Lawton (Editors), Agriculture in Semi-Arid Environ-
ments. Ecological Studies 34. Springer, Heidelberg, pp. 148-179.
Hiebsch, C.K. and McCollum, R.E., 1987, AreaxTime Equivalency Ratio: A method for eval-
uating the productivity of intercrops. Agron. J., 79: 15-22.
Khalfaoui, J.-L.B., 199 1. Determination ofpot.ential lengths of the trop growing period in semi-
arid regions of Senegal. Agric. Forest Meteorol., 55: 25 l-263.
Nageswara Rao, R.C., Wadia, K.D.R. and Williams, J.H., 1990. Intercropping short and long
duration groundnut (Arachis hypogeu) genotypes to increase productivity in environments
prone to end-of-season droughts. Exp. Agric., 26: 63-72.
Ng, N.Q. and Marechal, R., 1985. Cowpea taxonomy, origin and germplasm. In: S.R. Singh and
K.O. Rachie (Editors), Cowpea Research, Production and Utilization. John Wiley, New
York, NY, pp. 11-21.
Nicou, R. and Poulain, J.F., 1968. La fumure minerale du niebe au Senegal. In: Colloque Sur
La Fertilité des Sols Tropicaux, Tananarive, Madagascar, 19-25 November 1967. Agron-
omie Tropicale, Supplement to Nos. 2 and 5: 73 l-754.
Rattunde H.F., Ramraj, V.M., Williams, J.H. and Gibbons, R.W., 1988. Cultivar mixtures: a
means of exploiting morpho-developmenta.1 differences among cultivated groundnuts. Field
Crops Res., 19: 201-210.
Reddy, K.C., Mahamadou, 1. and Van der Ploeg, J., 1988. Appropriate cropping systems for the
Nigerian semiarid tropics. In: P.W. Unger, T.V. Sneed, W.R. Jordan and R. Jensen (Edi-
tors), Challenges in Dryland Agriculture - A Global Perspective. Texas Agricultural Exper-
iment Station, College Station, TX, pp. 767-769.
Shutz, W.M. and Brim, C.A., 1971. Intergenotypic competition in soybean. III. An evaluation
of stability in multiline mixtures. Crop Sci., 11: 684-689.
Singh, S.R., Jackai, L.E.N., dos Santos, J.H.R. and Abdalla, C.B., 1990. Insect pests of cowpea.
In: S.R. Sir@ (Editor), Insect Pests of Food Legumes. John Wiley, New York, NY, pp. 45-
89.
Thiaw, S.. 1990. Stability of dryland production of cowpeas ( Signa unguicuhtu (L. ) Walp. )
with varietal intercrops. MS. Thesis, University of California, Riverside, 57 pp.
Walker, A.K. and Fehr, W.R., 1978. Yield stability of soybean mixtures and multiple pure stands.
Crop Sci., 18: 7 19-723.
Ziska, L.H. and Hall, A.E., 1983. Seed yield and water use of cowpea subjected to planned-
water-deficit irrigation. Irrig. Sci., 3: 237-245.