|
African Crop Science Journal
African Crop Science Society
ISSN: 1021-9730 EISSN: 2072-6589
Vol. 8, Num. 3, 2000, pp. 243-249
|
African Crop Science Journal, Vol. 8. No. 3, pp. 243-249
African Crop Science Journal, Vol. 8. No. 3, pp. 243-249
AN APPRAISAL OF IRRIGATED TEMPERATE AND TROPICAL MILLET VARIETIES IN THE SEMIARID REGION OF SENEGAL
Saliou Diangar, Tanou Ba and C. F. Yamoah1 Institut Senegalais de Recherches Agricoles, ISRA CNRA, Bambey, BP 53 Senegal 1Soil Institute, Kumasi, Ghana
(Received 6 November, 1998; accepted 2 May, 2000)
Code Number: CS00026
INTRODUCTION
Pearl millet (Pennisetum glaucum (L). R. Br.) is the
main staple in the diet of people in northern Senegal. Yields are generally
low because it is entirely rainfed and rainfall in this region is presently
low (<300 mm/year) and fluctuates from month to month and even within months
(Dancette, 1974; INTSORMIL, 1993). Limited rainfall in northern Senegal is partly
attributed to advancing desertification of the Sahelo-Sudanian region. Furthermore,
soils of the semiarid region are sandy, infertile and contain low levels of
organic matter. Lack of improved varieties as well as poor crop and soil management
also contribute to the low productivity under peasant conditions in this environment.
Researchers in Senegal, Niger and other semiarid countries
have found that increased productivity of millet is possible with judicious
management under favourable climatic conditions (Bationo et al., 1993).
Management decisions under farmers control include moderate application of
manure and compost, a mixture of low doses of inorganic fertilisers with manure
or compost, rotation with legumes, varietal choices, appropriate land preparation
techniques and proper sowing dates.
Farmers do not have any direct control on rainfall but indirectly
they manipulate it through moisture retention strategies such as local irrigation
canals, mulching and ditches. Added to that, in a move to intensify land use
and boost cereal production in the light of growing human and livestock populations,
the government has provided irrigation facilities on 240,000 ha in northern
Senegal. Since millet has much lower water requirement than flooded rice, a
much larger area of land could be cultivated with millet. This could be a more
efficient way to use scarce irrigation water and provide significantly greater
food production than flooded rice and utilise irrigation scheme resources at
periods unfavourable for other crops. A millet hybrid from Nebraska (USA) and
an early local improved variety were tested under various cultural practices.
The objective of the study was to determine the response of improved millet
varieties to crop management (fertilisation, plant population and land preparation)
under irrigation.
MATERIALS AND METHODS
The study was conducted in 1991 and 1992 with double cropping
each year (i.e., four test periods) at the Thiago and Fanaye villages located
in Senegal River Valley in north Senegal. The two villages belong to the Sahelo-Sudanian
agroecological zone of Senegal (Bèye, 1977). The soils are Alfisols (USDA,
1975) and generally deficient in major nutrients (Nicou, 1976; Ndiaye, 1989;
Piéri, 1989). Rainfall is low (250 to 300 mm) and uncertain; followed
by a long dry season usually lasting about 8 months. Mean annual temperature
is 20°C to 35°C. Additional information on the Sahelo-Sudanian agro-ecological
zone of Senegal is provided by Sagna (1976).
Six field trials under irrigation were conducted, four in 1991
with two planted on 13 of February at Thiago (experiments 1 and 2) and two on
21 of August at Fanaye (experiments 3 and 4), and two in 1992,both planted on
11 of February (experiments 5 and 6). The design of the first trial in 1991
was a split plot with four replications. The main plots were millet genotypes,
early dwarf hybrid 68 A x MLS from University of Nebraska, Lincoln, Nebraska,
USA and a local dwarf synthetic variety, GAM 8201. Subplots were either three
land preparation techniques or fertiliser at three levels. The genotype differences
were important and were deliberately chosen, 68A x MLS and GAM 8201 were both
dwarf and likely to respond to higher plant densities; GB87-35 is medium tall.
All are shorter season, reaching physiological maturity in about 80 days, 10-15
days earlier than indigenous Senegalese varieties. Land preparation techniques
were: 1) flat; 2) single ridges (width = 30 cm and height = 40 cm); and 3) double
ridges (width = 100 cm and height=30 cm). All land preparation treatments received
22 N, 21 P and 39 K kg ha-1 at planting and 67 N kg ha-1 urea as a top dressing 2 weeks after planting. The fertiliser subplot treatments applied in 1991 were: 1) 15 N, 14 P and 26 K kg ha-1; 2) 22 N, 21 P and 39 K kg ha-1 + 67 kg N ha-1 as urea; and 3) 30 N, 28 P and 52 K kg ha-1 + 90 kg N ha-1 as urea all applied at planting.
The third and fourth experiments in 1991 minor season were
modified to include plant density as an additional treatment and the synthetic
variety GAM 8201 was replaced with variety GB 87-35. Research protocols of the
third and fourth experiments were split-split-plot with four replications. Main
plots were three fertiliser levels: 1)15 N, 14 P and 26 K kg ha-1 + 45 N kg
ha-1 as urea (divided into equal halves for top dressing at first weeding and
30 days later), 2) 22 N, 21 P and 39 K kg ha-1 at planting and 67 N kg ha-1 as urea (divided into equal halves for top dressing at first weeding and 30 days later) and 3) 30 N, 28 P and 52 K kg ha-1 + 90 kg N ha-1 as urea (divided into equal halves for top dressing at first weeding and 30 days later). Subplots were land preparation on flat, simple ridges and double ridges. Sub-sub plots
for variety GB 87-35 were plants spaced 80 x 80 cm, 80 x 60 cm, and 80 x 40
cm and sub-subplot size was 11.52 m2. Spacings for variety 68A x MLS were 70
x 30 cm; 70 x 20 cm; 60 x 30 cm, and 60 x 20 cm.
The rest of the studies conducted in 1992 examined the use
of reduced levels of fertilisers on hybrid 68 A x MLS and frequency and quantity
of water applied through irrigation. Fertiliser treatments were controlled with
no fertiliser, 5 N, 4 P and 8 K kg ha-1, 5 N, 4 P and 8 K kg ha-1 + 22 N kg ha-1 as urea, and 10 N, 9 P and 19 K kg ha-1 + 45 N kg ha-1 urea. Experimental design was a randomised block design with four replications. The weekly or twice weekly irrigation treatments consisted of a total of 693 mm water (maximum seasonal amount of water used by millet) or 526 mm water (75% of the maximum consumption). The design was a 2 by 2 factorial in a randomised complete block with four replications. Maximum water consumption (Mc) was calculated as: Mc = Kc * Ev, where Kc = crop coefficient and Ev = evapotranspiration. Statistical analyses were performed using SPSS (Norusis, 1997) and MSTAT-C (MSU, 1988).
RESULTS AND DISCUSSIONS
Land preparation did not significantly (P>0.05) affect average
yields of the Nebraska (68 A x MLS) and the local synthetic (GAM 8201) varieties
(Table 1). However, mean yields of the local synthetic variety planted
on the flat increased by about 58% and 26%, relative to planting on single and
double ridges, respectively. This finding is noteworthy, because flat planting
is cheaper than ridge planting. Conversely, the Nebraska variety performed better
in single ridges, showing a 14% increase over the flat. Average yields of both
millet genotypes, the Nebraska (4025 kg ha-1) and the local synthetic (4018
kg ha-1) were not significantly different (P>0.05) under irrigation,
but yields of the two varieties were much higher than yields on farmers fields
under rainfed conditions (430 - 1230 kg ha-1) in the semiarid environment (Bationo et al., 1993).
The main effect of fertiliser was not significant (P>0.05)
nor was average varietal difference (Table 1), even though the yield of the
local synthetic outperformed the Nebraska hybrid by 15 %. However, doubling
the fertiliser rate of 15 N, 14 P and 26 K kg ha-1 + 45 N kg ha-1 urea, resulted in a significant (P<0.05) 20% yield increase of 3295 kg ha-1 to 3911 kg ha-1 for the Nebraska hybrid whereas the local synthetic variety had a 10% yield reduction. Apart from genetic differences of the two genotypes, reasons for
the varietal differences are unclear; we suggest further research on this subject.
Results of yields for the new improved variety (GB 87-35)
and the Nebraska hybrid (68 A x MLS) in separate experiments are given in Tables
2 and 3. Low yields in Tables 2 and 3 were principally due to infestation of
stalk borers (Coniesta ignefusalis) and flower-sucking insects (Psalydollita
fusca and Cylindrothorax dussaulti). These insect pests caused about
30 % yield reduction. Fertiliser affected yield of GB 87-35 (P<0.05) but
methods of land preparation and plant density did not (P>0.05) (Table 2).
The yield of 1525 kg ha-1 with 22 N, 21 P and 39 K kg ha-1 and 67 N kg ha-1 urea was not different from 1601 kg ha-1 with 30 N, 28 P and 52 K kg ha-1 + 90 N kg ha-1 as urea but both of these were significantly higher than yields from 15 N, 14 P and 26 K kg ha-1 + 45 N kg ha-1 as urea. Thus, from this the recommended fertiliser rate is 22 N, 21 P and 39 K kg ha-1 and 67 N kg ha-1 urea in combination with any of the planting arrangements and land preparation methods for variety GB 87-35.
The results shown in Table 3 indicate that fertiliser did not
(P>0.05) affect yields of the Nebraska variety but land preparation techniques
and cropping patterns did (P<0.05). Also, ridges, both single and double,
improved yields of the Nebraska millet variety relative to planting on the flat,
consistent with the other study mentioned above. Yields of closer spacings between
rows (60 cm x 30 cm or 60 cm x 20 cm) were superior to wider (70 cm x 20 or
30 cm) spacings. It appears that the best cultural practices to attain optimum
yields of the Nebraska hybrid are ridges (large or small) and a high plant population
density. Average yields were higher with the Nebraska hybrid.
Results of the effect of reduced fertiliser rates on the Nebraska
variety are presented in Table 4. Both grain and stover yields of millet responded
to fertiliser (P<0.05). Highest yields for grain (4880 kg ha-1) and stover
(6330 kg ha-1) were obtained with the application of 10 N, 9 P and 17 K kg ha-1 + 45 N kg ha-1 urea. The lowest yields came from 5 N, 4 P and 8 K kg ha-1. The 36% difference in yield between the control (3300 kg ha-1) and the 5 N, 4 P and 8 K kg ha-1 (2440 kg ha-1) plots is not significantly different (P>0.05).
Yields of the Nebraska hybrid in Table 1 compared favourably
with the reduced yields in Table 4, thus, it may be safe to recommend less fertiliser
(10 N, 9 P and 17 K kg ha-1 + 45 N kg ha-1 as urea) to make this technology affordable to many farmers. Recent and past experiences in the semiarid region
of Senegal have demonstrated that a combination of mineral fertilisers and animal
manure or compost improved rainfed millet and peanut yields better than either
manure or fertiliser alone (Charreau and Nicou, 1971; Badiane, 1988). Additional
research is needed on irrigated millet to determine whether adding organic amendments
could further improve yield in the presence of inorganic fertiliser.
Data on yield and yield components of the Nebraska hybrid as
affected by frequency and amount of irrigation water are given in Table 5. Neither
the frequency nor the amount of water significantly (P>0.05) affected millet
yield. However, plots irrigated with 75% maximum water needed by millet tended
to yield more than plots with full water dose. Millet is a crop adapted to low
moisture conditions, particularly unsaturated soils, thus having lower water
requirements (Vachaud et al., 1978; Sarr et al., 1998). Also,
irrigation once per week yielded slightly higher than twice per week (Table
5). Governments and private organisations of the sub region should consider
developing low level irrigation facilities to promote higher and more stable
yields of millet and other cereal crops. These crops may offer a more efficient
way to use scarce irrigation water to produce more food per unit area than high
water-requiring crops such as lowland rice. However, environmental impact assessment
studies should be done before implementing such programmes.
ACKNOWLEDGMENTS
This project is a contribution of Natural Resource Based Agricultural
Research project and was supported by a grant from USAID/CID/Senegal project
#685-0285-C-00-2329-00. Appreciation is expressed to Dr. David Andrews of University
of Nebraska-Lincoln, for supplying 68 A X MLS and his advice on the design of
the studies and Dr. Richard P. Dick, Oregon State University, for editing the
manuscript.
REFERENCES
- Badiane, A.N., 1988. Courbe de réponse à des doses croissantes
de fumier (Thilmakha). Essai travail du sol (Sole III Nord du CNRA de
Bambey). Essai régénération des sols Ndiémane.
Résultats de 1987. ISRA-CNRA de Bambey.
- Bationo, A., Christianson, C.B. and Klaij, M.C. 1993. The effect of crop
residues and fertilizer on use of pearl millet yields in Niger. Fertilizer
Research 34:251-258.
- Bèye, G. 1977. Dégradation des sols au Sénégal.
Situation actuelle et perspectives. ISRA-CNRA. 23pp.
- Charreau, C. et Nicou, C. 1971. Lamélioration du profil cultural
dans les sols sableux et sablo argileux de la zone tropicale sèche
Ouest-Africaine et ses incidences agronomiques. Agronomy Tropicale
26:1183-1247.
- Dancette, C. 1974. Les besions en eau des plantes des grandes culture au
Senegal. In: Isotopes and radiation techniques in soil physics and irrigation
studies. Vienne, AIEA. pp. 351-371.
- INTSORMIL. 1993. Annual report 1992. Senegal Country Report. University
of Nebraska-Lincoln, NE. pp. 277-230.
- MSU. 1988. Microcomputer Statistical Program. Michigan State University,
East Lansing, MI.
- Ndiaye, J.P. et Sagna, A. 1989. La fertilisation des cultures au Sénégal.
Bilan diagnostic et perspectives. Ministère du Développement
Rural. 99pp.
- Norusis, M. J. 1997. SPSS 7.5. Guide to data analysis. Prentice Hall, Upper
Saddle River, NJ.
- Piéri, C. 1989. Fertilité des terres des savanes. Bilans de
30 ans de recherche et de développement agricole au Sud du Sahara.
Min. Coopération et CIRAD-IRAT. 444 pp.
- Sagna, A. 1976. Le bilan de pluies au Senegal de 1944 a 1973. Dakar, Faculte
des Lettres et des Sciences Humaines, Dept. de Geographie, 267 pp.
- Sarr, B., Boggio, D., Annerose, D. and Macauley, R.H. 1998. In: Approach
in modeling environmental factors and genotypes interaction: a case study
of millet (Pennisetum Glaucum). European Society of Agronomy, 5th ESA
Congress, Nitra, Rep. Slovaque, 28 Juin - 3 Juillet 1998.
- USDA. 1975. Soil taxonomy. A basic of soil classification for making and
interpreting soil surveys. U.S. Soil Conservation Service.
- Vachaud, G., Dancette, C., Sonko, S. et Thony, J.L. 1978. Méthodes
de caractérisation hydrodynamique in situ dun sol non saturé
du Sénégal en vue de la détermination des termes du bilan
hydrique. Ann. Agron. 29:1-36.
TABLE 1. Effects of land preparation and fertiliser on grain
yields of irrigated millet varieties in the semiarid area of Senegal (Experiments
1 and 2) |
Variety |
Land preparation |
Yield (kg ha-1) |
Fertiliser |
Yield (kg ha-1) |
68A x MLS |
Flat |
3,700 |
60 N, 14 P and 26 K kg ha-1 |
3,300 |
Simple ridge |
4,230 |
90 N, 21 P and 39 K kg ha-1 |
3,680 |
Double ridge |
4,145 |
120 N, 28 P and 52 K kg ha-1 |
3,920 |
GAM 8201 |
Flat |
4,974 |
60 N, 14 P and 26 K kg ha-1 |
4,390 |
Simple ridge |
3,146 |
90 N, 21 P and 39 K kg ha-1 4,060 |
|
Double ridge |
3,930 |
120 N, 28 P and 52 K kg ha-1 |
3,903 |
|
|
Level of significance |
|
Level of significance |
Variety (V) |
|
ns |
Variety |
0.079 |
Land preparation (LP) |
|
ns |
Fertiliser |
ns |
V x LP |
|
0.043 |
Variety x fertiliser |
ns |
TABLE 2. Effect of fertiliser, land preparation and plant density
on grain yields of irrigated millet (variety GB 87-35) in the semiarid area
of Senegal (Experiment 3) |
Fertiliser |
Density |
Flat |
Simple ridge |
Large ridge |
Yield (kg ha-1) |
60 N, 14 P and 26 K kg ha-1 |
80 cm x 80 cm |
1,120 |
1,170 |
1,330 |
80 cm x 60 cm |
1,290 |
1,040 |
1,590 |
80 cm x 40 cm |
1,170 |
1,250 |
1,490 |
90 N, 21 P and 39 K kg ha-1 |
80 cm x 80 cm |
1,950 |
1,410 |
1,700 |
80 cm x 60 cm |
1,490 |
1,270 |
1,610 |
80 cm x 40 cm |
1,300 |
1,660 |
1,350 |
120 N, 28 P and 52 K kg ha-1 |
80 cm x 80 cm |
1,770 |
1,510 |
1,800 |
80 cm x 60 cm |
1,640 |
1,460 |
1,610 |
80 cm x 40 cm |
1,630 |
1,610 |
1,370 |
Level of significance |
Density (D) |
ns |
|
|
|
Fertiliser (F) |
0.013 |
|
|
|
Land preparation (LP) |
ns |
|
|
|
LP x F |
ns |
|
|
|
LP x D |
0.047 |
|
|
|
LP x D x F |
ns |
|
|
|
TABLE 3. Effect of fertiliser, land preparation and plant density
on grain yields of irrigated millet (hybrid 68 A x MLS) in the semiarid
area of Senegal (Experiment 4) |
Fertiliser |
Density |
Flat |
Simple ridge |
Large ridge |
kg ha-1 |
60 N, 14 P and 26 K kg ha-1 |
70 cm x 30 cm |
2,030 |
2,340 |
2,030 |
70 cm x 20 cm |
2,170 |
1,790 |
2,380 |
60 cm x 30 cm |
2,060 |
2,680 |
2,420 |
60 cm x 20 cm |
2,730 |
2,630 |
2,330 |
90 N, 21 P and 39 K kg ha-1 |
70 cm x 30 cm |
1,850 |
2,250 |
2,200 |
70 cm x 20 cm |
1,680 |
2,170 |
2,420 |
60 cm x 30 cm |
2,370 |
2,730 |
2,830 |
60 cm x 20 cm |
1,890 |
2,530 |
2,280 |
120 N, 28 P and 52 K kg ha-1 |
70 cm x 30 cm |
1,280 |
2,650 |
2,820 |
70 cm x 20 cm |
1,790 |
2,420 |
2,680 |
60 cm x 30 cm |
2,370 |
2,680 |
2,980 |
60 cm x 20 cm |
1,740 |
2,630 |
3,130 |
Level of significance |
Plant density (D) |
0.0001 |
|
|
|
Fertiliser (F) |
ns |
|
|
|
Land preparation (LP) |
0.004 |
|
|
|
LP x F |
0.094 |
|
|
|
LP x D |
ns |
|
|
|
LP x D x F |
ns |
|
|
|
TABLE 4. Effects of reduced rates of mineral fertiliser on grain
and stover yields of irrigated millet (hybrid 68 A x MLS) in the semiarid
zone of Senegal in 1992 (Experiment 5) |
Fertiliser |
Grain yield (kg ha-1) |
Stover yield (kg ha-1) |
Control (without fertiliser) |
3,300 |
4,090 |
5 N, 4 P and 8 K kg ha-1 |
2,440 |
3,020 |
27 N, 4 P and 8 K kg ha-1 |
3,810 |
4,740 |
55 N, 9 P and 25 K kg ha-1 |
4,880 |
6,330 |
Level of significance |
Fertiliser |
0.062 |
0.004 |
TABLE 5. Millet (hybrid 68 A x MLS) grain yield and yield components
as affected by irrigation frequency and amount of water in 1992 (Experiment
6) |
Frequency |
Amount |
Yield |
Stover |
Non productive tillers (no. of tillers
ha-1) |
Productive tillers |
Head g/m2 |
kg ha-1 |
1/week |
Maximum |
3,820 |
5,423 |
172,590 |
637,135 |
3,155 |
75% maximum |
4,680 |
6,915 |
274,210 |
727,465 |
3,841 |
2/week |
Maximum |
3,580 |
5,202 |
220,980 |
582,290 |
2,974 |
75% maximum |
4,390 |
6,472 |
196,786 |
640,361 |
3,644 |
Level of significance |
Frequency (F) |
|
ns |
ns |
0.17 |
0.13 |
ns |
Amount (A) |
|
0.12 |
0.02 |
ns |
0.12 |
0.13 |
F x A |
|
ns |
0.05 |
ns |
ns |
ns |
About 75 % of 693 mm water per year |
|