ABSTRACT

Pearl millet [Pennisetum glaucum (L.) R. Br.] is the most important food crop in Niger and most other West African Sahelian countries. The study was conducted as a factorial combination of three pearl millet varieties and two management levels at Kollo, Niger. Low management consisted of 10,000 hills ha^-1 with no fertiliser application, while high management consisted of 20,000 plants ha^-1 with manure, and N and P fertiliser application. The three varieties used were the improved varieties ‘Zatib’ (tall) and ‘3/4 HK’(short), and a landrace variety ‘Heini Kirei’ (tall). Two plants per plot were sampled bi-weekly, partitioned into plant parts, dried, and weighed. Pearl millet under high management produced 517 to 551 kg ha^-1 greater yield and 200 to 300 g m^-2 more dry matter than with low management. In 1995, the shorter variety ‘3/4 H K’ produced 147 to 256 g m^-2 less dry matter and 155 to 485 kg ha^-1 less grain yield than the other varieties. Mean crop growth rates (CGRs) increased up to a maximum of 18.5 g m-2 day^-1 with high management and 2 g m^-2 day^-1 with low management in 1995, while in 1996 maximum CGRs were 6.4 g m-2 day^-1 with high management and 1.7 g m-2 day^-1 with low management. In this study, environmental variability due to years had the greatest effect on crop growth and grain yield; management had an intermediate effect; and genotypes the least effect. Pearl millet producers in Niger should increase plant population and apply fertiliser to optimize pearl millet grain and stover yield.

Key Words: Crop growth rate, dry matter partitioning, Pennisetum glaucum, relative growth rate, sahel

Le petit mil [Pennisetum glaucum (L.) R. Br.] est la plus importante culture vivrière au Niger. L’étude a été conduite avec une combinaison factorielle de trois varétés de mil et deux niveaux de gestion à Kolo, Niger. Le faible niveau de gestion était constitué de 10.000 poquets ha^-1 avec sans apport de fumure, tandis que le haut niveau de gestion comportait 20.000 poquets ha^-1 avec l’apport de fumure organique et d’engrais minéraux N et P. Les trois variétés de mil utilisées étaient réparties entre deux variétés améliorées ‘Zatib’ (haute taille) et ‘3/4HK’ (courte taille), et une variété locale ‘Heini Kirei’ (haute taille). Deux plants par parcelle étaient échantillonnés chaque deux semaines et séparés eu parties de la plante, sèchés et pesés. Le haut niveau de gestion a produit 517 à 551 kg ha^-1 de plus de rendement grain et 200 à 300 g m^-2 de plus de matière sèche que le faible niveau de gestion. En 1995, la variété de courte taille ‘3/4 HK’ a donné 147 à 256 g m^-2 de moins de matière sèche et 155 à 485 kg ha^-1de moins de rendement grain que les deux autres variétés. Les taux de croissance moyens de la culture ont augmenté jusqu’à 20 g m^-2 jour^-1 pour le haut niveau de gestion et 3 g m^-2 jour^-1 pour le faible niveau de gestion en 1995, tandis qu’en 1996 les taux étaient de 9 g m^-2 jour^-1 pour le haut niveau de gestion et 2,4 g m^-2 jour^-1 pour le faible niveau de gestion. Dans cette étude, la variabilité environnementale liée à l’année a eu le plus grand effet sur la croissance et le rendement de la culture; tandis que les niveaux de gestion ont eu un effet intermédiaire; et les génotypes le moins d’effet. Les producteurs de petit mil au Niger ont besoin de choisir des génotypes à haut rendement en combinaison avec de bonnes pratiques de production pour optimiser le rendement grain du petit mil.

Mots Clés: Taux de croissance de la culture, répartition de matière sèche, Pennisetum glaucum, taux de croissance relative de la culture, sahel

Pearl millet [Pennisetum glaucum (L.) Br.] is the staple and predominant grain crop produced in Niger and most other West African Sahelian countries. It is grown on 4.6 million hectares, more than 90% of cultivated area in Niger. Pearl millet is grown under low soil nutrient and water levels that interact to influence grain yield and growth. An understanding of crop growth is essential for management since many critical decisions involve precise timing. Studies have been reported for pearl millet grain yield and dry matter production under different crop management conditions (Azam-Ali et al., 1984; Bationo et al., 1990; Payne et al., 1991; Payne, 1997). They found soil nutrient level to usually be the most limiting factor, but interacts with the quantity of water available. National research programmes promote the use of both manure and chemical fertilizer to rectify low soil nutrient levels, and recommend improved cultivars to increase yield potential. Farmers usually plant local varieties that are tall (2.5 - 3 m), partially photoperiod sensitive, and late maturing (110-120 days). They traditionally use very low plant populations of 5200 hills ha^-1 (3-5 plants hill^-1) in order to reduce the risk of potential yield loss due to water stress (Bationo et al., 1990).

In Niger and the rest of the Sahel, low pearl millet grain yields of 300-550 kg ha^-1 are common due to low rainfall and soil nutrient levels, use of unimproved varieties and low management levels. Increased plant population combined with fertilizer application and improved varieties have been used to increase pearl millet grain yield. Bationo et al. (1990) found that increasing plant population from 5,000 to 40,000 hills ha^-1 and application of N fertiliser increased grain yield of pearl millet in normal or above average rainfall years but slightly reduced grain yield in drought years. Stoop (1987) found that in Burkina Faso, increasing plant population from 20,000 to 40,000 hills ha^-1 had no effect on pearl millet yield. Payne (1997) reported that even in dry years, higher grain yields and water use efficiency are possible using a plant population > 20,000 hills ha^-1 with application of > 40 kg N ha^-1 and >18 kg P ha^-1.

Crop growth rate is a physiological trait associated with improvement of grain yield in cereal crops. General aspects of growth and development of pearl millet plants were reported by Maiti and Bidinger (1981) and Bramel-Cox et al. (1984). Dry matter accumulations of 42 to 200 g plant^-1 (Carberry et al., 1995; Craufurd and Bidinger, 1989) and 140 to 310 g m^-2 (Azam-Ali et al., 1984; Maman et al., 1998) have been reported. Crop growth rates ranging from 8.5 to 18.9 g m^-2 day^-1 have been reported (Bramel-Cox et al., 1984; Maman et al., 1999). Relative growth rates of 0.07 to 0.14 g g^-1 day ^-1 during early growth have been reported in Australia which declined to 0.01g g^-1 day ^-1 near physiological maturity (Coaldrake and Pearson, 1985). Improved varieties, increased plant population and fertiliser application should increase pearl millet yields and growth rates. The objective of this study was to determine grain yield, dry matter production and growth rates of three pearl millet varieties grown under low and high management levels in Niger.

A two-year experiment was conducted in 1995 and 1996 at the Kollo Research Station located at latitude 13^°15' N, longitude 2^°18’E, and altitude 210 m (Niger, West Africa). The soil at the site was a Psammentic Paleustalf with a sandy texture. Prior to fertiliser and manure application in 1995, soils were sampled at 0-40 cm depth in each replication at six different places and mixed, to determine physical and chemical parameters. Test results showed low organic matter, pH, and nutrient content, except for a moderate P level (Table 1). The experimental site had a Sahelien climate with 540 mm 30-year average rainfall from May to October. During experimental years, average rainfall was 521 mm in 1995 and 405 mm in 1996 (Table 2).

The experimental treatments consisted of a factorial combination of three pearl millet varieties with two levels of management based on plant population and fertiliser rate, in a randomised complete block design with three replications. The varieties used in the experiment were; ‘Heini Kirei’, a local land race variety, 2.8 to 3.0 m tall and a maturity classification of 110-115 days to physiological maturity; ‘Zatib’, tall variety, 1.80 to 2.0 m tall and a maturity classification of 95 to 105 days to physiological maturity; and ‘3/4 Heini Kirei’ (‘3/4 HK’), an improved dwarf variety, 0.80 to 1.20 m tall and a maturity classification of 85 to 95 days to physiological maturity. Low management was a typical farmer practice and consisted of a plant population of 10,000 hills ha^-1 and no fertiliser application. High management consisted of 5 t ha^-1 of manure, 18 kg P ha^-1 as single super phosphate broadcast applied and mixed into the upper 40 cm of soil before planting, 23 kg N ha^-1 as urea with split application at 14 and 35 days after planting (DAP), and a plant population of 20,000 hills ha ^-1. Ten pearl millet seeds were hand planted in hills at 1m x 1m for the low plant population, and 1m x 0.5m spaced hills for the high plant population. Seeds were planted at 5 cm depth, and thinned to three plants per hill 14 days after emergence. The manure consisted of a farmyard noncomposted mixture of cattle dung and urine, with pearl millet and grain sorghum [Sorghum bicolor (L.) Moench] crop residues that were dried before weighing and application. Four samples of manure from the farm were analysed in 1995 and had N, P, and K concentration on dry matter basis of 14.3, 2.1, and 10.1 g kg ^-1, respectively. Manure source and preparation were similar in 1996, and it was assumed that soil nutrient levels in 1996 were similar to those of 1995, but no analysis was conducted to confirm this.

Plot size was 12 m wide and 15 m long. Grain yield was determined from the centre 4 rows (60 m2), and plants were harvested for dry matter accumulation in nonadjacent parts of plots. Pearl millet was planted 12 June 1995 and 26 June 1996, and harvested 10 October 1995 and 15 October 1996. Weeds were controlled by hand hoeing. No serious pest problems were encountered during the experiment.

Data collection and analysis. Two plants per plot were harvested biweekly from the four-leaf stage until physiological maturity. Plants were separated into leaf, stem, and panicle, dried at 65^°C for 48 hours, and weighed to determine dry matter accumulation and partitioning. Dry matter accumulation per unit area, Mean Crop Growth Rate (CGR) and Mean Relative growth rate (CGR) as defined by Radford (1967) were calculated. The CGR was calculated as CGR =dW/dt = (W₂-W₁)/t₂-t₁) where dry weight at time t, and t = time as DAP. The RCGR was calculated as RCGR = dW/ W *dt = (log_e W2 - log_e W₁)/( t₂ -t₁). Polynomial curves were fitted against DAP for total dry matter accumulation.

The data were subjected to analysis of variance (ANOVA) and mean separation using single degree of freedom orthogonal contrasts, using the general linear model (GLM) procedure of the SAS software (SAS , 1988). Data for each year were analysed separately.

Grain yield. In both years, variety x management level interaction effects were not present for grain yield (Table 3), indicating that varieties responded similarly to crop management (plant population, fertilizer and manure application). The high management level increased the grain yield over low management by 551 kg ha^-1 ( 65%) in the favourable year of 1995 and by 517 kg ha^-1 (176%) in the drier year of 1996. Stover yields were increased by 1830 kg ha^-1 (74%) in 1995 and by 2420 kg ha^-1 (260%) in 1996. Number of panicles m^-2 produced was greater with high management than with low management. This result is consistent with those of Payne (1997) who found that even during drier years, plant population of 20,000 hills ha^-1 along with a minimum of 18 kg P ha^-1 and 23 kg N ha^-1 was necessary to optimise pearl millet grain yield in Niger, but conflicts with Bationo et al. (1990) who found that increasing plant population from 5,000 to 40,000 hills ha^-1 and application of N fertilizer slightly reduced grain yield in drought years. In this study, higher plant population combined with N and P application apparently did not deplete soil water nor make water stress more severe, and actually increased grain and stover yield.

Variety differences for grain and stover yield were more pronounced in the higher rainfall year of 1995 than in 1996 (Table 3). In 1995, the land race ‘Heini Kirei’ produced 31% more grain and 44% more stover than the improved tall variety ‘Zatib’, and 53% more grain and 119% more stover than the dwarf variety ‘3/4HK’. In the lower rainfall year of 1996, grain and stover yields for the three varieties were similar. Number of panicles m^-2 produced were similar for the three varieties in both years. The lower yield of ‘3/4HK’ than ‘Heini Kirei’ and ‘Zatib’ in 1995 may have been due to the fact that the tall varieties’ produced more dry matter and had more stem carbohydrates available for translocation during grain fill. Blum et al. (1997) reported that in grain sorghum, the grain weight per panicle was reduced by drought stress only in the short genotypes due to less carbohydrate translocation from stem to panicle.

Dry matter accumulation and partitioning. Dry matter accumulation increased in a cubic fashion with high management level and in a quadratic fashion with low management for all three varieties in both years (Fig. 1). The improved short variety ‘3/4HK’ produced less dry matter than the other varieties particularly in 1995 when seasonal rainfall was near normal. Even in the drier year of 1996, the local variety ‘Heini Kirei’ produced more dry matter than the short ‘3/4HK’ and the tall improved ‘Zatib’ varieties. In Niger, the advantage of improved varieties is largely due to shorter plant maturity allowing escape of water stress during a dry year. These results indicated that even for the relatively dry year of 1996, the growth of pearl millet under high management level was greater than with typical farmer practices. The increase of dry matter with management level confirms results found by Bationo et al. (1990) and Payne (1997) to increased plant population combined with fertilizer application. The dry matter accumulation curves in both years in low and high management level contrasted with results of Bramel-Cox et al. (1984) who found a linear increase from planting until flowering. Dry matter accumulation after flowering varied among varieties, consistent with Brammel-Cox et al. (1984) who found variable growth ranging from decreasing to no dry matter accumulation to increasing dry matter accumulation between flowering and physiological maturity depending on variety and environment.

Dry matter partitioning indicated that in 1995, the leaf and stem percentage of total dry matter accumulation decreased between 95 and 100 DAP while the panicle increased (Fig. 2). In 1996, the leaf percentage of dry matter remained constant while the stem percentage of dry matter increased from 90 DAP to 111 DAP (Fig. 3). At physiological maturity in both years, panicle dry matter accumulation was 64 to 67 % of the total for the short improved ‘3/4HK’ variety, 57 to 60% of the total for the tall improved ‘Zatib’ variety, and 53 to 55% of the total for the local ‘Heini Kirei’ variety. The stem dry matter was 25% of the total for ‘3/4HK’, 29 to 36% of the total for ‘Zatib’, and 32 to 35% of the total for ‘Heini Kirei’ in both years. Dry matter partitioning to the panicle in this experiment was higher than reported by Maiti and Bidinger (1981) who found high-yielding dwarf varieties with as much as 50% of the dry matter in panicles and some tall varieties with no more than 20 to 30% in the panicles. The results of dry matter partitioning based on the management level indicated that at the final harvest, the panicles contributed 60 to 61% of the total with high management, and 51 to 54% with low management in both years.

The two management levels resulted in different patterns of dry matter partitioning (Fig. 3). Under low management, the leaves accumulated relatively more dry matter than with high management in both years, probably due to more tillers resulting from lower plant population (Carberry et al., 1985; Craufurd and Bidinger, 1989). However, the number of panicles m^-2 (Table 3) indicated that even though there was some compensation for low plant population, this was not enough to completely make up for differences with the higher plant population under high management. After 90 DAP, the panicle proportion of total dry matter accumulation was higher with high management due to a better grain fill (Fig. 3).

Growth analysis. Variety differences in CGR were found only at 66 - 80 and 80 - 95 DAP in 1995 (Table 4) with ‘Zatib’ having greater CGR than ‘3/4HK’( 4.92 vs 2.45 g m^-2 day^-1 and 11.69 vs. 7.23 g m^-2 day^-1). In 1996 (Table 5) at 0 - 27 DAP, ‘Zatib’ had a greater CGR than ‘3/4HK’ (0.09 vs 0.08 g m^-2 day^-1) while between 27 - 42 days, ‘3/4HK’ had greater CGR (0.82 vs 0.61 g m^-2 day^-1). The high management level increased CGR during both years of the study. Coaldrake and Pearson (1985) reported that growth of pearl millet was reduced by low N supply, and that maximum growth rate before panicle initiation was achieved at a N concentration of 1.56 % of whole plant dry weight and 1.3 % during later development. The N concentrations were within these ranges (data not shown); thus, growth rates were likely not limited by N supply. In the higher rainfall year of 1995, the CGRs were higher with average maximum rate of nearly 10 g m^-2 day^-1 (Table 4) which is similar to growth rates reported by Bramel-Cox et al. (1984). In the dry 1996 year, CGRs were lower but increased linearly up to physiological maturity with a maximum of over 4 g m^-2 day ^-1 (Table 5).

Variety x management level interactions and variety main effects for RCGR were not present (Tables 4 and 5). Management level had a small, inconsistent effect on RCGR. Relative growth rates for all varieties and management levels in both years decreased with plant age as reported for pearl millet (Coaldrake and Pearson, 1985), maize (Zea mays L.), sunflower (Helianthus annus L.) and cotton (Gossypium arboreum L.) by Evans (1972).

Variety x management level interaction effects were generally absent for grain yield and dry matter production. Thus, the local land race pearl millet variety ‘Heini Kirei’, the improved tall variety ‘Zatib’, and short variety ‘3/4HK’ responded similarly to management levels in this study. High management level increased grain and stover yield in both years compared to the low management level (i.e., traditional farmer practices). These results show that farmers in Niger can increase pearl millet yields by increasing plant population in combination with application of N and P fertiliser.

Variety differences for grain yield were more pronounced in the normal rainfall year than in the drier year. The short variety ‘3/4HK’ always produced less dry matter and, consequently, had less translocation from leaf and stem to the panicle during grain filling than the tall varieties ‘Zatib’ and ‘Heini Kirei’. Pearl millet dry matter accumulation and crop growth rates were greater for high than low management conditions in both years. Dry matter accumulation increased with plant age up to physiological maturity but the CGR decreased after 80 DAP in 1995. The CGR decreased with plant age and were not influenced by management level. This study indicated that crop management and seasonal rainfall influenced pearl millet grain and stover yields, and growth more than variety in Niger. We recommend that to optimise grain and stover yield, producers should increase plant population and apply fertiliser.

Contribution of the Department of Agronomy, University of Nebraska, Lincoln, NE 68583-0915 and INRAN (Institut National de Recherches Agronomiques du Niger), B.P. 429, Niamey, Niger. Paper No. 12306 of the Journal Series of the Nebraska Agricultural Research Division. This research was supported by USAID Grant No. DAN 1254-G-0021 through INTSORMIL, the International Sorghum and Millet Collaborative Research Program. The authors appreciate the assistance of Dr. Samba Traore in calculating growth rates.

• Azam-Ali, S.N., Gregory, P.J. and Monteith, J.L. 1984. Effects of planting density on water use and productivity of pearl millet grown on stored water. II. Water use, light interception and dry matter production. Experimental Agriculture 20: 215-224.
• Bationo, A., Christianson, C.B. and Baethgen, W.E. 1990. Plant density and nitrogen fertilizer effects on pearl millet production in Niger. Agronomy Journal 82:290-295.
• Blum, A., Golan, G., Mayer, J. and Siinmena, B. 1997. The effect of dwarfing genes on sorghum grain filling from remobilized stem reserves under stress. Field Crops Research 52: 43-54.
• Bramel-Cox, P.J., Andrews, D.J., Bidinger, F.R. and Frey, K.J. 1984. A rapid method of evaluating growth rate in pearl millet and its weedy and wild relatives. Crop Science 24:1187-1191.
• Carberry, P.S., Campbell, L.C. and Bidinger, F.R. 1985. The growth and development of pearl millet as affected by plant population. Field Crops Research 11:193-205.
• Coaldrake, P.D. and Pearson, C.J. 1985. Development and dry weight accumulation of pearl millet as affected by nitrogen supply. Field Crops Research 11:193-205.
• Craufurd, P.Q. and Bidinger, F.R. 1989. Potential and realized yield in pearl millet as influenced by plant population density and cycle duration. Field Crops Research 22: 211-225.
• Evans, C.G. 1972. The quantitative analysis of plant growth. Studies in Ecology: Volume I. University of California Press, Blackwell Scientific Publication, Oxford, U.K. 734 pp.
• Maiti, R. K. and Bidinger, F.R. 1981. Growth and development of pearl millet plant. ICRISAT Research Bulletin, No 6. Hyderbad, India.
• Maman, N., Mason, S.C. and Galusha, T.D. 1999. Hybrid and N influence on pearl millet production in Nebraska: yield, growth, and N uptake and use efficiency. Agronomy Journal 91:737-743.
• Payne, W.A., Lascano, R.J., Hossner, L.R., Wendt, C.W. and Onken, A.B. 1991. Pearl millet growth as affected by phosphorus and water. Agronomy Journal 83:942-948.
• Payne, W. A. 1997. Managing yield and water use of pearl millet in the Sahel. Agronomy Journal 89:481-490.
• Radford, P.J. 1967. Growth analysis formulae - Their use and abuse. Crop Science 7:171-175.
• Stopp, W.A. 1987. Adaptation of sorghum/maize and sorghum/pearl millet intercrop systems to the top sequence land types in the north Sudanian zone of the West African savanna. Field Crops Research 16:255-272.
• SAS. 1988. SAS users guide: Statistics. SAS Institute., Cary, NC.