Combined effect of Millet-Cowpea intercropping and biopesticide application against Heliocheilus albipunctella De Joannis (Lepidoptera: Noctuidae) in Burkina Faso

Study location

The study was conducted in Burkina Faso, in the communes of Djibasso and Dori during the 2021 rainy season. These two communes, Djibasso and Dori, are located in the Kossi and Seno provinces, respectively. Pearl millet is the main cereal crop in these two provinces and covers almost 78.00 and 80.00% of the cultivated area, respectively (Direction Générales des Etudes et des Statistiques Sectorielles (DGESS), 2023), and it is often intercropped with cowpea (Boly et al., 2022). The experimental plots were established in the villages of Bouakuy, located about 10 km from Djibasso, and Hoggo Sambowel, located about 10.00 km from Dori (Fig. 1). The cumulative rainfall from May to December was 827.00 mm in Djibasso and 557.00 mm in Dori. The relative humidity fluctuated between 53.16–91.44% in Djibasso and between 44.25–83.00% in Dori. While average monthly temperatures fluctuated between 25.00 and 32.00 °C. The vegetation is mostly covered with annual grass species, with areas of woodland and shrubland in which the dominant trees are Acacia species (Lykke, Kristensen & Ganaba, 2004). The soil in both communes is sandy in nature (Fontès & Guinko, 1995).

Experimental design

The experimental design is composed of a divided plot (= Split Plot) made up of three sub-blocks (= Large plots). Each sub-block is composed of twelve (12.00) sub-plots (= Small plot). The sub-plots were each 9.60 m × 9.60 m in dimension (Fig. 2). This type of design provides the possibility of evaluating two factors in the system; the type of cropping association (primary factor) and the phytosanitary treatment (secondary factor). The cropping system consists of four combinations (i) millet (MP), (ii) cowpea monocultures as single main crop (NP), (iii) an intercrop consisting of two rows of millet and a single row of cowpea (2M-1N) and (iv) farmers’ practices (PP, one planting of cowpea between four plantings of millet). Such associations represent common practice among farmers in these regions. The four cropping systems form the main plot (Σ MP + NP + 2M-1N + PP). The phytosanitary treatment assigned to each main plot (Σ MP + NP + 2M−1N + PP) at each large plot level consisted of three combinations: (i) an aqueous neem seed extract (50 g/L; 5% w/v), containing a range of bioactive compounds, was used. Key constituents include azadirachtin (tetranortriterpenoid limonoid, C_{3 5}H_{4 4}O_{1 6}), meliantriol (triterpenoid limonoid, C_{3 2}H_{5 0}O₈), and salannin (triterpenoid limonoid, C_{3 4}H_{4 4}O₉), which are primarily responsible for the extract’s insecticidal and repellent effects (Norten, 1999; Djibril et al., 2015); (ii) a synthetic insecticide (Lambda-Cyhalothrin 15.00 g/l + Acetamiprid 20.00 g/l); and (iii) a no-treatment control. Each of the cropping systems was replicated three times in each of the sub-blocks. All plots within each treatment were subjected to natural infestation by H. albipunctella. The spacings between rows and pockets in each cropping system were 0.80 m × 0.80 m for millet and 0.80 m × 0.40 m for cowpea. A distance of two m and six m was allowed between two sets of cropping systems and between sub-blocks respectively in order to prevent effects of treatments. For the millet, Local Djibasso and Local Dori were sown in the communes of Djibasso and Dori, respectively. The cowpea variety Komcalle was sown 20 days later to align its flowering stage (45 days after sowing) with the heading stage of millet. These varieties are the most commonly preferred and used by farmers in these areas. It is at this stage that H. albipunctella females prefer to lay their eggs on the spikes (Kabore et al., 2023). The sub-plots were thinned to two (2.00) millet plants per pocket at the first weeding, three weeks after sowing. A microdose of fertilizer consisting of 5.00 g of NPK (14/23/14; 100 kg/ha) per pocket was applied to both millet and cowpea after weeding followed by an application of urea (46% N; 50 kg/ha) to 3.00 g/pocket of millet at the time of the millet bolting.

To prepare the neem aqueous extract, the neem seeds were collected under the trees, stored, and dried in the shade for 4 to 5 weeks. Next, they were ground into a fine powder, and this powder was macerated in water for 24 h. Finally, the macerated material was filtered to obtain the extracts (Dabiré-Binso, Ba & Sanon, 2008).

For the neem seed extract treatment subplots, a formulation of 500.00 g per 10 L (w/v) of water per 400.00 m² was applied three times weekly (corresponding to 115 g per 2.30 L of water per subplot; Dabiré-Binso, Ba & Sanon, 2008). For synthetic insecticide treatment plots, a mixture of lambda-cyhalothrin (15.00 g/L) and acetamiprid (20.00 g/L) was applied twice at a rate of 1 L/ha, with applications two weeks apart (approximately 10 mL per 3.5 L of water per subplot). The choice of two treatments for the synthetic pesticide and three for the aqueous neem seed extract was based on previous studies in Burkina Faso (Dabiré-Binso, Ba & Sanon, 2008).

Data collection

Data collection started at the doughy grain stage of millet from each sub-plot and sub-block until harvest (time in weeks). The number of H. albipunctella larvae per spike, the damage caused by H. albipunctella (number and length of mines per spike and number of spikes attacked) and the grain yield were recorded. To determine the number of larvae per spike, an area of one m² was delimited in each sub-plot and the number of larvae is counted on each spike contained in each surface and repeated four times. while the number of attacked spikes (spike bearing at least one mine), the number and length of mines per spike and the grain yield, an area of 9.00 m² was determined at each sub-plot. Thus, the number of attacked spikes, the number and length of mines per spike and the grain yield were determined. The length of each mine was measured using a measuring tape. At harvest, the millet spikes from each area were threshed and weighed. The percentage of attacked spikes (PAE) and the grain yield (GY) were calculated for each sub-plot of each sub-block using Eqs. (1) and (2) (Boly, 2024): (1)${\displaystyle PAE\left(\text{\%}\right)=\left(NSA/TNS\right)\times 100.00;}$

where NSA: Number of spikes attacked per delimited area and TNS: Total number of spikes.

(2)${\displaystyle GY\left(\frac{kg}{ha}\right)=\left(\frac{GW}{S}\right)\times 10,000.00;}$

where GW: Grain weight of millet in kg per delimited area and S: delimited area in m².

Data analyses

The data collected were analyzed using a two-way analysis of variance (Factorial ANOVA) to examine the influence of the phytosanitary treatments and the cropping systems (two independent variables) on the studied parameters. Post hoc tests were performed using the ’lsmeans’ package in RStudio (Lenth, 2016) when factorial ANOVA was significant between groups. The visualize the data, box plots and figure were plotted using R and Excel, respectively. All statistical analyses were carried out using R (R Core Team, 2022) and level of significance set at 5% for all statistical analyses.

Influence of cropping system and phytosanitary treatment on the parameters studied

We found that phytosanitary treatment has a significant influence on the number of larvae per spike, H. albipunctella damage and grain yield (Table 1). The average percentage of damaged and the average number of damaged per spike were influenced by the cropping system (Table 1). On the other hand, the interaction between cropping system and phytosanitary treatments did not affect these parameters (Table 1).

Number of H. albipunctella larvae per spike according to phytosanitary treatments

The average number of larvae per spike of millet varied significantly depending on the location of the treatment (Djibasso ANOVA, F_2.439 = 2.94; p = 0.003; and Dori ANOVA, F_2.416 = 2.39; p = 0.01; Fig. 3). Regardless of the commune, it was greater in the sub-block that received no treatment (Fig. 3). In Djibasso, the average number of larvae per spike obtained with neem aqueous extract (1.33 larvae/spike) was significantly lower compared to treatment with Synthetic insecticide (where to find the result ie which figure and then the statistics, see above (ANOVA, F_2.439 = 2.94; p = 0.003). In contrast, no significant difference was observed between the neem aqueous extract treatment and Synthetic insecticide in the commune of Dori (Fig. 3).

Damage by H. albipunctella according to phytosanitary treatments

The average percentage of millet spike attacked by H. albipunctella was significantly greater at the level of the control treatment compared to those of the aqueous extract of neem and synthetic insecticide in the communes of Djibasso and Dori (Table 2). In addition, treatments with aqueous neem extract and synthetic insecticide reduced the average percentage of millet spike attacks by about 50.00% in both study sites. Regarding the average number of mines per spike, it varied significantly with location (Table 2). In Djibasso and Dori, the average number of mines per spike was lower from neem aqueous extract treatments in comparison to those from Synthetic insecticide and the control (Table 2). The mean length of mines, which reflects the extent of H. albipunctella damage, varied significantly at the level both communes (Table 2). In addition, in Djibasso and Dori, the mean length of mines per spike was lower at the level in the neem aqueous extract treatment than those of the synthetic insecticide and the control (Table 2).

Effect of the cropping system on the percentage millet per spike attacked and the number of mines per spike in the commune of Djibasso

The percentage of millet spikes attacked and the number of mines per millet spike were significantly influenced by the cropping association system in the commune of Djibasso (Table 3).

The highest and lowest significantly different percentages of damaged millet spike were observed on millet grown alone and on the combination of two rows of millet and one row of cowpea (Table 3).

The average number of mines per spike was significantly higher in the farmer’s practice compared with millet grown alone and the combination of two rows of millet and one row of cowpea (Table 3).

Millet grain yield by phytosanitary treatments

Grain yield of millet varied significantly at the level of commune of Djibasso (ANOVA, F_2.24 = 3.53; P = 0.0096; and Dori: F_2.24 = 2.34; P = 0.00473; Fig. 4). In contrast, regardless of the commune, millet grain yield was statistically similar between the neem aqueous extract and Synthetic insecticide treatments (Fig. 4). The phytosanitary treatment, in both communes, made it possible to obtain more than one ton of millet per hectare compared to the control treatment which was less than one ton (Fig. 4).