The study was conducted in Maradi at the experimental site of the Regional Center for Agronomic Research of Maradi (CERRA/INRAN) during the 2020 and 2022 cropping seasons. The INRAN station of Maradi is located in the southern Sahelian zone of Maradi city.

During the two experimental years, the cowpea variety IT90K-372-1-2 was used. It is resistant to aphids, sensitive to thrips, but resistant to drought (300 - 600 mm) with a vegetative cycle of 70 to 75 days.

The cowpea plants were naturally infested by insect pests on site, including the pod bug (C. tomentosicollis), the cowpea pod borer (M. vitrata), and thrips (M. sjostedti).

The neem seed extracts and the entomopathogenic fungus Beauveria bassiana 115 were used in trials conducted in 2020 and 2022. The doses per hectare of the products used are presented in Table 1 .

A Fischer block with 4 treatments and 6 repetitions was established. The treatments were randomly assigned to plots. The seeds were sown at a density of 0.75 m × 0.30 m. Each plot consisted of 9 rows and 20 stands per row, totaling 180 stands per plot. The blocks are 3 m apart, with 2 m between plots. The nine rows are arranged for data collection as follows:

The two central rows were kept for grain yield evaluation;

The other two internal rows near the central lines were used for insect pest records. Five randomly selected stands were marked for insect pest data recording. The treatments were as follows: synthetic pesticide, neem seed extracts, Beauveria bassiana 115, and a control without pesticide application.

Land preparation was done with a tractor before the cropping season, and manure was applied for both the 2020 and 2022 campaigns. Sowing occurred on June 12^th, 2020, and June 14^th, 2022. Three weeding were conducted each year. A microdose fertilizer application of NPK 15:15:15 was used. Three phytosanitary treatments were applied in both campaigns. The first spraying occurred at the beginning of the flower buds’ emergence, and subsequent treatments were applied weekly. Harvesting involved manually collecting and weighing the pods from the two rows dedicated to yield and other rows.

Observations were conducted on 5 stands selected in each plot every three days. The following parameters were noted: total number of flowers per cowpea stand; for Thrips and larvae of M. vitrata, 10 flowers were picked per plot, including 2 flowers at each marked stand (5 stands/plot), and placed in a 90% alcohol solution, then brought to the laboratory and dissected to record the number of insect species; number of pod bug larvae and adults per observation; total number of pods per stand; number of pods damaged by M. vitrata; number of pods damaged by pods bugs; number of M. vitrata larvae in damaged pods. At crop maturity, the pods were harvested and dried. They were threshed, and the seeds obtained from each plot were weighed. The yield was calculated and extrapolated to a hectare. Data transformation was done using the formula:

$Tr=ASIN\sqrt{Proportion}$ (1)

The proportion was obtained by dividing the rate by 100 (Rate/100). The pod infestation rate by insect pests was calculated using the following formula:

$\text{T}=\left(\text{number of pods damaged}*100\right)/\text{Total number of pods}.$ (2)

The yield was calculated based on the following formula:

$\text{R}=\left(\text{Weight of seeds harvested in kg}\right)/\text{surface unit}.$ (3)

The means and associated standard errors of all variables were calculated using SPSS20 (Statistical Package for Social Science) software. The analysis of variance (ANOVA) for pod infestation rates by insect pests was conducted after the ASIN transformation to compare the average insect and yield. A non-parametric test (Kruskal-Wallis) was used to compare the rates of pod infestations by insect pests.

Thrips were noted from the 43^rd day after sowing (DAS) at the flowering stage during the 2020 cropping season and from the 30^th DAS for 2022 campaign. Maximum densities were observed at the 64^th DAS for 2020 and at the 52^nd DAS for 2022 campaign ( Figure 1 ). During these peaks, the density was highest in the control treatment for both years and lowest for the pesticide treatment, followed by the Neem seed treatment.

The larvae of M. vitrata were noted from the 28^th DAS with the first emerged flowers during 2020 and 2022 agricultural campaigns. The control treatment had the highest density, while the chemical treatment had the lowest population, followed by neem seeds extracts in 2020. In 2022 campaign, the peak of M. vitrata was recorded at the 40^th DAS for all the treatments.

The Adults and larvae of C. tomentosicollis were observed from the 70^th and 58^th days after sowing (DAS) in 2020 and 2022, respectively, but at low densities. The maximum infestations were noted on the 88^th DAS for the pesticide treatment in 2020 campaign and at the 76^th DAS in 2022 and the 79^th day after sowing for the neem seed extracts. A proliferation of the pod bug was noted during the harvest periods for both years, even after the biocides were sprayed.

In the 2020 agricultural campaign, the cumulative average number of thrips observed per 10 flowers varied between the treatments. The highest infestation was recorded in the control plots and those treated with B. bassiana. The second group consisted of treatment with neem seed extract, which had moderate infestations. The thrips population was 2.2 times lower compared to the untreated control plot ( Table 2 ).

Means followed by the same letter are not significantly different (LSD, 5%). The same below.

Plots treated with the synthetic chemical pesticide recorded the lowest level of infestation. The density of thrips was 10.4 times lower compared to the control treatment.

In the 2022 agricultural campaign, the density of Thrips was low with chemical treatment and was comparable to the other treatments ( Table 2 ).

For M. vitrata, the statistical analyses identified three groups during the 2022 agricultural campaign: The first group consists of the control treatment, where the highest levels of infestation were noted. The second group consists of the B. bassiana treatment, which showed a moderate infestation with densities 1.5 times lower compared to control treatments.

The third group consists of neem seed extract and synthetic chemical pesticide treatments. The densities of M. vitrata for these treatments ranged from 3.2 to 5.9 times lower compared to the control treatment.

For the 2022 agricultural campaign, all treatments were statistically similar.

For C. tomentosicollis, pest density per treatment was higher in the control plots ( Table 3 ). The number of larvae and adults of the pest was 3.5 to 8.1 times lower with the biocide compared to the control. In the 2022 cropping season, all treatments were comparable but differed from the control, where the density was maximal ( Table 2 ).

The C. tomentosicollis infestations varied depending on the treatments in 2020 and 2022. In 2020, the infestation level in the control plot was the highest, followed by B. bassiana, which was slightly lower than the control. The infestation was moderate with neem seed extracts and low with the synthetic pesticide treatment.

In 2022, infestations were higher in the control treatment and similar between the treated plots ( Table 3 ).

The seed production of the plots varied between treatments. The yield obtained in plots treated with the entomopathogenic fungus B. bassiana was comparable to that obtained in the control treatment. The neem seed extracts recorded a production 3.1 times higher than that obtained in the control plots. The highest production was from plots treated with the synthetic chemical pesticide, with a yield 5.6 times higher than the control plots.

In 2022, the yield of fungus treatment B. bassiana was comparable to neem seed extracts and greater than the control. The highest yield was obtained with pesticide treatment, which was 8.6 times higher than the control ( Table 4 ).