Investigations were carried out in the field from May to September 2018 and 2019 at Dang (latitude 7˚42.264'N, longitude 13˚53.945'E and altitude 1106 m above sea level) in Ngaoundere III Subdivision, Vina Division, Adamaoua Region in Cameroon. The site belongs to the high-altitude Guinean savannah agroecological zone. The climate is characterized by a rainy season (April to October) and dry season (November to March), with an annual rainfall of approximately 1500 mm. The mean annual temperature is 22˚C, while the mean annual relative humidity is 70% [15]. The animal material was mainly represented by insects naturally present in the environment and 35 colonies of Apismellifera Linnaeus (Hymenoptera: Apidae) found close to the experimental field. The flora surrounding G. max field had various unmanaged and cultivated species. Compost was produced in the Composting Unit established and monitored at the Faculty of Science of the University of Ngaoundere. Approximately 1000 g of compost per hole were applied as a layer into sowing hole before sowing. Glycine max seeds (Figure 1) were provided by the Institute of Research for Agricultural Development (IRAD) at Wakwa-Ngaoundere (life cycle of 155 to 160 days). The fertilizer-NPK used was of the formula 20:10:10, purchased from a local phytosanitary store. It was applied 14 days after sowing, at a rate of 10g within the rhizosphere of each plantlet.

On May 7, 2018 and May 15, 2019, experimental soil was plowed and divided into nine subplots of 3 m² each. Three subplots were applied with compost, three with chemical fertilizer-NPK and three others left unapplied neither with compost nor with fertilizer-NPK. Two seeds were sown per hole on three lines per subplot, for a total of 12 holes per line. Holes were separated 25 cm from each other, while lines were 30 cm apart. Weeding was performed manually as necessary to maintain subplots weed-free.

On July 24, 2018, six subplots carrying 216 plants with 16,150 flowers at the bud stage were labeled. Three subplots carrying 108 plants with 8246 flowers were left open to be pollinated by insects (treatment 1) (Figure 2), while three others carrying 7904 flowers were protected with white gauze cages (1 mm² mesh) to prevent insect or other pollinating animals visits (treatment 2) (Figure 3). On July 27, 2019, the experiment was repeated. For treatment 3, there are three subplots carrying 108 plants with 10,130 flowers and for treatment 4, there are three subplots carrying 108 plants with 9873 flowers. Twenty days after shading of the last flower, the number of pods was assessed in each treatment. The fruiting index (P_i) was then calculated as described by [16]: P_i = F₂⁄F₁, where F₂ is the number of pods formed and F₁ the number of viable flowers initially set. The allogamy rate (Alr) from which derives the autogamy rate (Atr) was expressed as the difference in fruiting indexes between treatment X (unprotected flowers) and treatment Y (bagged flowers) as follows [17]: Alr = [((P_iX − P_iY))/(P_iX)]*100, Where P_iX and P_iY are respectively the podding average indexes of treatments X and Y Atr = 100 − Alr.

For each uninoculated subplot (treatment a), subplot inoculated with compost (treatment b), and (subplot applied with fertilizer-NPK (treatment c), sampling for the assessment of plant biomass was done on 15 randomly selected plants per elementary plot at 60 days after planting (DAP), enumerated, sun dried, stored in envelopes, and weighed. Plants were dried in an oven at 72˚C for 12 hours and weighed [18]. Plant biomass and nodulation were evaluated on the same 45 individual plants of treatments a, b and c.

The frequency of insect visits on G. max flowers was evaluated based on observations scheduled on four daily time frames (09:00-10:00 am, 11:00-12.00 am, 13:00-14:00 pm and 15:00-16:00 pm) in all treatments. From July 27^th to August 27^th 2018 and from July 30^th to August 31^th 2019, flowers were entirely opened at 09:00 am and closed before 16:00 pm, corresponding to the period of insects activity. All insect visits were recorded on flowers of treatment 1. Specimens of all insect taxa (3 to 5 per species) caught with an insect net on flowers were conserved in 70% ethanol, except Lepidoptera that were kept in curls for subsequent taxonomy determination. All insects observed on flowers were noted, and the cumulated results were expressed in number of visits to determine the relative frequency of each insect species in the anthophilous entomofauna of G. max [19]. In addition to the determination of the floral insects’ frequency, direct observations of the foraging activity on flowers were made on each insect species in the experimental field. Nectar or pollen harvested by insects during each floral visit was registered based on their foraging behavior [11].

In the morning of each sampling date, the number of opened flowers was counted, whereas the duration visits of each insect were recorded (using a stopwatch) for at least three times during each of the following daily time frames: 10:00 am-11:00 am, 12:00 am-13:00 pm and 14.00 am-15:00 pm. Moreover, the number of pollinating visits [19], the abundance of foragers [20] and the foraging speed referring to the number of flowers visited by an insect per minute [21] were determined.

Abundance of insects per flower was recorded following direct counts. For the abundance per 1000 flowers (A₁₀₀₀), the number of foragers was counted at blooming flowers on the same dates and daily periods as for the registration of the duration of visits. (A₁₀₀₀) was calculated by the formula: A₁₀₀₀ = [(A_x/F_x)*1000], where F_x and A_x are the number of opened flowers and the number of insects effectively counted on these flowers at time x [19]. The foraging speed was calculated by the formula: V_b = [(F_i/d_i)*60], where d_i is the time (sec) given by a stopwatch, and F_i, the number of flowers visited during d_i.

Around experimental plot, the disruption of the activity of each insect forager by competitors and the attractiveness exerted by other plant species on G. max insect foragers were assessed. Ambient temperature and relative humidity were recorded at each observation date after every 30 minutes, using a portable thermo-hygrometer (HT-9227).

From start of the flowering of the first flower to the wilting of the last flower, bloomed flowers of treatment 1 were counted. Data obtained were compared with the number of insect visits on the corresponding flowers.

The estimation of this parameter was based on the effect of compost on G. max yield. The comparison of productivity (fruiting rate, mean number of seeds per pod and percentage of normal seeds) of treatments 2 (bagged flowers) and 5 (bagged flowers inoculates by compost) for the first year, 4 (bagged flowers) and 6 (bagged flowers inoculates by compost) for the second year were assessed as influenced by compost on soybean plants.

This evaluation was based on the impact of both compost and insects on G. max yield. The comparison of yields (fruiting rate, mean number of seed per pod and percentage of normal seeds) of treatment 9 and 10 with those of treatments 2 and 4 were assessed. The contribution of cumulative action of insects and compost on soybean fruiting rate, mean number of seeds per pod and the percentage of normal seeds was calculated using data of treatment 9 or 10 (inoculated flowers open to insects) and those of treatment 2 or 4 (bagged flowers).

To analyze the data we used Microsoft Excel 2010 software and four test: Student’s (t) for comparison of means of two samples, correlation coefficient (r) to determine the linear relationship between two variables, Chi-square (χ²) to compare two percentages and Statgraghics Centurion for the comparison of means of more than two samples.

Table 1 indicates that the allogamy rate was 14.01% and 12.73%, respectively in 2018 and 2019, whereas the autogamous rate was 85.99% and 87.28% respectively in 2018 and 2019. For the two cumulative years, the allogamy rate was 13.37% and the autogamy rate was 86.63%. Thus, soybean variety used in this experiment has a mixed autogamous allogamous reproduction mode with the predominance of autogamy.

Plants inoculated with compost at sowing produced a significantly greater number of nodules, nodule dry weight and plant biomass compared to uninoculated plants in 2018 and 2019 (Table 2).

At Dang in 2018 and 2019, 948 and 593 visits of six insect species belonging to three orders were counted on 8246 and 10130 flowers respectively. Table 3 presents the list of the insects with their percentages of visits. This table shows that Hymenoptera was the most important order with 82.09% of 1265 visits. The most represented family was Halictidae, among which Lipotriches collaris ranked first with 43.87%; Diptera and Lepidoptera were poorly represented with 4.80% and 13.11% of visits each for all the treatment and the two years respectively.

A highly significant difference was obtained between flowers from the control and those from the compost plants (P < 0.001 in 2018 and 2019) and between the control and chemical fertilizer (P < 0.001 in 2018: 2019).

In each column, the means followed by the same letter are not significantly different at 5% level.

PC: subplot with compost; PE: subplot with fertilizer-NPK; PN: uninoculated subplot; n₁ and n₂: number of visits on 8246 and 10,130 flowers in 23 and 25 days respectively; sp.: undetermined species; P₁ and P₂: percentages of visits P₁ = (n₁/929)*100 and P₂ = (n₂/593)*100.

The abundance, the foraging speed and the duration of insect visits were focused on the two major flowers insects visiting of Lipotriches collaris and Ceratinasp.

The number of insect visits is proportional to the number of opened flowers on untreated, compost and chemical subplots of soybean (Figure 4). A positive and significant correlation was found between the numbers of opened flowers and the number of insect visits on flower of uninoculated plants (r = 0.44; P < 0.05) in 2018 and 2019 (r = 0.42; P < 0.05), compost affixed plants in 2018 (r = 0.64; P < 0.05) and 2019 (r = 0.86; P < 0.05), and fertilizer-NPK applied plants in 2018 (r = 0.57; P < 0.05), and 2019 (r = 0.47; P < 0.05).

Insects foraged were abundant on soybean flowers in the afternoon, with a daily pic of activity situated between 01:00 pm and 02:00 pm (Table 4). This activity was influenced by ambient temperature, but not by hygrometry. The correlation between the number of insect visits and the temperature was positive and significant on untreated (r = 0.89; P < 0.05), applied compost (r = 0.96; P < 0.05), and fertilizer-NPK applied (r = 0.88; P < 0.05) plants for both years. As for the relative humidity, the correlation with the number of insect visits was negative and not significant on untreated (r = −0.32; P > 0.05), applied compost (r = −0.43; P > 0.05), and chemical fertilizer applied (r = −0.40; P > 0.05) subplots of this crop.

n: number of visits in 23 and 25 days; p: percentage of visits; p = (n/A)*100; A: total number of insect visits; ^*: daily peak of visit. PC: subplot with compost; PE: subplot with fertilizer-NPK; PN: uninoculated subplot.

In 2018 and 2019, the largest number of individuals simultaneously active on a flower was 1 for all the treatments. For both years, the mean abundance per 1000 flowers (MATF) ranged from 35.87, 51.57 and 30 for Ceratina sp. to 37.09, 41.46 and 33.71 for Lipotriches collaris respectively on negative control, compost and fertilizer-NPK applied plants (Table 5).

In 2018 and 2019, the mean duration of insect’s visit varies from 2.19, 1.75 and 3.15 with Ceratina sp. to 2.99, 2.70 and 2.58 with Lipotriches collaris on untreated, compost and chemical fertilizer subplots respectively (Table 6).

The foraging speed’s mean of insects on soybean flowers was 10.21, 10.49 and 6.18 with Ceratina sp. to 57.35, 8.81 and 24.73 with Lipotriches collaris on untreated, compost and fertilizer-NPK subplots respectively (Table 7).

PC: subplot with compost; PE: subplot with fertilizer-NPK; PN: uninoculated subplot; m: average; s: standard deviation; n: sample size.

PC: subplot with compost; PE: subplot with fertilizer-NPK; PN: uninoculated subplot; m: average; s: standard deviation; n: sample size.

PC: subplot with compost; PE: subplot with fertilizer-NPK; PN: uninoculated subplot; m: average; s: standard deviation; n: sample size.

During anthesis of soybean, there was low nectar harvesting activity of A. mellifera workers on flowers. This result reveals the low attractiveness of soybean nectar to A. mellifera workers and consequently allows the classification of this plant species among the low bee plant species.

During nectar harvest on soybean, foraging insects always shook flowers and were regularly in contact with anthers and stigma, thus, increasing cross-pollination possibility of G. max fruiting rate, number of seeds per pod and percentage of normal seeds in different treatments (Table 8).

There was a highly significant between fruiting rate of free opened flowers (treatment 1) and that of bagged flowers (treatment 2), the first year (P < 0.001) and the second year (P < 0.001). The fruiting rate of treatment 1 (unprotected flowers) was higher than treatment 2 (protected flowers) in 2018 and in 2019. The fruiting rate due to the action of insects was 19.09 and 18.54% in 2018 and 2019 respectively. For the two cumulated years, the fructification rate due to the influence of insects was 18.81%.

The mean number of seeds per pod was highly significant between treatments 1 and 2 (P < 0.001). Consequently, a high mean number of seeds per pod in treatment 1 (unprotected flowers) was noticed compared to treatments 2 (bagged flowers). The number of seeds per pod attributed to the activity of insects was 19.27% in 2018 and 21.77% in 2019, giving an overall mean of 20.52%.

There was a highly significant difference between the percentage of normal seed of treatment 1 and that of treatment 2 in the first year (P < 0.001), as well as the second year (P < 0.001). Thus, the percentage of normal seeds in opened flowers was higher than that of protected flowers in 2018 and 2019. The percentage of the normal seeds due to the action of insects was 14.09% in 2018 and 17.58% in 2019. For all the flowers studied, the percentage of the normal seeds due to flowering insects was 15.83%.

NF: Number of flowers; NFP: Number of formed pod; FrR: Fruiting rate; TNS: Total number of seeds; NS: Normal seeds; % NS: Percentage of normal seeds; m: Mean; s: Standard deviation.

The comparison of the fruiting rate (Table 9) showed that the differences observed were highly significant between treatments 2 and 5 (P < 0.001) and treatments 4 and 6 (P < 0.001). The fruiting rate due to compost was 26.86% in 2018, 24.57% in 2019 and 25.71% for the two cumulated years.

The comparison of the mean number of seeds per pod revealed that differences observed were highly significant between treatments 2 and 5 (P < 0.001) and between treatments 4 and 6 (P < 0.001). The mean number of seeds per pod due to compost was 31.07% in 2018, 33.83% in 2019 and 32.45% for the two cumulated years.

Hence, in 2018 and 2019, the percentage of the normal seeds from flowers of plants protected and applied with compost (treatments 5 and 6) was higher than that of Bagged flowers and uninoculated (treatments 2 and 4). The percentage of the normal seeds due to compost was 19.40% in 2018, 21.60% in 2019 and 20.50% for the two cumulated years.

NF: Number of flowers; NFP: Number of formed pod; FrR: Fruiting rate; TNS: Total number of seeds; NS: Normal seeds; % NS: Percentage of normal seeds; m: Mean; s: Standard deviation.

The comparison of the fruiting rate (Table 10) showed that the differences were highly significant between treatments 7 and 2 (P < 0.001) as well as 8 and 4 (P < 0.001). the fruiting rate from flowers of plants protected and applied with fertilizer-NPK (treatment 7 in 2018; treatment 8 in 2019) was higher than that from flowers of plants protected and uninoculated (treatment 2 in 2018; treatment 4 in 2019). The fruiting rate due to fertilizer chemical was 20.36% in 2018, 21.73% in 2019 and 21.04% for the both years of study.

The comparison of the mean number of seeds per pod (Table 10) showed highly significant differences between treatments 7 and 2 (P < 0.001), as well as 8 and 4 (P < 0.001). Pod and seed yields from flowers of plants protected and applied with fertilizer-NPK (treatment 7 in 2018; treatment 8 in 2019) were higher than that from flowers of plants protected and uninoculated (treatment 2 in 2018; treatment 4 in 2019). The mean number of seeds per pod due to fertilizer chemical was 22.58% in 2018, 34.95% in 2019 and 28.76% for both years of study.

The comparison of the percentages of normal seeds shows that the differences were highly significant between treatments 7 and 2 (P < 0.001), and treatments 8 and 4 (P < 0.001). The percentage of normal seeds from flowers of plants protected and applied with fertilizer-NPK (treatment 7 in 2018; treatment 8 in 2019) was higher than those protected and uninoculated (treatment 2 in 2018; treatment 4 in 2019). The percentage of the normal seeds due to fertilizer chemical was 13.25% in 2018, 16.55% in 2019 and 14.90% for both years of study.

NF: Number of flowers; NFP: Number of formed pod; FrR: Fruiting rate; TNS: Total number of seeds; NS: Normal seeds; % NS: Percentage of normal seeds; m: Mean; s: Standard deviation.

The comparison of the podding rate (Table 11) showed that the differences observed were highly significant between treatments 9 and 2 (P < 0.001) and treatments 10 and 4 (P < 0.001). Therefore, in 2018 and 2019, the podding rate from plants applied compost and opened to insects (treatments 9 and 10 respectively) was higher than that of flowers protected during their flowering period (treatments 2 and 4 respectively). The cumulative effect of flowering insects and compost on the podding rate was 29.95% in 2018, 26.59% in 2019 and 28.27% for the two years of study.

As far as the mean number of seeds per pod is concerned, there were highly significant between treatments 9 and 2 (P < 0.001) and treatments 10 and 4 (P < 0.001). As a matter of fact, in 2018 and 2019, the mean number of seeds per pod from flowers of applied compost subplot and opened to insect pollinators (treatment 9 in 2018, treatment 10 in 2019) was higher than that of flowers bagged (treatment 2 in 2018, treatment 4 in 2019) during their flowering period. The combined effect of flowering insects and compost activity on the number of seeds per pod was 44.42 % in 2015, 36.42 % in 2016 and 40.42 % for the two cumulative years. For the percentage of normal seeds, the differences observed were highly significant between treatments 9 and 2 (P < 0.001) and treatments 10 and 4 (P < 0.001). Hence, in 2018, as well as 2019, the percentage of normal seeds from flowers opened to insects on applied compost plants (treatment 9 in 2018, treatment 10 in 2019) was higher than that of flowers protected from insect visits (treatment 2 in 2018, treatment 4 in 2019). The percentage of normal seeds due to cumulative effects of flowering insects and compost activity was 23.99 % in 2018, 24.95 % in 2019 and 24.47 % for the two years of study.

NF: Number of flowers; NFP: Number of formed pod; FrR: Fruiting rate; TNS: Total number of seeds; NS: Normal seeds; % NS: Percentage of normal seeds; m: Mean; s: Standard deviation.

The podding rates were 87.30%, 62.67%, 88.37% and 65.49% in treatments 11, 2, 12 and 4 respectively (Table 12). When treatments were compared two by two, the difference observed was highly significant between treatments 11 and 2 (P < 0.001) and between treatments 12 and 4 (P < 0.001). Hence, in 2018 and 2019, the podding rate from plants applied with fertilizer-NPK and opened to insects (treatments 11 and 12) was higher than that from flowers of plants protected (treatments 2 and 4).

The mean numbers of seeds per pod were 3.73, 2.64, 3.69 and 3.09, 2.96 in treatments 11, 2, 12 and 4 respectively (Table 12). The difference observed was significant between treatments 11 and 2 (P < 0.001) as well as between treatments 12 and 4 (P < 0.001). Thus, in 2018 and 2019, the mean number of seeds per pod from plants applied with fertilizer-NPK and opened to insects was higher than that from flowers of plants protected.

The percentages of normal seeds were 90.88%, 70.05%, 72.36%, 91.89% and 69.51% in treatments 11, 2, 12 and 4 respectively (Table 12). Pairwise comparisons showed that the difference observed was highly significant between treatments 11 and 2 (P < 0.001) as well as between treatments 12 and 4 (P < 0.001). For both cropping seasons, the percentage of normal seeds from plants applied with chemical fertilizer and opened to insects was higher than that from flowers of plants protected.

In 2018, the contribution of cumulative effects of flowering insects and fertilizer-NPK in the podding rate, the mean number of seeds per pod and the percentage of normal seeds were 28.21%, 29.22% and 22.92% respectively. In 2019, the corresponding figures were 25.89%, 16.26% and 24.35%. For the two cumulated years, the numeric contribution of combined effects of flowering insects and chemical fertilizer were 27.05%, 22.74% and 23.63% for the podding rate, the mean number of seeds per pod and the percentage of normal seeds respectively.

NF: Number of flowers; NFP: Number of formed pod; FrR: Fruiting rate; TNS: Total number of seeds; NS: Normal seeds; % NS: Percentage of normal seeds; m: Mean; s: Standard deviation.

Analysis of the results of the effect of organic manure enrichment on soybean growth parameters showed a significant correlation between the number of nodules, weight of dry nodules and biomass of plants enriched with compost, as compared to those enriched with chemical fertilizer or untreated. Successful nodulation of leguminous crops largely depends on compost, which helps improve the soil’s ability to hold water, an essential factor in the nodulation process [22]. Compost also created a favorable environment for root development, improve the accessibility of plant roots to phosphorus by making the nutrient available, and reduce crop stress related to factors such as soil acidity [23]. This result is in line with that of [24], who reported an increase in the number and weight of nodules obtained with the addition of organic fertilizer in the form of chicken droppings.

At Dang, during the two cropping season, Hymenoptera were the most important order with 82.09% of 1541 visits. They were mainly represented by Halictidae family, the most important being Lipotrichescollaris (43.87%). Diptera and Lepidoptera were poorly represented with 4.80% and 13.11% of visits each. These results are similar to those recently obtained by [25] indicating that at Maroua, among soybean entomofauna, Hymenoptera were the most important Order (with 52.59% of visits in 2015 and 24.62% in 2018), among which, Halictidae ranked first (23.58% in 2015 and 5.86% in 2018).

The high frequency of Halictidae on soybean flowers could be explained by the good attractiveness of its nectar and/or its pollen vis-à-vis of these insects, by its accessibility, its availability and also by the presence of their nests in the experimental field. In addition, the fact that at the flower level, Halictidae only harvests nectar and/or pollen suggests that the floral products of this plant may have stimuli responsible for the attractiveness exerted on them. These are mainly olfactory and taste stimuli. Indeed, for a given plant, the attractiveness of the nectar depends in part on the average concentration of total sugars it has at the level of the flowers, whereas the pollen has its own clean odor [26], which can be detected by insects, using antennae and palps [27] [28]. [29] further indicated that the smell of pollen was involved precisely in its localization by the insect, its regular harvest being mainly under the influence of taste stimuli. These results confirm those already reported by [25], who revealed L.collaris as the most frequent insect on flowers of the same plant in Maroua. On the other hand, [30] and [31] reported that Apismellifera was the main floral visitor of this Fabaceae, and was able to collect nectar exclusively on G. max in USA and Brazil respectively. This finding shows on one hand that plants have specific food resources available to insects through flowers, and on the other hand that the diversity of soybean insects may vary in time and space.

The abundance of insect visits on flowers was higher on subplots enriched with compost than on uninoculated subplot. This could be explained by higher number of flowers on subplot enriched with compost. The peak activity of insects on soybean flowers was between 01:00 pm and 02:00 pm, which corresponds probably to the period of higher availability of nectar and/or pollen on this crop. Between 7:00 am and 8:00 am time slot, no insect visits were recorded. This period could correspond to the time when flowers of this Fabaceae are not yet well bloomed. The reduction of insects activity observed on flowers after 4 pm could be linked to the low quantity and/or quality of their respective floral products and to the drop in ambient temperature in the experimental field, thus causing the closure of flowers. In fact, several foraging prefer hot periods for their activity on flowers [32]. As in all legumes [33], the leaflets of soybean are arranged vertically in broad daylight, this facilitating the exposure of flowers to insects. Such flowers can therefore receive a higher number of visits. This is true, as the localization of a flower by an insect depends more or less on the visual stimuli emanating from this organ [26] [29]. Similar observations were reported by [25] on soybean by foraging insects in Maroua. In fact, these insect species does not visit G. max flowers when they are poor in nectar after 03:00 pm. Moreover, according to [34], a higher temperature along with a very weak relative humidity has a negative influence on the activity of pollinators on flowers.

When harvesting pollen and/or nectar on soybean flowers, most insects were frequently in contact with the anthers and the stigma of visited flowers. They could therefore be directly involved in self-pollination, by putting pollen of one flower on to the stigma of the same flower. The fruiting rate, the percentage of seeds per pod and the percentage of normal seeds due to the influence of flowering insects were 25.71%, 32.45% and 20.50% respectively. These results indicate that flowering insects are not only important in the improvement of pod and seed yields of this plant, but they also play an important role in the production of seeds of good quality. According to [35], the more a flower receives pollen grains, the more it has the potential to turn into a bulky fruit containing many seeds. For this purpose, [11] pointed out that pollination by insects increases the fruiting rate, the percentage of the number of seeds per pod and the percentage of normal seeds of soybean by 35.87%, 73.09%, 31.1%, respectively. These percentages are high compared to those obtained in this study and could be explained by the presence of more pollinating species in their experimental plots.

The positive and significant contribution of compost in fruit and seed yields of soybean could be justified by its richness in nutrients such as phosphorus, nitrogen and potassium. In fact, these nutrients are involved in the correction of nutrient deficiencies, ensuring adequate nutrition and maintaining optimal soil fertility conditions, while improving the quality of crops [36]. Similarly, it could also be attributed to reduction of certain plant disease symptoms that have been reported to field application of compost and derived products [37].

The yields of the subplots that received chemical fertilizer exceeded those of the control plots. This suggests that the fertilizer applied under the crop had positive effects on yields. Indeed, this increase would be due to the improvement of the properties of the soil (in mineral elements), leading to a good development of the roots and a good assimilation of the nutrients released by them [38]. Similarly, it could also be explained by a difference between the nutrient balances provided by each treatment [39].

In our experience, compost and pollinating insects greatly increased the fruiting rate, number of seeds per pod and number of normal seeds by 28.27%, 40.42% and 24.47% respectively. Indeed, compost improves plant growth and the diffusion of nutrients to plants through microbiological processes [40]. Thus, legumes will satisfy, under the right conditions of symbiosis, most of their needs to ensure their growth, flowering and increase their production [41]. Moreover, to increase the possibilities of pollination, insects facilitate the release of pollen from the anthers for occupation of the stigma [42].