The study was conducted on the experimental plot of Nangui Abrogoua University (UNA). Geographically the experimental field was located at: 5˚17'N and 5˚31"N latitude and 3˚45'W and 4˚31'W longitude. The site was characterized by moderate and high temperatures (24.51˚C to 27.67˚C) and relative humidity of 80% during the experiment.

Damani violet of the Galmi type acquired at semi-variety was used as plant material for this study. This variety reaches maturity between 100 and 110 days.

On the other hand, pathogenic Fusarium sp. fungal material was isolated from onion bulbs collected at the Adjame wholesale market showing white mycelium and rotting symptoms.

The preparation of the microbial solution (EM) was essentially carried out in three steps following the method described by [14] .

The first step consisted of precooking the six starch sources separately and spreading them in perforated trays. The trays were then covered with fine cloth and placed individually in pits 20 to 30 cm deep, protected from disturbance and covered with Chinese bamboo leaf, then black bag and finally soil, for five days. After this incubation time, the contents of each tray were collected after eliminating areas with a grey, brown or black coloration characteristic of the presence of harmful microorganisms, while retaining those with a pink, green, white, blue, yellow or orange coloration for EM preparation.

The starch sources colonized by the microorganisms (15 kg) were mixed with 15 kg of brown sugar and stored in a hermetically sealed 70 L plastic barrel under a shed for 15 days, after which the mixture became compact.

Next, 2 kg of unbleached local rice were washed with rainwater. After five days in a closed container, one liter of this washing water was added to 10 L of fresh cow’s milk in a plastic barrel. The container was hermetically sealed and stored for seven days away from heat and light for fermentation, after which only the liquid part was extracted.

Finally, the dough and liquid obtained in steps 1 and 2 respectively were mixed, to which 100 g of baker’s yeast and 20 liters of rainwater (chlorine-free) were added. Seven days later, the mixture was filtered and stored in a 200-liter plastic container, out of direct sunlight.

The isolation and enumeration of fungi, bacteria and yeasts from the EM solution was carried out using the suspension dilution method [15] .

From 10 mL of EM extract, vortexed for 20 min and adjusted to 100 mL in a sterile Beaker, the stock solution (1/10, i.e. 10⁻¹) is obtained. Using a sterile graduated pipette, 1 mL of the stock suspension is removed and transferred to a sterile test tube containing 9 mL of sterile distilled water, numbered according to the 10⁻² dilution.

All further dilutions up to 10⁻⁸ are obtained by the same procedure. One µL of each dilution was taken and spread on the surface of the agar culture medium (PDA) contained in the Petri dish and incubated for 48 h at 28˚C. Only Petri dishes with a dilution of 10⁻⁸ were used.

As for Fusarium, isolation was carried out on PDA culture medium according to the method used by [16] . Explants were collected in the laboratory from the growth front of infected onion bulb parts and disinfected for three minutes with a 2% sodium hypochlorite solution. The fragments obtained were then rinsed three times in succession with sterile distilled water for 5 min, then placed on sterile blotting paper to remove excess water. These were then inoculated onto PDA culture medium at a rate of three fragments per dish. They were sealed with stretch tape and incubated for 48 to 72 hours at room temperature. Strains were purified after several successive subcultures on PDA medium.

Fungi were identified 7 days after seeding on PDA medium. Macroscopic and microscopic characteristics were recorded using the identification keys of [17] - [19] . As for microscopic characteristics, Gram staining was used to classify bacteria according to their Gram, morphology and mode of association. Gram-positive and catalase-negative bacteria are presumed to be lactic acid bacteria [20] . Bacterial cells were observed at ×1000 magnification using a Motic light microscope [21] .

Agar medium (PDA) at pH 3.5 (adjusted with 10% tartaric acid) was used for its ability to support colony formation by Saccharomyces cells that were contained in the EM solution.

The sample of EM solution was serially diluted in 0.1 M phosphate buffer (pH 7.0) before being spread in recovery agar using the casting plate technique. Colonies were counted after five days’ incubation at 29˚C [22] .

Nursery sowing was carried out under glass in three honeycombed trays, each with seventy-two holes filled with potting soil, purchased from SEMIVOIRE. Three seeds were then placed in each hole and covered with potting soil. Each plate received 600 mL of tap water per day until emergence, after which the seedlings were watered as required until transplanting. The insecticide Cypercal was applied once at a rate of 3.2 mL/600mL water for the three plates when caterpillar damage was observed three weeks after sowing.

The experimental set-up used was a Fisher block with complete randomization, comprising three blocks, i.e. three replicates. Each block was made up of nine sub-plots with nine treatments. The treatments were as follows: T0 (no inoculation), T01 (2.5% of EM + Fusarium sp.), T1 (1% of EM), T2 (5% of EM + Fusarium sp.), T3 (2.5% of EM), T4 (Fusarium sp.), T5 (1% of EM + Fusarium sp.), T6 (5% of EM) and T7 (Fusarium sp. + chemical fungicide) with a concentration of Fusarium sp. (F: 106 spore/mL).

Each subplot contained eleven (11) plants, for a total of 297 plants. Plants were spaced 0.33 m apart in width and 0.5 m apart in length. Subplots were spaced 0.5 m apart and 1 m between blocks.

The trial was conducted in a 7 × 6 × 3 (length × width × height) greenhouse, and nursery seedlings were exposed to the open air in a 14 m × 8 m plot. Forty-five days later, they were transplanted into two-liter plastic pots containing a substrate (sterilized in an oven for 30 minutes) made up of compost (pig droppings, legumes, poultry droppings, grass clippings) and UNA forest soil at a ratio of 1:2 (v/v). Plants received tap water as needed. Yara Mila fertilizer was applied at a dose of 3 g/pot two weeks after transplanting. Insecticide was applied at a rate of 40 mL/16L.

Inoculations were carried out in the greenhouse one month after transplanting, according to the treatments, while avoiding watering the plants on the day of inoculation. Only one inoculation per plant was made with the 106/mL spore suspension of the pathogen Fusarium sp. and four inoculations with EM. A single inoculation was made 24 h after inoculation of the plants with the pathogen, and then every week for three weeks. During the trial, the T0 control was treated with tap water only. Banko-plus was applied four times (30 mL/16L).

$\text{Incidence rate}\left(\%\right)=\frac{{{\displaystyle \sum }}^{\text{}}n}{N}\times 100$(1)

where: ${{\displaystyle \sum }}^{\text{}}n$: sum of diseased plants; N: total number of plants observed.

$\text{Severity}\left(\%\right)=\frac{{{\displaystyle \sum }}^{\text{}}\text{note}}{Nt\times 3}\times 100$(2)

with: ${{\displaystyle \sum }}^{\text{}}\text{note}$: sum of different notes for infected plants; Nt: total number of infected plants.

The agronomic parameters assessed were growth (plant height, number of leaves, plant collar diameter) and yield of fresh onion bulbs calculated according to the following formula:

$\text{Rendement}\left(\text{kg}\cdot {\text{ha}}^{-1}\right)\text{=}\frac{\text{Rendement parcellaire}}{\text{Superficie parcellaire}}\times 10000$(3)

All data collected was entered into an Excel spreadsheet. Graphs were designed for illustrative purposes. For comparative analysis, data were tested for normality and homogeneity using the Shapiro-Wilk and Leven tests. If these conditions were not met, the Kruskal-Wallis test was performed. Otherwise, the Newman-Keuls multiple comparison test was used to compare means when the analysis of variance (ANOVA 1) revealed significant differences. XLSTAT version 2016 software was used for these analyses, and statistical tests were performed at the 5% probability threshold.

Four fungal genera were identified from the EM solution using the key [18] . These were the genera Trichoderma, Aspergillus, Penicillium and Rhizopus. The genera Trichoderma and Aspergillus presented several morphotypes with different cultural aspects ( Figures 1-4 ).

Rhyzopus sp. colonies were white filaments with black spores at the tip. Microscopically, they showed an unpartitioned, unbranched mycelium with spores and branched sporangiophores at the end of the mycelium. Spores are found on the aerial part of the mycelium and surrounded by vesicles ( Figure 3 ).

The colony was flat and flaky, green-grey in color. Microscopically, the thallus, formed of hyaline septate mycelial filaments, bears conidiophores. Phialides are arranged in whorls at the tips of the conidiophores, and the latter are grouped in bundles ( Figure 4 ).

Colonies grown on MRS medium were small, white and round in shape. They were Gram-positive and did not produce hydrogen peroxide (H₂O₂) ( Figure 5 ).

Colonies were white, creamy, smooth and shiny. Microscopically, the cells were ovoid ( Figure 6 ).

Incidence rates varied according to the treatment applied ( Figures 7-8 ). Plants inoculated with Fusarium spores and treated had a lower incidence rate than those inoculated with Fusarium spores only. However, treatment T2 inoculated and treated with 5% EM recorded an incidence rate of 17.90%, statistically identical to that of T7 (19.25%) inoculated and treated with chemical fungicide. The highest incidence rates were recorded in treatments T4 (54.42%), T5 (31.18%) and T01 (22.56%). Low incidence rates were recorded in treatments T6, T3, T1 and T0.

Figure 9 shows the different health conditions of the plants while Figure 10 shows the severity of the disease depending on the treatments. Treatments showed different levels of disease severity, ranging from 16.67 to 70.31%. Treatment T2 showed the lowest severity rate (35.23%) compared with control T4 (70.31%), i.e. a reduction of 35.08%. On the other hand, disease severity was higher in treatments receiving no dose of EM or chemical fungicide. In these

treatments, disease severity was between 65 and 70%. As for treatments that received only the various doses of EM, their severity ranged from 12% to 18%.

The effect of treatments on the three growth parameters is listed in Table 1 .

In the same column, the averages for number of leaves, plant height and plant diameter followed by the same letter show no significant difference at the 5% threshold (Newman-Keuls test). NF: number of leaves; H: leaf length; D: diameter at collar; T0: control; T01: Fusarium sp. + 2.5% EM; T1: 1% EM; T2: Fusarium sp. + 5% EM; T3: 2.5% EM; T4: Fusarium sp.; T5: Fusarium sp. + 1% EM; T6: 5% EM; T7: Fusarium sp. + chemical fungicide. Mean ± standard deviation.

Production parameter data are reported in Table 2 . The means of the parameters (bulb diameter and bulb weight) between treatments show a significant difference (p < 0.035 and p < 0.041) at the 5% threshold.

In the same column, the means of bulb diameter, bulb weight and yield followed by the same letter show no significant difference at the 5% threshold (Newman-Keuls test). Dm Bulb: bulb diameter; Pds bulb: bulb weight; T0: control; T01: Fusarium sp. + 2.5% EM; T1: 1% EM; T2: Fusarium sp. + 5% EM; T3: 2.5% EM; T4: Fusarium sp.; T5: Fusarium sp. + 1% EM; T6: 5% EM; T7: Fusarium sp. + chemical fungicide. Mean ± standard error.