**4. Discussion**

#### *4.1. Farmers' Knowledge and Perceptions of S. frugiperda Attacks*

In this study, most maize farmers recognized the damage of *S. frugiperda* and were able to identify it at its larval stage. Some farmers (7.8 to 31.7%) were able to identify other development stages of *S. frugiperda*, including eggs, pupae, and adults. The identification of these development stages requires a better knowledge of the biology of the pest [4].

The first attack of *S. frugiperda* was recorded in 2015 by some farmers. According to Goergen et al. [2], the attacks of *S. frugiperda* were first reported in West and Central Africa in early 2016. From farmers' perception, it could be inferred that *S. frugiperda* was present in Benin before 2016, but its damage became significant from 2016.

As well, the study showed that farmers' knowledge and perceptions of *S. frugiperda* were associated with their membership in a farmer organization and their contact with research or extension services. Therefore, the institutional environment of farmers could play a crucial role in the sustainable managemen<sup>t</sup> of *S. frugiperda*.

#### *4.2. Damage Caused by S. frugiperda*

*S. frugiperda* was present throughout Benin despite the geo-climatic contrasts. The life cycle of the pest lasts, on average, 30 days and the optimal temperatures for adults and larvae are 25 ◦C and 30 ◦C respectively [32]. In Benin, the temperature varies between 24 and 31 ◦C depending on the climatic zones [29]. Thus, the country offers favourable climatic conditions for the permanent reproduction of this pest.

Yield losses caused by *S. frugiperda* averaged 797 kg of maize per hectare, or 49% of the average maize yield obtained by farmers before the invasion of *S. frugiperda*. This result corroborates forecasts by the Centre for Agriculture and Bioscience International (CABI), indicating that *S. frugiperda* could cause a loss of 40% of the average annual maize production in Benin [33]. In Kenya and Ethiopia, yield losses were greater. They ranged from 0.8 to 1 ton of maize per ha [28]. The differences in yield losses between countries could be explained by the levels of infestation which may depend on climatic factors, managemen<sup>t</sup> practices used by farmers and insecticide availabilities. In Nicaragua, Hruska and Gould [34] demonstrated a positive relationship between yield losses and levels of *S. frugiperda* infestation. For them, infestations of 55 to 100% of maize plants could cause yield losses ranging from 15 to 73%.

The severity of the pest attacks was not significantly associated with cropping systems. Andrews [35] showed that intercropping was less severely attacked by *S. frugiperda* than monocropping and that intercropping could reduce damage by up to 30%. Furthermore, Baudron et al. [36] demonstrated that frequent weeding and no-till sowing reduced the damage of *S. frugiperda*. As well, yield losses due to *S. frugiperda* attacks have been shown to vary with planting dates. Some farmers

in Kenya reported significant yield losses on late-planted maize plots compared to plots planted earlier [3].

Moreover, the severity of *S. frugiperda* attack was associated with the types of grown maize varieties. The attacks were more severe for modern varieties than traditional varieties. This means that the modern varieties used in Benin were not resistant to *S. frugiperda* attacks. Some papers reported that yield loss of modern varieties due to *S. frugiperda* was not significant when they received adequate fertilizers or when they were planted on rich soils [36,37]. Certainly in this research, farmers lacked in supplying enough fertilizers that may reinforce modern varieties defense against *S. frugiperda*. The perception of farmers about the resistance of traditional varieties to *S. frugiperda* suggests that these varieties may contain resistant genes that need to be investigated. In terms of managemen<sup>t</sup> strategies, increasing the diversification of varieties could be one of the means of effective managemen<sup>t</sup> of *S. frugiperda*, in addition to identify the resistance traits of traditional varieties and breed them into modern varieties.

#### *4.3. Farmers' Knowledge of Natural Enemies and Host Plants of S. frugiperda*

Some farmers identified the francolin, the village weaver, and the common wasp as natural enemies of *S. frugiperda*. The francolin and the village weaver are known in the literature as bird species that feed on a wide variety of plants and insects. Thus, they could truly be natural enemies of *S. frugiperda*. However, they are classified as the main grain-eating birds in maize field [38]. In general, insectivorous birds play an important role in reducing pest abundance in various agro-ecological systems. Some bird species may be able to extract *S. frugiperda* larvae from whorls and husks [39]. These birds are able to cause significant additional damage to plants [38]. Regarding the natural enemy function of the common wasp, studies confirm the perception of farmers. In a study in Brazil, wasps picked an average of 1.54 larvae per colony per hour and predated 77% of *S. frugiperda* present in maize plots (1 colony per 25 m2), providing effective control [39].

Other natural enemies of *S. frugiperda* have been recorded in West Africa. In Benin and Ghana, ten species were found parasitizing *S. frugiperda* among which two egg parasitoids (*Telenomus remus* Dixon and *Trichogramma* sp.), one egg–larval (*Chelonus bifoveolatus* Szépligeti), five larval (*Coccygidium luteum* (Brullé), *Cotesia icipe* Fernandez-Triana and Fiaboe, *Charops* sp., *Pristomerus pallidus* (Kriechbaumer) and *Drino quadrizonula* (Thomson)), and two larval–pupal parasitoids (*Meteoridea* cf. *testacea* (Granger) and *Metopius discolor* Tosquinet) [40]. Three predator species, namely *Pheidole megacephala* (F.), *Haematochares obscuripennis* Stål and *Peprius nodulipes* (Signoret), were recorded in Ghana [41].

As well, some farmers reported pest attacks in the sorghum, cotton, cowpea, and tomato fields. All these crops are among the 353 host plants of *S. frugiperda* larvae inventoried in Brazil [5].

#### *4.4. Farmers' Knowledge of Repellent Plants of S. frugiperda*

Farmers identified yellow nutsedge, chan, shea tree, neem, tamarind, and soybean as repellent plants of *S. frugiperda* through their experiments. They considered that the presence of these plants near or in the maize fields coincided with the low infestations of *S. frugiperda*. Some of these plants may act as a trap plant as a push pull system [4]. Peruca et al. [42] studied the harmful effects of soybean plants on *S. frugiperda*. They confirm that soybean plants could activate chemical defence mechanisms that alter the developmental cycle of *S. frugiperda*, suggesting effective cultural control options. Several other studies showed the effectiveness of neem extracts against *S. frugiperda*. Magrini et al. [43] concluded that neem derivatives had potent and adverse antifeedant effects on all stages of larval development of *S. frugiperda*. Tavares et al. [44] recommended neem oil to manage *S. frugiperda* due to its high toxicity. Zuleta-Castro et al. [45] formulated a botanical product active against *S. frugiperda* using neem extracts. Adeye et al. [46] found that neem oil at 4.5 l.ha−<sup>1</sup> reduced the incidence of pest attacks, the severity of damage and the loss of maize yield by 42.8% and 57.0%. Regarding the other repellent plants identified by farmers (yellow nutsedge, chan, shea, and tamarind), future studies should be carried out to evaluate the accuracy of farmers' perception.

#### *4.5. Farmers' Management Practices*

Most farmers in Benin used synthetic pesticides to manage *S. frugiperda*. The same was observed in other African countries such as Ghana, Zambia, Nigeria, Kenya, and Ethiopia [7,28,47]. Synthetic pesticides are indeed easily accessible for farmers. Institutions like USAID and other organizations such as FAO advocate low use of synthetic pesticides. However, there is no evidence that farmers comply with the recommendations of these organizations which advocate the rational and threshold use of synthetic pesticides. This implies that an effort remains to be deployed by the extension services concerning the use of synthetic pesticides against *S. frugiperda*.

In Benin, farmers found that chemical control was relatively effective. Farmers' perceptions of the effectiveness of chemical control differ from country to country. For example, in Kenya about 60% of farmers found synthetic pesticides ineffective, while Ethiopian farmers claimed that chemical control was effective against *S. frugiperda* [28].

Some farmers who used botanical pesticides thought they were more effective than synthetic pesticides against *S. frugiperda*. However, these botanical pesticides were little used. This could be explained by the lack of knowledge on the raw materials and the manufacturing process. The botanical pesticides used by farmers against *S. frugiperda* were usually made from neem leaves or seeds, vernonia leaves, pepper, and ash. Some of the farmers also added soaps, detergents, or petroleum to botanical pesticides. Vernonia is one of the African pesticidal plants selected to improve botanical-based pest managemen<sup>t</sup> in smallholder agriculture in Africa [48]. However, its effectiveness against *S. frugiperda* has not ye<sup>t</sup> been studied. It is the same with pepper, ashes, soaps, detergents, and petroleum. It is up to agricultural research institutions and scientists to refine and standardize botanical pesticides made by farmers for their scaling up.

The results showed that farmers' managemen<sup>t</sup> practices were significantly associated with their knowledge of the pest and their socio-economic characteristics such as membership of a farmer organization and contact with research or extension services. It is inferred that farmers' managemen<sup>t</sup> practices are the result of their knowledge of the pest. Farmer organizations and extension services have the potential to improve farmers' knowledge and induce behavioural changes in their pest managemen<sup>t</sup> strategies [49], and thus influence their pest managemen<sup>t</sup> decisions. As more than half of the farmers surveyed were uneducated, extension services should consider disseminating relevant information in the local language and doing demonstrations directly in the fields to improve farmers' knowledge and pest managemen<sup>t</sup> skills.
