Efficacy of homemade botanical insecticides based on traditional knowledge. A review

Homemade botanical insecticides are widely used by subsistence and transitional farmers in low-income countries. Their use is often driven by the limited availability or cost of commercial pesticides. Homemade botanical insecticides are often recommended by agricultural extension services and some development organizations. However, this could be questioned because scientific evidence of their efficacy and safety may not be available or accessible. Although botanicals with insecticidal properties have been widely studied, a synthesis focusing specifically on homemade preparations used in realistic field or storage conditions is missing. In this paper, we review efficacy assessments of botanicals used to prepare homemade insecticides. This covers twelve botanicals recommended by national extension partners in 20 countries within the global agricultural Plantwise program. These are as follows: garlic (Allium sativum), neem (Azadirachta indica), chili pepper (Capsicum spp.), Siam weed (Chromolaena odorata), mother of cocoa (Gliricidia sepium), chinaberry (Melia azedarach), moringa (Moringa oleifera), tobacco (Nicotiana tabacum), clove basil (Ocimum gratissimum), tephrosia (Tephrosia vogelii), tree marigold (Tithonia diversifolia), and bitter leaf (Vernonia amygdalina). This review shows that (1) all the selected botanicals contain active ingredients with insecticidal, antifeedant, or repellent properties, and (2) homemade insecticides based on all the selected botanicals have been used with some success to control pests or prevent damage, although efficacy was variable and often lower than the positive controls (synthetic pesticides). Factors affecting the efficacy of homemade botanical insecticide include variation in active ingredient content and concentration in plant material, as well as variation in the preparation process. In conclusion, there is some evidence that homemade botanical insecticides could contribute to reducing losses in food production. Since further research is needed to better understand their variable efficacy and potential health and environmental risks, those who promote the use of homemade botanical insecticides should also communicate those “unknowns” to the farmers who use such products.


Introduction
Globally, yield losses due to arthropods, diseases, and weeds are estimated to account for about 35% in major crops. Losses may exceed 50% in developing regions where pest control options are limited (Oerke 2006). In some cases, damage by pests, and arthropods in particular, can lead to even higher losses or total crop failure (Abate et al. 2000;Grzywacz et al. 2014;Lingappa et al. 2004). This highlights the key role of crop protection in safeguarding yields and thus ensuring food security. Although components of integrated pest management are now widely applied in developed countries, reliance on pesticides to control pest outbreaks remains high (Farrar et al. 2016;Vasileiadis et al. 2017). Synthetic pesticides are intensively used in developed and transitional countries (FAO 2013). In developing countries, many subsistence and transitional farmers do not have access to synthetic pesticides, or cannot afford them (Abate et al. 2000;Nyirenda et al. 2011). Similarly, commercial preparations of alternatives, such as biological control agents or botanical pesticides ("botanicals"), are often not available and may also be expensive (Amoabeng et al. 2014;Dougoud et al. 2018).
Botanicals were used for control of agricultural pests in ancient China, Egypt, Greece, and India already two millennia ago (Isman 2006). Still today, traditional pest control using botanicals for the protection of field crops or during storage is widespread and popular among subsistence and transitional farmers (Belmain and Stevenson 2001;Gerken et al. 2001). For example, up to 100% of the farmers in some regions of Zimbabwe and Uganda report using botanicals or have used them (Makaza and Mabhegedhe 2016;Nyirenda et al. 2011). These typically involve simple preparations, such as ground or whole plant material, and aqueous extracts thereof (Fig. 1).
Globally, over 2500 plant species belonging to 235 families have been reported to have biological activities against pests (Isman 2006;Roy et al. 2016;Saxena 1998;Stevenson et al. 2017). More specifically, the use of a wide variety of botanicals for insect pest control is highlighted in many farmer surveys, such as with 10 botanicals being used by farmers in Northern Malawi, 7 in Zambia (Nyirenda et al. 2011), 34 in the Lake Victoria basin in Uganda (Kamatenesi-Mugisha et al. 2010), or 11 in one district of the Tamil Nadu State in India (Kiruba et al. 2008).
In light of the limited availability and prohibitive cost of synthetic pesticides for subsistence and transitional farmers, some consider botanicals to be a valid alternative to synthetic pesticides (Isman 2008). A number of government agricultural departments actively promote botanical preparations in their advisory materials. As such, national extension partners in Plantwise (www.plantwise.org), a global agricultural development program led by CABI (Centre for Agriculture and Bioscience International), sometimes include homemade pesticide preparations in their recommendations and extension materials. An analysis of 811 pest management decision guides developed by national extension partners in Plantwise (www.plantwise.org/KnowledgeBank/home.aspx) showed that botanical preparations are frequently recommended in these extension materials in African countries (particularly in Zambia, Tanzania, Sierra Leone, Kenya, Ethiopia, and Mozambique), in Central, South, and Southeast Asian countries (particularly in Afghanistan, India, Myanmar, Nepal, Cambodia, and Sri Lanka), and to a lesser extent in the Americas (particularly in Nicaragua). The most widely recommended botanical was, by far, neem, followed by chili pepper, garlic, and tephrosia; however, 25 other botanicals were also recommended in the extension materials reviewed (J. Dougoud, unpublished).
The appropriateness of the recommendation and use of botanicals for pest control can be questioned. In general, the supporting evidence for the use of botanicals is very old and their efficacy needs to be reevaluated. Some of the botanicals that are being used for pest control may lack active ingredients, which would make their use by smallholder farmers a waste of time. Moreover, results may be unpredictable due to varying active ingredient content and concentration in the used plant material (Sarasan et al. 2011), as well as differences in the preparation method. Finally, their toxicity to nontargets (species that are not the intended target) has often not been evaluated. Although there is gathering evidence that some of the botanicals used for pest control are less toxic to nontargets than synthetic pesticides (Tembo et al. 2018), others may be hazardous to users, livestock, or the environment. Yet, the use of botanicals for pest control is so widespread that it cannot be ignored. Over the last decades, the efficacy of botanicals used in traditional pest management has been widely investigated in research trials. However, a synthesis of the scientific information on homemade botanical insecticides used by subsistence and transitional farmers is missing.
This review was therefore conducted to investigate the scientific basis of homemade insecticides for 12 botanicals with regard to their efficacy and reliability at reducing arthropod pest populations or their impact. Findings show the potential and limitations of the selected homemade botanical insecticides as alternatives to pesticides, with a discussion on factors affecting efficacy. Human and environmental safety, as well as practical and economic aspects, are also briefly discussed.

Methodology and scope of information search
The selection of the botanicals was based on (i) their recommendation in pest management decisions guides developed by agricultural extension and plant protection partners in the Plantwise program and (ii) availability of at least 5 literature references documenting the efficacy of homemade botanical insecticides in conditions similar to farmer practice. The 12 botanicals reviewed are as follows: garlic, Allium sativum L. This review was based on a literature search using the online repository CAB Direct (https://www.cabdirect.org) and Google Scholar (https://scholar.google.com). The search considered all the studies published over the last 30 years, i.e., from 1987 to 2017. It focused on homemade botanical insecticides used for arthropod control. Mixtures of botanicals were not covered in this review. Homemade botanical insecticides were reviewed for their efficacy in field and postharvest applications. Information on the economic viability of these homemade plant protection products was also collected, where available. Human, animal, and environmental safety is only briefly addressed as this complex topic should be covered in another review.
Only efficacy trials which used botanicals in their raw form or which used simple preparation methods (grinding, pounding, aqueous extraction) were considered for the development of Tables 1 and 2. For field applications, only field trials were considered. For storage applications, laboratory trials under realistic storage conditions were included, such as botanicals added to stored grains or beans, due to the limited availability of on-farm trials. Results obtained using commercial extracts and solvents other than water, as well as results obtained in laboratory trials (for field applications), or results obtained in trials that do not reproduce storage conditions (for storage applications) were used as complementary information for discussion.
Over half of the publications identified during this literature search as documenting trials with homemade botanical insecticides did not follow basic scientific procedures. For instance, the methodology was not sufficiently described, no control treatment was included, no exact figures were provided (e.g., results presented in charts with no numerical value), or no statistical analysis was performed. Publications with any of these critical flaws were not considered for this review. Table 1 summarizes efficacy trials performed using homemade botanical insecticides in field and storage applications. Results were grouped by botanical and plant material used and then by host crop. It was not possible to group results by target pest species or type because a substantial number of studies were performed under natural conditions and did not distinguish damage done by different pest species. Whenever a study assessed efficacy using multiple parameters and results were consistent, only one parameter was selected for this review; however, two or more parameters were selected when results were inconsistent. These parameters included the number of seeds produced; yield; pest mortality; pest population reduction; number of eggs laid; damage done to leaves, fruits, or whole plant; number of damaged grains; and grain weight loss during storage. Whenever available, data on yield were included.

Summarizing efficacy trial results
Not all plant material dosages and application frequency could be included. When there was no statistical difference in efficacy between plant material dosages, the lowest effective dosage was included; however, when there was a statistical difference between dosages, two dosages were selected to underline a dose-dependent effect. In a few cases, two different dosages were included even though they were not statistically different because efficacy varied by a factor of two or more. This was taken as an indication that dosage deserves further investigation. When a botanical was considered ineffective, only the highest dosage was selected to reduce the likelihood that the lack of effect was due to inadequate dosage.
In order to facilitate the comparison among trials, results were transformed into percentages, which were calculated based on the data given. When relevant, Abbott's formula (Abbott 1925) was used. The efficacy percentages shown in Table 1 for both homemade botanical insecticides and positive controls (i.e., a reference product such as a synthetic pesticide) are always relative to the negative controls (i.e., untreated or sprayed with water and adjuvants).   The pesticidal activity of garlic cloves has been attributed to sulfur-containing compounds that arise from the enzymatic degradation of allicin (Huang et al. 2000;Prowse et al. 2006;Yang et al. 2012;Zhao et al. 2013). Garlic extracts have been shown in laboratory trials to have acaricidal properties (Dąbrowski and Seredyńska 2007;Roobakkumar et al. 2010) and insecticidal properties against coleopteran, dipteran, lepidopteran, and hemipteran pests (Abdalla et al. 2017;Denloye 2010;Prowse et al. 2006;Yang et al. 2012;Zhao et al. 2013).
In field application trials, garlic aqueous extracts resulted in a varying level of control of hemipteran pests (Bahar et al. 2007;Baidoo and Mochiah 2016;Fening et al. 2013;Oparaeke et al. 2007;Said et al. 2015), lepidopteran pests (Baidoo and Mochiah 2016;Fening 2013;Oparaeke et al. 2007) as well as mites (Attia et al. 2011) (Table 1). Results were not significantly different in 3 out of 9 trials (Table 2), yet, in 1 trial, the 2 synthetic pesticides used as a positive control had no significant effect either. In comparison with positive controls, the efficacy of garlic aqueous extracts was statistically lower in half of the comparisons (Table 2). In 5 out of 7 studies, adjuvants were added to the preparations and those may have influenced their efficacy (see discussion for details). Other studies suggest that garlic-based homemade pesticides may be used to control mites on tomato (Kaputa et al. 2015) and fruit flies on watermelon (Degri and Sharah 2014), but the authors did not perform relevant statistical tests.

Neem (Azadirachta indica)
The insecticidal activity of neem has been attributed to limonoids. It is considered that azadirachtin A is the most active compound, but other limonoids may contribute to the efficacy of neem insecticides (Boursier et al. 2011;Isman et al. 1990;Lynn 2010;Nathan et al. 2005) and may even prevent resistance build-up against azadirachtin A (Feng and Isman 1995). Commercial extracts of neem are widely used for control of a wide range of insects as well as mites. The insecticidal and acaricidal properties of commercial neem-based products have been largely demonstrated (Morgan 2004).
Homemade aqueous extracts based on neem plant material (leaves, seeds, seed cake, and unformulated oil) have been successfully used for the control of blattodean pests (Ibrahim and Demisse 2013), hemipteran pests (Aziz et al. 2013;Degri et al. 2013;Gupta and Pathak 2009), lepidopteran pests (Abate 2011;Attia et al. 2011;Kumar et al. 2015;Okrikata et al. 2016), and thysanopteran pests (Shiberu et al. 2012) in field application (Tables 1 and 2). Out of 8 trials against lepidopteran pests, neem aqueous extracts showed efficacy in 7 trials. Foliar or soil application of neem aqueous extracts for insect pest control gave results superior to negative controls in 15 out of 18 instances and resulted in a yield increase in all 9 trials where yield was assessed. In comparison with synthetic pesticides, neem aqueous extracts were comparable in 10 out of 15 instances, but were inferior in 5 instances. The efficacy of neem aqueous extracts or of an oil emulsion against mite pests in field applications was documented only by Patil and Nandihalli (2009); both preparations reduced mite population, but did not impact yield. Efficacy of neem oil against fruit flies attacking watermelon has been reported, but appropriate statistics were not provided (Degri and Sharah 2014).
Ground neem plant material successfully and consistently controlled coleopteran pests in storage applications trials Boeke et al. 2004b;Ileke and Oni 2011;Ilesanmi and Gungula 2010;Kemabonta and Falodu 2013;Kossou 1989). It did not provide any control in only 1 instance out of 8, although this may be explained by the low quantity of neem leaves used (Tables 1 and 2). The efficacy of ground neem in storage is supported by farmer participatory trials conducted by Paul et al. (2009) and earlier studies (Lale and Abdulrahman 1999;Pereira and Wohlgemuth 1982).

Chili peppers (Capsicum spp.)
Capsaicin is the main compound that gives chili peppers their spiciness. Capsaicin-rich commercial insecticide formulations are widely available. Capsaicin has repellent and insecticidal properties, for example, against hemipterans (Bergmann and Raupp 2014;Dayan et al. 2009). Antonious et al. (2006Antonious et al. ( , 2007 indicate that other compounds may contribute to the insecticidal activity of preparations based on chili peppers. In field application trials, chili pepper aqueous extracts have been used to control hemipteran pests Baidoo and Mochiah 2016;Fening et al. 2013Fening et al. , 2014, lepidopteran pests Baidoo and Mochiah 2016;Fening et al. 2013Fening et al. , 2014Okrikata et al. 2016), and thysanopteran pests (Fening et al. 2014), yet the results obtained were inconsistent. Chili pepper aqueous extracts were superior to negative controls in 5 out of 10 instances. However, in 4 of the 5 trials where chili pepper aqueous extracts were ineffective, the synthetic pesticides used as positive control were also ineffective (Fening et al. 2013(Fening et al. , 2014; therefore, those trials are not conclusive (Table 1). The positive controls were superior to chili pepper aqueous extracts only in 1 out of 14 instances, underlining that further research is needed for a conclusion.
In storage application trials, ground chili pepper fruits controlled the cowpea weevil Callosobruchus maculatus in 2 studies (Onu and Aliyu 1995;Yusuf et al. 2011) but was not effective in a third study (Boeke et al. 2004b) (Table 1). However, the efficacy of this practice is supported by farmer participatory trials conducted over 5 years in Ghana (Belmain and Stevenson 2001). Moreover, a study by Belmain et al. (1999) has shown that chili pepper was effective in killing and repelling various species of weevil attacking stored grains, although this publication does not provide exact figures.

Siam weed (Chromolaena odorata)
One of the main constituents of Siam weed's essential oil, the monoterpenoid α-pinene, has insecticidal and repellent activities against coleopteran storage pests (Avlessi et al. 2012;Huang et al. 1998;Kim et al. 2010;Kossouoh et al. 2011;Owolabi et al. 2010). The insecticidal properties of the essential oil (Bouda et al. 2001) and extracts (Lawal et al. 2015) of Siam weed have been demonstrated on coleopteran pests in the laboratory.
In field applications, aqueous extracts of Siam weed controlled coleopteran pests (Onunkun 2012), lepidopteran pests (Amoabeng et al. , 2014Ezena et al. 2016), and hemipteran pests (Amoabeng et al. , 2014Degri et al. 2013;Ezena et al. 2016) (Table 1). Siam weed aqueous extracts were consistently (8 out of 9 instances) superior to negative controls and often comparable (7 out of 11 instances) to positive controls (Table 2). A study by Devi et al. (2013) suggested that Siam weed is effective in controlling mites on tea, but the authors did not provide statistical evidence.

Mother of cocoa (Gliricidia sepium)
Major secondary metabolites of mother of cocoa, coumarins, were reported to have insecticidal properties against coleopteran, dipteran, and lepidopteran pests as well as antifeedant properties against fall armyworm Spodoptera frugiperda (Kaniampady et al. 2007;Moreira et al. 2007;Vera et al. 2006). In the laboratory, extracts of mother of cocoa have been shown to possess insecticidal activities against coleopteran, dipteran, and lepidopteran insects (Parvathi and Jamil 1999;Sharma et al. 1998) as well as acaricidal activities (Sivira et al. 2011). (Tables 1 and 2). In storage applications, ground leaves of mother of cocoa controlled coleopteran pests, according to Ojo et al. (2013) and Rendón-Huerta et al. (2013).
Aqueous extracts of both leaves and fruits controlled lepidopteran pests (McKenna et al. 2013;Singh et al. 2013) and mites (Attia et al. 2011) in the field, although efficacy was inferior to positive controls (Tables 1 and 2). Other studies indicated that aqueous extracts of chinaberry may be used to control serpentine leaf miners on Swiss chard (Abou-Fakhr Hammad et al. 2000) and cabbage aphids on cabbage (Kibrom et al. 2012), but the authors did not provide statistical details.
In storage applications, Chinaberry ground plant material has been successfully used for control of coleopteran storage pests in 2 studies (Espinoza et al. 2012;Hafez et al. 2014) and led to a reduction of grain damage (despite a nonsignificant effect on the pest population) in a third study conducted by Rendón-Huerta et al. (2013) (Tables 1 and 2).

Moringa (Moringa oleifera)
Lectins in moringa seeds have larvicidal activity on the mosquito Aedes egypti and on the flour moth Anagasta kuehniella The efficacy of moringa homemade botanical insecticide has been primarily assessed in storage applications, except in 1 study by Alao and Adebayo (2015) in which aqueous extracts of moringa controlled dipteran and coleopteran pests in the field (Tables 1 and 2

Tobacco (Nicotiana tabacum)
Tobacco owes its insecticidal properties to nicotine and other related alkaloids. Nicotine pesticides have a long history, but their use is decreasing because of its high toxicity to humans (Isman 2006).
Tobacco aqueous extracts controlled hemipteran pests (Amoabeng et al. , 2014Bahar et al. 2007), lepidopteran pests (Amoabeng et al. , 2014, and thysanopteran pests (Shiberu et al. 2012), and their efficacy was comparable or superior to positive controls (Tables 1 and 2). In storage applications, the efficacy of ground tobacco leaves was documented in only by Boeke et al. 2004b andLonge 2016, with an efficacy of 100% against cowpea weevils in both studies.

Clove basil (Ocimum gratissimum)
Clove basil essential oil and of some of its constituents have insecticidal and repellent activities against coleopteran pests and are repellents to houseflies in laboratory trials (Kéita et al. 2001;Ogendo et al. 2008;Ouedraogo et al. 2016;Singh and Singh 1991).
Homemade aqueous extracts of clove basil controlled hemipteran pests Oparaeke 2006) and lepidopteran pests  in field applications (Tables 1 and 2). Results of storage trials are inconsistent, with 1 out of 3 trials showing no efficacy against coleopteran pests (Law-Ogbomo and Enobakhare 2007; Mlambo et al. 2017;Osipitan et al. 2014). However, the efficacy of clove basil in storage applications is supported by farmer participatory trials (Belmain and Stevenson 2001), indicating the adequacy of this botanical for storage applications.
In field applications, tephrosia aqueous extracts have been used successfully to control coleopteran pests ( (Table 1). In comparison with positive controls, the efficacy was often similar or inferior (3 instances each), but was superior in 1 instance (Table 2).
In storage applications, tephrosia ground leaf material controlled coleopteran pests (mortality rates of 69% and above) in 4 out of 5 instances Mkenda et al. 2015b;Ogendo et al. 2003). The 1 trial with a negative result had used a rotenoid-poor chemotype, underlining the importance of plant material selection ). The potential of tephrosia for storage application is supported by trials performed by Ogendo et al. (2004). In comparison with the positive controls, ground tephrosia leaves were less effective in 2 out of 3 comparisons. Results of this study were provided as charts with no numerical values and could thus not be included in the summary tables.

Marigold tree (Tithonia diversifolia)
A major constituent of marigold tree essential oil, the monoterpenoid α-pinene, has insecticidal and repellent activities against coleopteran storage pests (Huang et al. 1998;Moronkola et al. 2006). Sesquiterpene lactones in tree marigold have been shown to be toxic to the coleoptera Callosobruchus maculatus (Green et al. 2017).

Bitter leaf (Vernonia amygdalina)
Sesquiterpene lactones with insecticidal activities against coleopteran pests and with repellent activities against lepidopteran pests have been isolated from bitter leaf (Ganjian et al. 1983;Green et al. 2017).
In field applications, bitter leaf aqueous extracts were used to control coleopteran pests (Mkenda et al. 2015a;Tembo et al. 2018), hemipteran pests (Degri et al. 2012;Mkenda et al. 2015a;Oparaeke 2006;Onunkun 2012;Tembo et al. 2018), and lepidopteran pests (Degri et al. 2012), with a positive outcome in 6 out of 8 instances and with a level of control inferior or similar to the positive controls (Tables 1 and 2). However, the application of ground plant material to the soil was not effective against termites (Ibrahim and Demisse 2013). In storage applications, bitter leaf ground plant material was used to control coleopteran storage pest (Law-Ogbomo and Enobakhare 2007; Mkenda et al. 2015b;Musa and Adewale 2015), with a positive outcome in all 4 instances (Tables 1 and 2).

Assessing efficacy
Efficacy of commercial pesticide products is assessed as part of the registration process. According to the guidelines published by the Food and Agriculture Organization of the United Nations (FAO 2006), this should include the testing of pesticides in comparison with a negative control and to a positive control, i.e., a reference product such as a synthetic standard pesticide or a standardized botanical extract. For major pest/crop combinations, these guidelines recommend at least 8 fully supportive trials (or more should the results be inconsistent) conducted over at least two seasons, but fewer trials are acceptable for minor pests or crops. Minimal efficacy levels should not be cast in stone, and low efficacy levels may be acceptable as long as they provide benefits to the grower, in particular if the product has no or low risk to nontarget species. Negative impacts on the target crop (yield or quality reductions), risks of resistance build-up, and other risks for agronomic sustainability should be included in the efficacy evaluation. Commercial pesticides based on some of the botanicals covered in this review (e.g., neem, garlic, chili pepper) are available and are generally tested this way. However, no homemade botanical insecticide has been tested in this way. As shown in Table 1, the efficacy studies compiled in this review at best included trials conducted over two growing periods or in two separate locations.
As Table 1 shows, there is a limited number of publications documenting the efficacy of a given botanical on a specific pest or pest/crop combination. The efficacy of the control achieved by homemade botanical insecticides was generally inconsistent among trials (Tables 1 and 2). This may be attributed to the variability in secondary metabolite (active ingredients) content or concentration in the plant material used or the processing and application methods, as illustrated by Kamanula et al. (2017) and by Stevenson et al. (2012). Yet, this should not obscure that, in a local context, a homemade botanical insecticide produced using a defined procedure and using plant material from a constant source could possibly produce consistently reliable results.
In the compiled studies, the chemical characterization of the plant material used was rarely performed and quantification of the active ingredients was performed only once, and this is in line with earlier observations made by Isman and Grieneisen (2014). Inconsistencies may also have arisen from the misidentification of botanicals. Indeed, plant names vary across locations, and the same local name may be used for similar yet distinct botanicals. This could be avoided by comparing gathered botanicals with type specimens documented in herbariums (Belmain and Stevenson 2001). For each botanical considered in this review, the trial methodology varied among the studies cited, e.g., plant part used, dosage, adjuvants, target host and pest, or the preparation method. Hence, the cited references are difficult to compare and can only be considered as indicative. This underlines that, although this review gives a highly valuable insight into the potential of homemade botanical insecticide for arthropod pest control, further research would be needed to validate which specific pests can effectively be targeted by a selected botanical.

Absence of evidence is not evidence of absence
This analysis suggests that homemade insecticides based on all 12 selected botanicals may or may not provide, under different circumstances, control of arthropod pests. Homemade botanical insecticides based on other botanicals have provided control of pests, such as Mexican tea, Dysphania ambrosioides L. (Asterales: Asteraceae) (Mazzonetto et al. 2013;Mkenda et al. 2015b;Pamela Nuñez et al. 2010;Paul et al. 2009;Tapondjou et al. 2003) or billygoat weed, Ageratum conyzoides (L.) Mosyakin & Clements (Caryophyllales: Amaranthaceae) (Amoabeng et al. , 2014Kar et al. 2008;Onunkun 2012;Singh et al. 2013). However, this review was conditioned by the availability of literature. Negative results may not have been published, and this may have biased our analysis in two different ways: by overevaluating the efficacy of the botanicals included in this review or by making it impossible to include further and potentially ineffective botanicals because of a lack of literature. This last hypothesis is supported by farmer participatory trials conducted by Belmain and Stevenson (2001) and laboratory trials performed by Boeke et al. (2004b) indicating that a significant proportion of the botanicals used by farmers may have no or little efficacy.

Variation in active ingredient in botanical plant material
A wide range of factors affects secondary metabolite content and concentration in plants and thus the concentration of active ingredients in botanical insecticicdes. Different plant parts are highly variable in active ingredient content and concentration. The plant's genotype, a wide range of environmental factors, and the development stage of the plant can strongly affect both content and concentration (Canter et al. 2005;Figueiredo et al. 2008;Gahukar 2014). Among the 12 botanicals included in this study, this has been particularly documented for neem (Elteraifi and Hassanali 2011;Gahukar 2014;Prakash et al. 2005;Sidhu et al. 2003;Yakkundi et al. 1995), Siam weed (Avlessi et al. 2012;Kossouoh et al. 2011), and tephrosia ), but can be expected to be relevant for all botanicals. Such variations can dramatically impact the efficacy of homemade botanical insecticides. For example, at least one chemotype of tephrosia is totally ineffective . Furthermore, active ingredients may break down over time, and this may be affected by storage conditions, as reported with neem seeds (Yakkundi et al. 1995). Three strategies have been suggested to reduce variation in active ingredient content and concentration in plant material: the collection of plant material from a large number of plants (Mkenda et al. 2015a); the selection and propagation of plant material with an appropriate content and an elevated concentration of secondary metabolites Canter et al. 2005); or the identification of optimal harvest timing (Yakkundi et al. 1995).

Processing and adjuvants
The method of processing homemade botanical insecticides can impact their efficacy in numerous ways, yet this is not well documented. Boursier et al. (2011) showed that a neem seed aqueous extract whose active ingredient concentration is much higher than recommended for commercial products could be obtained using a seed dosage typical of a traditional Malian recipe. Yet, they also showed how azadirachtin content in the extract may be influenced by factors such as extraction time and process or the shelling of the kernels prior to grinding. Some traditional preparation methods may be suboptimal. For instance, whole leaves of botanical source plants are often mixed directly with grains for protection during storage even though pulverization is considered to enhance efficacy (Belmain and Stevenson 2001).
The addition of adjuvants such as surfactants or stickers to pesticides is a common practice and is aimed at enhancing their efficacy through better coverage and longer persistence (Witt 2012). Similarly, small quantities of vegetable oil and/or soap or starch are often added during the preparation of aqueous extracts and before or after extraction. These are thought to improve extraction or coverage of the foliage (Kaputa et al. 2015;Mochiah et al. 2011). Authors added adjuvants to aqueous extracts in about half of the reviewed studies, and this may have had an influence on efficacy.
The most widely used adjuvant was soap, either as bar flakes or liquid soap. Laboratory trials have shown that small concentrations of household soap diluted in distilled water may have an outstanding insecticidal activity (Lee et al. 2006;Liu and Stansly 2000). Although field trials indicate that small concentrations of household soap (1 ml/l) do not statistically significantly affect pest populations (Amoabeng et al. 2014;Mkenda et al. 2015a), higher amounts of soap will likely do so. Besides its direct action on insects, soap also affects surface tension of spray mixture and should thus logically improve coverage and therefore efficacy of the active ingredient. Furthermore, enhanced extraction has also been reported using a surfactant (10 ml/l) for extraction of rotenoids from tephrosia plant material, indicating that the addition of an easily available surfactant such as soap would enhance the extraction of less polar compounds, such as azadirachtin Morgan 2009).

Complex interactions and implications
The efficacy of insecticides is commonly understood as their ability to kill a target pest. Some botanicals, such as pyrethrum or tobacco, contain compounds that have a neurotoxic activity, causing the rapid death of arthropod pests. However, a large number of botanicals and their compounds act in a more subtle way. For instance, azadirachtin, the main active ingredient of neem, affects the metabolism of insects, leading to female infertility and disruption of the molting process. Neem and chinaberry, as well as other botanicals of the Lamiaceae and Asteraceae family, have been shown to possess antifeedant properties. Other botanicals, such as citronella (Cymbopogon spp.), have repellent properties (Isman 2006). Some field trials with homemade botanical insecticides show that lower pest mortality may not always mean lower efficacy (Aziz et al. 2013;Charleston et al. 2006;Mkenda et al. 2015a;Montes-Molina et al. 2008a;Tembo et al. 2018). Natural enemies also play an important role in the reduction of pest populations, and the application of pesticides of either synthetic or natural origin can harm them (Cloyd 2012;Pimentel et al. 1992). Some botanical insecticides, such as bitter leaf, neem, tephrosia, and tree marigold, may have a low toxicity towards natural enemies (Aziz et al. 2013;Mkenda et al. 2015a;Mkindi et al. 2017), and the use of such products can favorably shift the pest/natural enemy balance and result in a prolonged efficacy of the control intervention (Naranjo and Ellsworth 2009).
Some plant species used for the preparation of homemade botanical insecticides also have fungicidal and/or bactericidal properties, such as tephrosia (Owolade et al. 2004), neem (Hassanein et al. 2010), and Siam weed (Avlessi et al. 2012). Others may simultaneously act as a foliar fertilizer, such as mother of cocoa (Montes-Molina et al. 2008b). These properties may positively impact yields. On the other hand, some botanicals have been reported to have allelopathic properties, such as Siam weed (Sahid and Sugau 1993), so their use may negatively affect crop growth. These observations underline the importance of full-season field trials that take yield into account, as this allows a better understanding of how homemade botanical insecticides can contribute to preventing yield losses.

Human safety
Risks linked to pesticide use depend on their toxicity and on the exposure of applicators or consumers. Pesticides are normally assessed as part of the registration process. According to the guidelines published by the Food and Agriculture Organization of the United Nations and the World Health Organization, assessments shoud include the acute toxicity of formulated product in order to identify appropriate protective measures. The acute toxicity of the active ingredient and its metabolites or degradates should be assessed in order to identify health hazards linked to short-term exposure. Subchronic and chronic effects, mutagenicity, carcinogenicity, and reproductive and developmental toxicity should be assessed to identify the risks related to a long-term exposure. Moreover, exposure of farm workers and applicators, as well as residue in the crop produce, should be evaluated to define whether the risks linked to pesticide use are acceptable WHO 2013, 2016). With the exception of neem products, such safety assessments have not been conducted with homemade botanical insecticides, or only partially. A major difference between homemade botanical insecticides and commercial pesticides is that the former contains a cocktail of active ingredients with unknown concentrations, as well as a long list of compounds with unknown properties in variable concentrations. Moreover, although concentration in plant material may be low, exposure during processing has not been assessed and may be high. As a result, even when safety assessments exist, the risks identified in laboratory trials are difficult to extrapolate to real-life situations. Plant protection legislation in many countries does not allow the use of homemade preparations, something that often contradicts the reality in farming. Therefore, some countries legally allow the use of such preparations, at least for noncommercial farming (Belmain and Stevenson 2001;Klein et al. 2015).
Neem products, in particular neem oil and aqueous extracts, have a low subchronic and chronic toxicity. Results of acute toxicity trials are more ambiguous but generally point to a low toxicity to mammals (Boeke et al. 2004a). Scattered data are available for other botanicals covered in this review. Aqueous ethanol extracts of Siam weed had a low mammalian toxicity in acute and subchronic toxicity trials (Ogbonnia et al. 2010). Clove basil and marigold tree essential oils and ethanolic and aqueous extracts have been shown to have an extremely low acute mammalian toxicity, and subchronic toxicity trials suggest that aqueous extracts of tree marigold are relatively safe (Kamatenesi-Mugisha et al. 2010;Passoni et al. 2013). Although these data need to be validated and complemented, the above plants have been used for centuries as a traditional medicine, and this supports the idea that they have a relatively low toxicity. Bitter leaf, chili pepper, garlic, and moringa are consumed as food or spice. The long use history of these botanicals indicates that their use as a pesticide represents minimal risk. Compounds occurring widely in food are granted the status of "Generally Recognized As Safe" by the Food and Drug Administration of the United States of America (FDA 2019). In contrast, tobacco contains nicotine, which has a high acute toxicity. Nicotine is classified as highly hazardous (class Ib) by the World Health Organization (WHO 2009), and nicotine pesticides are now banned in most countries. For this reason, Plantwise discourages the use of tobacco-based homemade botanical insecticides. Tephrosia contains rotenone, which is classified as moderately hazardous (class II) because of its acute toxicity and it has been linked to Parkinson's disease (Tanner et al. 2011;WHO 2009). Finally, ingestion of chinaberry is reported to have caused human and animal poisonings. However, some authors argue that active ingredient concentrations in plants like tephrosia and tobacco are low, and that the use of these botanicals for pest control by subsistence farmers is unlikely to cause intoxications Isman 2008).
Smallholder farmers who use homemade botanicals to control agricultural pests do so primarily for economic reasons but are also worried about potential health issues resulting from the use of synthetic pesticides (Belmain and Stevenson 2001;Isman 2017;Kamatenesi-Mugisha et al. 2008). Smallholder farmers who cannot afford to buy synthetic pesticides also will not be able to buy appropriate protective equipment. This underlines that more safety assessments should be conducted so that safe botanicals and preparation methods can be identified. Yet, certain homemade botanical insecticides may represent a relatively safe alternative. This is particularly relevant when considering that highly hazardous pesticides are still often used in low-income countries, resulting all too often in farmer poisonings (Grzywacz et al. 2014;Kesavachandran et al. 2009;Ngowi et al. 2007;Weinberger and Srinivasan 2009), thus incurring significant hidden societal costs (Bourguet and Guillemaud 2016;Soares and Porto 2012). Yet, the principle of precaution should be applied to homemade botanical insecticides. Exposure to botanicals known to pose a significant risk to human health should generally be avoided.

Environmental safety
In parallel to human health risks, adverse effects of pesticide use on nontarget organisms-such as natural enemies of pests, pollinators and also birds, fish, or mammals-depend on their toxicity and on exposure. These risks should be assessed as part of the registration process to define whether they are acceptable WHO 2013, 2016). Data on environmental fate is also normally required for the registration of pesticides. Bioaccumulation is generally considered less likely to occur with homemade botanical insecticides as they contain naturally occurring substances, which are known to degrade more quickly than many synthetic compounds, as illustrated by natural pyrethrins vs. synthetic pyrethroids (Smith and Stratton 1986).
Thorough environmental safety assessments have not been conducted with most botanical insecticides as they are not as heavily regulated. The toxicity of commercial neem pesticides to natural enemies has been reviewed by El-Wakeil et al. (2013). The authors conclude that neem toxicity is usually significantly lower compared with synthetic pesticides, although some nontarget species may be particularly susceptible. The few available studies on the impacts of homemade botanical insecticides on nontarget species suggest that aqueous extracts often have a relatively lower impact on natural enemies compared with broad-spectrum insecticides. Aqueous extracts of Siam weed and tobacco had lower impacts on nontarget ladybirds, hoverflies, and spiders compared with emamectine benzoate . In a trial conducted by Mkenda et al. (2015a), aqueous extracts based on tephrosia and tree marigold had no impact on ladybirds and limited or no impact on spiders. An aqueous extract based on bitter leaf had no impact on spiders but suppressed ladybird population to a degree similar to lambda-cyhalothrin. More recently, trials conducted by Mkindi et al. (2017) confirm the relatively low toxicity of tephrosia, tree marigold, and bitter leaf aqueous extracts on hover fly, lacewing, lady bird, and spider populations. Another example is provided by Singh et al. (2013), who observed that an aqueous extract of chinaberry had a lower impact on ladybird populations in comparison with malathion.
Data on pollinator toxicity of homemade botanical insecticides are also lacking (IOBC 2018). Laboratory trials indicate that commercial neem extract may be harmful to bees and wild pollinators (Bernardes et al. 2017), but no impact could be detected in the field (Naumann et al. 1994). Pure azadirachtin is classified to be moderately toxic to bees, and the use of the pesticides falling in this category is usually not recommended on blooming plants or areas that are visited by bees (Cluzeau 2002;Maciorowski 1994). A commercial garlic extract as well as rotenone-based product proved to have lethal and sublethal effects on bees but not on wild pollinators in laboratory trials (Xavier et al. 2010;Xavier et al. 2015). Rotenone is classified as relatively nontoxic to bees, and the use of the pesticides falling in this class is usually not restricted (Cluzeau 2002;Devillers 2002).
Other nontarget organisms include other arthropods, fungi, molluscs, aquatic organisms, mammals, or birds, yet toxicity of homemade botanicals to these organisms is often not known. However, the use of tephrosia products for poison fishing illustrates the potential risk posed by homemade botanical insecticides to the environment (Neuwinger 2004;Pubchem 2018).
These data highlight that, despite the fact that some homemade botanical insecticides may have lower toxicity to nontarget species compared with broad-spectrum insecticides, harmful effects have been observed, underlining the need for further research. The application of botanicals should thus be guided by necessity and done with care, taking their potential negative impact to nontarget species into consideration. Likewise, just as pesticides should not be used as the only pest management intervention, botanicals should not be used in isolation either. Botanicals can fit into an IPM system. For example, botanicals can be used in combination with crop diversification, habitat management, and other nonpesticide tools.

Risks to biodiversity
Some of the plants used for the preparation of homemade botanical insecticides are invasive species, such as tree marigold and Siam weed (CABI 2017). The collection in cultivated areas or in the wild may perhaps contribute to reduce the populations of these invasive species. Yet, their cultivation for the purpose of producing botanical insecticides may contribute to their expansion and further increase the negative impact of these invasive species on biodiversity and on the livelihoods of farmers. This also means that botanicals used in homemade insecticides and known to be invasive should never be introduced into areas where they are presently absent.

Economic viability and practicality
Before botanicals can be processed, they must be grown, collected in nature, or bought from the market. Processing may require a heavy workload, and homemade botanical insecticides may require more frequent applications than synthetic pesticides. Benefit/cost impact studies, which take into account the total costs, including labor of homemade botanical insecticide preparation and application, give an insight into their economic viability.
The total costs are often reported to be substantially lower compared with the cost of buying and applying a commercial chemical pesticide (Gupta 2005;Gupta and Pathak 2009;Mkenda et al. 2015a). In 1 case, the total costs of homemade botanical insecticides were comparable with the synthetic pesticide (Amoabeng et al. 2014); however, this is mainly because the local daily wage used in this study was 8.33 USD, which is considerably more than the daily income of many subsistence farmers in the world (FAO 2015).
Among the botanicals selected for this review, benefit/cost ratios have been calculated for field applications use of homemade botanical insecticides based on 7 of the 12 botanicals. This is most well documented in a number of studies on neem homemade botanical insecticides that found profitable benefit/ cost ratios: using leaves, seed kernels, seed cake, or oil for control of (1) the green leafhopper Nephotettix virescens in rice (Rajappan et al. 2000), (2) the aphid Sitobion aveanae in wheat (Aziz et al. 2013), (3) the aphid Lipaphis erysimi in mustard (Gupta 2005), (4) the whitefly Bemisia tabaci and the pod borer Maruca testulalis in black gram (Gupta and Pathak 2009), (5) the pod bug Clavigralla gibbosa in pigeon pea (Narasimhamurthy and Ram 2013), and (6) the Sesamia calamistis stem borers in sorghum (Okrikata et al. 2016). The economic viability of homemade botanical insecticides based on other plant species is less well documented. Nonetheless, in all the reviewed studies, profitable benefit/ cost ratios have been obtained with homemade botanical insecticides. This includes Siam weed or tobacco for control of the diamondback moth Plutella xylostella and aphids Brevicoryne brassicae in cabbage (Amoabeng et al. 2014). Similar results have been obtained by Mkenda et al. (2015a) with aqueous extracts based on tephrosia, tree marigold, and bitter leaf in bean against the aphid Aphis fabae, the bean flower beetles Epicauta albovittata and E. limbatipennis, and the bean foliage beetles Ootheca nutabilis and O. bennigseni. A profitable benefit/cost ratio has also been obtained using ground chili pepper for control of Sesamia calamistis stem borers in sorghum (Okrikata et al. 2016).
These results underline the economic viability of this management practice, but should not detract from the fact that commercial pesticides, although costly, may provide a better control of pests and a better net gain for the farmer at the end of the season. Nevertheless, independently of the time required for their production, some botanical preparations may represent an interesting alternative for resource-poor subsistence and transitional farmers who often simply cannot access or afford to buy synthetic pesticides.

Conclusions
The pest control methods using traditional knowledge are based on centuries-long empirical observations, but may also be tainted with belief. Our review shows that active ingredients with insecticidal properties have been isolated in all the botanicals covered by this study. The data indicate that homemade insecticides based on the 12 selected botanicals have the potential to lower arthropod pest populations or to reduce the losses they cause. Although synthetic pesticides may often be more effective, all existing benefit/cost studies support the economic viability of homemade botanical insecticides. This means that homemade botanical insecticides could be, in some cases, an acceptable alternative to commercial pesticides, particularly where availability of and access to synthetic pesticides are limited. Nonetheless, it must be acknowledged that the results of using homemade botanicals are highly unpredictable, as their effectiveness and safety have not been fully tested.
For every botanical included in this study, the efficacy varied among trials, and, in some cases, the use of homemade botanical insecticides neither resulted in a reduction of pest populations nor prevented crop losses. This can be attributed to variation in active ingredient content or concentration in plant material, variation in the preparation method, or also variation in the conditions and the way in which they were tested. National researchers in some countries are already working on improving homemade botanical insecticide efficacy through plant material selection and optimized processing. These efforts should be encouraged and supported so that the locally appropriate, optimized, and standardized homemade botanical insecticide preparation methods can be disseminated to subsistence and transitional farmers. In addition, research on the possible health risks related to the use of botanicals ought to be better financed, and national pesticide legislation could be adapted to more formally address the benefits and risks associated with homemade botanicals.
Neem products have a low human toxicity and an acceptable environmental toxicity, and their use can thus be considered as relatively safe. However, human and environmental risks of other botanicals often have not been sufficiently assessed using standardized procedures, and this deserves further investigation. Some botanicals used for the preparation of homemade insecticides are consumed as food or traditional medicine, which suggests that their use should pose limited risk. Yet, the possible consequences of extensive and prolonged exposure to homemade botanical insecticides are unknown. Use of botanicals known to be toxic, e.g., tobacco, should be avoided. As a precaution, personal protective equipment should be worn during the preparation and application of homemade botanical insecticides, irrespective of which botanical is used.
The use of homemade botanical insecticides is so widespread in low-income countries that it should not be ignored. Homemade botanical insecticides may be less effective than synthetic pesticides but, in particular when used within an IPM approach, constitute an option for farmers who have no access to commercial pesticides or who cannot afford them. Thus, they contribute to reducing losses during food production in the most underprivileged regions of the world. It is just important that those promoting the use of these homemade botanical insecticides are aware of and also communicate the uncertainties around the use of these products (i.e., variable efficacy and potential effects on health and the environment). This review provides an in-depth analysis of the potential and limitations of homemade botanical insecticides and defines areas for further research.