Statistical analysis

Descriptive statistics (means) were employed for the different variables assessed. Pesticides were evaluated for toxicity using the Chi-Square test (p < 0.05). The problems, in order of importance, were analyzed using the Kruskal-Wallis H test (p < 0.05). The dosages used and recommended were compared using the Wilcoxon test (p < 0.05).

Cucurbitaceae

The major pest problems farmers mentioned were the aphid Aphis gossypii Glover (Hemiptera: Aphididae) and whiteflies of the Bemisia tabaci (Gennadius) complex, followed by larvae of the genus Diaphania (Lepidoptera: Crambidae), the melon-worm moth Diaphania hyalinata L., and the pickle-worm Diaphania nitidalis Stoll (table 3; p< 0.05). Producers control these pests spraying insecticides at an average of 2.6 times weekly (range: 2-3; table 3). These spraying frequencies are much higher than those reported by Valarezo, Cañarte, Navarrete, Guerrero and Arias (²⁰⁰⁸), who indicated that 100 % of melon and watermelon producers in the province of Santa Elena, Ecuador, applied insecticides approximately once a week to control whiteflies, which they describe as an “extremely high use of insecticides”.

Farmers indicated that to handle sucking insects, they mainly used neonicotinoid insecticides, which agrees with what has been reported by Valarezo et al. (²⁰⁰⁸). Diaphania species were controlled with methomyl, methamidophos, profenofos, and fipronil; the first three are very toxic and have the highest use percentage (table 4) compared to the last (table 5). The weekly frequency applications obtained in this study represent 23 to 34 in total throughout one crop cycle that lasts from 60 to 90 days. ^{Vargas-González et al. (2016)} indicated 21 pesticide applications during a melon cultivation cycle in Comarca Lagunera, Mexico, which is lower compared to what has been estimated here.

Table 3. Number of weekly sprays (Ap) (mean ± standard deviation) and main pest problems identified in different crops 

% of referred importance. Means with the same letter do not differ significantly. Comparisons made with the Kruskal-Wallis H Test (p < 0.05).

Elaborated by the authors

Fabaceae: Phaseolus vulgaris: common bean

The pests referred to by farmers for bean cultivation were the leaf miners Liriomyza sativae and Liriomyza huidobrensis (Blanchard) (Diptera: Agromyzidae), as well as the black aphid Aphis craccivora Koch (Hemiptera: Aphididae), followed by the whitefly species B. tabaci and Trialeurodes vaporariorum (Westwood) (Hemiptera: Aleyrodidae) (table 3; p < 0.05). The leaf miners and the whiteflies species are considerably polyphagous (^{Tong-Xian, Kang, Heinz, & Trumble, 2009}; ^{Abdul-Rassoul & Al-Saffar, 2014}; ^{Chirinos et al., 2014}; ^{Romay, Geraud-Pouey, Chirinos & Demey, 2016}); however, independently of the host plants, during this study, differences regarding their altitudinal distribution were found. Thus, L. sativaeand B. tabaciwere found in low areas of El Oro and Guayas; meanwhile L. huidobrensis and T. vaporariorum, were found in high areas of Chimborazo, Loja, and El Oro. for the control of leaf miners, farmers mainly use abamectin and cypermethrin plus lambda-cyhalothrin (table 5); however, they also use organophosphates (methamidophos) and carbamates (methomyl) (table 4), performing on average, a weekly application, gene- rally mixing some of the insecticides mentioned above. In the case of the black aphid and whiteflies, farmers indicated applying acetamiprid and buprofezin, respectively (table 7).

The way in which farmers reported using insecticides in common beans, especially to control leaf miners, is controversial, given the presence of natural biological controllers. Among these, we found some parasitoid species, such as Closterocerus sp. and Chrysocharis sp. (Hymenoptera: Eulophidae) (larval parasitoids), as well as Ganaspidium sp. (Hymenoptera: figitidae) and Halticoptera sp. (Hymenoptera: Pteromalidae) (parasitoids larva-puparium). This diversity of parasitoids associated with Liriomyza is frequently mentioned in the literature (^{Abdul-Rassouly & Al-Saffar, 2014}; ^{Hernández, Guo, Harris, & Liu 2011}; ^{Osmankhil, Mochizuki, Hamasaki, & Iwabuchi, 2010}).

Table 4. Extremely dangerous (Ia) and very dangerous (Ib) toxic insecticides (Class I) with the dosage (g or mL of active ingredient/L) used (DU) and recommended (RD), and the use percentage in major crops in the provinces of Ecuador assessed 

Elaborated by the authors

Field experiments conducted by Chirinos et al. (²⁰¹⁷) with the leaf miner L. sativae in Ecuador, showed that in common bean plots free of insecticide applications, there was high parasitism associated with low populations of this phytophagous. This contrasts with what was observed in other plots treated with continuous applications of chemical insecticides, where high levels of pest populations were associated with a low percentage of parasitism. These results, together with research conducted in other latitudes (^{Chirinos et al., 2014}; ^{Tran, 2009}; ^{Tran, Tran, Konishi, & Takagi, 2006}), suggest that insecticide applications for insect control are unnecessary and interfere with the natural biological control of these species, which can generate resistance risks.

Solanaceae: Solanum tuberosum: potato

Farmers mentioned the Andean potato weevil Premnotrypes vorax Hustache (Coleoptera: Curculionidae) as the most relevant problem for potato cultivation. The flea beetles Epitrix sp. (Coleoptera: Chrysomelidae) and several species of mining moths (Lepidoptera: Gelechiidae) were referred to in second and third order of importance (p < 0.05), respectively. for the control of these insects, farmers carried out the application of insecticides every 15 days, which is similar to what has been found in other areas of the country (^{Aldás, 2012}). Premnotrypes vorax is considered an essential problem for this crop (^{Gallegos, Avalos, & Castillo, 1997}). To control it, despite some legal restrictions (^{Agencia de Regulación y Control Fito y Zoosanitario [Agrocalidad], 2013}), farmers used carbofuran. furthermore, the use of insecticides containing mixtures of lambda-cyhalothrin and thiamethoxam, as well as others used individually (lambda-cyhalothrin, cypermethrin, methamidophos or profenofos) was also highlighted. These results are similar to those obtained by Aldás (²⁰¹²), who observed that the insecticides mostly used in the control of potato pests were mixtures of commercial formulations of lambda-cyhalothrin and thiamethoxam, or carbofuran and lambda-cyhalothrin.

Table 5. Moderately hazardous insecticides (Class II) and dosage (g or mL of active ingredient/L) used (DU) and recommended (RD) in some crops in the southwestern provinces of Ecuador assessed 

Elaborated by the authors

Given the importance of P. vorax as a pest species of this crop in Ecuador (^{Landázuri, Gallegos, & Barriga, 2005}) and the other species mentioned, as well as the impact of the insecticides used to control these, it is necessary to conduct studies of the population dynamics and the damages caused in observation lots with no application and interference of pesticides. This will define the real limiting factors of production that allow the evaluation and propose more rational management strategies. There are previous experiences in which alternatives of lower environmental impact have been tested both for P. vorax (^{Landázuri et al., 2005}) as well as for Tecia solanivora (Povolny) (Lepidoptera: Gelechiidae) (^{Bosa et al., 2008}).

Solanaceae: Capsicum annuum: pepper

The major entomological problems reported for pepper cultivation were the aphids A. gossypii and Myzus persicae Sulzer (Hemiptera: Aphididae), the mite species Polyphagotarsonemus latus Banks (Acari: Tarsonemidae) and Tetranychus urticae Koch (Acari: Tetranychidae) (Tetranychidae) 0.05), differing significantly from the gall midges Prodiplosis longifila (Gagné) (Diptera: Cecidomyiidae). To control these problems, farmers made an average of two weekly sprays of insecticides (table 3). During the field inspections carried out to corroborate the presence of the mentioned species, no infestation or damage by P. longifila was observed in any of the pepper crops examined (92 pepper fields), despite being referred to as the second cause of insecticide applications. Valarezo, Cañarte, Navarrete and Arias (²⁰⁰³), reported that in interviews with farmers from different provinces of Ecuador, they listed at least 15 host crops for this species, including pepper. However, the same research indicated that an appreciable percentage of farmers (up to 70 %) mentioned that they did not know other hosts for “this pest” other than tomato.

Prodiplosis longifila is a critical problem in tomato, but not in pepper. According to observations made in pepper fields, the most notorious damages in this crop are those caused by aphids and mites, which prior to evaluation, could easily be controlled with selective applications of neonicotinoids and specific acaricides, respectively. In other words, for pepper, no major phytosanitary limitations were observed and, consequently, the amount of pesticide applications is not justified, including some of high environmental impact and toxicity such as carbofuran, methomyl, and methamidophos (table 4).

Table 6. Slightly toxic insecticides (Class III) and dosage (g or mL of i.a.L-1) used (DU) and recommended (RD) in some crops in the southwestern provinces of Ecuador 

Elaborated by the authors

Solanaceae: Solanum lycopersicum: tomato

In this crop, although farmers mentioned whiteflies (Hemiptera: Aleyrodidae) and Liriomyza spp. as problems, P. longifila was the main reason for insecticide application (p < 0.05), initiated a few days after transplantation, and reaching an average of 2.8 applications per week (table 3) (range: 2-4 applications/week). Valarezo et al. (²⁰⁰³) pointed out that for about three decades, P. longifila is the species considered as the principal tomato pest of importance in Ecuador. This species was first detected in the country in 1986 in the Arenillas canton (El Oro province), and it was later found in 12 provinces of the coastal zone and inter-Andean valleys, with an altitudinal distribution from the sea level to 1,800 m a.s.l.

These same researchers mentioned that this insect had restricted the tomato production (about 6,000 ha), so all producers have indiscriminately applied chemical products, and in some areas, farmers left their crops without management due to total loss of profitability. The 2.8 weekly applications detected in this study would represent a total of 45 applications per crop cycle, which is much higher than what Valarezo et al. (²⁰⁰³) reported. These authors referred to a range of 21 to 31 applications made per crop cycle in several provinces of the country and agree that it is an indiscriminate use of pesticides. In summary, the high use of insecticides to control P. longifilais probably accentuating phytosanitary, as well as human and environmental health problems (^{Lindao, Jave, Retuerto, Erazo, & Echeverría, 2017}).

Chemical products used

The results showed frequent applications of chemical insecticides made by farmers, which corroborates what was mentioned in previous research carried out throughout the country (^{Aldás, 2012}; ^{Crissman et al., 2002}; ^{Valarezo et al., 2003}, ²⁰⁰⁸) and in Latin America (^{Chirinos & Geraud-Pouey, 2011}; ^{Ruiz, Ruiz, Guzmán, & Perez, 2011}; ^{Wesseling et al., 2003}). None of the farmers stated that they follow the recommendations written in the label of the insecticide containers regarding the waiting period or confidence interval before harvest, necessary to ensure the decomposition of the pesticide.

Of the insecticides used, 31.2, 46.8, 12.9, and 8.4 % are classified in the Toxicity Class I, II, III, and IV, respectively (tables 4-7) (χ²: 96.89; p < 0.09). These results show that the most toxic insecticides, belonging to classes I and II, constituted approximately 80 % of the applications in crops mentioned by farmers. Of the four categories, Class I products referred to by producers, i.e., methamidophos, methomyl, and profenofos, are restricted; moreover, carbofuran was banned in the country (^{Agro-calidad, 2013}). Except for profenofos, these insecticides are proven to be the major causes of acute poisoning in humans in other countries (^{fernández et al., 2010}).

All farmers (table 8) mentioned mixing two or more insecticides in one spraying session. This is similar to the results obtained by Vargas-González et al. (²⁰¹⁶), who detected that 95 % of the farmers surveyed in the melon-producing region of Comarca Lagunera, Mexico, mixed at least two pesticides during one spraying session. When more than one pesticide is used in a spray, the individual toxicity scales lose validity due to the additive effects of the mixtures, which is especially relevant because all the dosages used were significantly higher (W : 4207; p < 0.05) to those recommended on the product label (tables 4-7). The additive effects of pesticides occur because these exert a joint action and will be the result of the sum of each of them (^{Stephens, Maige, Beltran, & González, 2005}).

Table 7. Use percentage of insecticides that probably do not cause any risks (Class IV) and dosage used (DU) and recommended (RD), expressed in grams or milliliters per liter 

Elaborated by the authors

Table 8. Percentage of farmers who mix products and source of pest management consultancy 

Elaborated by the authors

It was common to observe empty chemical insecticide containers scattered in the crop fields (figure 3). As for the preparations and spraying sessions, these were carried out by agricultural operators without body protection, which could compromise their health, especially as most were relatively young (^{Lindao et al., 2017}). furthermore, the manipulation of fruits covered with pesticide residues during the collection, selection, and cleaning using their bare hands, and the accommodation of these in the transport baskets, are operations that are usually carried out by women and children. Studies in Ecuador detected the presence of chlorinated pesticide residues in the final potato, pepper, and tomato products, but with quantities that did not exceed the maximum permitted limits (^{Ministerio de Ambiente, 2004}). Other research carried out on the basic family basket in tomato samples collected in four provinces of the country, showed that methamidophos pesticide residues had values eight times above the Maximum Residue Limit (MRL) (^{Crissman et al., 2002}) for the control of agricultural pests, an average of 58 % (range: 42-69 %) of the interviewees reported that representatives of pesticide trading houses advise them on this topic (table 8). Valarezo et al. (²⁰⁰³) reported that chemical recommendations to farmers come from sellers of agricultural inputs, and, on other occasions, they rely on their own experience.

The results show the high use of chemical insecticides in the locations and crops studied. These are widely used in agriculture because they are effective, have a rapid effect and are flexible in adapting to many agronomic and ecological conditions. Besides, it is the only pest management tool that can be used when damage levels are excessive and are an effective alternative. for these to be used harmoniously within the context of pest management, it is necessary to replace the application schedules with the treatments that are required and recognize that 100 % insect control is not necessary to prevent economic losses (^{Luckmann & Metcalf, 1975}).

Additionally, the efficiency of sprays should be improved and, where the situation allows, localized and targeted applications should be performed. It is also vital to reduce the dosages to the minimum necessary and respect the waiting periods or application times established before harvest. These results demand studies of pesticide residuals in the final agricultural products, as well as the impact on the health of operators and consumers, and the effects on the environment due to the use of chemical pesticides.

Given this serious situation of high application of chemical insecticides, it is essential to propose alternatives, for which knowledge about the ecological dynamics of agroecosystems is essential, especially with regard to arthropod fauna and their trophic interactions, to define the real problems that need to be managed and how to do it, without generating new problems (^{Chirinos & Geraud- Pouey, 2011}). The first alternative to be considered within an integrated pest management program is the action of the natural biological controllers that, in some cases, such as the species of the Liriomyza genus, is generally sufficient to maintain the population of this phytophagous insect controlled in different crops (^{Chirinos et al., 2014}; ²⁰¹⁷; ^{Tran et al., 2006}).

When natural biological control is not enough, other alternatives must be considered before chemical control is used. The biological control applied has been successful in management programs of some agricultural pests in several countries of the Neotropical region (^{Colmenarez, Corniani, Jahnke, Sampaio, & Vásquez, 2018}). Microbial insecticides are also part of low environmental impact management strategies, due to their biological selectivity, whose active ingredients based on fungi, bacteria, and baculovirus, among others, have demonstrated effectiveness for the control of different important pest species in various crops (^{Ayala & Henderson, 2017}; ^{Landaruzi et al., 2005}; ^{Portela-Dussán, Chaparro-Giraldo, & López-Pazos, 2013}). Likewise, botanical insecticides, as well as the use of pheromones, have shown their effectiveness in the control of agricultural pests (^{Bosa et al., 2008}; ^{Campos et al., 2018}).

Figure 3. Pesticide containers dispersed in crop fields in Ecuador. a. Arenillas, province of Gold; b. Pedro Carbo, Guayas province; C. Pallatanga, province of Chimborazo; d. El Morro, province of Santa Elena