Introduction

The common bean (Phaseolus vulgaris L.) has a high economic and social importance in Brazil, because of its nutritional properties and the labor employed in its cultivation (^{Pedrosa et al. 2015}). However, pest insects attacks cause losses in crop productivity (^{Jesus et al. 2010}a, ^b; ^{Janini et al. 2011}). Among the insect pests attacking common bean, the fall armyworm - Spodoptera frugiperda (J. E. Smith, 1797) (Lepidoptera: Noctuidae) is of major importance because the feeding damage results in leaf area reduction and destruction of the reproductive structures.

Spraying of insecticides is the most common means of controlling S. frugiperda. However, this practice contributes to unbalanced agroecosystems(^{Bernardi et al. 2012}) and the selection of insect resistant populations (^{Cruz et al. 2010}). The use of alternative pest control methods, such as plant resistance to insects, is an important component in an integrated pest management (IPM) approach(^{Jesus et al. 2010a}, ^2010b; ^{Janini et al. 2011}; ^{Silva et al. 2016}).

Plant resistance to insects (PRI) is considered an efficient method of pest control. PRI maintains the pest population below the level of economic damage, does not cause adverse effect on the environment, does not incur additional cost and may be compatible with other control methods (^{Lara 1991}; ^{Smith 2005};^{War et al. 2012};^{Seifi et al. 2013}).

Plant resistance mechanisms are classified as antixenosis, antibiosis and tolerance. Antixenosis occurs when there is a deterrence ofthe plant for feeding, oviposition or shelter. Antibiosisis characterized by detrimental effects on insectbiology and toleranceis the plant capacity to endure or recover from insect caused damage, through the production of new vegetative or reproductive structures (^{Gatehouse 2002}; ^{Cunningham 2012};^{War et al. 2012};^{Seifi et al. 2013}).

Several studies to select bean genotypes resistant to stored grain pests have been conducted. However, few studies have examined insects that attack bean plantsin the field. Considering the importance of bean crops for food security in Brazil, and the damage caused by defoliating caterpillars that injure reproductive structures, the selection of resistant genotypes is an important IPM strategy (^{Maldonado et al. 1996};^{Bottega et al. 2011};^{Boiça Júnior et al. 2015}).

In previous studies the bean genotypes IAC Jabola, Arcelina 1, IAC Boreal, Flor de Mayo and IAC Formoso have been reported to present antixenosis to oviposition, and Arcelina 4, BRS Horizonte, Pérola, H96A102-1-1-1-52, IAC Boreal, IAC Harmonia and IAC Formoso present edantixenosis to feeding by Chrysodeixis includes (Walker, 1858) (Lepidoptera: Noctuidae) (^{Morando et al. 2015}).Our study was conducted to identify resistance in common bean cultivars to S. frugiperda. In this preliminary study, leaves of the common bean were evaluated for antixenosis and antibiosis in the laboratory. Further field evaluations will be based on the results of this study.

Materials and methods

The experiments were conducted in a climate-controlled room (25 ± 1°C, 70 ± 10 % RH and a 12-h photophase) at the Laboratory of Agricultural Entomology of the Goiano Federal Institute, Urutaí Campus, Brazil. The test varieties were obtained from the germplasm bank of the National Center of Research Rice and Beans -EMBRAPA Rice and Beans, Santo Antonio de Goiás, Goiás, Brazil. The following common bean cultivars were evaluated: BRS Ametista, Pérola, BRS Notável, BRS Realce, Jalo Precoce, BRS Campeiro, BRS Agreste, BRS Cometa, BRS Executivo and BRS Pitanga. These varieties were selected because common bean producers in Brazil currently cultivate them.

To rear S. frugiperda, adults collected in the field were placed in cages of polyvinyl chloride (PVC) tubes (10 cm in diameter and 21.5 cm high) covered internally with white paper sheets, which served as oviposition sites (^{Campos et al. 2012}; ^{Jesus et al. 2014}). The newly hatched larvae were placed on artificial diet for rearing (^{Greene et al. 1976}). The common bean cultivars were sown in 5 L plastic pots containing soil, manure and sand (2:1:1) mixture as substrate. Following germination, plants were thinned to one plant per pot, and grown in a greenhouse until 30 days after emergence (DAE).

Results and discussion

The free-choice attractiveness test revealed differencesamong the common bean cultivars tested to S. frugiperda at 5 and 30 minutes, as well asat 3,6 and 12 hours (Table 1). At 5 min larvae of S. frugiperda were not attracted by cultivars BRS Ametista, Pérola, BRS Notável, BRS Agreste, BRS Cometa and BRS Pitanga. BRS Realce and Jalo Precoce were the most attractiveat 5 min. At 30 min Pérola, BRS Executivo and BRS Pitanga were not infested and Jalo Precoce and BRS Campeiro were the most attractive cultivars to S. frugiperda. At 3 h all genotypes were not infested by S. frugiperda except BRS Agreste and BRS Campeiro. At 3 h the cultivar BRS Campeiro showed the most attractiveness to S. frugiperda. Overall, the cultivars BRS Pitanga and BRS Executivo were the least attractive to S. frugiperda. BRS Campeiro was the most attractive.

The no-choice attractiveness test revealed significant differences between the cultivars at 5, 15 and 30 min. At 5 min BRS Notável, BRS Campeiro, BRS Executivo were not infested and BRS Agreste was the most attractive to S. frugiperda. At 15 min Pérola, BRS Notável, BRS Realce, Jalo Precoce, BRS Campeiro and BRS Pitanga were not infested and BRS Agreste was the most attractive and 30 min Pérola had low larval numbers and BRS Agreste again was the most attractive cultivar to S. frugiperda. Generally, BRS Notável, BRS Campeiro, BRS Executivo and BRS Pitanga cultivars were the least attractive to S. frugiperda and BRS Agreste was the most attractive.

BRS Pitanga, BRS Executivo and BRS Campeiro cultivars showed antixenosis to S. frugiperda (Tables 1 and 2). This indicates the presence of chemical compounds and/or morphological leaf characteristics in these cultivars that are repellent to S. frugiperda (^{Cunningham 2012}). Few studies have characterized the types of resistance present in common bean cultivars upon defoliating insects. ^{Morando et al. (2015)} reported antixenosis to C. includens in H96A102-1-1-1-52, Flor de Mayo and Arcelina 4 genotypes. ^{Boiça Junior et al. (2015)} observed antixenosisin IAC Una, IAC Uirapuru, IAC Diplomata and IAC Onixcultivars and tolerancein IAC Onix and IAC Una upon Diabrotica speciosa (Germar, 1924) (Coleoptera: Chrysomelidae).

The hierarchical cluster analysis performed for both attractiveness and antixenosis resistance againstthe third-instar larvae of S. frugiperda that were evaluated atdifferent time periods in free-choice and no-choice tests revealed significant differences between the tested cultivars, separating them into groups according to their degree of similarity (Fig. 1). BRS Pitanga, Perola, BRS Ametista and BRS Executivoat Euclidean distance 3.3, 3.5 and 4.7, respectively, were grouped indicating similarity among these cultivars (Fig. 1). The second group was formed by BRS Cometa and BRS Notável at Euclidean distance 4.7, and the third group included Jalo Precoce and BRS Realce at Euclidean distance 4.9. Finally, all the cultivars were grouped at 7.4 (Fig. 1).

Figure 1 Dendrogram based on attractiveness of third instar Spodoptera frugiperda in free and non-choice tests by common bean. The hierarchical cluster analysis was carried out using the Ward’s method with the Euclidean distance as dissimilarity measure. The arrow indicates the Euclidean distance used for the separation of groups (phenon line). 

Considering a Euclidean distance of 5.0 as the criterion for group division, the tested cultivars were separated according to the level of antixenosis resistance as follows: moderately resistant - BRS Campeiro and BRS Agreste; susceptible - Jalo Precoce, BRS Realce, BRS Cometa and BRS Notável; and highlysusceptible - BRS Executivo, BRS Pitanga, Pérola and BRS Ametista.

Comparing the results of the hierarchical cluster analysis with those of the free choice and no-choice attractiveness tests, BRS Campeiro was found to be one of the most attractive cultivarto S. frugiperda in free-choice test, despite its inclusion in the moderately resistant group (Table 1 and Fig. 1).

The common bean cultivars presented a significant effect on the biological parameters of S. frugiperda (Table 3). The larvae fed on cultivar BRS Realce presented a longer larval stage compared to the remaining cultivars, exceptfor BRS Ametista. The larval phase prolongation in S. frugiperda caused by BRS Realce is characteristic of antibiosis (^{Smith 2005};^{Seifi et al. 2013}). This data is in agreement with those of^{Jesus et al. (2014)} that found antibiosis on cotton cultivar Nuopal to S. frugiperda that completed the larval stage in 23.3 days. ^{Paiva et al. (2016)} observed that S. frugiperda completed the same stage in BX 1293 YG antibiosis resistant corn in 31.85 days. Larvae of S. frugiperda fed on artificial diet had their larval stage completed in 16 days (^{Silva et al. 2016}).

The pupae originating from larvae fed on BRS Realce also presented an extended pupal stage that was 6.8 days longer than those from Jalo Precoce. The adults presented a higher longevity after feeding on BRS Notável compared to Pérola (Table 4). The same authors mentioned previously found extension in the life cycle of S. frugiperda when fed on resistant plants. It is important to use as IPM tactic because over time there is a reduction in the population density of the pest (^{Gatehouse 2002}).

The observed extension of the larval and pupal stages on S. frugiperda fed on BRS Realce indicates a detrimental effect of the plant on their biology. S. frugiperda needed more time to complete the larval and pupal stages than those feeding on a susceptible cultivar. Delays on S. frugiperda development are usually associated with antibiosis. Antibiosis is characterized by an extension on the larval stage when the insect feeds on a resistant host plant. This results in negative effects on initial insect developmental stages, such as high mortality, deformities and decreased body weight (^{Lara 1991}; ^{War et al. 2012}; ^{Seifi et al. 2013}).

The principal component analysis separated the bean cultivars into four levels of antibiosis resistance: highly resistant - BRS Notável; moderately resistant - BRS Pitanga, BRS Campeiro, BRS Cometa and BRS Realce; susceptible - BRS Ametista, BRS Executivo and Pérola; and highly susceptible - Jalo Precoce and BRS Agreste (Fig. 2).

Figure 2 Distribution of the common bean cultivars according to principal component analysis based on biological parameters of Spodoptera frugiperda. 

Conclusions

The cultivars BRS Pitanga, BRS Executivo, BRS Notável and BRS Campeiro presented antixenosis and BRS Realce antibiosis to S. frugiperda under laboratory conditions. However, it is not known whether the levels of resistance, exhibited in the laboratory, are sufficiently high, to be of any economic value in farmers’ fields. Therefore, the next step is to evaluate, under field conditions, the cultivars showing the most resistance in the laboratory tests. If field evaluations indicate sufficient levels of resistance, to be of practical value, these cultivars may be used as donor sources in the breeding program or used directly by farmers.