The guava crop, Psidium guajava L. (Myrtaceae), despite its growing economic importance in Brazil (Pereira et al. 2017), has productivity reduced due principally to the occurrence of some key pests (Souza Filho & Costa 2003), such as the triozid Triozoida limbata (Enderlein) (Hemiptera: Triozidae) (Burckhardt & Queiroz 2012; Sá & Fernandes 2015; Queiroz et al. 2018).

Triozoid limbata immatures wrap the margins of the guava leaves forming something similar to galls, where they are found; they principally use the shoots (Queiroz et al. 2018). The attack intensity of this species to guava is favored by mild temperatures (18 to 20 °C) and a shorter photoperiod (10 to 11 h of light) (Martins 2008).

The presence of T. limbata occurs on new shoots in conventional orchards, favoring their occurrence yr round (Melo et al. 2009). However, the highest population densities of adults in semi-organic orchards occur during spring and summer (Duarte et al. 2015).

The main control tactic for T. limbata in guava is the use of chemical pesticides. According to Pazini & Galli (2011), the only insecticide registered to control T. limbata is imidacloprid. The use of guava cultivars less susceptible to attack by T. limbata should be explored in the integrated pest management programs of guava crops.

Plants naturally have a certain degree of insect resistance, and for many yr biosynthesis, and regulation of plant chemical compounds associated with these defenses has been studied. Defensive substances are found in various plant tissues; among these compounds are antibiotics, alkaloids, terpenes, and proteins. Among the proteins are enzymes such as chitinases, lectins, and inhibitors of digestive enzymes (Ryan 1990).

Varietal resistance is a phenomenon related to the insect-plant interaction, and can be examined or defined from the point of view of one or both of the interacting agents (Painter 1951). Resistant plants negatively affect the reproductive potential, survival, and development of phytophagous insects, restricting their abundance (São João & Raga 2016). The relationship between the genetic variation of plants and environmental factors act to develop mechanisms of resistance to the attack of herbivores (Aoyama & Labinas 2012). On the other hand, herbivores also evolve to adapt and break these resistance mechanisms (Mauricio et al. 1997), which are expressed by plants as constitutive defense and induced defense, with direct or indirect action on herbivorous arthropods (Aoyama & Labinas 2012, São João & Raga 2016).

The plants present 2 basic mechanisms of defense against herbivory: constitutive defense and induced defense. The former is represented by chemical compounds and morphological structures that hinder herbivorous access to plants, and may affect some parameters of the biological cycle, such as the development and reproduction of insects. The second refers to morphological or physiological changes resulting from insect actions upon plants, causing a non-preference (antixenosis) of such insects for such plant varieties. The principal physiological changes of plants to resist herbivores are the reduction of their nutritional quality (Coley & Barone 1996; Aoyama & Labinas 2012; São João & Raga 2016).

The process of co-evolution between plants and insects constitutes a system of dynamic interactions. Plant chemical compounds, such as attractants, feed stimulants, repellents, deterrents, hormone sources, pheromones, and kairomones, are involved in various metabolic and behavioral processes of insects (Lovatto et al. 2012).

The plants present several forms of defense (resistance) to the attack of pests. Three types of plant resistance to insects are reported: antixenosis (non-preference), antibiosis, and tolerance. Antixenose is the response of an insect, involving behavioral processes that result in avoiding certain plants as food or as a substrate for oviposition. Antibiosis occurs when the insect feeds normally from the plant, but this has an adverse effect on its biology. Tolerance is defined as the plant's own ability to withstand or recover from damage produced by an insect population, which would normally cause serious damage to a more susceptible host (Painter 1951; Kogan & Ortman 1978; Lara 1991).

There is a lack of information on the evaluation of strategies for population monitoring and control of T. limbata. There are several cultivars of P. guajava in Brazil. With the intense process of breeding of this species, the cultivars become very different among themselves: canopy format, productivity, production time (early, mid-season, and late), number, size and shape of fruit, pulp color, post-harvest shelf life, and resistance to pests and diseases (Fumis & Sampaio 2011; Pommer et al. 2012). This can affect the behavior of insects, among them T. limbata, principally regarding its colonization in the guava tree. Our hypothesis is that some cultivars of P. guajava present resistance against the triozid.

Considering the lack of information essential to the selection of strategies for population monitoring and control of T. limbata, this research evaluated their preference for accessions in guava cultivars, evaluated the incidence of their nymphs and adults among the positions of the branches in the plants, and detected less susceptible varieties.

Materials and Methods

STUDY AREA

The evaluations were carried out in an experimental orchard of guava at Embrapa Agropecuária Oeste, Dourados, Mato Grosso do Sul, Brazil (22.277847°S, 54.818502°W; 430 masl) during the period of May 2013 to Jul 2014. The local climate is Cwa type (humid mesothermal climate, warm rainy summers and dry winters), Jun and Jul having low temperatures (< 18 °C), and Jan the hottest month (> 22 °C) (Fietz & Fisch 2008). The soil is characterized as dystrophic purple latosol (Ker 1997).

The sample area comprised 0.4 ha with 8 cultivars of P. guajava distributed in 4 blocks with 3 plants of each cultivar in sequence. The plant spacing is 7 m × 5 m, and the cultivars evaluated were ‘Pedro Sato,' ‘Kumagai,' ‘Paluma,' ‘Cascuda,' ‘Século XXI,' ‘Sassaoka,' ‘Novo Milênio,' and ‘Tailandesa.’ At the beginning of this study, the plants were in the first cycle of fruiting (2 yr old).

Samplings were biweekly, quantifying the number of nymphs and adults of T. limbata in guava trees, presents in the upper third of the trees from the apex to the second pair of fully expanded leaves. We examined 3 leaves randomly per main branches, 4 branches per tree (between plants and between streets). We evaluated 96 plants in the area used, sampling 4 branches from each plant, and 12 leaves per plant, totalizing 1,152 leaves (by visual count or direct method).

Each plant was also monitored with a yellow adhesive trap (10 × 15 cm) fixed at the height of the canopy approximately 1.5 m above ground level (indirect method). These traps were placed in the same guava orchard evaluated by the direct method, and were exchanged every 2 wk. Then, traps were wrapped with transparent PVC film, packed in a plastic paste for subsequent quantification of the captured individuals. The counts were performed at the Laboratório de Insetos Frugívoros, Universidade Federal da Grande Dourados, Dourados, Mato Grosso do Sul, Brazil.

Management Practices in the Orchard

During the sampling period, routine cultural treatments were applied on the entire sampling area on the same date for all treatments, such as fertilization, control of invasive plants with mechanized brush-cutter, pruning for fruit production, and application of fungicides and insecticides when pathogens or pests reached the control level. Irrigation was through a micro sprinkler.

Fertilization occurred from Sep to Dec 2013, and Jan 2014, in a total of 9 applications. Six fungicide applications were made between Oct and Nov 2013 and May 2014. Twelve applications of insecticides were made from Oct 2013 to May 2014, aiming at the control of insect pests, particularly: defoliating, fruit flies and other phytophagous insects.

Experimental Design and Statistical Analysis

The design was in a completely randomized block, consisting of 8 treatments (8 cultivars) and 4 replications, with a sample size of 3 plants. For statistical analysis of the data, we used a controlled experimental design in the field. To reach the proposed objectives, 2 assays were prepared.

Assay I. The objective of this study was to determine if there are more attractive cultivars for the triozid (immatures and adults) through direct sampling in branches and indirect sampling with yellow traps.

The design of the first sampling consisted of the evaluation of each cultivar consisting of 4 plots with sets of 3 guava trees per cultivar, with observation made in the longitudinal and transverse positions. For the total number of immatures and adults quantified in the 4 plots of plants, the procedure was repeated 32 times, totaling 256 replicate samples for each cultivar.

For the second sample design, 96 yellow adhesive traps were installed to intercept the adults of T. limbata in flight. The sample design consisted of 8 cultivars, each cultivar represented by 4 plots (3 plants of each cultivar), adding 12 traps for each cultivar. This procedure was repeated 30 times, totaling 120 replicate samples for each cultivar.

Assay II. The objective of this study was to determine if the incidence of T. limbata (immatures and adults) in branches is equal between the positions. The design of this sampling consisted of 8 cultivars and 2 positions: longitudinal lines between plants (Q1–2), and transverse columns between streets (Q3–4). Each position consisted of the sum of 2 consecutive quadrants (Q1–2 = East–West, Q3–4 = South–North). For this evaluation the total number of immatures and adults quantified was as described in Assay I.

In both experiments, before applying the analysis of variance and multiple comparison tests, it was determined whether the assumptions of normality (Kolmogorov-Smirnov) and homogeneity (Levene) required in parametric models were attended by the variables measured. Because these pre-assumptions and independence within and between the variables were not met, Kruskal-Wallis nonparametric statistical tests were used to determine the variance analysis, and the Dunn-Bonferroni bilateral test was used to make the paired comparisons.

Table 1.

Antixenosis of nymphs and adults of Triozoida limbata (Hemiptera: Triozidae) evaluated by direct method in 8 different cultivars of guava, Psidium guajava L. (Myrtaceae), Dourados, Mato Grosso do Sul, Brazil (May 2013 to Jul 2014).

All the results were analyzed considering the level of significance α = 5%. The values were processed in the SPSS Predictive Analytics Software, version 21 (IBM Corp., Armonk, New York, USA) by Odival Faccenda at Universidade Estadual de Mato Grosso do Sul, Dourados, Mato Grosso do Sul, Brazil.

Results

INCIDENCE OF ADULT TRIOZOIDA LIMBATA ON PLANTS OF 8 GUAVA CULTIVARS IN YELLOW TRAPS (INDIRECT SAMPLING)

The cultivars ‘Cascuda,' ‘Kumagai,' and ‘Sassaoka' had the lowest mean number of adult T. limbata compared to the ‘Paluma' cultivar. The latter had the highest mean number of individuals. The cultivars ‘Tailandesa,' ‘Pedro Sato,' ‘Novo Milênio,' and ‘Século XXI’ are an intermediate group that did not differ among themselves. The high values of the standard deviation were due to the distinct preference or rejection of T. limbata to the different cultivars (Table 2).

Table 2.

Antixenosis of adults of Triozoida limbata (Hemiptera: Triozidae) evaluated by indirect method in guava cultivars, Psidium guajava L. (Myrtaceae), municipality of Dourados, Mato Grosso do Sul, Brazil (May 2013 to Jul 2014).

Discussion

Due to the fact that this research was conducted in an orchard under routine cultural practices (pruning, fertilization, and pesticide applications), no correlation with weather data was tested because the experiment should be free of cultural practices to isolate the effects of abiotic factors.

There was an increase in triozid abundance during the sample period, regardless of climatic variations. In this context, the results of the publications that correlated populations of T. limbata with weather data are conflicting. Several papers published in Brazil report this conflict in relation to the influence of weather. Some studies conclude that there is influence of temperature on the population density of the triozid, both in the ‘Paluma' cultivar (Colombi & Galli 2009) and in native guava trees (Dalberto et al. 2004). On the other hand, other studies report the absence of temperature and rainfall effects on the populations of T. limbata adults in orchards with a collection of guava genotypes (92 cultivars) (Duarte et al. 2015), and in orchards with the cultivar ‘Paluma’ (Pazini & Galli 2011).

Table 3.

Population distribution of nymphs and adults of Triozoida limbata (Hemiptera: Triozidae) in the branches of Psidium guajava L. (Myrtaceae) from 8 different cultivars, i.e., Longitudinal, between plants (East-West), versus Transverse, between streets (South-North); trees 2 yr old in an orchard in the municipality of Dourados, Mato Grosso do Sul, Brazil (May 2013 to Jul 2014).

Scientific publications widely report that the climatic factors affect the abundance of the triozid. According to Queiroz et al. (2018), T. limbata is restricted to Central and South America, although guava is currently cultivated in several warm regions around the world. These authors, through the modeling of potential distributions based on distributive data from the Americas, have inferred that the effects of global climate change will be more evident in the east and northeast of Brazil, where there will be an increase in the concentration of this pest species exerting more pressure on the production of guava, and increasing the potential for damage induced by the guava's triozid.

During the sampling period of this research, large temperature variations occurred (–0.7 to 35.8 °C). The beginning of sampling coincided with the low and high temperature periods, and with the pruning and opening of the 4 main branches of each plant to for the first yr of fruiting. The females of T. limbata lay their eggs along branches, new shoots, and new leaves (Nakano & Silveira Neto 1968). Due to this behavior, there were few immatures at the beginning of the sampling period because to a lack of new leaves in the plants, as a consequence of the pruning to induce flowering and fruiting. This fact was expected, because it is known that the weather influences insect population densities (Huffaker & Gutierrez 1999).

The pruning to induce fruiting of the plants was carried out between 9 and 20 September 2013. After this period, there was an increase in the T. limbata population density, constantly reaching the control level, generating the need to perform several actions for population suppression. However, after Apr 2014 (the 21st evaluation), control actions were no longer necessary due to the beginning of fruit harvest. In contrast to the previous yr, after harvesting there was neither fruit pruning nor a marked drop in temperature, which led to the rapid development of tree canopies, increasing the availability of new leaves, and promoting rapid growth of the T. limbata population. This shows that the abundance of herbivorous insects is affected by both the availability of food resources (Pinheiro et al. 2002; Araújo 2013) and the weather (Wolda 1988).

The occurrence of insect pests is among the factors that most influence the productivity of agricultural crops, because they directly consume parts of plants and can transmit diseases. Therefore, it is expected that when plants are well-nourished (adequate fertilization), they better tolerate pest attack (Bianchini et al. 2015). However, the effect of fertilization on the production of the 8 cultivars was not evaluated, but the spatial distribution of T. limbata in the evaluated cultivars was studied. This triozid presented aggregate behavior, demonstrating preference for some guava cultivars. According to Bianchini et al. (2015), the effect of plant fertilization on pest incidence often is not significant, but there is an effect on the plant that is better nourished, making it more tolerant to herbivores.

Plant species generally exhibit a certain degree of resistance to herbivorous insects (Ryan 1990). These also present defenses that can affect the growth and development of insects through toxic secondary metabolites (War et al. 2018). It is equally important to understand insect-pest adaptations to these defensive characteristics in order to develop and deploy management strategies to suppress insect pest populations (War et al. 2018). Preferences of insect pests to host plants often can be modified by experience and conditioning (Barron 2001; Santa-Cecilia et al. 2013). It is likely that the adult females of T. limbata define which plants will receive their offspring, because we observed the preference of adults for certain cultivars. As described by Painter (1951), when insect females are caged in resistant plants they deposit fewer eggs than females caged in plants of susceptible varieties.

Susceptible guava cultivars may offer better conditions for T. limbata to develop and reach adulthood. On the other hand, in resistant plants this species has difficulty in establishing. Such plants generally negatively affect the biology of insect pests (Gould 1998) because physiological changes in the induced defense mechanism of plants, such as antixenosis or non-preference, or also a decrease in the nutritional quality for their herbivores may occur. Thus, resistant plants may offer fewer resources to the herbivore than susceptible ones (Aoyama & Labinas 2012). In response to pest attack, they can activate defensive genes which lead to the production of physical barriers or synthesis of chemical compounds that may prevent or reduce the feeding of phytophagous insects (Guerrieri & Digilio 2008).

As phytophagous insects require adequate habitat for the development of their offspring, host plants are essential to provide water, nutrients, shelters, mating sites, and oviposition, among other prerequisites (Edwards & Wratten 1981). In this sense, T. limbata probably co-evolved in this interaction, selecting its host plants. Ouvrard et al. (2015) have highlighted the ability of Psylloidea to adapt to host plants probably depends on their ability to deal with a complex set of biochemical defenses. On the other hand, Rossiter (1996) pointed out that it is difficult to determine if differences in the levels of preference observed in the field have a genetic basis, or are the result of intervening factors, such as temperature, plant nutrition, conditioning, and adaptation inheritance to the environmental effects. But the impact of inherited environmental effects on offspring may be positive or negative, depending on the nature of the contribution and the ecological context in which the offspring exist.

Our results show that immature as well as adults of T. limbata present antixenosis to ‘Kumagai' and ‘Cascuda' cultivars. The quantification of T. limbata populations in situ (direct sampling) and the use of adhesive traps (indirect sampling) showed the same efficiency to evaluate infestation levels in the different guava cultivars tested, as well as in the detection of the levels of damage. However, in situ we recorded the adults on the plants only at the evaluation intervals (from 7:00 A.M. to 9:00 A.M.), whereas for adults caught in traps this period extended to full time (15 d), beyond the fact that adults are also attracted by color. Thus, it was verified that indirect sampling was adequate to demonstrate the antixenosis (sensu Kogan & Ortman 1978) of adults in ‘Cascuda' and ‘Kumagai’ cultivars. The direct sampling showed that the number of immatures in these 2 cultivars differs from the others. In this context it is considered that the adults (indirect sampling) avoided these cultivars, both for feeding and for oviposition.

From the results obtained in this study, it should be clear that the preference of immatures (sensu Burckhardt et al. 2014) and adults of T. limbata is for the ‘Paluma’ guava cultivar. This fact can be explained, at least in part, by the behavioral response of this triozid to the characteristics of the host plant. This cultivar, which produces fruits destined for consumption in natura and industrialization, presented the higher populations of T. limbata. The high populations of immatures, which are responsible for the hard damage levels in guava orchards (Semeão 2006) are easily visualized in the new leaves.

In this study, T. limbata presented distinct preference for the different evaluated guava cultivars, due to the low levels of infestations observed in some cultivars. The cultivars less susceptible to the attack of T. limbata immatures and adults, were ‘Cascuda' and ‘Kumagai.’ Therefore, both are promising cultivars for guava breeding programs, aiming at the resistance of P. guajava to the attack of T. limbata. This discovery is of great importance for genetic improvement programs on guava trees, aiming to incorporate resistance to control the guavas's triozid in commercial cultivars. Our results contribute to the integrated management of guava pests, making it possible to reduce the use of pesticides and the cost of production and, consequently, environmental impact and intoxication in humans.

The fact that there is no significant difference in the number of immatures and adults among plant or between streets, as verified in this research, allows producers and technicians to save time and money for population monitoring of T. limbata in the field. Thus, in order to carry out sampling, there will be no concern as to which of the main branches (among the 4 that leave the trunk) of the plant should have its new leaves sampled. There are no differences in levels of infestation by immature and adult T. limbata in the different guava tree branches, as pointed out by Moreira (2005). Therefore, in P. guajava, for sampling of T. limbata in the field, new leaves from any of the 4 main branches can be evaluated and will allow the detection of their occurrence, and of their population density in the orchards.