FOOD PREFERENCE OF Thyrinteina arnobia (STOLL, 1782) (LEPIDOPTERA: GEOMETRIDAE) ON NATIVE AND EXOTIC HOSTS

One of the factors that may aff ect and limit the production in eucalypt plantations is the attack of defoliating insects. Among those, the brown eucalypt caterpillar, Thyrinteina arnobia (Stoll, 1782) (Lepidoptera: Geometridae), stands out for being the major defoliating pest of Eucalyptus spp. in Brazil. Thus, the present study aimed to investigate the food consumption of T. arnobia, in its native host, guava (Psidium guajava L.), and in the diff erent E. urograndis clones (VE 41, I 144, TP 361 and VCC 865). To assess T. arnobia food consumption, choice and non-choice tests were carried out using the native and the exotic host, alone or in combination. In non-choice tests, it was observed a higher consumption for the VE 41 clone and the native host (guava). The food consumption evaluation in choice tests indicated no food preference of T. arnobia between guava and E. urograndis clones, with the exception for the TP 361 clone, which was signifi cantly less consumed than guava. In choice tests between the diff erent E. urograndis clones, the clone I 144 presented a tendency towards lower food preference, being consumed only after 48 hours. In addition, the leaf consumption was similar between the VE 41, I1 14 e VCC 865 clones. In choice tests using the E. urograndis clones in pairs, the VE 41 clone was more consumed while the I 144 clone was less consumed when compared to the TP 361 clone. The obtained results provide basic information for the indication of eucalypt clones, and the understanding of the interaction and ecological relationships, assisting in the development of Forest Integrated Pest Management (Forest IPM) programs for the control of T. arnobia.


1.INTRODUCTION
For purposes of forestry genetic improvement, the genotypic features of each genetic material may directly aff ect the insect-plant interaction. Thus, it is essential to identify and characterize possible resistance factors, contributing to the development of appropriate and effi cient pest management strategies (Jesus et al., 2015).
Among the insects native to Brazil, Thyrinteina arnobia stands out for being found all over the national territory and has migrated from its native hosts, such as guava (Psidium guajava L.), to the exotic host Eucalyptus spp., becoming the main defoliator on eucalypt crops in the country, requiring the constant use of pest control techniques (Barreto and Mojena, 2014). Therefore, several management tactics are being developed to mitigate the yield losses caused by T. arnobia, including the selection of less susceptible species in the genus Eucalyptus, as well as the search for resistant genotypes or clones.
According to Boiça Júnior et al. (2013), plant resistance is a feature depending on environmental and genetic factors, and for that, diff erent genetic materials, such as species, hybrids and/or clones of Eucalyptus spp. may show variations in susceptibility to biological agents, including defoliating insects, such as T. arnobia. Thus, the search for resistance features is essential, since the presence of morphological or chemical stimuli can aff ect herbivory and depress the consumption and food preference of insects, keeping the population density of the pests below the level of economic damage, having no or low adverse eff ect on the environment (Seifi et al., 2013). Nevertheless, studies related to the interaction of native and exotic host species and the food consumption of T. arnobia, in the search for the characterization of non-preference resistance are still rare. Jesus et al. (2015) in tests using species/genotypes of Eucalyptus spp., found results that consider Eucalyptus dunnii and the hybrid Corymbia citriodora x Corymbia torelliana presented antibiosis and/or antixenosis for T. arnobia. Similarly, clone C10 (C. citriodora x C. torelliana), proved to be less preferred by T. arnobia, corroborating that T. arnobia exhibits a distinction between the genetic materials off ered.
Thus, the present study aimed to investigate the food consumption of T. arnobia on its native host guava and on diff erent E. urograndis clones (VE 41,I 144,TP 361 and VCC 865) in order to contribute to the development of Integrated Forest Pest Management programs (Forest IPM).

MATERIAL AND METHODS
The research was conducted at the Laboratory of Agricultural and Forestry Entomology (LEAF) of the Engineering and Agricultural Sciences Campus, Federal University of Alagoas (CECA/UFAL) and at Embrapa Coastal Tablelands.

T. arnobia fi eld collection
T. arnobia was obtained through manual collections in E. urograndis crops, variety 1407 and 224, at Albuquerque Farm, municipality of Atalaia, Alagoas state, 9 ° 30 '27' 'S and 36 ° 1 '24' 'W. The biological forms, such as eggs, caterpillars, pupae and/or adults were collected manually or with the aid of an entomological net and then transported to the Laboratory of Agricultural and Forestry Entomology (LEAF).

T. arnobia rearing
T. arnobia eggs were immersed in copper sulfate solution (CuSO4) for 10 seconds, then washed in distilled water in order to avoid contamination and placed in Petri dishes, with moistened fi lter paper, until the eggs hatch. After hatching, the caterpillars were placed in plastic buckets of 20L, with two side openings and a lid with holes covered by voile fabric. For feeding, leaves of E. urograndis or guava were off ered, in branches, placed in 500 ml glass bottles with water, until the loss of turgor, when they were replaced by new branches. Pupae were sexed and coupled in 10 x 20 cm PVC tube cages, internally lined with sulfi te paper, for adult emergence and oviposition. The adults were daily fed with a 10% honey solution. Eggs were daily collected.
Guava leaves were collected from trees belonging to the Campus and the diff erent E. urograndis clones (TP 361, VCC 865, I1 44 and VE 41) were collected in the experimental clonal forest stand of the Engineering and Agricultural Sciences Campus (CECA). Then, the samples were labelled and taken to the Agricultural and Forestry Entomology Laboratory (LEAF), washed in running water and cut into 4 cm diameter discs (≈ 1378mm 2 ).
To determine the consumed area, the host leaves were previously drawn on bond paper and at each evaluation period, the consumed area was marked in its respective outline, with diff erent colours for each period. Thereafter, T. arnobia food consumption was determined using the Bioscentifi c Lad ADC-AM-3000 leaf area meter.
2.3.1 Food consumption of T. arnobia on E. urograndis clones and guava in non-choice tests.
The leaf discs of E. urograndis clones (VE 41, I 144, TP 361, VCC 865) and guava were individualized in arenas of 6.0 x 5.0cm plastic pots, lined with fi lter paper and moistened with distilled water. Subsequently, one T. arnobia caterpillar was released in each arena and the leaf consumption was evaluated after 24 hours and 48 hours.
The experimental design was completely randomized, with 10 repetitions and 05 treatments, 4 clones of eucalypt (TP 361, VCC 865, I1 44 and VE 41) and 1 of guava. The data obtained were subjected to analysis of variance (ANOVA) and the means compared by the Tukey test (P≤ 0.05), using the statistical package of the SAS version 9.0 program (SAS Institute, 2011). Graphics were created using the SigmaPlot Software version 11.0 (Systat software, 2006).

Food consumption of T. arnobia in free choice tests using paired E. urograndis clones with guava.
For the evaluation of T. arnobia food consumption in free choice tests with E. urograndis clones (TP 361, VCC 865, I1 44 and VE 41) and guava, plastic pots (26 x 16 x 4 cm) with an orifi ce in the lid, covered with voile fabric, and lined with polyethylene foam, moistened with distilled water and covered with fi lter paper, were used as arenas. In each arena, the leaves of the diff erent hosts were placed in pairs (E. urograndis clone x guava), equidistant from the center and from each other. One T. arnobia caterpillar was released in the center of the arena and after 24 and 48 hours, the consumed area of each disc was evaluated.
The experimental design was completely randomized, with fi ve treatments (VE41 × Guava; I144 × Guava; TP361 × Guava; and VCC865 × Guava) and fi ve repetitions for each treatment. The data were submitted to the Chi-squared test of Independence (P≤ 0.05), using the statistical package of the SAS version 9.0 program (SAS Institute, 2011). The graphics were created using the SigmaPlot Software version 11.0 (Systat software, 2006).

Food consumption of T. arnobia in free choice tests among E. urograndis clones.
For the evaluation of T. arnobia food consumption in free choice tests among E. urograndis clones, plastic pots (26 x 16 x 4 cm) with an orifi ce in the lid, covered with voile fabric, and lined with polyethylene foam, moistened with distilled water and covered with fi lter paper, were used as arenas. In each arena, the four clones (TP 361, VCC 865, I1 44 and VE 41) were arranged concomitantly, in a circle, equidistant from the centre and from each other. One T. arnobia caterpillar was released in the centre of the arena and after 30 min, 2, 4, 6, 24 and 48 hours, the leaf consumption was assessed.
The experimental design was completely randomized, with four treatments (VE 41, I 144, TP 361 and VCC 865) and 10 repetitions. The data were submitted to analysis of variance (ANOVA) and the means compared by the Tukey test (P≤ 0.05), using the statistical package of the SAS version 9.0 program (SAS Institute, 2011). Graphics were created using the SigmaPlot Software version 11.0 (Systat software, 2006).

Food consumption of T. arnobia in free choice tests using paired E. urograndis clones.
For the evaluation of T. arnobia food consumption in free choice tests using paired E. urograndis clone plastic pots (26 x 16 x 4 cm) with an orifi ce in the lid, covered with voile fabric, and lined with polyethylene foam, moistened with distilled water and covered with fi lter paper, were used as arenas. In each arena, the leaves of the diff erent clones were placed in pairs, equidistant from the center and from each other. One T. arnobia caterpillar was released and after 24 and 48 hours, the leaf consumption was evaluated. The data were submitted to the Chi-squared test of Independence (P≤ 0.05), using the statistical package of the SAS program version 9.0 (SAS Institute, 2011). Graphics were created using the SigmaPlot Software version 11.0 (Systat software, 2006).

RESULTS
3.1 Food consumption of T. arnobia on E. urograndis clones and guava in non-choice tests.
Data from food consumption of T. arnobia on E. urograndis clones and guava in non-choice tests, revealed signifi cant diff erences between the evaluated treatments after 24h (F = 9.25; P <0.0001) and 48h (F = 10.33; P <0.0001), following the same pattern of leaf consumption on both periods.
It was found that all treatments were consumed, however, greater consumption of leaf area was observed for the native host guava, with an average leaf consumption area of 480.6 ± 89.52 mm 2 and 841.1 ± 132.7 mm 2 , followed by the VE 41 clone with 455.6 ± 94.2 mm 2 and 560.1 ± 108.5 mm 2 after 24 and 48 hours, respectively, signifi cantly diff ering from the other treatments (Figure 1).

Food consumption of T. arnobia in free choice tests using paired E. urograndis clones with guava.
The results obtained for the food consumption of T. arnobia on paired E. urograndis clones with guava (VE 41 x Guava; VCC 865 x Guava; I1 44 x Guava) showed similar consumption for the hosts, after 24h and 48h.

Food consumption of T. arnobia in free choice tests among E. urograndis clones.
Regarding the food consumption in free choice tests among the E. urograndis clones, it was observed that after 2h of evaluation, only the VE 41 (4.2 ± 4.2mm 2 ) and TP 361 (7.0 ± 7.0mm 2 ) clones were consumed, despite not diff ering from the other unconsumed clones (F = 1.0; P = 0 .40).
In general, I 144 clone presented a lower food preference from T. arnobia, being consumed only after 48 hours of evaluation. In spite of this, statistical diff erences were not identifi ed for this evaluation period between any of the tested clones, with average values of area consumed of 31.3 ± 21.8 mm 2 for I 144; 37.4 ± 18.02 mm 2 for TP 361; 90.7 ± 52.1mm 2 for VE 41; and 94.0 ± 28.0mm 2 for VCC 865 (F = 1.05; P = 0.38) (Figure 3). 3.4 The food consumption of T. arnobia in free choice tests using paired E. urograndis clones.
For the food consumption in free choice tests using paired E. urograndis clones, a signifi cant diff erence was observed in both evaluation periods (24 and 48h), for the pairing between I 144 × TP 361 clones, showing lower consumption of I 144 clone, with average values of 7.8 ± 10.7mm 2 for I 144 and 94.4 ± 89.4mm 2 for TP 361, after 24h (F = 69.52; P = 0.00012); as well as 98 ± 44.7mm 2 for I 144 and 161.8 ± 85.6mm 2 for TP 361 after 48h (F = 10.21; P = 0.004).
For the other pairings, no signifi cant diff erences were observed in any of the evaluated periods.

DISCUSSION
According to the results, it is possible to observe signifi cant diff erences in the food consumption of T. arnobia among native and exotic hosts as well as among the tested genetic materials (clones).
The VE 41 clone and the native host guava were more consumed than the other genetic materials tested in non-choice food consumption tests. This fact suggests that the least consumed clones (I 144, VCC 865 and TP 361) may possibly present chemical, physical or morphological stimuli able to reduce T. arnobia feeding (Lima et al., 2018).
It is important to note that, although the native host guava has been more consumed than the I 144, VCC 865 and TP 361 clones, when off ered in nonchoice tests, the native host was not always the most preferred, when T. arnobia had a chance to choose. These results suggest that T. arnobia may present a similar attraction to the stimuli emitted by both the native host and the tested genotypes of the exotic host, with the exception of TP 361 clone. This fact is quite notable, since the hybrid E. urograndis was developed in Brazil in the mid-70s (Faria et al., 2013), revealing that in a relatively short period of time, T. arnobia no longer distinguishes between its native and exotic hosts. However, in spite of not presenting remarkable preference between the native and exotic hosts, there is evidence in the literature that suggests diff erences in T. arnobia development (Marinho et al., 2008;Holtz et al., 2003b;Santos et al., 2000). Holtz et al. (2003c) reported that T. arnobia had a higher intrinsic population growth rate (rm) in E. cloesiana than in its native host guava. However, Santos et al., (2000) assessing the development of T. arnobia in E. urophylla and guava, demonstrated a better performance in the native host, with 5% of larval mortality in guava against 46.5% in E. urophylla. In general, even though there is no consensus among the authors, these facts point to an adaptation of T. arnobia to the exotic host.
According to West and Cunnengham (2002), outbreaks of T. arnobia populations in forest stands of Eucalyptus spp. are often considered larger than in the native host guava crops, indicating that this fact may be due to the production method and not only to the genetic features of the hosts, as the forest plantations of Eucalyptus spp.
These are crops commonly presented in extensive and contiguous clonal monoculture, favouring the incidence of pests, demonstrating that the preference for a host may depend on many factors, such as the quality and amount of available food resources (West and Cunnengham, 2002). Among the possible causes for the succession of T. arnobia population outbreaks in Eucalyptus spp., Marinho et al. (2008) consider that the exotic host has not yet developed defensive mechanisms, which would have already happened with the native host, in co-evolution processes, this is because eucalyptus plants produce more protease inhibitors than guava, but are more attacked by caterpillars of the genus Thyrinteina that possibly adapted for the protection of those plants by increasing the production of digestive enzymes. Futuyama (2008) approaches that adaptations to introduced exotic genetic materials have been gradually occurring in several species of insects, increasing the host range and favouring the change and maintenance of the population level between native and exotic hosts, thus ensuring the survival of these individuals in the fi eld.
The lower consumption observed for the clone 144 in the present study is substantiated on the fact that in the state of Alagoas, Brazil, the same E. urograndis clone has shown a higher productivity when compared to the others, being the most used clone for forest stands implantations in the region. This higher productivity may be related to lower attack by pest insects, including T. arnobia, which defoliates Eucalyptus spp. directly aff ecting growth rates and in case of continuous attacks may cause the death of its host (Pereira 2007;Moreira 2013).
Nevertheless, when the E. urograndis clones were off ered in pairs, the I 144 clone was only signifi cantly less preferred when compared to clone TP 361, indicating that when there is few options of choice, T. arnobia tends to make low distinction between the off ered materials.
When it comes to food preference, the insect responses may vary according to the stimuli coming from the host plant, being of chemical (allelochemical), physical (colour) or morphological nature (hairiness, texture, hardness, structure dimension, among others). For Eucalyptus spp., physical-chemical features, in addition to the presence of secondary compounds, such as tannins, phenols, fats and essential oils, can provide phagostimulating or deterrent properties, infl uencing the herbivory process, showing a direct infl uence on host preference (Ohmart et al., 1985;Lara, 1991;Ohmart and Edwards, 1991).
The results of although it was not possible to affi rm the existence of non-preference resistance among the diff erent clones of E. urograndis (VE 41, I1 44, TP 361, VCC 865) and the native host guava for T. arnobia, the results of the present study suggest a higher consumption for the VE 41 clone and guava, as well as a low preference for the I1 44 clone.
In general, the investigation of the interactions of native and exotic hosts and T. arnobia is of fundamental importance for a better understanding of ecological relationships, assisting in the development, planning and use of appropriate methods for Forest Integrated Pest Management programs (Forest IPM).

CONCLUSIONS
There are diff erences in the food consumption of T. arnobia between its native and exotic hosts, in which the VE 41 clone and the native host guava were more consumed, when compared to the other tested genetic material. The TP 361 clone was less preferred when off ered paired to the native guava host and the I 144 clone presented less food preference, when the genetic materials were off ered together. The results obtained provide basic information for the indication of eucalypt clones and understanding of the interaction and ecological relationships, assisting in the development of Forest Integrated Pest Management (Forest IPM) programs for the control of T. arnobia.

AUTHOR CONTRIBUTIONS
Almeida, C. A. C. and Breda, M. O. was responsible for the conception and design of the work, data collection, data analysis and interpretation, drafting the article, revision and fi nal approval of the version to be published. Santos, J. M., Gonçalves, F. S. and Rodrigues, M. B. was responsible for the data collection.