Phenological asynchrony occurs when the periods of crop vulnerability and pest presence do not overlap (Messina & Jones, 1990). This condition can be induced by varying sowing time (in annual crops) or by choosing cultivars with appropriate growing cycles (in woody crops) if the pest dynamics is not too strongly dependent on that of the crop, i.e. the crop is not a limiting resource for the pest (Asch & Visser, 2007). Practices inducing or favoring phenological asynchrony are important to support integrated pest management and organic farming, since they guarantee fruit production and reduce the reliance on chemical inputs and, therefore, decrease their related environmental costs (Hammons et al., 2010). An assessment of the periods of crop vulnerability and pest presence is necessary to take advantage of phenological asynchrony for a given pest-crop association.

Anastrepha fraterculus (Wiedemann) (Diptera, Te­ph­ri­tidae) is a fruit pest of great economic importance because it attacks a wide variety of fruit plants and is widely distributed, from northern Argentina to sou­thern USA (Hernández-Ortiz & Aluja 1993). This species has a multivoltine reproduction pattern and do not have a diapause stage (Aluja & Mangan, 2008). Consequently the South America fruit fly ( A. fraterculus) needs to infest host plants continuously to survive because this species has no mechanisms to bridge any fruit-free period (Aluja et al., 2014). The females deposit their eggs in ripening fruits (Rosa et al., 2017) that consequently will rot and be unmarketable (Härter et al., 2015).

The Brazilian southwestern region has large exten­sions of natural pasture fields, with presence or absence of shrubs. In addition, there are isolated patches of native vegetation (Semi Decidual Forest) intermixed with commercial orchards. Among the most common and widespread cultivated A. fraterculus hosts in the temperate region are peach ( Prunus persica L.), apple ( Malus domestica L.), orange ( Citrus sp.) and grape ( Vitis sp.), whereas native plant hosts mainly belong to the Myrtaceae Family (Nava & Botton, 2010; Machota et al., 2016). The fruiting periods of all these tree species provide conditions for A. fraterculus to find plant hosts during all months of the year (Nava & Botton, 2010). However, abiotic factors, specifically temperature and rainfall, play an important role in tephritid outbreaks (Aluja et al., 2012). Extreme thermal limits can accelerate or delay the development cycle of tephritid fruit flies (Taufer et al., 2000).

The South American fly represents a threat for peach cultivars in southern Brazil (Botton et al., 2002). There, the main peach-producing areas are located in the state of Rio Grande do Sul where the municipalities surrounding the city of Pelotas produce, at least, 90% of the peach national production for the canning industry (Timm et al., 2007). In this region, peach orchards are commonly managed using a conventional protection strategy and trees are subjected to 4-5 broad-spectrum insecticide treatments against fruit flies per growing season. In addition, protection may be complemented by toxic baits containing hydrolyzed protein (3%) and an organophosphate insecticide (Malathion 1000 CE, 200 mL in 100 L) set on the orchard edges (Härter et al., 2015).

Pest control is one of the biggest challenges for organic farming, which is increasing in Brazil but still represented less than 2% of the overall production in 2006 (IBGE, 2006). Peaches are susceptible to damage by A. fraterculus during their ripening period, starting ca. 30 days before harvest (Salles, 1994). In Rio Grande do Sul, grown cultivars have harvest times extending from mid-October to mid-December. Then, periods of fruit susceptibility vary depending on the cultivar and favoring phenological asynchrony with A. fraterculus may be one option to reduce the application of chemical pesticides.

The present work aimed to assess fruit infestation by A. fraterculus at harvest in early, average and late maturing peach cultivars for one growing season (2012-2013). Another objective was to verify fruit infestation for non-cultivated native and non-cultivated wild exotic plant hosts around the peach orchards during 2013 and 2014. In addition, we determined the weekly abundance of A. fraterculus during three consecutive growing seasons, S1 (2011-2012), S2 (2012-2013) and S3 (2013-2014), in order to assess phenological asynchrony between peach cultivars and A. fraterculus. In particular, this study aimed to check the influence of meteorological variables on A. fraterculus abundance. The hypothesis to be tested was if A. fraterculus abundance at the time when fruits are susceptible could be different among cultivars. Eventually, we discussed the possibility of sustainable management of peach in southern Brazil.

The current work provides a first approach for determining if phenological asynchrony could be used as a means to control pest outbursts affecting peach cultivars in South Brazil. Under these conditions, the main findings of our research were that the highest level of infestation occurred in average-maturing cultivars, while early-maturing cultivars were not infected. In addition, we detected that monthly abundance of fruit flies increased with daily temperature and decreased with monthly rainfall. Moreover, non-cultivated fruit tree species near the orchards acted as hosts for the flies. Furthermore, in the present study, A. fraterculus was the main fruit fly species infesting peaches in Southern Brazil in accordance with Bisognin et al. (2015) and Rosa et al. (2017). All individuals ( i.e. females) that could be identified to the species level belonged to the A. fraterculus species, both in traps and among individuals that emerged from fruits. Consistently, males and females indistinctly infested fruits, as we observed sex-ratios close to 0.5.

Infested fruits were recovered only from average or late maturing cultivars, harvested in mid-December. During winter (June to September), few plant hosts are available in this region. The main and most abundant plant host around the experimental area is E. japonica, which has the peak of fruit production in August. Our data showed a higher infestation index of 2.23 and 0.21 fruit per pupae in August of 2014. These data highlight an important point to consider: weather conditions (temperature and relative humidity), because these are essential factors for the life cycle of insects (Hedström, 1992). The average temperature in July of 2013 and 2014 in the experimental area was 11.3 and 12.6 °C, respectively. Development and adult movements of A. fraterculus become faster as temperatures increase from 22 °C to 27 °C, remaining slow for temperatures below 18 °C (Salles, 1995). To complete the biological cycle (egg to adult) of the South America fruit fly takes approximately 88 days with temperatures around 15°C (Salles, 1995; Nava & Botton, 2010). If the temperature is lower, more days will be necessary to complete the biological cycle (Hedström, 1992; Nava & Botton, 2010). Therefore, if we consider that the oviposition occurred around 15 ^th August, the adults will emerge after 15 ^th November in the experimental area. After emergence, adults of fruit flies need to feed and they need some days to become sexually mature. This period (since egg stage until the female oviposit the first viable egg) is called period of pre-oviposition. The length of this period varies with temperature, being around 7 days at 25 - 30° C and 19 days at 20 °C. The average temperature in November of 2013 in the experimental area was 19.88 °C; this means that adults would become ready to infest fruits around the 2 ^nd week of December, as observed in the peach sampling carried out in the experimental area.

The trapping results reflect the observations of peach infestation recorded in the field. In the current study, average temperatures did not reach the favorable range for A. fraterculus before November (Fig. 2) and, consequently, fly populations peaked in December whatever the cultivar and year. The high proportion of females obtained in the trap samples may result from the trapping methodology since protein baited traps attract more females than males due to their need to feed on protein to mature sexually and to be able to mate (Houston, 1981; Aluja et al., 2012). The trapping results strongly suggest that the lower infestation in early-maturing cultivars resulted from a phenological asynchrony with fruit fly populations, since less flies were trapped during the fruit susceptibility period of early cultivars. A more favorable range of temperatures and the overlap of fruit period with different plant hosts, like P. cattleianum and E. uniflora, may have determined the higher abundance of flies late in the season. These native non-cultivated fruit trees are considered multiplying hosts of A. fraterculus and, thus, are sources of infestation in peach orchards (Bisognin et al., 2015).

In the current study, while fruit fly populations peaked in December or January during the three stu­died growing seasons, their actual abundance varied from year to year, as previously observed in Southern Brazil by Rosa et al. (2017), and for other species of the Anastrepha genus in Latin America (Aluja et al., 2012). In addition to direct effects of variations in climatic conditions, the abundance of fruit flies in orchards may be affected by the presence of fruiting wild host plants that may maintain or increase both fly populations and those of their natural enemies (Aluja et al., 2014; Schliserman et al., 2016). An interesting fact is that the peak of fruit flies captured (December-January) was particularly high in orchards with early-maturing cultivars. This should be interpreted cautiously because, although it demonstrates that A. fraterculus population exploited the site to obtain food and shelter, it is also possible that the traps were more attractive during that period due to the low abundance of fruits after harvest (Hendrichs & Hendrichs, 1990).

Although this study was carried out at a local scale, the results observed are relevant at a global level since early-ripening varieties of peach can suffer less attacks of tephritid flies worldwide. The fact that taking advantage of phenological asynchrony with fruit fly seasonality could be investigated in other regions and for fruit trees that have fruiting periods during the coldest months of the year, when it is too cold for tephritid fruit fly attacks. Similar results to those reported in the current study were found in peach orchards in Tunisia (Hafsi et al., 2016), with a very different climate and soil conditions than those of the current study. These authors demonstrated that early cultivars of peach suffered less attacks of Ceratitis capitate (Diptera: Tephritidae) than mid-maturing varieties (Hafsi et al., 2016). The trapping results of this research showed that, at the beginning of the season, the number of tephritid flies were low and it increased gradually as temperatures increased to about 29 °C and as host fruits became available in nearby fields (Hafsi et al., 2016). Therefore, the pattern of tephritid flies in this study was similar to that of the present survey. To demonstrate the moments when the lower densities of fruit flies occur in Southern Brazil is a great contribution to the design of integrated pest management (IPM) strategies, since some strategies for pest control, e.g. attract and kill bait station, are more efficient at lower densities than when the density of tephritid flies is heavy (Hafsi et al., 2016). Moreover, this would be particularly important for organic peach production for which fruit flies are an important limitation in Southern Brazil (Härter et al., 2015). The increase of organic crop production worldwide (Ozinci et al., 2017) has raised questions about the design of specific organic crop ideotypes in breeding programs (Löschenberger et al., 2008; Crespo-Herrera & Ortiz, 2015).

Resistance to herbivores and diseases is one of the main targets for organic crops (Bruggena & Termorshuizen, 2003). Selecting for different pheno­logies in crops is feasible and generally aims at adapting crops to new growing regions (frost escape) (Papanikolaou et al., 2005) or increasing the range of harvest dates for commercialization purposes (Raseira & Bonifacio, 2006). Few examples exist that specifically aim at phenological asynchrony with herbivores (Hammons et al., 2010). In Southern Brazil, most peaches are harvested for the canning industry and late-maturing cultivars are nowadays the most suited ones for this aim because of their better texture and lesser browning susceptibility (Techakanon et al., 2016). Our results indicate that a more sustainable peach farming needs that breeding programs aim at producing early-maturing peach cultivars that fulfill the aforementioned characters. As for any pest resistance that is bred into crops, evolutionary processes are expected to select pests that overcome this resistance, hereby increasing their phenological synchrony with hosts (Combes, 2001; Asch & Visser, 2007). However, since fruit flies are polyphagous and depend on a number of cultivated and wild hosts, selection for asynchrony with peach trees may not be very strong.

In summary, our results suggest that promoting phenological asynchrony between peach cultivars and A. fraterculus could help to develop a more sustainable peach production and, specifically, limit insecticide use during the growing season. This study constitutes a first assessment of the periods of crop vulnerability and pest presence in peach orchards in South of Brazil and provides necessary information for taking advantage of the phenological asynchrony phenomenon for this pest-crop association. Moreover, we detected that non-cultivated fruit tree species near the orchards acted as hosts for the flies. Therefore, a deeper research on the combined effects of peach cultivar characteristics and those from other hosts on fruit fly populations will be necessary to improve their control in fruit-tree orchards.

The authors would like to thank Prof. Lino B. Monteiro of Federal University of Paraná for enga­ging the collaboration between the authors and the Coordination for the Improvement of Higher Person­nel (CAPES) for the scholarships. We also thank Dr. José Manuel Mirás Avalos from the University of Santiago de Compostela for his useful comments on the manuscript. We thank the useful comments from three anonymous reviewers. Also, we thank the help of Prof. Msc Maria Victoria Calvo, (UDELAR), Msc Diana Valle (INIA Las Brujas); Msc. Alexandra Peter Krüguer; agronomist Ariomar A. Bartsch, and the students Marcos H. F. Telles and Victor Cieza Tarrillo for the contribution on the field and labwork. We thank in special to the agronomy student Alexandre M. Neumann for allowing us to sample peaches and other non-cultivated fruits in his farm, and by his contribution in the labwork.