Field evaluation of the susceptibility of mill and table olive varieties to egg‐laying of olive fly

The susceptibility of 20 widely distributed mill and table olive varieties to Bactrocera oleae (Rossi) as affected by irrigation, and fruit diameter and oil content was evaluated in a 3‐year trial in Southern Spain. Bactrocera oleae was bivoltine life cycle in the experimental site, with significant differences among population size throughout the study. Even though the olive fruit fly damaged all varieties, significant differences in susceptibility were detected. Among the mill olive varieties “Nevadillo Blanco de Jaén” was the most susceptible, with average infestation levels ranging between 6.7% and 52.2% and between 10.3% and 69.2% under rainfed and irrigated conditions, respectively, and “Arbequina” was the least susceptible, with average infestation levels ranging between 0.6% and 12.7% and between 2.3% and 18.5% under rainfed and irrigated conditions, respectively. Among the table olive varieties, “Gordal Sevillana,” “Ascolana Tenera” and “Ocal” were the most susceptible (with average infestation levels reaching 39.7%, 36.5% and 33.3%, respectively), while “Callosina” was the least susceptible (with infestation levels of only 8.4%). Irrigation tended to promote both B. oleae infestation and its earlier occurrence compared to the rainfed condition. Even though the diameter and oil content were positively correlated with B. oleae fruit infestation (correlation coefficients ranged between 0.5 and 0.95), the present work reveals that other yet‐unknown factors may influence B. oleae oviposition preferences. The results of this study can be useful for breeding programmes to develop olive varieties resistant to B. oleae and provide key information for wide‐area olive fly pest management decisions.

The species is native to the Mediterranean basin and has been recently introduced into California and Mexico, leading to many problems in quarantine protection and international trade (Nardi, Carapelli, Dallai, Roderick, & Frati, 2005;Neokosmidi et al., 2005;Tzanakakis, 2006). This tephritid lays its eggs in olive fruits. After hatching, the larvae feed and grow in the mesocarp, causing it to drop prematurely before maturation, making it unsuitable for eating (table olives) and reducing the quality of the oil produced (Kakani et al., 2010;Santiago-Álvarez & Quesada-Moraga, 2007). The olive fly has become unmanageable worldwide because it has developed resistance to chemical insecticides, which have been overused in bait sprays for B. oleae adult control since the early 1960s (Kakani et al., 2008(Kakani et al., , 2010Stasinakis, Katsares, & Mavragani-Tsipidou, 2001). Moreover, severe effects of pesticides on nontarget arthropod fauna in olive orchards have been found (Pascual, Cobos, Seris, & González-Núñez, 2010;Ruano et al., 2001). Other strategies against B. oleae like cultural and sanitary methods, mass trapping and preventative spraying with kaolin (physical barrier) and copper have a very limited application (Daane & Johnson, 2010;Haniotakis, 2005), although they can be combined with other components of an Integrated Pest Management (IPM) programme (Vargas, Piñero, & Leblanc, 2015). Consequently, there is an urgent need to develop alternatives to synthetic chemicals for olive fly IPM, with emphasis on both biological control and the development of new olive varieties which are resistant to B. oleae (Fernández Escobar et al., 2013; Quesada-Moraga, Campos-Aranda, & Santiago-Álvarez, 2010). The genetic resistance of olive germplasm is a key tool for controlling B. oleae because olive varieties with olive fruit fly resistance could be used to obtain quality products in both organic and conventional systems and they could also contribute to the reduction in insecticide applications (Rizzo, Virgilio, & Alberto, 2012).
There were two plots (1 ha each) in the experimental site. One Plot, under drip irrigation, contained 10 mill and 10 table olive varieties, with a total of 160-200 trees. Between eight and ten, randomly distributed 11-year-old olive trees per variety were planted with 7 × 7 m tree spacing. The second plot, under rainfed conditions, contained only the 10 mill varieties and also consisted of 8-10 randomly distributed 11-year-old olive trees per variety (total of 80-100 trees) and 7 × 7 m tree spacing. Thus, the mill varieties were evaluated under both irrigated and rainfed conditions, whereas the table varieties were evaluated only under irrigated conditions. This random distribution of many varieties in the same olive orchard, the experimental unit being the tree and not the plot, helps improving the evaluation of susceptibility of olive varieties to B. oleae attack by offering different olive varieties to females under agronomic and climatic uniformity (Supporting Information Figure S1). Drip irrigation was conducted once a week with the same volume from May 1 to September 30 reaching 2,000 m 3 /ha/year. No phytosanitary treatments were performed during the field trials.  (2008, 2009, 2010 and 2011) were: 601.7, 724.8, 1,167 and 504.8 mm, respectively.

| Susceptibility of table and mill olive varieties to Bactrocera oleae under rainfed and irrigated conditions
All trees of each variety (8-10) were sampled eight times at 10-day intervals from September to December to estimate the percentage of B. oleae punctures. For each tree and sampling date, 10 olives were randomly harvested at a height of 1.5 m, circling around the tree to include fruits at all orientations (MAGRAMA, 2014); thus, approximately 80-100 olives of each variety were sampled throughout the sampling period (approximately 24,000 fruits per year of all varieties). The sampled olives were transported in plastic bags to the laboratory and the possible presence of oviposition punctures was determined using a stereoscopic microscope (Nikon Instruments INC., Melville, USA).

| Relationship between olive fruit traits and oviposition preferences of Bactrocera oleae
To evaluate the influence of fruit traits on olive fly oviposition preferences, the relationships among oil content, fruit diameter and B. oleae infestation in each variety were studied. After examining the fruits for possible oviposition punctures as described above, the transversal fruit radius was measured with a caliper rule, resulting in 10 diameter values for each tree for each sampling date. To determine the oil content, two fruit samples of 70 g each were taken from each variety and each sampling date (16 samples per each variety). These samples were placed in Petri dishes and dried in an oven at 105°C for 42 hr. Then, the dry weight of each fruit was determined and the percentage of oil was measured by nuclear magnetic resonance following the methods described in Cimato and Attilio (2008).

| Statistical analyses
To compare the susceptibility of the different mill and table olive varieties to the olive fruit fly, the areas under B. oleae infestation curves (AUBIC) were calculated using the trapezoidal integration method of SAS (Campbell & Madden, 1990). The area under the pest progress curve is a very useful quantitative summary of pest's intensity over time, and it is used normally for comparison across years, locations, or management tactics, and for describing quantitative resistance of pests as confirmed by American Phytopathological Society (Jeger & Viljanen-Rollinson, 2001;Litsinger, 1991;Lyman, Bryan, Jeffrey, & Jesse, 1989;Ruesink & Kogan, 1994;Worner, 2014). All data collected over time (eight sampling dates) for each tree were used to calculate the AUBIC. The 8-10 values of AUBIC per variety were subjected to Kruskal-Wallis one-way nonparametric analysis of variance using Statistix 9.0 software (Analytical Software 2008).
Spearman's rank correlations were performed to detect the relationship between the infestation level and oil content and fruit diameter.

| Olive fly dynamics
The flight curves indicated that B. oleae was bivoltine in the sam-

| Susceptibility of table and mill olive varieties to Bactrocera oleae under rainfed and irrigated conditions
There were significant differences in B. oleae incidence among mill  Table S1).  Information Table S2).

| Relationship between olive fruit traits and oviposition preference of Bactrocera oleae
The correlation coefficients (Spearman) between B. oleae infestations of olive fruit and fruit diameter and oil content for all varieties are shown in Table 1.
The correlation coefficients between B. oleae infestation of olive fruits and fruit oil content varied greatly from year to year for the same variety and under the same irrigation conditions and were mostly significant (Table 1). In 2008-2009, all the correlation coefficients were significant except those of "Arbequina," "Cornicabra," "Hojiblanca," "Lechín de Sevilla" and "Picual" in irrigated conditions.
In 2010-2011, the correlation coefficient for the Arbequina and Frantoio varieties in the rainfed condition was not significant nor was the correlation coefficient for the Cornicabra variety in the irrigated condition. However, in 2011-2012, all the correlation coefficients between the infestation rate and the oil content were significant except for the Cornicabra variety in the rainfed condition (Table 1).
For mill olive varieties, the correlation coefficients between olive fly infestation and fruit diameter also varied greatly from 1 year to another, even for the same variety and under the same irrigation condition, and generally tended to be significant (Table 1). In 2008-2009, all the correlation coefficients were significant except for "Cornicabra," "Empeltre," "Lechín de Sevilla" and "Nevadillo Blanco de Jaén" in irrigated conditions (  were also significant except for those of "Arbequina," "Nevadillo Blanco de Jaén," "Pico Limón" and "Picual" in the rainfed condition and the Frantoio variety in the irrigated condition. In 2011-2012, the correlation coefficients were not significant for "Cornicabra," irrespective of irrigation regime nor for "Frantoio," "Lechín de Sevilla" and "Picual" in the rainfed condition and the Hojiblanca and Pico Limón varieties in the irrigated condition (Table 1).
A similar trend was discernible for the table olive varieties, which showed significant correlation coefficients between infestation level and fruit oil content in 2008-2009 for all varieties except "Uovo de Piccione" (Table 2). In addition, in 2010-2011, all the correlation coefficients between infestation level and fruit oil content were significant except for "Uovo di Piccione" and "Mollar de Cieza". Concerning correlation coefficients between infestation TA B L E 1 Correlation coefficients (Spearman) between Bactrocera oleae infestation level, fruit diameter, and oil content of the mill olive varieties under both rainfed and irrigated conditions for the three study years were obtained for "Ascolana Tenera," "Mollar de Cieza" and "Uovo di Piccione" (Table 2). In 2011-12, all the infestation-oil content correlation coefficients were significant except those for the Callosina and Barnea varieties, but the correlation coefficients between level of infestation by B. oleae and fruit diameter were not significant for any variety except for "Kalamon," "Ocal" and "Manzanilla de Sevilla" (Table 2).

| D ISCUSS I ON
The olive fly, B. oleae, is a multivoltine species that completes be-  (Tzanakakis, 2006). Indeed, at temperatures below 30°C, this tephritid reaches sexual maturation and reproduces normally, whereas temperatures above 30°C and low relative humidity lead to egg resorption and reproductive diapause (Santiago-Álvarez & Quesada-Moraga, 2007). Hence, the increased presence of B. oleae adults in the irrigated olive plot during this adverse period could be attributable to these ecological traits (low temperature and high relative humidity; Gutierrez & Cossu, 2009;Pappas, Broufas, Koufali, Pieri, & Koveos, 2010).
None of the 20 selected widely distributed mill and table olive varieties was resistant to B. oleae, confirming previous observations, which indicated than even wild olive trees are susceptible to this pest (Alvarado, Civantos, & Durán, 2008). It seems that the search and acceptance of olive fruit by B. oleae are determined by visual and chemical signals which remain poorly known (Burrack & Zalom, 2008).
Our results indicated that some fruit variables such as diameter and oil yield can partially explain the susceptibility of mill and table olive varieties to B. oleae; the percentages of fruit infestation increased as the olive fruit increased in diameter and oil yield.
Hence, among the mill varieties, the smallest variety, Arbequina, was the least susceptible one during all years of the study in both irrigated and rainfed conditions. Likewise, "Arbequina" was also the least susceptible to B. oleae among several commercially important olive varieties in California during three consecutive years (Burrack & Zalom, 2008). Nonetheless, other unknown variables may also influence fruit selection by adult fly females as revealed by the fact that several small-diameter varieties (i.e., "Nevadillo blanco de Jaén" and "Lechín de Sevilla") were more susceptible to B. oleae than were higher diameter ones (i.e., "Picudo" and "Pico Limón").
The relationship between olive fruit size and B. oleae oviposition preferences is a controversial issue. Research exists that indicates that these measures both correlate (Gümusay et al., 1990;Neuenschwander et al., 2009;Rizzo et al., 2012) and do not correlate (Iannotta, Perri, Tocci, & Zaffina, 1999). It has even been proposed that olive fruit size is the most important parameter during the early stages of fruit ripening but, subsequently, it becomes less important due to both the increase in the population density and the presence of particular substances in the olives that induce females to select the smaller drupes for ovipositing (Dominici, Pucci, & Montanan, 1986). In further study, Rizzo et al. (2012) (Kombargi et al., 1998).
More recently, it has been demonstrated that olive volatiles such as toluene and α-copaene could be B. oleae oviposition promoters and fruit mineral element content (Bononi & Tateo, 2017;De Alfonso, Vacas, & Primo, 2014;Garantonakis et al., 2016;Malheiro et al., 2015). In the present work, the B. oleae infestation level of each variety tended to be both earlier and slightly higher under the irrigated condition than under the rainfed condition. This result supports the previous report by Bjelis, Masten, and Imala (2008) and may be due to the earlier ripening and greater firmness of irrigated fruits (Rizzo et al., 2012).
The present study, which includes the 20 most widely distributed mill and table olive Spanish varieties, is an extensive survey of olive variety responses to B. oleae and it also provides key information for wide-area olive fly pest management decisions, because the most susceptible varieties can contribute to B. oleae population increases in olive variety mosaic landscape patterns, thereby improving the timing, sequence and intensity of monitoring efforts and spraying of insecticides determined by GPS-based forecasting models. Our results also reveal that fruit size and oil content at least partially explain the oviposition preferences of B. oleae, although yet-unknown factors may also operate. Likewise, irrigation tended to promote both B. oleae incidence and infestation and should also be considered in pest management decisions. Based on these results, the susceptibility of olive tree varieties to B. oleae must be considered in IPM programs because varieties play a very important role in making the populations of this key pest manageable.