Evaluation of Mass Trapping Devices for Early Seasonal Management of Ceratitis Capitata (Diptera: Tephritidae) Populations

Mass trapping is an environmentally safe alternative to insecticide application for the Mediterranean fruit fly management. The selection of effective trap-attractant combinations for monitoring and mass trapping control remains challenging. The current study explored the attractiveness of trapping devices during spring (early season) and summer (late season) in field cage trials. Five trapping devices were assessed: (a) the commercially available Decis® trap, (b) Tephri trap baited with Biodelear, (c) Tephri trap baited with BioLure, (d) International Pheromone McPhail trap (IPMT) baited with Biodelear, and (e) IPMT baited with BioLure. On a test day, 100 adults (50 males and 50 females) were released in each field cage wherein traps were placed individually. Trap captures were recorded at hourly intervals from 10:30 am to 5:30 pm. Our results showed that Tephri traps baited either with BioLure or Biodelear captured the most adults under low temperatures. Efficacy of Tephri traps baited with BioLure were higher than that of other trap-attractant combinations at high temperatures. Adult captures in Decis® trap were low during both seasons. More males than females were captured at low temperatures. Both efficacy and female selectivity of trapping devices are related to prevailing temperature regimes during spring and summer under semi-field conditions.


Introduction
The Mediterranean fruit fly (medfly), Ceratitis capitata (Wiedemann) (Diptera: Tephritidae) is a widely distributed pest that attacks numerous fruit and vegetables crops causing high yield losses [1,2]. Broad-spectrum insecticides (mainly organophosphates) applied either as bait or cover sprays have been extensively used to manage medfly populations for decades. However, most of them have been banned in the European Union (EU) due to environmental and public health concerns [3]. In recent years, novel control techniques, such as attract-and-kill approaches, have been developed and tested against fruit flies to reduce the dependency on insecticides [4][5][6][7]. Field trials revealed that they can either be as effective as conventional control method or should be used in combination with them, particularly when population density is high [3,8,9]. When insects are not retained in or on a device, the attract-and-kill method is called "lure-and-kill", while when a trap type that retains the target species is used, it is referred to as mass trapping [3]. Hence, the need for wide application and dispersion of synthetic insecticides is dramatically reduced following the attract-and-kill approaches, especially mass trapping. Bait stations that include the combination of an attractant and a toxic surface that kill, sterilize or infect the target insects with microorganisms are the most common lure-and-kill tools considered at this point, the medfly has already built up high population densities, making their control more difficult and less cost effective with mass trapping. The establishment of the mass trapping grid early in the season (spring) targeting low medfly population densities may provide an alternative, promising approach. The approach is called "off-season" management of insect pests and targets the population growth over a longer period of time and well before the ripening season arrives. Under this scheme, it is anticipated that less effort in terms of labor and trap density will be necessary for population suppression. However, limited information exists regarding the performance of trapping devices for "off-season" control under relatively low population densities that may be coupled with limited adult flight activity and dispersion [2,29]. The "off-season" application often in temperate areas coincides with non-optimal conditions for medfly population growth (i.e., low temperatures). However, there is limited information about the performance of different mass trapping devices under various temperature conditions, which may affect both the activity and behavior of adults as well as the performance of attractants.
Considering the importance of the mass trapping technique for medfly control and the existing knowledge gaps for its "off-season" application, the main objective of the current study was to examine the effectiveness of common, commercially available trapping devices baited with two types of female attractants at two seasonal temperature regimes. In particular, the performance of the different trapping devices was tested in field cages early in the spring under sub-optimal temperatures for medfly activity and during summer under optimal temperature conditions. The importance of seasonal performance of different trapping devices is discussed.

Insects and Field Cages
All flies used were "wildish" (F4, F5 laboratory generation from field infested fruits). To establish the laboratory colony, field infested citrus fruits were collected from the area of Volos (Magnesia, Greece) and transferred to the Laboratory of Entomology and Agricultural Zoology at the University of Thessaly where larvae completed development and pupated under constant laboratory conditions (temperature: 25 ± 2 • C and RH: 55 ± 5% 14L:10D). Pupae were collected twice a week and placed in wooden framed, wire-screened cages (30 × 30 × 30 cm 3 ) to allow adults to emerge. Upon emergence, adults had ad libitum access to water and a standard adult diet consisting of sugar, yeast hydrolysate and water at a ratio of 4:1:5. Females had access to oviposit into artificial oviposition substrates, consisted of a red, plastic, hollow hemisphere 5 cm in diameter with 50 evenly distributed holes (diameter 0.7 mm) through which females laid their eggs. The base of each hemisphere was fitted into a 5 cm in diameter hole performed on the cover of a plastic, 5.5 cm in diameter Petri dish. A plastic cup with 0.5 mL of orange juice was placed in the base of the Petri dish to stimulate oviposition. Eggs were collected from the inner surface of the dome, and larvae were reared on a standard artificial diet [30].
Trapping trials were performed using five cylindrical (2.9 m diameter × 2 m height) plastic-screen field cages housing one potted citrus or olive tree each (~1.80 m height and 1.3 m canopy diameter). Fruits, if any, were removed from the trees. Field cages were protected from the direct sunlight by placing a shading net one meter above their top.

Trapping Devices
The following five trapping devices were assessed: (a) Decis ® trap, (b) Tephri trap baited with Biodelear, (c) Tephri trap baited with BioLure, (d) the International Pheromone McPhail trap (IPMT) baited with Biodelear, and (e) IPMT baited with BioLure. The commercially available, ready-to-use Decis ® trap (Bayer CropScience, International, Monheim am Rhein, Germany) is a dry trap consisted of a top part, which is colorless, transparent and internally impregnated with a pyrethroid insecticide (0.015 g deltamethrin/dispenser) as a killing agent. The lower part is rather orange with four lateral holes (2 cm each) that serve as entrance to the trap. A mixture of 7.8 g ammonium acetate, 0.5 g chlorohydrate trimethylamine and 0.03 g 1.5 diamineopentane is used as attractant in Decis ® trap. The Tephri trap (Agro Alcoy, Alcoy, Spain) is consisted of an 11 cm deep, yellow, invaginated base covered with a 3.5 cm in high opaque lid. The diameter at the junction of lid and base is 12 cm. In addition to the bottom opening (3 cm diameter), the Tephri trap has four lateral openings of 2.1 cm in diameter each, being 4 cm from trap top. The plastic International Pheromone's McPhail trap (IPMT) has a 7 cm tall yellow base with a 4.5 cm diameter, funnel-like opening and a 12 cm tall clear top. The trap is 17 cm in diameter at its widest point (International Pheromone Systems Limited, South Wirral, England).
The female-specific attractants, BioLure and Biodelear, were used in both IPMT and Tephri traps. The BioLure (Unipack by Suterra) is a three-component food lure consisting of a mixture of 29.8% ammonium acetate, 12.4% trimethylamine hydrochloride and 0.2% 1,4-Diaminobutane. Biodelear as mentioned previously, is an innovative, environmentally friendly female attractant in liquid form with high viscosity. BioLure was properly placed in the IPMT trap by attaching the dispenser with paper tape on the inner walls of the transparent part of the trap. In the Tephri trap, BioLure was placed at the top of the device. Custom-made dispensers, consisting of pieces of Wettex sponge wrapped and placed in rolls tightly closed in a piece of sieve, were used for applying Biodelear attractant into the traps. The Wettex rolls were soaked in liquid Biodelear and allowed to drain under ambient conditions before being used. Each dispenser was impregnated with 17 g liquid Biodelear. In the Tephri trap, the Biodelear dispenser was placed in the existing basket. In the IPMT trap, the dispenser was hung from the trap ceiling with a short piece of wire. In both IPMT and Tephri traps, 250 mL of water with 1% detergent as surfactant agent was used to capture and retain the attracted flies.

Experimental Procedure
Field cage trials were conducted from 29 March to 14 April 2020 (spring season) and from 6 to 26 July 2020 (summer season) in the outdoor facilities of the Laboratory of Entomology and Agricultural Zoology in Volos area (Magnesia, Greece).
Two days before adult emergence, pupae were transferred from the standard laboratory conditions to ambient conditions of a non-heated warehouse until adult emergence. Upon emergence, adults were placed in wooden cages with ad libitum access to adult food (sugar: yeast in ratio 4:1) and water. Two days before conducting an experimental trial, adults were sexed and placed in groups of 50 individuals (males or females) into Plexiglas, wire-screened cages (20 × 20 × 20 cm 3 ) with ad libitum assess to water and granulated sugar.
At test day (9:00 a,m.), 50 females and 50 males (10-15 days old) were released into each field cage. Thirty minutes later, the respective mass trapping devices were placed in the field cages. A single trapping device was placed in each field cage at 1.7 m height, in the southwestern side of the canopy of each potted tree. Flies had ad libitum access to non -proteinaceous food source (drops of sugar syrup) and water (water-soaked cotton pads fastened to tree branches). Trap captures (number of males and females entering the trap) were recorded and removed at hourly intervals from 10:30 a.m. to 5:30 p.m. (8 records/replication). To do so, each mass trapping device was placed into a plastic bag that sealed to avoid escape of captured flies and replaced with a new one of the same type. The experiment was repeated 10 times (trial days). Trapping devices were rotated within field cages in a clockwise manner per replication. The temperature was hourly recorded using an analog thermometer, placed near to the field cage arena in each replication. Data for daily relative humidity were obtained from the meteorological station of Volos city located 3 km away from the experimental area (https://stratus.meteo.noa.gr/, date: 10 August 2020).
In case of dead flies found within the field cage arena, they were replaced with new ones of the same cohort before the next record.

Statistical Analysis
The Kolmogorov-Smirnov test was used to assess the normality of data distributions. Generalized Linear Models (GLMs) with Poisson loglinear distribution was used to determine the effects of treatment (mass trapping device), sex, temperature and their interactions (excluding that of treatment *sex * temperature) on adult captures in each experimental day. Given that the interaction between treatment and temperature on adult captures was significant (p < 0.001) (see "Results" section), we explored the effects of treatment and sex, and their interaction on adult captures for each season using the GLMs with Poisson loglinear distribution. The Bonferroni test was carried out to adjust for multiple comparisons at α = 0.05. To assess the relationship between temperature and adult captures (males and females) quadratic regression analysis was performed. The proportion of female captures was estimated by dividing the total female captures by the sum of males and females captured. One-Way ANOVA followed by the Tukey HSD test for multiple comparisons was used to determine the effect of treatment on the proportion of female captured in spring and summer trials. All analyses were performed using SPSS v.26.0 (SPSS Inc., Chicago, IL, USA).

Climatic Data
The average temperature inside the field cages ranged from 11 to 21.9 • C and from 26.4 to 33.5 • C during the spring and summer experimental trials, respectively. The absolute minimum and maximum temperature ranged from 11 to 19 • C and 14 to 24 • C, respectively in spring trials, while in the summer trials it the range was from 25 to 31 • C and 29 to 36 • C, respectively ( Figure 1). The mean daily relative humidity ranged from 40 to 70% in both seasons.

Efficacy and Female Selectivity of Trapping Devices
The GLMs model reveals trapping device, temperature and sex as significant predictors of adult captures ( Table 1). The significant interaction between the trapping device and temperature highlights the importance of the season for the performance of the different mass trapping devices. (Table 1) Likewise, the significant interaction of sex by temperature indicates the differential response of males and females to different temperatures. At lower temperatures, captures were more skewed towards males than at higher

Efficacy and Female Selectivity of Trapping Devices
The GLMs model reveals trapping device, temperature and sex as significant predictors of adult captures ( Table 1). The significant interaction between the trapping device and temperature highlights the importance of the season for the performance of the different mass trapping devices. (Table 1) Likewise, the significant interaction of sex by temperature indicates the differential response of males and females to different temperatures. At lower temperatures, captures were more skewed towards males than at higher temperatures ( Figure 2). On the other hand, both males and females appear to respond equally among the different trapping devices as the interaction between these two parameters was not significant (Table 1).   Considering the data from each season, average adult captures (males and females) in spring trials ranged from 19.5 ± 3.4 to 59.6 ± 8.9 adults and in summer trials from 22.2 ± 3.2 to 55.7 ± 3.2 adults in Decis ® and Tephri traps baited with BioLure, respectively ( Table 2). Figure 3 gives the male and female captures for each trapping device during spring and summer trials. The distribution range of capture values for both males and females was wider in the spring trials compared to the summer ones. Additionally, an increased number of females responded to traps during summer regardless of the trapping device. The graph also depicts the higher performance of the Tephri tap baited with BioLure and the lower of the Decis ® trap. The GLMs model shows that both the trapping device (x 2 = 251.8, df = 4, p < 0.001) and sex (x 2 = 101.7, df = 1, p < 0.001) were significant predictors of the adult captures during spring trials. The interaction between trapping device and sex was not significant (x 2 = 7.1, df = 4, p = 0.125), suggesting similar sex ratio of the captured adults among trapping de-  The GLMs model shows that both the trapping device (x 2 = 251.8, df = 4, p < 0.001) and sex (x 2 = 101.7, df = 1, p < 0.001) were significant predictors of the adult captures during spring trials. The interaction between trapping device and sex was not significant (x 2 = 7.1, df = 4, p = 0.125), suggesting similar sex ratio of the captured adults among trapping devices. Tephri traps baited with either BioLure or Biodelear captured the highest number of adults followed by the IPMT traps baited with either BioLure or Biodelear (Table 2, Bonferroni test, p < 0.05). The Decis ® trap captured the lowest number of adults (p < 0.05). For summer trials, GLMs analysis reveals trapping device as the only significant predictor of adult captures (x 2 = 210.5, df = 4, p < 0.001), as opposed to sex (x 2 = 2.5, df = 1, p = 0.11) and their interaction (x 2 = 6.5, df = 4, p = 0.164). The Tephri trap baited with BioLure outperformed all other trapping devices during summer trials, followed by Tephri trap baited with Biodelear (p < 0.05). The IPMT trap baited with BioLure captured higher number of adults compared to both the IPMT trap baited with Biodelear and the Decis ® trap (p < 0.05) that captured similar number of individuals (p > 0.05).

Relationship of Treatment Captures and Temperature for Males and Females
Quadratic regression analysis on the effects of temperature on either male or female captures for each mass trapping device, demonstrates a peak male response at temperatures ranging from 20 to 25 • C (Figure 2). A slight shift of female response to higher temperatures has also been revealed with optimum captures recorded at temperatures ranging from approximately 22 to 28 • C (Figure 2). Responses of both males and females dramatically decrease at temperatures lower than 15 • C and higher than 35 • C. The relationship between temperature and adult captures was significant (p < 0.05) in all cases except for female response to IPMT-BioLure (p > 0.05) (Figure 2).

Discussion
Our results demonstrated that the performance of trapping devices differ between seasons. The Tephri traps outperformed the IPMT McPhail-type traps under the low temperatures of spring, regardless of the baited attractant. In summer trials, Tephri traps captured more adults when they were baited with Biolure than Bioldelear, followed by the IPMT trap baited with BioLure. It seems that the ambient temperatures in each season affect both the trap type and the attractant efficacy. Female selectivity of the five trapping devices considered in the current study is also affected by the temperatures. More males than females were captured in all trapping devices under low temperature conditions. The proportion of female captures increased during summer trials but barely exceeded that of males. Although the performance of the Decis ® trap was the lowest as far as adult captures is considered, it was the most female selective at high temperatures, as opposed to IMPT traps baited with each attractant. Hence, at least under our experimental conditions, the seasonal temperature regimes affect the performance of traps and attractants as well as the sex ratio of captured medflies.
The proportion of captured adults during both spring and summer trials never exceeded the 60% of the released individuals, albeit more flies are likely to have been attracted and landed on the traps [27]. This indicates that a considerable proportion of adults do not respond even to the best existing trapping devices and under confinement in field cages. In line with a previous comparative study conducted in citrus orchards [29], Tephri traps baited with BioLure were more attractive than IPMT ones at low temperatures. In our study Tephri traps outperformed IPMT traps at higher temperatures in contrast to Katsoyannos et al. [15] and Katsoyannos and Papadopoulos [31]. The above discrepancies may be attributed to the different experimental approaches followed and to the performance of trapping devices in divergent field sites and conditions. For example, the Open Bottom Traps were more attractive for medflies in humid tropical environments than Mediterranean orchards [13]. The performance of both Tephri and IPMT traps baited with BioLure as well as that of Decis ® traps was similar between spring and summer trials. In contrast, Tephri and IPMT traps baited with Biodelear were much more efficient in spring compared to summer trials. Apparently, the commercially available dispensers (e.g., BioLure) exhibit a rather controlled and constant emission rates at variable environmental conditions compared to Biodelear. Indeed, the custom-made dispensers considered for the novel attractant Biodelear do not assure any controlled release of the attractive compounds. This also indicates that development of a proper dispersion device could increase the performance of Biodelear.
Both BioLure and Biodelear perform better in Tephri than IPMT traps under both spring and summer conditions in field cage trials. It seems that the dispersion of the attractants was more efficient in Tephri traps compared to IPMT ones, at least in close range. Tephri traps, besides the typical lower, funnel-like opening of all McPhail type traps such as IPMT, include four lateral holes that serve as both emission and entrance gates. Other aspects that may contribute to the higher performance of the Tephri traps are related with visual stimuli, which are quite important for the short-range (within tree) attraction of the Mediterranean fruit fly and that of other fruit flies [32,33]. Visual stimuli related to trap type are expected to be more important in our field cage study since flies kept confined in proximity. However, the visual properties of such traps might not be equally important in field conditions. In such conditions, there is a list of other factors that affect the performance of the traps, such as the physiological and age structure of feral flies (e.g., age, mating status, feeding history), the structure of orchards as well as the prevailing environmental conditions that may affect both dispersion of odor attractants and response of adult flies. The efficacy of traps may be also related to retention method [34]. Wet traps that employ a liquid retention system (aqueous solution of propylene glycol or surfactant) like the Tephri traps used in our trials may suffer of high escape rates of the attracted individuals. On the other hand, under relatively dry environmental conditions, the humid environment within trap may facilitate emission of the attractant and hence have a synergistic contribution at low temperatures [15,29]. Dry traps bearing insecticide killing agents such as the Decis ® trap are expected to have lower escape rates and high knockdown effects. Repellent properties of killing agents such as DDVP (2,2-dichlorovinyl dimethyl phosphate) have been also reported in the past [15,35]. Hence, the overall efficacy of the mass trapping devices involves a complex interaction between trap design, attractant combination, retention method and temperature under semi-field conditions. Results from field cage studies testing the attractiveness of different trapping devices should be carefully interpreted considering some of the limitations mentioned earlier. However, they can generate much needed information for the performance of trapping devices during periods of adverse environmental conditions, when fruit fly populations are extremely low in field conditions like the winter/spring period in temperate areas [2].
Even though the proportion of female captures remain low (<60%), female selectivity of mass trapping devices was significantly affected by temperature. Under low temperatures, adult captures were male biased in all mass trapping devices, supporting previous results for a similar response of medfly to BioLure regardless of trap type (Tephri or IPMT), [29]. For BioLure-baited traps, the female selectivity remained high (>70%) and similar between McPhail type and Tephri traps in deciduous orchards of West Australia during summer [36]. The proportion of female captures in IPMT traps baited with BioLure was higher than that of males throughout the season in earlier detection studies in deciduous fruit orchards [2]. In field, seasonal patterns of female captures are aligned with the presence of ripe or ripening host fruits. Females are often "clustered" around host trees bearing fruits, as opposed to males which are more randomly dispersed across orchards and the fruiting season [2,[37][38][39]. Mated females are more likely to be attracted to traps that mimic the shape, size, and color of host fruits when searching for mates and oviposition sites [32,40]. Whether the visual properties of the traps account for the variability in the proportion of female captures in our summer trials (e.g., higher performance of the Decis ® traps) needs to be further explored. It is well demonstrated that female medflies respond strongly to yellow traps (such as IPMT and Tephri traps) [32], and sphere-like objects (such as the orange Decis ® trap) [31,33,41,42]. For increasing female selectivity, female-targeted traps are suggested to be placed on or near host trees bearing ripe or semi-ripe fruits [43]. Hence, along with the trap type, the presence of ripe or ripening host fruits (early maturing or not) may be crucial for female captures under different temperature conditions, and related field verification studies are needed.
Low medfly populations remain a prerequisite for success of mass trapping control in deciduous fruit orchards [25,44,45], underlying the need for both timely detection of small population and the use of efficient trap-attractant combinations for mass trapping control. Trimedlure-based traps are used worldwide for early detection purposes based on their species specificity and their effectiveness under low temperature and low population conditions [29,46]. Males were caught earlier in male lure-based traps than females in food-based traps [19,29,36], although exceptions exist [2,13,14,37]. Inclusion in medfly monitoring networks of food-based lures that capture both females and males will increase the probability of precisely detecting medfly [19,29], excluding in any case any potential negative interactions between traps when placed in proximity [47]. Our results that adult captures were higher in Tephri traps baited either with BioLure or Biodelear at low temperatures, highlighting the potential of both Tephri combinations to be used for early detection purposes. In line with our results, Miranda et al. [29] proposed Tephi traps baited with BioLure for monitoring wild female medfly populations under low temperatures. The same combination of traps and attractants could be considered to target low medfly populations for mass trapping purposes early in the season and under low temperature conditions. On the other hand, female selectivity of mass trapping devices is also equally important for success of mass trapping control since maximum number of females need to be removed from field populations before attacking fruits [48]. The low female selectivity of all trapping devices tested under low temperature conditions should be considered in early season mass trapping efforts. Nevertheless, the efficacy of mass trapping to control medfly populations might remain high since very low populations are targeted. Considering also the efficacy of mass trapping devices [17], Tephri traps baited with BioLure as a first choice and Tephri traps baited with Biodelear as a second choice could be used for mass trapping control at higher temperatures. Under such conditions, the female selectivity in the "ready-to-use" Decis ® trap was substantially higher compared to that of Tephri and IPMT traps, but the low numbers of total captures are a matter of concern. Along with the extensively reported superior performance of BioLure for suppressing medfly populations at high temperatures [17,24,31,39,49], our results indicate also the promising choice of Biodelear in Tephri trap, instead the IPMT trap as previously selected for pilot trials [21].

Conclusions
The selection of effective trap-lure combinations for monitoring and mass trapping control of medfly populations especially early in the season and under low population densities remain challenging. In this study, we showed that both efficacy and female selectivity of mass trapping devices are related to prevailing temperatures during spring and summer seasons under semi-field conditions. However, the field cage trials provide an approximation of the performance of trap-attractant combinations [50,51] and, consequently, the field verification of the most promising mass trapping devices should be conducted.