Tachinidae ( Diptera ) collected in traps used for mass-trapping of Bactrocera oleae ( Rossi ) ( Diptera : Tephritidae ) in olive groves in Central Spain

Tachinid flies are important for biological control of pests, because most species are parasitoids of insects. The objectives of this work were 1) to describe the species of tachinid flies captured by mass trapping devices against the olive fruit fly, Bactrocera oleae, in olive groves in Central Spain, and 2) to report on the selectivity of the different devices for this important group of insects. The study was carried out in two olive groves in the province of Madrid during 2005 and 2008. The number of trapping devices was five in 2005 and twelve in 2008. A total of 66 species of tachinid flies was captured. Comments on some of them are provided because of their special interest. Three tachinid species (Clemelis massilia, Schembria meridionalis and Ceromya flaviseta) are recorded from the Iberian Peninsula for the first time. It is remarkable that captures of C. massilia accounted for 72.3% in 2008. Comparison of the different mass-trapping devices indicates that none of them is selective for this important family of parasitoids. Numbers of tachinid flies captured per trap and season ranged between 1 and 13 in 2005, and 24 and 283 in 2008. This is the first study carried out in Spain on tachinids captured by mass-trapping devices and it reports the abundance and species diversity of this group of insects in olive groves as well as the poor selectivity of mass-trapping on them. Additional key words: Ceromya flaviseta; Cestonia cineraria; Clemelis massilia; Mintho compressa; Schembria meridionalis; side-effects.

In Spain pest control programs in olive groves are focused mainly on the olive fruit fly, Bactrocera oleae (Rossi, 1790), which is the key pest of this crop.One of the control methods available to control B. oleae is mass-trapping.However, the use of traps may have an adverse effect on non-target insects, especially on insects belonging to the order Diptera, in which the family Tachinidae is considered of highest importance among the beneficial ones.
Preliminary results on undesired side effects of mass-trapping of B. oleae have shown that the different devices commonly used for this control method capture a large number of non-target arthropods (Seris et al., 2010).Predators of 35 different families and parasitoids of 26 families were captured by the traps.The most important family of parasitoids captured was Tachini-dae (about 80% of the parasitoids were Tachinidae).Other studies have reported that the family Tachinidae forms the largest component of the predator-parasite guild present in surveillance traps used for Mexican fruit flies (Thomas, 2003).
To better understand an agroecosystem we must know not only the amount of non-target arthropods affected by control methods, but also the species to which they belong.It is important to determine the species of tachinid flies present in our agroecosystem in order to know to what extent they could contribute to natural control of insect pests in olive groves.Thus, the objective of this work was two-fold: first, to carry out a taxonomic study to determine the tachinid species captured by mass-trapping devices, and second, to compare different devices regarding their negative impact on tachinids as a whole.To our knowledge, this is the first taxonomic study on tachinids captured by traps in olive groves carried out in Spain.

Material and methods
Specimens were captured in two olive groves near Villarejo de Salvanés (southeastern Madrid, Spain).The area sampled in 2005 is located about 8 km south of Villarejo de Salvanés (40.106766N, 3.273411W, 690 m) and it is surrounded by agricultural landscape and wasteland (Figure 1a).In 2008 the olive grove was about 3 km northeast of Villarejo de Salvanés (40.178841N 3.249507W, 760 m).It is a mixed vine and olive field (Figure 1b).
Five and twelve different mass-trapping devices were used in 2005 and 2008, respectively.These were  combinations of traps and attractants, as described in Table 1.The traps used in the study (Figure 2) are designed for mass-trapping and monitoring of different fruit flies, including B. oleae.The attractants are commonly used for mass-trapping of tephritids and they were prepared according to the manufacturer's directions.Nulure ® and Tephri-lure ® are protein hydrolysates.Diammonium hydrogen phosphate was added at a 4% rate.The period sampled in both years was from June to November and attractants were replaced every other week.One trap was hung per tree and four and three replicates were set up for each combination of trap and attractant in 2005 and 2008, respectively.In 2005, an additional trap (Easy trap ® + Nulure ® ) was placed on a holm oak (Quercus ilex Linnaeus, 1753) next to the olive grove.In both years traps were hung in contiguous trees in a homogeneous area in the centre of the olive groves.
Trapped arthropods were collected by filtering the bait when it was replaced.Specimens were determined to family level at the Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA), Madrid, Spain.The majority of the captured tachinids were sent for determination to the first author, at the Staatliches Museum für Naturkunde (SMNS), Stuttgart, Germany.Easy trap ® Tephri-lure ® x 5 Probodelt ® (Fig. 2c) Nulure ® x 6 Olipe (Fig. 2d) Tephri-trap ® (Fig. 2e) Tephri-trap Ecological ® (Fig. 2f) Tephri-trap Ecological ® Nulure ® x 14 Tephri-trap Ecological ® Tephri-lure ® x Tachinidae captured in devices for olive fly mass-trapping in Central Spain Some specimens were not sent because they were in poor condition and, although recognizable as tachinids, they were not suitable for species identification.

Statistical analysis
Analysis of variance was carried out on number of tachinids per trap and season captured by the different mass-trapping devices.After analysis, mean values were compared by a Tukey HSD test (p ≤ 0.05).Statistical tests were performed using the software Statgraphics ® Centurion XV (StatPoint, 2005).

Captured species
Total of 155 specimens of Tachinidae, belonging to 32 species, and 3,465 specimens, belonging to 55 species, were captured in 2005 and 2008, respectively.The total number of species reported in this study is 66.From the whole numbers of arthropods collected, 1.7% in 2005 and 2.4% in 2008 were tachinids.They were the most important group of parasitoids captured both in 2005 (65.6%) and 2008 (83.2%).
The relative abundance of the different species of tachinids was low, with percentages from the total numbers of tachinids lower than 20%, except C. massilia, captures of which accounted for 72.3% of all tachinids in 2008.Table 2 lists the species in the order of the relevant catalogue of Herting and Dely-Draskovits (1993).The number of males and females is given for each year.Additionally, summarised information on the hosts is provided, based on published and unpublished data compiled by Tschorsnig for many years.A single host species mentioned in this table does not necessarily mean that the respective tachinid is a specific parasitoid of that host species because it might simply be the only host record known so far.In case of Lepidoptera, the larval stage is the host; for other host orders the parasitized stage is detailed in Table 2.Only 11 specimens out of the 155 captured in 2005 were captured in the holm oak trap.
Members of three of the four subfamilies of Tachinidae (Exoristinae, Tachininae and Dexiinae) were well represented in the traps, whereas only two specimens of a single species belonged to the fourth tachinid subfamily, Phasiinae (parasitoids of adult Heteroptera).It remains unknown why the traps were not attractive for members of this subfamily.Small species (7 mm in body length or less) accounted for 93% of the collected specimens and only 7% were larger ones (up to 12 mm).
It is striking that parasitoids of typical olive tree pests were nearly absent among the collected Tachinidae.There are only two specimens of P. nigrina, a species which is known from P. oleae, among other microlepidopterous hosts.Similarly only four specimens of N. maculosa, and 40 specimens of Pales pavida (Meigen, 1824) were caught, both being unspecialised species including P. vitrealis as host.C. massilia is a special case (see below under the heading "Remarks on species of special interest").
Tachinidae are robust, excellent flyers.Therefore, there is no guarantee that the hosts of the collected species develop in the experimental areas.In certain cases this is obviously true, e.g. for Phryxe caudata (Rondani, 1859) which needs pine trees for its host Thaumetopoea pityocampa (Denis and Schiffermuller, 1775).

Remarks on species of special interest
All but three species of Tachinidae captured have already been recorded from the Iberian Peninsula (Tschorsnig and Báez, 2002).For more detailed information see Tschorsnig (1992) and Tschorsnig et al. (1997).
The following five species deserve further comment.

Cestonia cineraria Rondani, 1861
The last record of this species from the Iberian Peninsula dates from more than 100 years ago (Czerny and Strobl, 1909).
Clemelis massilia (Herting, 1977) This species was by far the most common tachinid found during this investigation (72.3% of all specimens ) an important role on the experimental area or at least nearby.The ptilinum was improperly retracted in about 20% of the collected specimens, which indicates that they most probably had entered the traps soon after emergence.This also implies that their host must have developed not far away from the traps.Furthermore it may be concluded from the more or less uniform dimension of this tachinid, i.e. variation of the body length between 4.5 and 5.5 mm (only single specimens being smaller, up to 3.5 mm, or larger, up to 7 mm), that C. massilia probably develops in a single host species.The pyralid moth P. vitrealis, regularly found in low numbers in an olive plantation near the study area, would perhaps be a potential host, but there was no rearing or direct observation to support this suggestion.Wood-boring hosts, such as Euzophera pinguis (Haworth, 1811), can be ruled out because the small eggs of Clemelis are laid on leaves and must be swallowed by the feeding caterpillars.
C. massilia was described from southern France by Herting (1977) and later it was also found in Tenerife (Tschorsnig and Báez, 2002) and Italy (Cerretti, 2004).The species was not yet recorded from the Iberian Peninsula, but it is probable that some older identifications of the very similar species Clemelis pullata (Meigen, 1824) might refer to C. massilia.The separation of both species is sometimes difficult because several characters (head width, distance between the posterior ocelli, thickening of the arista) overlap.The most reliable character is the male genitalia as described by Herting (1977).Females often remain doubtful (but were assignable to C. massilia in the present investigation).

Schembria meridionalis Rondani, 1861
This is a rare species which was collected in the traps in relatively large numbers (32 specimens).S. meridionalis was only known before as a few specimens from Malta, Sicily and Israel (Herting and Dely-Draskovits, 1993;Cerretti, 2005).This is the first record of this species from the Iberian Peninsula.Nothing is known on its hosts.
The collected specimens of S. meridionalis have a remarkably variable body length (between 3.9 and 8.0 mm).Also the petiole of the wing cell R4+5 is of variable length, between slightly less than diameter of vein M to slightly longer than half of crossvein r-m.
Ceromya flaviseta (Villeneuve, 1921) This species is mainly distributed in temperate Europe and this is the first record from the Iberian Peninsula.
Mintho compressa (Fabricius, 1787) Only the holotype of M. compressa was previously known from Spain, described by Fabricius (1787) without exact location, so this is only the second record for the Iberian Peninsula in more than 200 years.However, the species is widely distributed and common in other Mediterranean countries and it is not known why M. compressa is so rare in the Iberian Peninsula.

Comparison of trapping devices
The mean number of tachinid specimens collected by the different combinations of trap + lure tested in 2005 and 2008 are given in Table 3.As commented in the materials and methods, in 2008 a number of specimens were determined as tachinids but it was not possible their determination at species level.
Other studies of the side effects of mass-trapping on non-target arthropods have reported no captures of tachinids.Porcel et al. (2009) studied the effect of the Olipe mass-trapping (Figure 2d) on olive non-target arthropods, and they found a high proportion of nontarget Diptera captured but they did not report their families.On the other hand, in the study by Thomas (2003) on non-target insects captured in Mexican fruit fly surveillance traps, tachinids accounted for 58% of the predator-parasite guild.
The number of tachinids captured in 2005 was much lower than that in 2008.The two field trials were carried out, apart from different years, at different locations (Figure 1), and the two olive groves differed in features relevant to the arthropod fauna expected to be present in the canopy, such as surrounding landscape, crop age, soil type and cover, etc.
When comparing the devices, in 2005, the McPhail trap baited with diammonium hydrogen phosphate was the one that captured the lowest number of tachinids (Table 3).No significant differences were observed amongst the catches in other four devices studied, the Probodelt trap baited with the attractant Nulure capturing the highest number of specimens (Table 3).When comparing numbers captured in the three traps baited with the Tachinidae captured in devices for olive fly mass-trapping in Central Spain same attractant Nulure, the Probodelt trap captured a significantly higher number than the Olipe and Easy traps.
In 2008, Tephri-trap Ecological baited with Nulure was the device that captured the highest number of tachinids, while the lowest value was for Tephri-trap baited with diammonium hydrogen phosphate (Table 3).The number of specimens captured by Olipe traps never reached those observed in the other traps with some of the attractants.On the other hand, Tephri-trap Ecological captured very high numbers of tachinids, regardless of the attractant.Tephri and Easy traps showed intermediate values.Regarding the attractants, no general pattern can be detected.
It is difficult to conclude which devices are most environmentally friendly.We have previously reported that there is a strong variation in the captures of B. oleae and other non-target arthropods in each individual trap (Seris et al., 2007(Seris et al., , 2010)), which indicates a high heterogeneity amongst the trees of the same olive grove.In this study we have avoided including trees which are on the edge of the groves, because of this heterogeneity, and we have used contiguous trees within a relatively homogeneous area inside the olive grove.However, there are individual features associated with each tree that cannot be avoided.To minimize this strong variation it would be desirable to carry out studies including a larger number of trees per treatment.
In spite of this, there are general trends that agree with the results of captures of other insects.Protein hydro-lysates capture higher numbers of individuals than diammonium hydrogen phosphate.This same result has also been reported for B. oleae and non-target arthropods (Seris et al., 2010).Thomas (2003) reported also higher numbers of tachinids captured by proteinaceous attractants (Torula yeast) compared to synthetic lures (ammonium acetate and putrescine) and Gómez-Gómez et al. (2010) reported the attractiveness of carrion for tachinids.But there are still obvious knowledge gaps on lures which might be specifically attractive for Tachinidae.General trends concerning traps are that Olipe trap captures low numbers of arthropods, while the number of specimens captured by Probodelt trap is high, as reported previously (Seris et al., 2007(Seris et al., , 2010)).
Tephri-Trap Ecological is designed to reduce the captures of non-target arthropods.Based on this aim, it incorporates nets covering the entrance openings of the trap (Figure 2f).For some beneficial arthropods there was indeed a reduction in the number of captures by this device, such as chrysopids (Seris et al., 2007(Seris et al., , 2010)).However, in this study Tephri-Trap Ecological did not exclude tachinids (Table 3), since all captured specimens were small enough to go through the net.Olipe and McPhail traps never caught high numbers of tachinids, and it is possible that the colour of the traps might have played a role here.Olipe and McPhail traps are transparent while Tephri-traps and Easy traps are yellow.Yellow colour is well-known as an attractant for many flying insects and is used for Tachinidae in yellow pan traps (Tschorsnig, 2002(Tschorsnig, , 2008)).It is important to bear in mind that the main objective of these traps is capturing as many B. oleae specimens as possible.Thus, the balance between efficacy and side effects must always be taken into account, along with the fact that mass-trapping is a more environmentally friendly control method for B. oleae than conventional chemical control (Haniotakis et al., 1991;Porcel et al., 2009).This study has shown that there is a large number of tachinid species in the olive agroecosystem in central Spain, some of which have been reported here for the first time.The number of tachinids captured suggests that they could play an important role in the balance of the agroecosystem.However, further work should be carried out to clarify the impact of these species on natural pest control as well as to determine the impact of different mass-trapping devices on those relevant species.

Table 1 .
Combinations of traps and attractants used in the mass-trapping field trials in 2005 and 2008

Table 3 .
Number of tachinid specimens captured per trap and season during 2005 and 2008 Data are means of three replicates.Data followed by different letters for each year are significantly different according to Tuckey HSD test (p ≤ 0.05).