Introduction
Carabid beetles are important polyphagous predators in agroecosystems (Lövei & Sunderland, Reference Lövei and Sunderland1996; Lövei, Reference Lövei2008). Most of them consume other insects, molluscs or millipedes, and a range of plant material such as seeds, or are scavengers (Kromp, Reference Kromp1999; Symondson et al., Reference Symondson, Erickson and Liddell1999; Honek et al., Reference Honek, Martinkova and Jarosik2003; Foltan, Reference Foltan2004; Wallace, Reference Wallace2004; Wallinger et al., Reference Wallinger, Sint, Baier, Schmid, Mayer and Traugott2015). Due to their predatory behavior, carabid beetles can be important natural control agents of crop pests (Kromp, Reference Kromp1999).
The olive fruit fly, Bactrocera oleae (Rossi) (Diptera: Tephritidae) is considered the major pest of olives in most commercial olive growing regions worldwide (Nardi et al., Reference Nardi, Carapelli, Dallai, Roderick and Frati2005; Daane & Johnson, Reference Daane and Johnson2010). Damages caused by this insect include the premature fall of infested fruits, pulp consumption by developing larvae and a general reduction in olive oil quality (Pereira et al., Reference Pereira, Alves, Casal and Oliveira2004). Although control options for this pest are still based on insecticides, recent efforts intend to promote biological control. So far, the use of natural enemies, mainly parasitoids, is still unsuccessful (Daane & Johnson, Reference Daane and Johnson2010). Moreover, the action of predators on larvae is difficult once this stage develops inside the olive fruit. However, pupation occurs on the soil making this developmental stage the most susceptible to the attack of edaphic predators (Civantos, Reference Civantos1999; Orsini et al., Reference Orsini, Daane, Sime and Nelson2007).
Predation is a biotic interaction that can alter the distribution and abundance of both organisms involved in the relationship (Begon et al., Reference Begon, Townsend and Hayer2006) and should be promoted in integrated pest management programs as a mortality factor for reducing pest populations (DeBach & Rosen, Reference DeBach and Rosen1991). Such programs have been receiving increased attention because of the current need to reduce the use of synthetic insecticides for pest control (Directive 2009/128/EC). Although, successful biocontrol is critically dependent on the consumption rate of the predator in order to maintain pest density at a low level, which can vary with preferences and availability of alternative prey (Sengonca et al., Reference Sengonca, Al-Zyoud and Blaeser2005). In this context, carabid beetles are considered voracious feeders and studies of their feeding preferences and consumption rates are essential to understand basic trophic relationships and their potential efficacy as natural control agents. However, their efficiency may be affected by the simultaneous occurrence of alternative prey resulting in a decreased predation on the target pest species. Another important factor regulating population dynamics of predator–prey systems is the functional response of a predator. It represents the relationship between prey density and the number of prey consumed by an individual predator (Solomon, Reference Solomon1949) and an accurate description is important for practical and applied aspects of biological control (Van Leeuwen et al., Reference Van Leeuwen, Jansen and Bright2007). Predatory functional responses are typically described by three types of curves depending on prey density. Thus, for types I, II and III functional responses, the number of prey consumed increases linearly, asymptotically to a plateau and sigmoidally with increasing prey density, respectively (Holling, Reference Holling1966).
In the olive grove, carabid beetles can have an important role as natural enemies of the olive fruit fly population. Previous studies showed that they are abundant insects among the edaphic arthropod community of the olive grove (Santos et al., Reference Santos, Cabanas and Pereira2007; Gonçalves & Pereira, Reference Gonçalves and Pereira2012), mainly in autumn (Oliveira, Reference Oliveira2013) coinciding with the increase of pupae on the soil. Moreover, generalist carabid beetles (i.e., common species such as Carabus banonii Dejean and Pterostichus creticus (Frivaldsky) were referred to predate pupae of the olive fruit fly in the laboratory as well as in field experiments (Neuenschwander et al., Reference Neuenschwander, Bigler, Delucchi and Michelakis1983; Orsini et al., Reference Orsini, Daane, Sime and Nelson2007; Odoguardi et al., Reference Odoguardi, Bonnacci, Bruno, BrandMayr and Zetto2008). However, no studies were performed in order to understand the potential of carabid beetles as natural enemies of the olive fruit fly. Thus, the main objective of this work was to evaluate the feeding preference and functional responses of two carabid species, Calathus granatensis (Vuillefroy) and Pterostichus globosus (Fabricius), fed on pupae of B. oleae in laboratory conditions. C. granatensis and P. globosus were dominant species in olive groves, mainly in Northeastern Portugal (Oliveira, Reference Oliveira2013; Dinis et al., Reference Dinis, Pereira, Pimenta, Oliveira, Benhadi-Marín and Santosin press), representing interesting species for evaluating predation on pupae of the olive fruit fly under laboratory conditions. We tested the hypothesis that specimens belonging to the largest species, P. globosus have higher predation rates than the smallest species, C. granatensis.
Material and Methods
Test organisms
Laboratory rearing of B. oleae was initiated in October/November 2013 with pupae obtained from infested olive fruits collected in several olive groves in the region of Mirandela (Northeastern Portugal). Adult flies were kept in poly-methyl-methacrylate cages (40 × 30 × 30 cm3) and every 2 days, around 100 healthy olive fruits were provided as oviposition places. Larvae were collected daily from the infested olives and stored in plastic boxes to pupate. Rearing was maintained under controlled conditions at 21 ± 1°C, 70 ± 5% relative humidity (RH), and a photoperiod of 16:8 (L:D) at the School of Agriculture (ESA), Bragança. Pupae from the 2nd to the 5th generation were used in the experiments.
Ceratitis capitata (Wiedemann) pupae were originally collected from the stock colony in the Unidad de Protección de Cultivos of Technical University of Madrid, and rearing has been maintained at the ESA, Bragança since September 2012. Adult flies were kept in poly-methyl-methacrylate cages (40 × 30 × 30 cm3) under controlled conditions at 24 ± 2°C; 60 ± 5% RH and a photoperiod of 16:8 (L:D). Larvae were reared on an artificial diet according to González-Núñez (Reference González-Núñez1998). Both adults of B. oleae and C. capitata were fed ad libitum with water and an artificial diet composed by a mixture of sucrose and yeast hydrolysate at a ratio 4:1 (based on dry weight). C. capitata pupa was used as model alternative prey in preference experiments and was selected due to its similarity with B. oleae trying to mimic less mobile prey items present in the olive grove.
Adult specimens of C. granatensis and P. globosus were hand collected in an organic olive grove in the region of Mirandela (Northeastern Portugal) between September 2013 and May 2014. Specimens found on the ground, leaf litter or under the stones were placed in plastic boxes (7.5 cm in diameter × 4.5 cm height) and carried out to the laboratory where the identification of the species was confirmed and each species was transferred to different rearing plastic cages (15 × 37 × 53 cm3) containing dry natural soil (a layer of about 8 cm height) and several stones scattered on the surface to provide shelter. The soil used in the rearing cages was collected in the olive grove, sieved to <2.0 mm and air dried. Beetles were fed every 5 days with different food items such as C. capitata larvae and dead adults, B. oleae dead adults, and cat food; water was provided in wet acrylate spheres. Specimens were acclimatized, at least, for 2 months before the beginning of the experiments.
Feeding preferences, predation efficiency and functional responses
Specimens of C. granatensis and P. globosus were transferred from the rearing cages and placed individually in plastic containers (10.7 cm diameter and 4.0 cm height) with a layer of dry natural soil, a small stone for sheltering and one wet acrylate sphere for water supplying. A hole of 6.0 cm in diameter was made on the lid of the containers and substituted by a permeable piece of cloth (1.0 mm mesh) to ensure ventilation. Experiments were performed under controlled conditions at 21 ± 1°C, 70 ± 5% RH, and a photoperiod of 16:8 (L:D). Carabid beetles were starved for seven days prior to the start of each experiment.
Feeding preferences
Eight treatments, corresponding to eight different prey ratios, were offered to each carabid beetle species. The following prey ratios were tested: (1) 20 B. oleae pupae, (2) 18 B. oleae pupae and 2 C. capitata pupae, (3) 15 B. oleae pupae and 5 C. capitata pupae, (4) 12 B. oleae pupae and 8 C. capitata pupae, (5) 10 B. oleae and 10 C. capitata pupae, (6) 8 B. oleae pupae and 12 C. capitata pupae, (7) 5 B. oleae pupae and 15 C. capitata pupae and (8) 2 B. oleae pupae and 18 C. capitata pupae. For each prey ratio, pupae were randomly allocated in a Petri dish (6.0 cm diameter and 1.0 cm height), and then placed inside each testing container. A total of 25 individuals of each species were tested in each treatment for 24 h. Pupae consumed by each beetle were calculated by counting the non-consumed pupae of each prey and subtracting them to the initial number. Each specimen was used once in each combination.
Predation efficiency
The weight of 30 randomly selected individuals of C. granatensis and P. globosus was recorded in order to calculate the average weight of each species. Individuals were starved for 5 days to guarantee equal conditions; they were cleaned with a moisten paint-brush to remove soil particles and weighted individually in plastic tubes. The weights of 50 pupae of B. oleae and C. capitata were also recorded to calculate the average weight of each prey. Data obtained were used to evaluate the biomass of prey consumed by each predator, by multiplying the average weight of pupae by the number of pupae consumed by each individual and was also used to measure a predator weight/prey weight ratio.
Functional responses
Experiments were conducted using ten adult specimens of each carabid beetle species as replicates in each density. Different densities of the prey (pupae of B. oleae) were offered to each species. C. granatensis were exposed to seven densities (2, 5, 10, 15, 20, 25 and 30 pupae of B. oleae), whereas P. globosus were exposed to 11 densities due to their bigger size (2, 3, 5, 8, 10, 15, 20, 25, 30, 40 and 50 pupae of B. oleae). After 24 h, the number of pupae consumed was recorded.
Data analysis
Statistical analyses were performed using R (R Core Team, 2015). Firstly, feeding preferences of C. granatensis and P. globosus were assessed by performing a multivariate two-sample Hotelling's T 2 test (Lockwood, Reference Lockwood1998) using the hotelling.test function from the Hotelling package. The percentage of biomass of prey consumed was used as explanatory variables in order to reduce the noise introduced by the different total percentage of prey consumption within each arena and 10,000 permutations were carried out to lead with the lack of independence on data. The consumed ratios of B. oleae pupae were calculated by dividing the number of B. oleae pupae by the total number of pupae consumed. Then, Manly's preference index (Manly et al., Reference Manly, Miller and Cook1972) was calculated; this is a method to evaluate preference that takes into account the prey densities depletion by predation during experiments (Cock, Reference Cock1978) as following:
$$a = \displaystyle{{r_1 /n_1} \over {r_1 /n_1 + {\rm} r_2 /n_2}} $$
where r 1 represents the proportion of prey 1 in the predator diet (B. oleae pupae), and n 1 the proportion of prey 1 available (0.9, 0.75, 0.6, 0.5, 0.4, 0.25, 0.1); r 2 represents the proportion of prey 2 in the predator diet (C. capitata pupae) and n 2 the proportion of prey 2 available (0.1, 0.25, 0.4, 0.5, 0.6, 0.75, 0.9).
The predation efficiency was evaluated using the total number of consumed pupae, the total biomass of consumed pupae (calculated as the weight of the pupae × number of consumed pupae) and the percentage of biomass of consumed pupae (calculated as the total biomass consumed/total biomass offered, in percentage).
The consumed ratios of B. oleae pupae, the Manly's preference index values, the total number of pupae consumed, the total biomass and the percentage of biomass of consumed pupae were compared between species of carabid beetles using two-sided t-tests and subsequently a pairwise procedure was followed correcting the alpha threshold (0.05/21 = 0.0024) in order to uncover differences between the offered ratios of B. oleae pupae within each beetle species and the same alpha threshold was used all along the analyses. All statistical outputs were summarized in the Appendix.
Functional response – A logistic regression analysis was used to determine the shape of the functional response with the proportion of predated pupae versus initial density of pupae (Trexler et al., Reference Trexler, McCulloch and Travis1988). In the regression, it was fitted a polynomial function (Juliano, Reference Juliano, Scheiner and Gurevitch2001) as the following:
$$\displaystyle{{N_e} \over {N_0}} = \displaystyle{{{\rm exp}\left( {{\rm \beta} _0 + {\rm \beta} _1 N_0 + {\rm \beta} _2 N_0 ^2 + {\rm \beta} _3 N_0 ^3} \right)} \over {1 + {\rm exp}\left( {{\rm \beta} _0 + {\rm \beta} _1 N_0 + {\rm \beta} _2 N_0 ^2 + {\rm \beta} _3 N_0 ^3} \right)}}$$
where N e represents the number of B. oleae pupae consumed, N 0 is the initial density of B. oleae pupae, β0, β1, β2 and β3 are, respectively, the constant, linear, quadratic and cubic parameters related to the slope of the curve that were estimated using the method of maximum likelihood (Juliano, Reference Juliano, Scheiner and Gurevitch2001). A negative linear coefficient means a better adjustment to type II, whereas a positive linear coefficient and a negative quadratic coefficient imply that the data fit a type III functional response. Significance level was established at P = 0.001.
Discrimination between types I and II responses has previously been carried out by comparing proportional mortality at different prey densities (Rogers, Reference Rogers1972; Juliano, Reference Juliano, Scheiner and Gurevitch2001). The data indicated type II functional responses for both carabid species and because there was no prey replacement, the random predator equation was fitted (Rogers, Reference Rogers1972), excluding those densities in which all prey were consumed in 24 h, following equation (3).
$$N_e = {\rm} N_0 \{ 1 - {\rm exp}\,[a(T_h N_e -{\rm} T)]\} $$
where a represents the attack rate (searching efficiency per time), T h the handling time (time to attack, kill and eat each prey) and T the time of the experiment (24 h).
Rogers’ random predator equation was primarily fitted using the nlsLM and lambertW functions of the minpack.lm and emdbook packages respectively in R, but an overestimated result for the parameter a was obtained. Therefore, a second fitting was performed establishing an upper limit for this parameter. The value used was the asymptote of the model, i.e., the inverse of the handling time (1/T h ) estimated after the first model fitting (Bolke, Reference Bolke2007).
Estimated T h were used to calculate maximum attack rates T/T h , which is the maximum number of prey that can be attacked by a predator during the time interval considered. Data are presented as mean values ± 1 standard error (SE).
Results
Feeding preferences
The Hotelling's T 2 test showed statistically significant differences for the proportional consumption of the two prey species (T 2 = 441.93, P < 0.001). For C. granatensis, the consumed ratio decreased when the offered ratio of B. oleae pupae decreased but was always superior to the offered ratio of B. oleae (fig. 1). For P. globosus, the consumed ratio was superior to the offered only when the number of B. oleae available was higher than the number of C. capitata. For this species, the consumed ratio of B. oleae pupae decreased with the decrease of the offered ratio of B. oleae pupae (fig. 1).
Fig. 1. Consumed ratio of Bactrocera oleae pupae (mean – standard error of the mean, SE) for Calathus granatensis and Pterostichus globosus. Means with different letters for each carabid species were significantly different (P < 0.0024). In all cases, the different consumed ratios were significantly different between the two carabid beetle species (P < 0.0024).
Manly's preference index was significantly different over the offered ratio of B. oleae pupae for each carabid beetle species (Appendix 1). Comparing both species, C. granatensis showed significantly higher Manly's preference indexes than P. globosus for all offered ratios of B. oleae pupae (table 1). For C. granatensis, the Manly's preference index increased as the offered ratio of B. oleae pupae also increased, being higher than 0.80 (80%) for all the offered ratios. On the other hand, for P. globosus, the Manly's preference index decreased when the offered ratio of B. oleae decreased. When the offered ratio was higher than 0.5, this species showed a preference above 50%, however, when the offered ratio reached 0.5 the index decreased rapidly reaching 24%.
Table 1. Manly's preference indexes (mean ± SE) for different ratios of offered Bactrocera oleae pupae for adult Calathus granatensis and Pterostichus globosus.
SE, standard error of the mean.
Means within a column with different letters were significantly different at P < 0.0024.
The asterisks (*) mean that, within the row, carabid species were significantly different for the same ratio of B. oleae pupae at P < 0.0024.
Predation efficiency
The average (±SE) weight of C. granatensis was 47.5 ± 2.2 mg and the average weight of P. globosus was 248.8 ± 7.4 mg. For the prey, the average weight of B. oleae pupae was 5.0 ± 0.2 mg, whereas that of C. capitata was 8.0 ± 0.1 mg. The ratio between the weight of the predator and that of B. oleae pupae was 9.5 for C. granatensis and 49.8 for P. globosus.
The total number of consumed pupae was significantly different for C. granatensis over the offered ratios of B. oleae pupae (table 2) and also between species for each offered ratio of B. oleae pupae (table 2). C. granatensis consumed a significantly higher number of pupae in the 0.9 ratio of B. oleae pupae when compared with the 0.1 ratio of B. oleae pupae while no differences were detected for P. globosus (table 2).
Table 2. Total number of pupae, percentage of total biomass and total biomass consumed (mean ± SE) of Bactrocera oleae pupae plus Ceratitis capitata pupae (Bo + Cc) in 24 h by Calathus granatensis and Pterostichus globosus for different ratios of offered B. oleae pupae.
SE, standard error of the mean.
For each carabid species, means within a column with different letters were significantly different at P < 0.0024.
The asterisks (*) mean that, within a column, carabid species were significantly different for the same ratio of B. oleae pupae at P < 0.0024.
The percentage of biomass consumed over the offered ratios of B. oleae pupae differed significantly for C. granatensis but it was not different for P. globosus (table 2). There were significant differences between the percentage of biomass consumed by both species; the number of pupae and the percentage of total biomass consumed by P. globosus were about three times higher than that consumed by C. granatensis (table 2).
The total biomass consumed over the offered ratio of B. oleae pupae differed significantly for P. globosus (increasing the total biomass of prey consumed as the offered ratio of B. oleae pupae decreased (table 2). However, for C. granatensis the total biomass consumed did not differ significantly with the decrease of the offered ratio of B. oleae pupae. The total biomass consumed was significantly higher for P. globosus than for C. granatensis (table 2).
Functional responses
The estimated parameters from the logistic regression analysis of the proportion of B. oleae pupae consumed by C. granatensis and P. globosus indicated a type II functional response for both species based on the respective linear coefficient obtained, β1 = −3.425 ± 0.736, P < 0.001 and β1 = −0.312 ± 0.090, P < 0.001.
Both carabid beetle species showed an increase in predation with the increase of the density of B. oleae pupae, although C. granatensis reached a plateau at lower prey densities (fig. 2). For P. globosus, the estimated handling time (T h ) was 1.223 ± 0.118 h and the coefficient of attack rate (a) was 0.281 ± 0.165 h−1, resulting in a maximum attack rate (T/T h ) of 19.6 pupae. For C. granatensis, the estimated handling time was 3.230 ± 0.627 h and the coefficient of attack rate was 0.300 ± 0.939 h−1, resulting in a maximum attack rate of 7.4 pupae.
Fig. 2. Functional responses of adult Calathus granatensis (a) and Pterostichus globosus (b) fed for 24 h on increasing densities of pupae of Bactrocera oleae. Circles represent the number of prey eaten in each density of offered pupae and lines represent the sketched fitted values.
Discussion
This study shows that both C. granatensis and P. globosus were able to feed on B. oleae pupae, although they had significantly different feeding preferences and abilities to respond to increasing prey densities. Thus, C. granatensis had a preference for B. oleae pupae irrespectively of the offered proportion of prey, and consumed more pupae and more percentage of biomass at high ratios of B. oleae. On the other hand, P. globosus preferred the alternative prey and showed some degree of switching since B. oleae was disproportionately less eaten when it was present at low ratios. In this context, P. globosus seemed to be more polyphagous than C. granatensis since the former was able of exploiting both resources. This characteristic was previously noted by Hengeveld (Reference Hengeveld1980) referring that species of the genus Pterostichus eat whatever prey they can ingest. Diverse prey items, such as slugs (Oberholzer et al., Reference Oberholzer, Escher and Frank2003), lepidopteran pests (Suenaga & Hamamura, Reference Suenaga and Hamamura1998) and dipteran pupae including B. oleae pupae (Neuenschwander et al., Reference Neuenschwander, Bigler, Delucchi and Michelakis1983; Odoguardi et al., Reference Odoguardi, Bonnacci, Bruno, BrandMayr and Zetto2008) are commonly present in the diet of these carabid beetles.
P. globosus consumed a significantly higher amount of total biomass than C. granatensis probably because the former is larger (14–22 mm in length) than the latter with 9–12 mm in length (Aguiar & Serrano, Reference Aguiar and Serrano2012) and larger carabid beetles have larger guts and consequently are able to consume more biomass (Wallace, Reference Wallace2004). Such differences can also justify the results obtained in the food preference experiments as P. globosus seemed to select prey items that were more valuable in terms of energy intake per unit of handling time. In previous studies conducted to evaluate feeding preferences of carabid beetles on different slug species, the weight of the slug was considered the main factor influencing the choice of the predator (Thiele, Reference Thiele1977; Ernsting & Vanderwerf, Reference Ernsting and Vanderwerf1988; Wheater, Reference Wheater1988; Ayre, Reference Ayre2001; McKemey et al., Reference Mckemey, Symondson, Glen and Brain2001; Hatteland et al., Reference Hatteland, Haukeland, Roth, Brurberg, Vaughan and Symondson2013) followed by the slug species (Foltan, Reference Foltan2004). Thus, in our study, P. globosus could select C. capitata pupae because it is the heaviest prey item representing the most profitable prey in terms of gained energy. Moreover, the apparent switching behavior showed by P. globosus, which started when both prey items were equally present, demonstrates that this species can be more opportunistic in its feeding habits, switching to the most abundant prey available, which is a common behavior for carabid beetles (Hengeveld, Reference Hengeveld1980; Barney & Pass, Reference Barney and Pass1986). On the other hand, the smaller size of C. granatensis may determine its ability to efficiently exploit one prey instead of the other. Several morphological constrains, such as the mandible size (Hengeveld, Reference Hengeveld1980), can limit C. granatensis of easily exploiting C. capitata pupae that mainly fed on the alternative prey at lower ratios of B. oleae. This idea is reinforced by the fact that the total biomass consumed by C. granatensis did not differ with the decrease of the offered ratio of B. oleae pupae which suggests that the presence of higher densities of the alternative prey did not significantly influence the choice of the predator.
As far as we know, there are no other studies considering the feeding preferences and efficiency of these carabid beetle species, although they are quite well distributed in the Iberian Peninsula. P. globosus can be found in many agro-forestry environments (e.g., forests of oaks and pines and olive groves) and grasslands, usually found under stones and in the leaf litter (Cárdenas & Bach, Reference Cárdenas and Bach1988; Ortuño, Reference Ortuño1990; Oliveira, Reference Oliveira2013); C. granatensis is an Iberian endemism, is also a lapidicolous beetle, commonly found in olive groves (Cárdenas & Bach, Reference Cárdenas and Bach1985, Reference Cárdenas and Bach1993; Zbyšek, Reference Zbyšek2012; Oliveira, Reference Oliveira2013).
Both carabid beetle species exhibited a type II functional response in which the consumption rate of B. oleae pupae rose with prey density, but gradually decelerated until a plateau was reached and the consumption rate remained constant with the increase of B. oleae pupae density. The plateau was reached at lower numbers of consumed prey by C. granatensis than by P. globosus meaning that they differ in their maximum consumption rates. This kind of response is the most frequently observed in many arthropod predators (Hassell et al., Reference Hassell, Lawton and Beddington1977; Sueldo et al., Reference Sueldo, Bruzzone and Virla2010) and is characterized by a predation rate that is limited by the handling time that a predator needs to devote to each prey item it consumes (Sueldo et al., Reference Sueldo, Bruzzone and Virla2010). Thus, as prey density increases, searching for prey takes shorter time and limits less the predation rate because prey is easier to find, becoming the predation rate affected by the handling time, which causes a decelerating rate of increase in the predation rate (Sueldo et al., Reference Sueldo, Bruzzone and Virla2010). The estimated handling time of B. oleae pupae for C. granatensis was, in average, 2.6 times longer than for P. globosus. Thus, although both carabid beetles exhibited the same kind of functional response, the time required for handling B. oleae pupae may indicate different abilities to deal with increasing B. oleae densities and different levels of satiation, voracity or digestive rates between them. P. globosus can consume more pupae before satiation and can be more efficient in handling pupae than C. granatensis.
According to these results, both species can be natural control agents of B. oleae in the field since both were able to decrease the abundance of pupae. However, the ability of a predator to control pests is dependent on the predator's functional response, on the presence of alternative prey and on the interactions between predator species (Lester & Harmsen, Reference Lester and Harmsen2002). C. granatensis showed higher preference for B. oleae pupae in detriment of the alternative prey, thus, for this species, the presence of alternative prey items in olive groves might affect less its consumption on B. oleae pupae. However, other studies need to be done using prey items smaller than B. oleae pupae in order to clarify the feeding habits of this species. Although P. globosus consumed more B. oleae pupae than C. granatensis, the presence of alternative prey items in olive groves can originate a decrease of the consumption of B. oleae pupae due to switching to more energetic prey items and, consequently, to higher levels of satiation given by that prey (Murdoch, Reference Murdoch1969; Murdoch & Oaten, Reference Murdoch and Oaten1975), which can be considered a short-term negative impact on biological control of that pest (Settle et al., Reference Settle, Ariawan, Astuti, Cahyana, Hakim, Hindayana and Lestari1996).
On the other hand, the presence of alternative prey items and switching behavior can be seen as positive factors contributing to biological control by increasing the abundance of the predator when pest levels in the agroecosystem are low (Settle et al., Reference Settle, Ariawan, Astuti, Cahyana, Hakim, Hindayana and Lestari1996; Symondson et al., Reference Symondson, Glen, Ives, Langdon and Wiltshire2002). Thus, the ability for consuming other prey items can be more advantageous for P. globosus that will have conditions to reach high populations. Moreover, in olive groves, both carabid beetle species have peaks of activity in autumn, coinciding with the peak of abundance of B. oleae pupae on soil. Thus, in this period, both P. globosus and C. granatensis can significantly contribute to reduce pest levels, the former because the prey is present in high proportion and the latter because it prefers this prey. Further studies (e.g., semi-field assays or PCR-based gut content analysis for tracking B. oleae predation) will need to be done in order to confirm the contribution of both species as biocontrol agents of B. oleae. Moreover, the development of B. oleae biocontrol strategies should take into account the conservation or enhancement of these species of carabid beetles in olive groves. Therefore, management practices such as tillage or herbicide application should be avoided whereas non-crop habitats (plants and stones) may be established in the agroecosystem since they provide shelter and alternative food resources for carabid beetles.
Acknowledgements
This study was financially supported by FEDER Funds throughout Programa Operacional Factores de Competitividade – COMPETE and National Funds throughout FCT – Fundação para a Ciência e Tecnologia, within the project EXCL/AGR-PRO/0591/2012: Olive crop protection in sustainable production under global climatic changes: linking ecological infrastructures to ecosystem functions.
APPENDIX
Statistics of comparisons between carabid beetle species for consumed ratios, Manly's preference index, number of pupae consumed, percentage of biomass consumed and total biomass consumed for each offered ratio of Bactrocera oleae pupae.
Comparison between each offered ratio combination of Bactrocera oleae for consumed ratios, Manly's preference index, number of pupae consumed, percentage of biomass consumed and total biomass consumed. Within each parameter (first column), P-values for t-tests are provided on the upper side of diagonals for Calathus granatensis and the lower side corresponds to Pterostichus globosus.
