Introduction
In biological control programmes, there is a hypothesis that releasing multiple parasitoid species that attack different host stages would improve pest control because a later acting species would attack the fraction of host population that escaped from early acting species (Knipling, Reference Knipling, Ridgway and Vinson1977; Sivinski, Reference Sivinski, McPheron and Steck1996). However, both mathematical models and empirical evidence indicate that those parasitoid species that attack earlier stages of a common host exhibit competitive advantages over those parasitoids that parasitize later host stages, reducing the effectiveness of later acting species (Pedersen & Mills, Reference Pedersen and Mills2004; Wang & Messing, Reference Wang and Messing2004). Contrasting results indicate that a superior competitor can take advantage of parasitized host and feed on immatures of both the host and of the early acting parasitoid (Briggs, Reference Briggs1993). Furthermore, interaction between species could be altered by the coevolutionary history of species. Thus, when native and introduced species interact, there is no recognition between them, and defensive responses are not stimulated eliminating the advantage of an early acting species (Wang et al., Reference Wang, Bokonon-Ganta and Messing2008; Harvey et al., Reference Harvey, Poelman and Tanaka2013).
A low host availability may transform a stable coexistence of two parasitoid species into a competence by increasing superparasitism (i.e., parasitism of a host by more than one parasitic individual of the same species) and hyperparasitism (i.e., the capability of a parasitoid to attack another parasitoid on or within the natural host), consequently interfering in the control of target pests (Godfray, Reference Godfray1994; Harvey et al., Reference Harvey, Poelman and Tanaka2013; Xu et al., Reference Xu, Yang and Wan2013). There are two types of hyperparasitoids: (a) obligatory hyperparasitoids, which can develop only on or within a primary parasitoid, and (b) facultative hyperparasitoids, which can develop on an herbivore or in a parasitoid (Sullivan, Reference Sullivan1987; Sullivan & Völkl, Reference Sullivan and Völkl1999). The latter trait is considered detrimental in species used as natural enemies in biological pest-control programmes (e.g., Pérez-Lachaud et al., Reference Pérez-Lachaud, Batchelor and Hardy2004; Wang & Messing, Reference Wang and Messing2004).
Some evidence indicates that there is no cumulative insect pest control when multiple species of parasitoids are released, suggesting possible negative interactions between natural enemies (Briggs, Reference Briggs1993; Denoth et al., Reference Denoth, Frid and Myers2002). Conversely, other studies indicate that the addition of two or more biocontrol agents may substantially increase host mortality compared with single releases (May & Hassell, Reference May and Hassell1981; Stiling & Cornelissen, Reference Stiling and Cornelissen2005; Bader et al., Reference Bader, Heinz, Wharton and Bográn2006). The controversy of using single or multiple species of parasitoids to control an insect pest goes further when using augmentative releases that favour the probability of interactions between the parasitoids. Augmentative releases of braconids attacking larvae have been successfully used to control pest tephritid flies (Wong et al., Reference Wong, Ramadan, McInnis, Mochizuki, Nishimoto and Herr1991; Sivinski et al., Reference Sivinski, Calkins, Baranowski, Harris, Brambila, Diaz, Burns, Holler and Dodson1996; Montoya et al., Reference Montoya, Liedo, Benrey, Cancino, Barrera, Sivinski and Aluja2000a). However, when larvae develop in large fruits that function as refuges, they are less likely to be parasitized by these guild of parasitoids (Sivinski et al., Reference Sivinski, Smittie and Burns1991, Reference Sivinski, Aluja and López1997; Hawkins, Reference Hawkins1992; Montoya et al., Reference Montoya, Cancino, Zenil, Santiago, Gutierrez, Vreysen, Robinson and Hendrichs2007). Therefore, the control reached with larval parasitoids can be reinforced adding pupal parasitoids that attack pupal host that escaped from larval parasitism, since larvae pupate in the soil, they are no longer protected by the host fruit (Cancino et al., Reference Cancino, Liedo, Ruiz, López, Montoya, Barrera, Sivinski and Aluja2012). Thus, understanding the effect of competition and multiparasitism is relevant for augmentative release decisions to minimize undesirable interactions.
Diachasmimorpha longicaudata (Ashmead) (Hymenoptera: Braconidae), a solitary larval endoparasitoid is one of the natural enemies most frequently used to control fruit flies. This parasitoid originated in the Indo-Australian region and was introduced to Mexico during the 1950s (Jiménez y Jiménez, Reference Jiménez y Jiménez1956). In Mexico, D. longicaudata has been reported to parasitize various species of the genus Anastrepha (Aluja et al., Reference Aluja, Guillen, Liedo, Cabrera, Ríos, de la Rosa, Celedonio and Mota1990). Because it can be easily reared in the laboratory (Wong & Ramadan, Reference Wong and Ramadan1987; Cancino & Yoc, Reference Cancino, Yoc, Nicoli, Benuzzi and Leppla1993), D. longicaudata is an ideal candidate for pest-control programmes using augmentative releases (Sivinski et al., Reference Sivinski, Calkins, Baranowski, Harris, Brambila, Diaz, Burns, Holler and Dodson1996; Montoya et al., Reference Montoya, Liedo, Benrey, Cancino, Barrera, Sivinski and Aluja2000a, Reference Montoya, Cancino, Zenil, Santiago, Gutierrez, Vreysen, Robinson and Hendrichs2007). Coptera haywardi (Ogloblin) (Hymenoptera: Diapriidae), a native pupal parasitoid, is a good candidate for use in biological pest control because it is the only parasitoid reported to attack Anastrepha spp. pupae from southern Mexico to northern Argentina (Sivinski et al., Reference Sivinski, Vulinec, Menezes and Aluja1998; López et al., Reference López, Aluja and Sivinski1999; Ovruski et al., Reference Ovruski, Aluja, Sivinski and Wharton2000). Additionally, this parasitoid exhibits great discrimination ability against young (3–5-day-old) Anastrepha ludens (Loew) pupae previously parasitized by D. longicaudata (Cancino et al., Reference Cancino, Liedo, Ruiz, López, Montoya, Barrera, Sivinski and Aluja2012). However, some preliminary observations suggest that under specific conditions, C. haywardi can also attack pupae already parasitized by D. longicaudata using hyperparasitism as a strategy (Guzmán-Salinas & Montoya, Reference Guzmán-Salinas and Montoya2008).
The objectives of the present study were to determine the level of hyperparasitism by C. haywardi on the primary parasitoid D. longicaudata through: (1) studying the patterns of oviposition and emergence of C. haywardi in A. ludens pupae of different age and previously parasitized by D. longicaudata, (2) examining the emergence of the two parasitoid species when variable proportions of parasitized and unparasitized pupae are exposed to different numbers of females of C. haywardi, and (3) to analyse some parameters of fitness of C. haywardi adults emerging as hyperparasitoids. The results can be used to infer the feasibility of using simultaneously both parasitoid species in augmentative biological pest-control projects.
Materials and methods
Biological material and workplace
The experiments were carried out in the Biological Control Laboratory within the Methods Development Unit of the Moscafrut Program (SAGARPA-IICA), located at Metapa de Dominguez, Chiapas, Mexico. The bioassays were conducted using 8-day-old A. ludens larvae and 5–7-day-old D. longicaudata females provided by Moscafrut Facility where these species are mass reared (Cancino et al., Reference Cancino, Ruiz, López, Moreno, Montoya, Toledo and Hernández2010; Domínguez et al., Reference Domínguez, Artiaga-López, Solís, Hernández, Montoya, Toledo and Hernández2010). Seven-day-old C. haywardi females were obtained from the rearing colony of the Biological Control Laboratory. The bioassays were conducted at 22 ± 2°C and 75 ± 5% relative humidity.
Exposure of pupae of different ages to C. haywardi under non-choice conditions
Anastrepha ludens unparasitized or parasitized pupae by D. longicaudata were obtained from mass rearing facilities. Parasitized pupae present scars left by the ovipositor of D. longicaudata (Montoya et al., Reference Montoya, Liedo, Benrey, Barrera, Cancino and Aluja2000b). Pupae from 1–11-day old were separately exposed to C. haywardi females. Eight groups of 50 pupae of every age and condition studied (i.e., parasitized/unparasitized) were exposed during 48 h to the attack of four C. haywardi females of 7-day old and sexually mature. Twenty-four hours before the test, C. haywardi females received oviposition experience being exposed to 3-day-old A. ludens pupae. The experiment was conducted in plastic trays (26.5 × 16.5 × 7 cm) filled with moist vermiculite to simulate soil. The trays were fitted with lids with an 18 × 6.5 cm window covered with tulle mesh to prevent parasitoid escape while providing ventilation.
The following parameters were measured: (1) percentage of adult emergence of D. longicaudata and C. haywardi, (2) number of C. haywardi oviposition scars per pupa (determined on 10% of the attacked pupae) and (3) the number of immature C. haywardi parasitoids per pupa (recorded after dissecting the pupae). The experiments were repeated eight times. Anastrepha ludens and D. longicaudata pupae that were not exposed to C. haywardi were used as controls.
Exposure of parasitized and unparasitized A. ludens pupae to female C. haywardi under choice conditions
This experiment tested the effect of the simultaneous presence of pupae that had or had not been parasitized by D. longicaudata during their larval stage, on the performance of different numbers of C. haywardi females. Based on the results of the previous experiment on pupal age, where we observed when hyperparasitism of C. haywardi begins (see fig. 1), we used 7-day-old A. ludens pupae previously parasitized by D. longicaudata, and unparasitized 3-day-old A. ludens pupae were used as a control (Cancino et al., Reference Cancino, Liedo, Ruiz, López, Montoya, Barrera, Sivinski and Aluja2012). The proportions of previously parasitized and unparasitized pupae were: (1) 100% unparasitized A. ludens pupae (n = 50) (control), (2) 80% unparasitized A. ludens pupae and 20% pupae previously parasitized by D. longicaudata, (3) 60% unparasitized A. ludens pupae and 40% parasitized pupae, (4) 40% unparasitized A. ludens pupae and 60% parasitized pupae, (5) 20% unparasitized A. ludens pupae and 80% parasitized pupae, and (6) 100% A. ludens pupae previously parasitized by D. longicaudata. Each group was exposed to two, four, six or eight 7-day-old C. haywardi females with oviposition experience.
(a) Number of scars (mean ± SE) on pupae after being hyperparasitized by Coptera haywardi. (b) Number of immatures (mean ± SE) on pupae after being hyperparasitized by C. haywardi.
Similar to the previous experiment, the pupae were subjected to attack by C. haywardi females in plastic trays (26.5 × 6.5 × 7 cm) filled with moist vermiculite to simulate soil. The following parameters were tested: (1) the number of C. haywardi oviposition scars on 10% of the pupae (five pupae in each replicate); (2) the number of immature C. haywardi per pupa (obtained after dissection); and (3) the percentage of adult emergence of A. ludens, D. longicaudata and C. haywardi. Eight replicates were used, including control groups of unparasitized and previously parasitized A. ludens pupa that were not exposed to C. haywardi.
Fitness parameters of C. haywardi adults emerging from previously parasitized pupae
The body size, average fertility and longevity were determined in 17 C. haywardi pairs (males and females) that emerged from pupae previously parasitized by D. longicaudata. Once they reached sexual maturity (7-day), 20 3-day-old A. ludens pupae were offered daily to each female until death (Cancino et al., Reference Cancino, Liedo, Ruiz, López, Montoya, Barrera, Sivinski and Aluja2012). The following parameters were tested: (1) fertility, expressed as the number of male and female descendants per female per day, and (2) longevity, expressed as the number of days that the males and females survived with water and food.
The body size was determined measuring the length of antennae, wings, femur and the width of the abdomen. All measurements were taken using a stereomicroscope NIKON, SMZ745T with camera ProgRes® CT3 equipped with software ProgRes® CapturePro 2.9.0.1. Male and female C. haywardi adults that emerged from unparasitized A. ludens pupae by D. longicaudata were used as controls.
Statistical analysis
For binary data of parasitism, a generalized linear model (GLM) with binomial distribution and logit link function were used. Small integer counts recorded for two of the response variables (i.e., number of scars and number of immatures) were also analysed using GLM with Poisson errors, a log-link function and type III significance tests (Crawley, Reference Crawley1993; Agresti, Reference Agresti1996). Contrasts were used to test for differences in levels within a variable. For the continuous response variables, such as body size, t test was used. Longevity was analysed through the non-parametric log-rank test (Francis et al., Reference Francis, Green and Payne1993). Analyses were performed using the software JMP v.7, SAS Institute, Cary, NC, USA and Statgraphics Centurion XV (Statgraphics, 2008).
Results
Pupal age
The analysis of the patterns of oviposition indicates that the number of scars vary depending on the condition of exposed pupae, parasitized/unparasitized, and according to the age of pupae unparasitized young pupae received higher numbers of ovipositions (χ2 = 195.2, df = 10, P < 0.0001; χ2 = 31.5, df = 10, P = 0.0005, respectively). The interaction between factors was significant indicating that females presented a different pattern of oviposition in parasitized and unparasitized pupae according to the pupae age (χ2 = 81.1, df = 10, P < 0.0001) (fig. 1a).
The number of immatures developed inside the pupae was also affected by the pupal condition, parasitized/unparasitized, exposed to C. haywardi and for the age of pupae. This result correlates with the number of scars and followed a similar pattern (χ2 = 186.3, df = 1, P < 0.0001; χ2 = 51.7, df = 10, P < 0.0001, respectively). The interaction between factors was also significant (χ2 = 109.1, df = 10, P < 0.0001) (fig. 1b).
The number of C. haywardi adults emerging from A. ludens pupae previously parasitized by D. longicaudata was significantly lower than the number of C. haywardi adults emerging from unparasitized A. ludens pupae (χ2 = 1247.1, df = 1, P < 0.0001). The effects of C. haywardi attack and of pupal age on the emergence of D. longicaudata were also significant (χ2 = 73.9, df = 10, P < 0.0001). Again, the interaction term was significant, indicating that pupal age also influenced the attack (χ2 = 123.1, df = 10, P < 0.0001) (fig. 2).
Emergence (mean proportion) of Coptera haywardi from Anastrepha ludens pupae of different ages parasitized or not by Diachasmimorpha longicaudata.
These data indicate that C. haywardi parasitized mostly young unparasitized hosts. In parasitized pupae by D. longicaudata, the interactions between the immature of the two species of parasitoids of the two species changed over time. In young pupae (1–4 days), young parasitoids of both species were observed feeding simultaneously within the puparium. Older C. haywardi larvae (6–11 days) fed on third-instar larvae, pre-pupae and pupae of D. longicaudata (table 1). Developmental stages of immature D. longicaudata parasitoids were identified following methods of Carabajal-Paladino et al. (Reference Carabajal-Paladino, Papeschi and Cladera2010).
Interaction between immatures of Coptera haywardi and Diachasmimorpha longicaudata within Anastrepha ludens pupae parasitized as larvae by D. longicaudata and hyperparasitized by C. haywardi at different ages (dissected 72 h after hyperparasitized).
1 The developmental stages of immature D. longicaudata parasitoids were identified according to the descriptions of Carabajal-Paladino (Reference Carabajal-Paladino, Papeschi and Cladera2010).
Proportions of pupae and female density
Emergence of C. haywardi was affected by two factors: the proportion of parasitized/unparasitized pupae exposed and female density. Emergence was higher when the proportion of parasitized pupae was lower and when the density of females increased (χ2 = 717.4, df = 5, P < 0.0001; χ2 = 25.9, df = 3, P < 0.0001, respectively). The interaction between these factors was significant (χ2 = 74.0, df = 15, P < 0.0001). Additionally, emergence of D. longicaudata was influenced by the proportion of parasitized pupae and by the density of females (χ2 = 642.0, df = 4, P < 0.0001; χ2 = 41.9, df = 3, P < 0.0001, respectively). The interaction between these two factors was significant (χ2 = 116.1, df = 12, P < 0.0001) (fig. 3). Furthermore, the parasitism by C. haywardi reduces the emergence of D. longicaudata (Wilcoxon test of the emergence proportion of the exposed and unexposed pupa of D. longicaudata, Z = −10.76, df = 156, P < 0.0001) (mean ± SE: 0.29 ± 0.0119 exposed; 0.36 ± 0.015 unexposed).
Emergence (mean proportion) of the three species: Anastrepha ludens, Coptera haywardi and Diachasmimorpha longicaudata from Anastrepha ludens pupae (unparasitized and parasitized by D. longicaudata (UP/P) available at different proportions to different densities of females of C. haywardi).
Fitness parameters
The survival of C. haywardi females emerging from unparasitized A. ludens pupae did not differ significantly from that of females emerging from pupae previously parasitized by D. longicaudata (log-rank χ2 = 0.06, df = 1, P = 0.805) (fig. 4a). However, the survival of hyperparasitic males emerging from pupae previously attacked by D. longicaudata (fig. 4b) was significantly higher than that of males emerging from unparasitized A. ludens pupae (log-rank χ2 = 14.81, df = 1, P < 0.0001).
Percentage of survival of females (a) and males (b) of Coptera haywardi emerging from Anastrepha ludens pupae previously parasitized or not by Diachasmimorpha longicaudata.
The fertility of the two groups of females did not differ significantly (t = 0.732, df = 52, P = 0.467). Females produced an average of 0.25–6.1 offspring per day of exposure, and their fertility was higher during the first days of exposure. There were non-significant differences on none body sizes measures between females (t test, α = 0.05). In the case of males, wings and abdomen length values were higher in males emerging from hyperparasitic conditions (t test, α = 0.05) (table 2).
Size attributes in mm (mean ± SE) of Coptera haywardi males and females of emerging from hyperparasitized and non-hyperparasitized conditions.
t = t test; * = significant differences between columns; NS = not significant, α = 0.05.
Discussion
We found that C. haywardi is a primary parasitoid of A. ludens that compete with early D. longicaudata immatures in parasitized A. ludens pupae; however, it is also able to hyperparasitize on advanced-immature stages of D. longicaudata (table 1). This facultative hyperparasitism strategy was observed only under conditions of high host competition and under the presence of advanced immature stage of D. longicaudata. Furthermore, those C. haywardi adults that emerged from hyperparasitized hosts presented fitness parameters very similar to those emerged from unparasitized ones.
Our results indicate that when C. haywardi attacks young pupae previously parasitized by D. longicaudata, interspecific competition takes place, and under this condition, D. longicaudata has advantage over immatures of C. haywardi. Conversely, when C. haywardi attacked 6-day-old or older fly pupae previously parasitized by D. longicaudata, it is susceptible to be attacked by first- and second-instar larvae of C. haywardi that becomes then a hyperparasitoid. It is possible that advanced developmental stages (third-instar larvae, prepupae and pupae) of D. longicaudata had consumed most of the host, forcing C. haywardi larvae to feed on the immatures of D. longicaudata. Thus, C. haywardi can be considered an indirect facultative hyperparasitoid because it initially attacks a phytophagous host, but it is able to hyperparasitize depending on the age of D. longicaudata larvae (Sullivan & Völkl, Reference Sullivan and Völkl1999).
Facultative hyperparasitoids can occupy different trophic levels because they can develop as primary parasitoids or as parasitoids of other parasitoid species attacking the same host (Powell et al., Reference Powell, Walton, Jervis, Jervis and Kidd1996). Parasitoid strategies exhibit a continuum, where the two extremes are represented by obligate primary parasitism and obligate secondary parasitism (hyperparasitism) (Ehler, Reference Ehler, Mackauer, Ehler and Roland1990). A facultative secondary parasite occupies an intermediate position, which is advantageous when unparasitized hosts are scarce. Some facultative parasitoids develop as koinobionts and can survive in two hosts that are not taxonomically related (Godfray, Reference Godfray1994). This strategy is possible when the different primary parasitoids feed on the same host. Therefore, the host and the primary parasitoid exhibit physiological similarities as shown in the present study. Pupal parasitoids may use facultative hyperparasitism because of interspecific competition with larval parasitoids (Grandgirard et al., Reference Grandgirard, Poinsot, Krespi, Nénon and Cortesero2002). It has been proposed that early attack of larval parasitoids represents a competitive advantage. However, reports of attacks by Pachycrepoideus dubius (Ashmead) (Hymenoptera: Pteromalidae) on Delila radicum (L) pupae previously parasitized by the larval parasitoid Trybliographa rape (Westwood) (Hymenoptera: Figitidae) indicate that hyperparasitism can be an advantageous strategy for pupal parasitoids (Grandgirard et al., Reference Grandgirard, Poinsot, Krespi, Nénon and Cortesero2002).
Adult C. haywardi emerging from hyperparasitized pupae were similar in those fitness parameters measured to those emerging from unparasitized A. ludens pupae, since body size, longevity and fertility of the two types of adults did not differ significantly. Similar results were observed in pupal hyperparasitoid species, such as Pachrycrepoideus vindemniaea (Rondani) (Hymenoptera: Pteromalidae) (Wang & Messing, Reference Wang and Messing2004), which can successfully develop on four different braconid species: D. longicaudata, Fopius arisanus (Sonan), Diachasmimorpha kaussii (Viereck) and Psyttalia concolor (Szépligeti). Pachrycrepoideus vindemniaea has also been reported to attack pupae previously parasitized by other parasitoids of fruit flies, such as Diachasmimorpha tyroni (Cameron), Psyttalia humilis (Silvestri) (Hymenoptera: Braconidae) and Dirhinus giffardii (Silvestri) (Hymenoptera: Chalcidicae), but detailed information concerning these interactions lacks (Wang & Messing, Reference Wang and Messing2004).
In the present study, the reduction in various parameters of C. haywardi analysed (i.e., the number of oviposition scars and the number of immature parasitoids and adult emergence) indicated that A. ludens pupae previously attacked by D. longicaudata was not a first option for oviposition. However, when all of the available pupae had been previously parasitized by D. longicaudata, C. haywardi females preferentially attacked upon 6–7-day-old pupae. Adult emergence was higher in these hosts category than in the others (6 and 6.25%, respectively) (fig. 1a). A large decrease in the emergence of D. longicaudata after being attacked by C. haywardi in experiments using Anastrepha suspensa (Loew) as the primary host has been reported previously (Sivinski et al., Reference Sivinski, Vulinec, Menezes and Aluja1998).
When two species of parasitoids use different stages of the same host, their coexistence is possible because their life histories are different. Commonly, the larval parasitoids would have more hosts to oviposit and thus, their fecundity will be higher than in pupal parasitoids that would confront lower host availability (Price, Reference Price1972; Bonsall et al., Reference Bonsall, Jansen and Hassell2004). Furthermore, the generalist–specialist continuum strategies also play important role that may permit the coexistence of two species of parasitoids since generalists would possess more available hosts than the specialists (Price, Reference Price1972; Bonsall et al., Reference Bonsall, Jansen and Hassell2004). However, the coexistence could be broken when the availability of hosts fall and the probability of finding a host gets reduced. For example, switching from primary parasitism to a hyperparasitism in two coexisting species may be triggered by host availability and the dispersal abilities of females. For example, in two solitary secondary hyperparasitoids, Lysibia nana (Gravenhorst) and the wingless Gelis agilis (Fabricius), which concur in cocoons of a primary parasitoid, Cotesia glomerata L., G. agilis is the winner when competition for a host takes place (Harvey et al., Reference Harvey, Pashalidou, Soler and Bezemer2011). Thus, the wingless condition reduces the searching capacity and hence the availability of hosts for G. agilis. In our study however, both species are highly mobile; thus, the low hyperparasitism by C. haywardi females could be explained by their high host choosiness behaviour (Cancino et al., Reference Cancino, Liedo, Ruiz, López, Montoya, Barrera, Sivinski and Aluja2012).
Our findings suggest that hyperparasitism of C. haywardi when used as biocontrol agent could be relevant if the proportion of parasitized pupa by the primary parasitoids is higher, which can be influenced by the density of the larval parasitoid, but also by the host fruit size that contained the larvae initially attacked by this species. Large fruit can protect the host larvae and facilitate their escape from their natural enemies, bring about large numbers of unparasitized pupae available for the pupal parasitoid (Sivinski et al., Reference Sivinski, Vulinec, Menezes and Aluja1998; Montoya et al., Reference Montoya, Cancino, Zenil, Santiago, Gutierrez, Vreysen, Robinson and Hendrichs2007, Reference Montoya, Ayala, López, Cancino, Cabrera, Cruz, Martínez, Figueroa and Liedo2016). The hyperparasitism by C. haywardi was neglected when the proportion of unparasitized hosts was high, confirming that this species possess a high discrimination ability (Cancino et al., Reference Cancino, Liedo, Ruiz, López, Montoya, Barrera, Sivinski and Aluja2012).
The impact of facultative hyperparasitoids and their interactions on biological pest-control programmes are complex and poorly understood but are generally thought to be detrimental (Rosenheim, Reference Rosenheim1998; White et al., Reference White, Bernal, Gonzáles and Triapitsyn1998; Brodeur, Reference Brodeur, Hochberg and Ives2000). Some studies have shown that facultative hyperparasitoids can prevent primary parasitoids from reaching their full potential as natural enemies, thus interfering with the success of biological pest-control programmes (Mills & Gutierrez, Reference Mills, Gutierrez, Hawkins and Cornell1999). Facultative hyperparasitoids can play an important role in biological pest control; however, their exact impacts are uncertain. When a female discovers a host that has been previously parasitized by a primary parasitoid, it may oviposit on or within the host; consequently, the offspring will develop as secondary parasitoids of the primary parasitoid (Ehler, Reference Ehler1979; Rosenheim et al., Reference Rosenheim, Kaya, Elher, Marois and Jaffee1995; Heinz & Nelson, Reference Heinz and Nelson1996). This phenomenon is classified as intra-guild predation rather than competition. Some authors consider this phenomenon to be common in parasitoid guilds (Hawkins, Reference Hawkins1992). However, the prevalence of this type of predation among parasitoids is low.
This study adds evidences of the complexity of interspecific interactions of parasitoids. There are intrinsic factors such as the development stage of early acting species that clearly reduce the window of susceptibility for being hyperparasitized. Extrinsic factors such as host availability may trigger hyperparasitism in the later acting species; hence, the probability of negative interactions between these two species is scarce. Thus, C. haywardi can be considered as a low-risk facultative hyperparasitoid of D. longicaudata that could be used as a complementary biocontrol agent in augmentative biological-control programmes against Anastrepha fruit flies.
Acknowledgements
The authors thank Jorge Cancino, Patricia López and technicians at the Biological Control Laboratory of the Programa Moscafrut SAGARPA-IICA for technical support during experiments. Gabriela Pérez-Lachaud (ECOSUR) reviewed an earlier version of this manuscript and Javier Valle-Mora (ECOSUR) provided statistical advice. The Programa Nacional Moscas de la Fruta DGSV-SENASICA-SAGARPA funded the study.
