Introduction

Host selection by parasitic wasps entails perceiving a wide range of cues, including visual, chemical and vibratory stimuli (Vinson 1998; Meyhöfer and Casas 1999). The success of the host selection process is closely related to the quality of the host. Parasitized hosts are often of a worse quality than healthy ones, and females of many parasitoid species are able to recognise and reject them through the identification of specific host-borne physical or chemical stimuli (Goubault et al. 2011). Many parasitoids have evolved an ability to discriminate between unparasitized and parasitized hosts, including Aphidiidae (Chow and Mackauer 1996), Eucoilidae (Varaldi et al. 2005), Eupelmidae (Darrouzet et al. 2007), Ichneumonidae (Ueno and Tanaka 1994; Zhang et al. 2010), Mymaridae (van Baaren et al. 1994; Santolamazza-Carbone et al. 2004), Pteromalidae (Goubault et al. 2004), Scelionidae (Rabinovich et al. 2000; Mahmoud and Lim 2008), Torymidae (Tepedino 1988) and Braconidae (Cloutier et al. 1984; Moore and Ridout 1987; Outreman et al. 2001; Wang et al. 2010).

Despite this discrimination ability, superparasitism—the laying of one egg (by solitary parasitoids) or a number of eggs (by gregarious parasitoids) in an already parasitized host—is common among several parasitoid species (van Alphen and Jervis 1996). Long considered the result of mistakes made by ovipositing females, superparasitism has now been recognised as adaptive in several contexts (van Alphen and Visser 1990; Dorn and Beckage 2007). It has been shown that the ability of female wasps to recognise parasitized hosts does not necessarily mean that superparasitism is avoided (Gu et al. 2003; Dorn and Beckage 2007). Theoretical models predict that parasitoids rely on the trade-off between the benefits and costs of oviposition when they decide whether to avoid or accept an already parasitized host (Gu et al. 2003; Godfray 1994; Plantegenest et al. 2004). Among solitary parasitoids, superparasitism can affect the size of the wasp’s offspring (Harvey et al. 1993; Mackauer and Chau 2001), induce a longer developmental time (Vinson and Sroka 1978; Harvey et al. 1993; Mayhew and van Alphen 1999) and/or a decrease survivorship of immature stages (Vinson and Sroka 1978). Although in some cases superparasitism seems to be maladaptive in terms of individual offspring fitness, this is advantageous in several contexts, especially when unparasitized hosts are rare in a habitat patch (e.g. tephritid larvae in fruit tree groves) and parasitoid females have a high egg load (Weisser and Houston 1993; Yamada and Miyamoto 1998; Gu et al. 2003). Even if there has been extensive research on the mechanism of host discrimination and the role of superparasitism in many parasitoid species, only little evidence is still available on tephritid parasitoids. This highlights that more knowledge is needed to gain insight into the evolution of ovipositional decisions in these parasitic wasps.

Psyttalia concolor (Szépligeti) (Hymenoptera, Braconidae, Opiinae) is a koinobiont larval-pupal endoparasitoid of many Tephritidae. It is able to attack at least 14 tephritid species on different wild and cultivated plants, including pests of great economic importance, such as the Mediterranean fruit fly, Ceratitis capitata (Wiedemann) (Diptera: Tephritidae), and the olive fruit fly, Bactrocera oleae (Rossi) (Diptera: Tephritidae) (Wharton 1997; Benelli and Canale 2012). Psyttalia concolor has been used in Italy and other Mediterranean climate areas for the biological control of B. oleae by inundative and propagative releases with limited results (Daane and Johnson 2010). It has also been released in Californian olive groves as part of classical biological control programs (Wang et al. 2011).

The P. concolor female inserts the ovipositor and probes a potential host before the acceptance or rejection of oviposition (Canale 1998; Canale and Benelli 2012). Such behaviour suggests the existence of a host discrimination ability in this parasitoid. Psyttalia concolor females can also superparasitize their hosts; the parasitoid/host ratio, as well as the duration of exposure of the host to the parasitoid, is important to determine the magnitude of superparasitism rates (Canale 1998; Raspi and Canale 2000). However, it still remains unclear whether females are able to actively discriminate between healthy tephritid hosts and already parasitized ones, as well as parasitized versus superparasitized larvae. Information obtained on the host discrimination ability of P. concolor lead to a better understanding of its host location behaviour to enhance its efficiency in biological control programs and to improve its mass-rearing technique. In this study, two-choice bioassays were conducted to evaluate (i) the preferences of P. concolor females for healthy or single-parasitized/superparasitized C. capitata larvae and (ii) the host discrimination ability of females among medfly larvae superparasitized by the same wasp (self-superparasitism) or by a conspecific one (conspecific-superparasitism).

Materials and methods

Parasitoid and host rearing

The parasitoid P. concolor and the host C. capitata were reared as described by Benelli et al. (2012a, b). The C. capitata production unit consisted of cylindrical PVC cages, each containing about 2,000 adults (sex-ratio 1:1). Adults were fed on a dry diet of sugar and yeast extract (10:1 w:w). Eggs were collected every 2 days and distributed into plastic bowls (50 × 15 × 2 cm), each containing 0.5 kg of artificial culture medium. The cages for P. concolor breeding were made of transparent Plexiglas tubes (40-cm diameter, 50-cm length) into which 1,500 adults were introduced (male:female ratio 0.3:0.5). Nylon mesh bags, each containing up to 600–800 third instar C. capitata larvae, were placed in the cage for the parasitisation phase (15 min).

After emergence from the host and until the test, parasitoids were stored in cylindrical Plexiglas cages (40-cm diameter, 50-cm height) at a density of 100 specimens (males and females, sex ratio 1:2) per cage [21 ± 1 °C, 48 ± 10 % relative humidity and 16:8 (L:D) photoperiod]. A semisolid diet (honey mixed with pollen) and water was offered to the parasitoids. Eight-day-old third instar C. capitata larvae were used as host in the experiments.

Host discrimination experiments

Psyttalia concolor females were tested for their ability to discriminate between healthy, parasitized or superparasitized C. capitata (MF, hereafter) larvae. The following bioassays were conducted in two-choice conditions:

  1. (i)

    a healthy MF larva versus a MF larva self-parasitized through one oviposition event;

  2. (ii)

    a healthy MF larva versus a MF larva self-superparasitized through two oviposition events;

  3. (iii)

    a MF larva self-parasitized through one oviposition event versus a MF larva self-parasitized through two oviposition events;

  4. (iv)

    a MF larva self-parasitized through one oviposition event versus a MF larva parasitized once by a conspecific P. concolor female;

  5. (v)

    a MF larva self-parasitized through two oviposition events versus a MF larva parasitized twice by a conspecific P. concolor female.

To obtain C. capitata larvae self-parasitized once by a P. concolor female, each host was subjected to a single parasitisation act with a mean duration of at least 20 s by the same female used in the testing phase. In detail, the host larva was placed in an oviposition dish (PVC cylinder, 8-cm diameter; 1-cm depth) inside a hole (0.8-cm diameter; 0.5-cm depth) created in the centre of the dish. The top of the hole was covered with an organdy screen which was tightly fixed by inert glue around the rim of the hole (Canale and Benelli 2012). A naïve P. concolor female (i.e. a wasp without any oviposition experience) was released directly onto the rim of the screen cover of the oviposition dish using a cylindrical glass tube (1-cm diameter). Subsequently, the dish was covered using a cylindrical transparent Plexiglas lid (6-cm diameter; 1-cm height), creating a still-air arena, and the female’s behaviour was directly tracked until oviposition occurred. Hosts superparasitized with two eggs were produced through two subsequent parasitisation events (each lasting at least 20 s) by the same female used in the testing phase. Hosts parasitized once or twice by a conspecific female were obtained following exactly the procedure reported above, but using different females over the testing phase. In all bioassays, self- and conspecific-parasitisation events occurred in a clean oviposition unit about 20 min before the testing phase.

During the testing phase, a similar oviposition unit as described above was used with the only difference that the oviposition unit in these two-choice bioassays had two distinct holes (Ø 0.8 cm, 1-cm depth) in the centre of the dish. The two holes were spaced at 1 cm apart. Thirty replicates were carried out for each two-choice bioassay. For each replicate, (a) the latency time (i.e. time between the wasp’s release and initiating the search), (b) the wasp’s first choice of one of the two given hosts, and the following (c) oviposition or (d) rejection time were measured.

A host location was recorded as successful when the searching female, stationary on the patch where the host was placed, raised its abdomen and drove its ovipositor into the host for at least 20 s, which is when a wasp usually lays an egg. In contrast, ovipositor insertions for 5–10 s were considered as host location followed by host rejection (Canale and Benelli 2012). In order to make a definitive assessment of the presence/absence of parasitoid eggs, at the end of the experimental period, all the located larvae were preserved in 70 % ethanol, subsequently dissected in Rüngen’s solution and inspected. A replicate was considered finished when the wasp had successfully oviposited or rejected the host, or when 5 min had elapsed without a successful host location. Each female was tested only once.

Between each replicate, the odour-cleaning procedure was as follows: the oviposition unit was first washed for about 30 s with warm water at 35–40 °C, then cleaned in a water bath with mild soap for about 5 min, rinsed with hot water for about 30 s and finally rinsed with distilled water at room temperature (Carpita et al. 2012). All trials were conducted in a 12-m2 room, illuminated with daylight fluorescent tubes placed so as to guarantee that the intensity of light was as even as possible. The temperature was set at 23 ± 1 °C and the relative humidity at 50 ± 5 %.

Data analysis

Measured variables (i.e. latency times and probing durations) were analysed through a general linear model with one fixed factor (i.e. the treatment) using JMP 7®. Nominal variables (i.e. the female first choices and the number of successful ovipositions in each two-choice bioassay) were analysed through Chi-square tests with Yates correction (Sprinthall 2003).

Results

When P. concolor females were simultaneously exposed to healthy and self-parasitized larvae, longer latencies were recorded in females which preferred unparasitized hosts (F = 7.474; df = 29; P < 0.05) (Table 1). Female wasps showed longer latencies before choosing a self-parasitized larva over a host parasitized by a conspecific female (F = 5.724; df = 29; P < 0.05). Similarly, females showed longer latencies before choosing a self-superparasitized larva over a host superparasitized by a conspecific female (F = 6.993; df = 29; P < 0.05).

Table 1 Latency time spent by Psyttalia concolor females on each combination of cues from healthy and already parasitized Ceratitis capitata larvae (treatments A and B) in two-choice bioassays conducted in a still-air arena
Full size table

Owing to host location, no preferences were registered by P. concolor naïve females for larvae which had self-parasitized only once, compared to healthy ones. Females significantly chose more unparasitized C. capitata larvae than hosts superparasitized by the same wasp (χ 2 = 7.50, df = 1, P < 0.05) (Fig. 1a). In addition, they significantly preferred hosts self-parasitized once rather than the ones self-parasitized twice (χ 2 = 7.50, df = 1, P < 0.05). Parasitoids did not discriminate between hosts self-parasitized once and hosts conspecific-parasitized, nor were they able to distinguish larvae superparasitized by other females from self-superparasitized ones.

Fig. 1
figure 1

Number of first choices (a) and successful ovipositions (b) of Psyttalia concolor females in two-choice bioassays conducted in a still-air arena with different combinations of cues from healthy and already parasitized Ceratitis capitata larvae. Thirty wasps were tested for each two-choice bioassay. For each two-choice bioassay, an asterisk indicates a significant difference (χ2 test with Yates correction, P < 0.05). H healthy host, P1 host parasitized with one egg, P2 host parasitized with two eggs, self host previously parasitized by the same female, not self host previously parasitized by a conspecific female, n.s. not significant

Full size image

Female parasitoids successfully oviposit more in unparasitized larvae than hosts parasitized once or twice by the same wasp (χ 2 = 5.76, df = 1, P < 0.05; χ 2 = 16.00, df = 1, P < 0.05, respectively) (Fig. 1b). Females also preferred to oviposit in larvae self-parasitized once than in hosts self-parasitized twice (χ 2 = 4.27, df = 1, P < 0.05). Also, the larvae self-parasitized once by the same female were preferred in terms of ovipositions to hosts parasitized once by a conspecific female (χ 2 = 3.70, df = 1, P < 0.05). No differences were detected in female ovipositions on hosts parasitized twice by another female or twice by the same wasp.

Finally, in all trials, no differences were observed in host acceptance and rejection times (Tables 2, 3).

Table 2 Host acceptance time spent by Psyttalia concolor females on each combination of cues from healthy and already parasitized Ceratitis capitata larvae (treatments A and B) in two-choice bioassays conducted in a still-air arena
Full size table
Table 3 Host rejection time spent by Psyttalia concolor females on each combination of cues from healthy and already parasitized Ceratitis capitata larvae (treatments A and B) in two-choice bioassays conducted in a still-air arena
Full size table

Discussion

Among solitary braconids, intraspecific host discrimination is quite common (van Alphen and Visser 1990). For instance, females of Aphidius rhopalosiphi De Stefani Perez and Spathius agrili Yang discriminate between parasitized and healthy hosts, mainly through specific chemical cues (Outreman et al. 2001; Wang et al. 2010). Similarly, Aphidius nigripes Ashmead females are able to discriminate between parasitized and unparasitized potato aphids even though they do not consistently avoid superparasitism (Cloutier et al. 1984), while Chasmodon apterus Nees females avoid superparasitism when attacking Oscinella spp. larvae (Moore and Ridout 1987). In contrast, no host discrimination ability has been detected in other solitary braconids, such as Asobara tabida Nees (van Alphen and Nell 1981). Our results showed that P. concolor prefers to oviposit in a unparasitized C. capitata larva than in a self-parasitized one, and spend longer latencies on the former. Moreover, females prefer healthy medfly larvae over superparasitized hosts, as already observed for the braconids Cardiochiles nigriceps Viereck and Microplitis croceipes (Cresson) (Vinson and Guillot 1972). This host discrimination ability of P. concolor females may be due to the perception of variations in the physiological parameters of the host, such as differences in haemolymph composition (Thompson 1986). The sensory structures on the P. concolor ovipositor (Canale and Raspi 2000) may be implicated in the detection of parasitoid eggs within the host and of the host quality in general, as has been suggested for other parasitoid species (Fisher 1971; Greany et al. 1977; Goubault et al. 2011).

Psyttalia concolor females show an innate ability to discriminate between C. capitata larvae parasitized twice or only once, preferring the latter, as already observed in other solitary parasitoids (Bakker et al. 1989). This ability can help the P. concolor females to optimise their oviposition decisions by avoiding superparasitized hosts since they give a lower return in offspring number and quality than do unparasitized or singly parasitized hosts (van Alphen and Visser 1990; Canale 1998; González et al. 2007). However, P. concolor females do not avoid depositing eggs in hosts already parasitized once by themselves or conspecific members, despite their ability to distinguish between parasitized and unparasitized hosts. In P. concolor, superparasitism occurs even when unparasitized hosts are available (Zhang and Raspi 1999; Raspi and Canale 2000). The occurrence of moderate superparasitism seems to represent a conditional strategy, which increases the reproductive success in this species, as already highlighted for the braconid Diachasmimorpha longicaudata (Ashmead) (González et al. 2007). Indeed, on fully grown C. capitata larvae, two P. concolor eggs/host represents the optimal conditions to obtain the maximum quantity of offspring (Canale 1998), as noted for other braconid wasps (Kitano and Nakatsuji 1978; Gu et al. 2003). By contrast, second instar medfly larvae only require one oviposition event to achieve the highest success in P. concolor offspring emergence (Raspi and Canale 2000). Robust hosts, such as fully grown medfly larvae, can encapsulate the parasitoid eggs through a strong cellular immune response (Hegazi and Khafagi 2008). Therefore, superparasitism may help P. concolor females to overcome the physiological defence of the host since the virus-like particles contained in the poison glands are related to the P. concolor capacity to elude the haemocytic encapsulation mechanism of the host (Jacas et al. 1997).

Although the C. capitata larvae self-parasitized once were not visited more by P. concolor compared to conspecific-parasitized ones, we observed that female wasps performed more successful ovipositions on the former. Self-superparasitism for solitary parasitoids usually produces zero or negative fitness returns (van Alphen and Visser 1990; Darrouzet et al. 2007; Dorn and Beckage 2007). However, there are also some exceptions; for some species, self-superparasitism is recognised as an adaptive strategy to maximise offspring (Yamada and Miyamoto 1998; Yamada and Sugaura 2003). In P. concolor, self-superparasitism probably does not have any downsides because on fully grown medfly larvae two subsequent ovipositions help to increase the survival chances of the parasitoid’s offspring (Canale 1998). Similarly to our results, Campoletis chlorideae Uchida (Hymenoptera: Ichneumonidae) females have been shown not to hesitate in carrying out additional ovipositions on self-parasitized hosts (Zhang et al. 2010). Finally, in our study P. concolor females did not show any preference for hosts parasitized twice by another female or by the same wasp probably because the cues arising from self-superparasitized hosts and conspecific-superparasitized ones are quite similar. Previous researches conducted on different solitary and gregarious braconids (Outreman et al. 2001; Gu et al. 2003), as well as on trichogrammatids (van Dijken and Waage 1987), reported comparable results.

Overall, we believe that our findings could contribute to a better understanding of the host location behaviour of P. concolor and also to improve its mass-rearing technique through a rational handling of the main rearing parameters, such as the host/parasitoid ratio and the host exposure time. Indeed, the proper setting of these parameters allows to reduce the fraction of single-parasitized and heavily superparasitized larvae so enhancing the rearing in terms of parasitoid offspring. Further research is needed to determine whether natural P. concolor populations show similar ovipositional decisions, as well as to understand how previous ovipositional experience can affect the female’s preferences.