Classical biocontrol programs usually employ multiple species to control a single pest; however, the beneficial effects are not always additive due to competition between the introduced species. Knowledge of these potentially negative interactions is crucial when determining whether the introductions were successful and the extent to which they influence pest suppression. Here, we report the results of such competition between 2 exotic wasp species [Encarsia bimaculata Heraty & Polaszek and Eretmocerus sp. nr. emiratus (Hymenoptera: Aphelinidae)] introduced into Florida for control of the circumglobal super pest currently known as Middle East-Asia Minor 1, a cryptic species within the Bemisia tabaci complex (Hemiptera: Aleyrodidae). Levels of parasitism, progeny production, and host feeding were evaluated in the laboratory under various parasitoid combinations on 2 host plant species (collard and eggplant) that differed in leaf pilosity (i.e., number of trichomes on the abaxial leaf surface). Significant differences in parasitism were observed by treatment, but not host plant. Encarsia bimaculata produced fewer progeny when introduced before Er. sp. nr. emiratus on collard; however, this trend was reversed on eggplant. Eretmocerus sp. nr. emiratus produced less progeny in all combinations involving En. bimaculata on collard and when introduced before En. bimaculata on eggplant. Mortality caused by host feeding was atypical for the whiteflyparasitizing aphelinids (<5% in all treatments). Based on differences in parasitism and progeny production, no additional pest suppression would be gained by releasing En. bimaculata into an environment already under control by Er. sp. nr. emiratus.
Classical biological control programs typically involve the introduction of multiple exotic species for control of a single target pest (Denoth et al. 2002; Gould et al. 2008). Such introductions undoubtedly lead to competition between the introduced organisms for the shared resources vital to their survival and perpetuation. Competition for resources is commonplace in parasitoid Hymenoptera, many of which develop on a single host and display a high degree of niche overlap. Knowledge of the dynamics between competing parasitoid species, in relation to use of their shared host(s), has direct relevance to their utilization in biological control programs.
Whiteflies and their parasitoids are a model host-parasite system to evaluate the outcomes of such competition. Many whiteflies become pests after escaping their natural boundaries and establishing populations in new areas (Martin 1987). Such was the case following the invasion of a member of the Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) species complex in Florida, and later many other parts of the world, in the latter decades of the last century (Stansly & Naranjo 2010). These novel environments may not have a natural enemy cohort capable of controlling this whitefly species below the economic threshold, resulting in the need for the introduction of exotic natural enemy species.
Parasitoid wasps in the family Aphelinidae (Hymenoptera: Chalcidoidea) are one of the most successful groups of insects used to control sternorrhynchous Hemiptera (Greathead 1986), and are regarded as the most important natural enemies of whiteflies, including B. tabaci, used in biological control (Lahey & Stansly 2014; Liu et al. 2015). Over the past century, numerous species have been introduced for control of adventitious pest whiteflies with many projects ending in complete success (e.g., Rose & DeBach 1992; van Lenteren et al. 1996; Pickett & Pitcairn 1999; Liu et al. 2015). A recent example was a program aimed at establishing parasitoids of B. tabaci Middle East-Asia Minor 1 (MEAM1; formerly biotype B or Bemisia argentifolii Bellows & Perring) in the United States (Goolsby et al. 2005; Gould et al. 2008). This whitefly is established worldwide and causes annual economic losses in the billions (US dollars). Florida was the site of the first reported MEAM1 invasion into the United States, where it remains a key pest of tomato, watermelon, and many other plants of agricultural and economic importance (Hamon & Salguero 1987).
An effort to establish exotic biological control agents was initiated in the early 1990s in an attempt to control MEAM1 in Florida through the mass rearing and release of parasitoid wasps (Nguyen & Bennett 1995). Seven parasitoid species were chosen as the best candidates for long-term control of MEAM1 and were released in numerous counties across the state. Recent recoveries of parasitoid species from the B. tabaci species complex in Florida (Lahey 2014) have confirmed the establishment of two species released during this program, namely, Encarsia bimaculata Heraty & Polaszek imported from Guatemala, India, and Sudan (Nguyen & Bennett 1995; Heraty & Polaszek 2000), and Eretmocerus sp. nr. emiratus imported from Sudan (= Eretmocerus sudanensis Zolnerowich & Rose in Castillo & Stansly 2011). Host records for both species appear to be relatively narrow, with En. bimaculata known from two whitefly species, B. tabaci and Trialeurodes vaporariorum (Westwood) (Hemiptera: Aleyrodidae) (Heraty & Polaszek 2000), and Er. sp. nr. emiratus only from B. tabaci (Castillo & Stansly 2011).
The parasitoid genera Encarsia and Eretmocerus have different reproductive biologies that may influence their ability to suppress whitefly populations within the same environment. In most Encarsia species, females develop as primary parasitoids, whereas males develop as hyperparasitoids on the larvae and pupae of conspecific or heterospecific individuals, including Eretmocerus (Walter 1983; Hunter & Woolley 2001). Eggs are deposited internally after the host is punctured by the hypodermic-like ovipositor (Gerling et al. 1998). By contrast, both sexes of Eretmocerus develop as primary parasitoids, and egg deposition by the spatulate ovipositor is always external to the host whitefly in the space between the leaf surface and nymphal venter. Both species kill additional nymphs by host feeding. One of the goals of the current study was to determine the competitive influence of each species on the other's survival early in the parasitization process (i.e., <48 h after each species parasitizes the same host), providing insight into interactions between similar development stages of these parasitoids.
Recent field observations indicated a clear difference in the host plant preferences for each species of parasitoid with Er. sp. nr. emiratus favoring hirsute plant species, whereas En. bimaculata was found more often on plants with fewer plant trichomes (Lahey 2014). Previous studies suggest that it may be easier for Eretmocerus to oviposit under whitefly nymphs on hairy plants, or those with uneven surface texture, due to larger gaps underneath the nymph than on plants with smooth leaves (e.g., McAuslane et al. 1995; Headrick et al. 1996; De Barro et al. 2000). Such gaps may facilitate insertion of the blunt-tipped, slightly curved ovipositor of Eretmocerus (Gerling et al. 1998), an action that may be hampered on smooth leaf surfaces (McAuslane & Nguyen 1996). Encarsia are less likely to be influenced by this factor, but may experience reduced walking speed and host finding capacity on hirsute plants (van Lenteren et al. 1996). Based on these observations and life history traits, all experimental treatments were performed on a glabrous (collard) and hirsute (eggplant) host plant in the hopes of determining the extent to which plant preference plays a role in mediating biological control efficiency and species coexistence in these introduced parasitoid wasps.
Materials and Methods
HOST PLANTS
Collard [Brassica oleracea L. ‘Georgia’ (Brassicales: Brassicaceae)] and eggplant [Solanum melongena L. ‘Nadia’ (Solanales: Solanaceae)] seeds were germinated in 200-unit seedling trays and transferred into 15.2 cm diameter pots filled with Fafard 2S or 4S Mix growing media (Conrad Fafard Inc., Agawam, Massachusetts). Plants were fertilized after the appearance of the first 2 true leaves with 20-20-20 (N-P-K) general purpose fertilizer. Cultivation occurred under conditions of natural temperature, humidity, and light in mesh cages (Live Monarch, Boca Raton, Florida) located outdoors (average conditions: 22 °C, 80% RH, 11:13 h L:D photoperiod) at the Southwest Florida Research and Education Center (SWFREC; 26.7667°N, 81.7167°W) in Immokalee, Florida. Plants with at least 3 fully expanded true leaves were used in experiments.
INSECTS
Whitefly and parasitoid cultures were maintained on collard. The whitefly B. tabaci MEAM1 served as the parasitoid host used in the experiments. The whitefly colony was established from collections made on Sonchus asper (L.) Hill (Asterales: Asteraceae) located at the SWFREC during Jun 2013. Laboratory cultures of En. bimaculata and Er. sp. nr. emiratus were established by rearing the parasitoids from MEAM1-infested collard and eggplant grown outdoors at the SWFREC. All insects were kept in separate climate-controlled rooms maintained at 26.5 ± 2 °C, 70 ± 10% RH, and a 14:10 h L:D photoperiod. The whitefly colony was housed in a wood-framed cage (60 × 60 × 60 cm) with walls made of transparent plexiglass and antithrips polyethylene mesh to allow for ventilation. Parasitoid colonies were reared in separate Bug-Dorm (Megaview Science Co., Ltd., Taichung, Taiwan) cages (60 × 60 × 60 cm) with front and back panels made of clear plastic and sides of polyester netting for airflow. Overlapping generations of the whitefly and parasitoids were maintained in their respective cages with fresh plants being introduced as needed.
Parasitoids used in experiments were collected from each colony by removing collard leaves that contained wasp pupae, cutting the leaves into pieces, and sequestering them in separate clear plastic emergence containers (Tri State Plastics, Dixon, Kentucky; 18.0 × 12.5 × 10.5 cm). The top of each container was modified to allow removable plastic vials to be inserted; emerged parasitoids moved upwards into the vials, which were then removed to obtain even-aged individuals for experiments. Each container was provisioned with a moist, braided cotton dental roll (Richmond Dental, Charlotte, North Carolina) to provide moisture and a thin strip of paper towel impregnated with honey to nourish the emerged insects. Single pairs of newly emerged (i.e., <72 h) En. bimaculata and Er. sp. nr. emiratus females and males were aspirated directly into 0.5 mL PCR tubes with tips removed and a small piece of cotton inserted in the opposite side for air exchange. The intake was sealed with Parafilm M (Bemis Company, Inc., Neenah, Wisconsin) to prevent escape. Wasps were held for <1 h in their respective colony rooms before being used in experiments. If a wasp of either sex died at any time during the parasitization period, the experimental unit was discarded and repeated.
TREATMENTS
The competitive influence of one parasitoid species on the other and the effect of competition on suppression of MEAM1 were determined by comparing 6 treatments: 3 heterospecific (Eb/Ee, Ee/Eb, Ee+Eb) and 2 conspecific combinations (Ee only, Eb only) and a whitefly only control (used to test for parasitoid contamination). The 3 heterospecific treatments were either synchronous (i.e., a single female and male of both parasitoid species released at the same time; Ee+Eb) or asynchronous [i.e., a single female and male of En. bimaculata released first followed by a single female and male of Er. sp. nr. emiratus (Eb/Ee) and vice versa (Ee/Eb)]. The ability of combinations of these parasitoids to suppress MEAM1 populations was compared with treatments consisting of each parasitoid species alone. Statistically significant increases in numbers of parasitized nymphs or nymphs killed by host feeding would indicate greater control potential compared with the single species introductions.
EXPERIMENTAL DESIGN
The experiment was set up as a factorial design with 2 host plants (collard and eggplant) and 6 treatments. Approximately 50 unsexed MEAM1 adults were confined to the abaxial leaf surface under a clip-on cage (Stansly & Liu 1997) for a 24 h oviposition period. Clip-on cages were made from plastic medicine cups (top and bottom diameters of 4 and 2.5 cm, respectively) attached to metal hair clips. The top of each cup was lined with yarn to prevent damage to leaf tissue, and the bottom of the cup was removed and replaced with 60 mesh nylon screen. Experimental plants contained at least 3 fully expanded leaves onto which a single clip-on cage was fastened at a maximum of 3 cages per plant. After 24 h, the whiteflies and clip-on cages were removed and the eggs were monitored daily for emergence. After all the insects had hatched and settled on the host plant, all but approximately 30 nymphs were removed with a size 0 insect pin affixed to a wooden applicator stick. When the 3rd instar was reached (i.e., the preferred stage for both parasitoid species) (Qiu et al. 2007; Castillo & Stansly 2011), the clip-on cages were replaced, and the parasitoids were released according to treatment. In the asynchronous treatments, both wasp species had 24 h access to the whitefly nymphs regardless of their order of release, making for a total access period of 48 h (i.e., 24 h for Er. sp. nr. emiratus and 24 h for En. bimaculata). Conversely, the synchronous treatment consisted of both wasp species released together and lasted for 24 h. In the single species introductions, each pair of En. bimaculata and Er. sp. nr. emiratus male and female had a 48 h exposure period to the nymphs within the clip-on cage.
After the final exposure period, the clip-on cages and wasps were removed, and the plants were kept in a room maintained at 26.5 ± 2.0 °C, 70 ± 10% RH, and 14:10 h L:D photoperiod for 10 d to allow for development and differentiation of the wasps by morphological differences in their larval and pupal forms. Eretmocerus sp. nr. emiratus larvae are circular in shape and take on a milky-white appearance within their host (Castillo & Stansly 2011). Encarsia bimaculata larvae are hymenopteriform, sickle shaped, and translucent with a visible ileo-labial gland (Antony et al. 2004). Pupae of Er. sp. nr. emiratus and En. bimaculata can be differentiated by the absence and presence of meconial pellets near the vasiform orifice of the whitefly, respectively. Pupae of Er. sp. nr. emiratus are completely yellow and have red eyes, whereas En. bimaculata pupae are covered in black cuticle.
After 10 d, the leaves were removed and examined under a stereomicroscope. Counts were made of the number of unparasitized (i.e., empty whitefly pupal cases), En. bimaculata parasitized, and Er. sp. nr. emiratus parasitized nymphs, and those killed by host feeding. In all treatments where Er. sp. nr. emiratus was released, nymphs were turned over to check for the presence of hatched and unhatched parasitoid eggs. Each treatment was replicated 5 times on each host plant. All treatments in a replicate were conducted simultaneously, although not all replications were performed at the same time.
STATISTICAL ANALYSES
Mean percentage of parasitism, numbers of progeny produced by En. bimaculata and Er. sp. nr. emiratus, and numbers of nymphs killed by host feeding on each plant relative to the treatment type (excluding the control), were subjected to a 2-way analysis of variance (ANOVA) using the package lme4 (R version 3.1.2; R Development Team 2014). Means were separated using the difflsmeans function within the lmerTest package. Results were considered significant at an alpha value of 0.05.
Results
TOTAL PARASITISM
The total level of parasitism varied significantly among treatments (F = 11.5; df = 4,40; P < 0.001) but not among plant species (F = 0.10; df = 1,40; P = 0.751) (Fig. 1). There was no interaction between treatments and host plants (F = 1.51; df = 4,40; P = 0.218). The percentage of nymphs parasitized was similar in all treatments on collard with the exception of the single introduction of En. bimaculata, where parasitism was significantly lower (16.0 ± 4.3%). On eggplant, percentage of parasitism was more variable, being greatest in the synchronous parasitoid release treatment (58.9 ± 6.8%) and lowest when En. bimaculata was released alone (14.5 ± 1.4%).
PRODUCTION OF EACH PARASITOID SPECIES
Significant differences in the number of progeny produced by En. bimaculata were found among parasitoid release treatments (F = 4.68; df = 3,32; P = 0.009) but not between plant species (F = 0.14, df = 1,32; P = 0.697) (Fig. 2). There was no treatment by host plant interaction (F = 1.20; df = 3,32; P = 0.274). Encarsia bimaculata produced significantly fewer progeny (1.2 ± 0.6) when introduced before Er. sp. nr. emiratus on collard; similar numbers of progeny were produced in all other treatments. On eggplant, En. bimaculata produced significantly fewer progeny (1.8 ± 0.7) when introduced after Er. sp. nr. emiratus compared with the single introduction of En. bimaculata (4.4 ± 0.2).
Significant differences in the number of progeny produced by Er. sp. nr. emiratus were found among treatments (F = 4.21; df = 3,32; P = 0.013) but not between plants (F = 0.01; df = 1,32; P = 0.920) (Fig. 3). There was no treatment by host plant interaction (F = 1.73; df = 3,32; P = 0.181). Eretmocerus sp. nr. emiratus produced significantly more progeny when introduced alone (21.8 ± 2.1) than in any combination with En. bimaculata on collard. On eggplant, Er. sp. nr. emiratus produced significantly fewer progeny (8.0 ± 1.3) when introduced before En. bimaculata compared with being introduced singly (17.2 ± 3.0) or in synchrony with En. bimaculata (17.2 ± 4.0).
HOST FEEDING
The number of nymphs killed by host feeding did not differ among treatments (F = 0.83; df = 4,40; P = 0.508) or between host plants (F = 1.53; df = 1,40; P = 0.220). There was no significant interaction between treatments and host plants (F = 2.55; df = 4,40; P = 0.053).
Across all treatments, the total number of nymphs killed by host feeding never exceeded more than 5% of those offered in a particular treatment. Host feeding was lowest in the single introduction of Er. sp. nr. emiratus, where 0.9% (n = 3) of nymphs were killed by host feeding. The introduction of Er. sp. nr. emiratus before En. bimaculata resulted in the greatest number of nymphs host fed upon at 2.8% (n = 9). When the host plant was taken into consideration, host feeding was both lowest and highest on collard, lowest when no host feeding was observed in the single species introduction of Er. sp. nr. emiratus, and highest when 4.4% (n = 7) of nymphs were killed in the synchronous treatment. On eggplant, 0.6% of nymphs were killed by host feeding when En. bimaculata was introduced before Er. sp. nr. emiratus (n = 1), as well as in the synchronous treatment (n = 1); host feeding was greatest in the single introduction of En. bimaculata at 2.6% (n = 4).
Discussion
We evaluated the ability of 2 parasitoid wasp species to contribute to the suppression of a shared host, and the impact each has on the other, on a glabrous (collard) and hirsute (eggplant) host plant. Our results suggest that in terms of parasitism, the order of release of the parasitoids did not affect the number of hosts parasitized on either host plant; however, differences in incidence of parasitized whiteflies were observed in previous studies relative to order of introduction. Pang et al. (2011) found significant decreases in the numbers of B. tabaci nymphs parasitized by En. formosa Gahan and En. sophia (Girault & Dodd) on tomato when the wasps were introduced simultaneously, or in sequence, compared with each species alone. In competition experiments under 2 host densities, Xu et al. (2013) found that the sequential introduction of En. sophia, followed by Er. hayati Zolnerowich & Rose, resulted in greater levels of parasitism than En. sophia alone when 30 B. tabaci nymphs were present on tomato. Conversely, sole access of Er. hayati to 10 nymphs contributed to greater levels of parasitism compared with allowing Er. hayati access first, followed by En. sophia.
The only differences we observed in parasitism levels were in the individual release treatments with En. bimaculata on both host plants, and when En. bimaculata was introduced after Er. sp. nr. emiratus, where the numbers of parasitized B. tabaci were significantly lower than in other treatments. This result is expected based on life history differences between the genera. Encarsia species mature more of their egg complement after emergence of the adult (i.e., synovigeny), whereas Eretmocerus emerge as adults with nearly their full egg complement (i.e., pro-ovigeny) (Jervis et al. 2001). As a result, Eretmocerus are able to oviposit in significantly more hosts than Encarsia earlier in their lifespan resulting in the observed differences in parasitism between treatments.
Progeny production of each parasitoid species was reduced in certain competition treatments relative to treatments in which each species was released individually. On collard, the number of progeny produced by En. bimaculata was significantly reduced (by 45.4% or 1 offspring) when introduced before Er. sp. nr. emiratus. The opposite combination, Er. sp. nr. emiratus followed by En. bimaculata, reduced the progeny production of the latter on eggplant (by 7% or 0.6 offspring). All 3 competition treatments significantly reduced the number of Er. sp. nr. emiratus progeny produced on collard. Differences on eggplant were confined to the treatment in which Er. sp. nr. emiratus was followed by En. bimaculata. A reduction in the number of progeny is the expected result when parasitoids compete for hosts. Collier & Hunter (2001) found that Er. eremicus Rose & Zolnerowich reduced the progeny production of En. sophia by up to 50%, and En. sophia reduced the progeny production of Er. eremicus by up to 92%, when allowed access to 8 B. tabaci nymphs on cotton. Similarly, Er. hayati and En. sophia negatively affected the number of progeny produced by each other in sequential release combinations regardless of the number of available hosts (i.e., 10 or 30) (Xu et al. 2013).
The cause for such reductions appears to be due to 2 factors: multiparasitism and host feeding. A combination of these factors accounted for the negative influence En. sophia had on Er. eremicus, and multiparasitism was implicated in Er. eremicus' effect on En. sophia (Collier & Hunter 2001). In this study, we did not investigate the mechanism by which each species interfered with the other; however, host feeding across treatments occurred at an unusually low rate, decreasing the likelihood that this factor contributed to our results. It is possible that our method of mass rearing the parasitoids may have allowed early emerging parasitoids the opportunity to host feed, so that they became satiated prior to their use in experiments.
In both treatments in which each parasitoid species was introduced sequentially (i.e., Ee/Eb and Eb/Ee), we observed an empty Eretmocerus egg underneath a nymph containing an En. bimaculata pupa. In all cases (n = 4), circular entry scars were observed in the ventral integument of the whitefly where the Eretmocerus larva had penetrated. This observation indicates some capacity for En. bimaculata to successfully compete for the same host early (<24 h) in the parasitism process. In addition, this phenomenon was observed in both sequential treatments, suggesting that neither parasitoid has a capacity to discriminate between nymphs recently (<24 h) parasitized by the other.
The different host plants used in this study did not have any effect on total parasitism, progeny production, or the number of nymphs killed by host feeding by either parasitoid species. Previous studies indicated that Eretmocerus will parasitize more hosts on hirsute plants, or those with rough leaves (e.g., rockmelon), where whitefly nymphs do not lie flat against the leaf surface presumably making oviposition easier (Headrick et al. 1996; De Barro et al. 2000). Therefore, we expected that the single species treatment of Er. sp. nr. emiratus on eggplant would result in significantly greater parasitism than the same on collard; however, no differences in total parasitism were detected. Qiu et al. (2005) found Er. sp. nr. furuhashii to parasitize more hosts on glabrous collard leaves than on the increasingly pilose leaves of eggplant, cucumber, and tomato. In a similar study, Er. sp. nr. furuhashii and En. bimaculata parasitized significantly fewer whitefly nymphs on cucumber and tomato compared with less hairy eggplant and glabrous collard (Qiu et al. 2007). A potential caveat of these studies, and our own, is that they were run under no-choice conditions and may not reflect host plant choice by these parasitoids in the field. Clearly, more research is needed to elucidate the relationships between whiteflies, their parasitoids, and the host plant structures that may enhance whitefly biological control.
The results of this study suggest there would be no need to introduce En. bimaculata, or any other species of Encarsia, into an environment already under control by Er. sp. nr. emiratus. Although multigenerational studies should be conducted to say this definitively, the introduction of En. bimaculata would almost certainly reduce the positive effect of Er. sp. nr. emiratus, especially after the first generation when heteronomous hyperparasitism becomes possible (Walter 1983; Williams 1996).
Acknowledgments
This work was completed as part of the first author's graduate studies. Z. J. L. thanks his co-authors for financial support and constructive comments throughout his coursework.