The dynamics of predation by the green lacewing Chrysoperla externa Hagen (Neuroptera: Chrysopidae) on Enneothrips flavens Moulton (Thysanoptera: Thripidae) was investigated by placing embryonated eggs and first-instar larvae of C. externa on potted peanut plants (Arachis hypogaea L.; Fabales: Fabaceae) in a greenhouse. The plants that received either embryonated eggs or larvae of C. externa showed significant reductions in the mean numbers of thrips, about 9 days after the release of the predator on the plants. The potential of C. externa as a biological control agent for thrips is discussed.
Among the pests of peanuts (Arachis hypogaea L.; Fabales: Fabaceae), the thrips, Enneothrips flavens Moulton (Thysanoptera: Thripidae), has received increasing attention because of the serious economic damage that it causes (Moraes et al. 2005; Dalastra et al. 2011; Michelotto et al. 2013). Enneothrips flavens lives in the closed buds or enclosed parts of the plant, and punctures and sucks the cell contents. Consequently, peanut buds are deformed and distorted, exhibiting streaks and discolorations, which result in major crop losses (Gallo et al. 2002). Although chemical control is frequently used, its intense and increasing application contributes to environmental contamination (Bhanti & Taneja 2007), decline of pollinators, and the development of pesticide resistance (Fournier 2005; Henry et al. 2012; Whitehorn et al. 2012). Biological control has been used as an alternative to chemical control, and in some instances it is an efficient tool for pest management (Jonsson et al. 2012).
The green lacewing, Chrysoperla externa Hagen (Neuroptera: Chrysopidae), is an important natural enemy of several pest species, because it is tolerant to some pesticides and is a voracious predator (Brettell 1982; Freitas 2001a; Rimoldi et al. 2012; Silva et al. 2012). It has been found in different agroecosystems and has shown significant potential as a biological control agent of phytophagous insects (Carvalho & Souza 2000; Freitas 2002; Bonani et al. 2009).
A growing body of research has shown the importance of releasing green lacewings as control agents for the management of pests, including thrips (Carvalho & Souza 2000). By releasing second-instar larvae of Chrysoperla carnea Stephens, Hassan (1978) demonstrated successful control of Myzus persicae Sulzer (Hemiptera: Aphididae) on eggplant (Solanum melongena L; Solanales: Solanaceae) grown in greenhouses. The apple aphid, Aphis pomi De Geer (Hemiptera: Aphididae), has been controlled by releasing eggs of C. carnea on apple cultivars (Hagley 1989). The control of various pests in North America by augmentative release of C. carnea has been reported. In cotton, C. carnea has suppressed Helicoverpa zea Boddie (Lepidoptera: Noctuidae) and Heliothis virescens (F) (Lepidoptera: Noctuidae) (Ridgway & Jones 1969), and, as also demonstrated with C. rufilahris, has significantly suppressed the thrips Scritothrips citri Moulton (Thysanoptera: Thripidae) on mango (Khan & Morse 2001).
This study investigated the efficacy of using C. externa against E. flavens, for reducing the latter's population size in response to the release of C. externa eggs and larvae onto peanut plants grown in a greenhouse.
Materials and Methods
Chrysoperla externa Populations
Green lace wing adults were collected by means of an entomological net in a grass field. The field is located near a plantation of Pinus sp. in the municipality of Jaboticabal, São Paulo, Brazil. The insects were identified in the taxonomy laboratory at the Universidade Estadual Paulista, Jaboticabal, São Paulo, Brazil. The green lace wings were allowed to mate and males and females were maintained at 25 ± 2 °C, 65 ± 10% RH and 12:12 h L:D. The insects were reared using the methodology developed by Freitas (2001b). The eggs were placed in individual glass bottles (4×1 cm) and the newly hatched larvae were used in the experiments.
Growing Peanut Plants
Peanut plants were grown in a greenhouse in 5 L plastic containers containing soil and sand in a 3:1 ratio. Ten seeds of the variety ‘Runner IAC 886’ were sown per container. Fifteen days after germination, the peanut seedlings were thinned, leaving only 1 plant per container. No pesticide was applied to the plants.
Enneothrips flavens Populations
Twenty-day-old peanut plants were infested with E. flavens by placing branches containing thrips on the plants. Taking into account that only 1 infestation might not be enough, a new infestation was performed after 5 days. The thrips used for infestation were obtained from 25 day-old peanut fields at the Universidade Estadual Paulita, Jaboticabal, São Paulo, Brazil. When the experimental plants were 44 days old, the first sample was obtained. The number of thrips on each plant was recorded by observing 6 closed buds on the central branch of the plant. After this initial sample, the predator was released.
Bioassay
The experiments were set up using a fully randomized design with 20 replicates and 3 treatments as follows: Control (no C. externa eggs or larvae), T2 (C. externa eggs) and T3 (C. externa larvae). Each experimental unit was 1 peanut plant, observed after 0, 4, 9 and 15 days, totaling 360 observations. The release of C. externa occurred as follows. The control consisted of plants that received no C. externa individual. Treatment T2 was composed of plants receiving 4 C. externa eggs/plant, and treatment T3 of plants that received 3 newly hatched C. externa first-instar larvae/plant. The eggs were placed in a plastic container (height 4 cm × diam 5 cm) with shredded paper to minimize cannibalism. The larvae were released by catching them with a brush and placing them on the plants. All experimental units (containers) were covered with voile bags tied over the plant, to prevent contamination with other plants or insects.
Before the release of the predator, one sampling was done. After the release of the green lace wings, samples were obtained after 4, 9 and 15 days, for a total of 4 samples. The selection of days for sampling was based on the lacewing life cycle: the larva requires 4 days to hatch, and the 1st, 2nd and 3rd stadia each last for 3 days under laboratory conditions.
Statistical Analysis
There was wide variability among the treatments and within each treatment (Fig. 1). Given the dependence of the observations taken in the same experimental unit over time, the nonlinear behavior of the data, as well as the assumption that the mean number of thrips per plant decreases over time, an asymptotic mixed-effects regression model (Pinheiro & Bates 2000) was used. This model can be written as:
where yijk is the mean number of thrips (Table 1) for the i-th treatment, j-th replicate, and k-th time period, Φ1i is the asymptote for the i-th treatment, Φ2i is a scaling parameter for the i-th treatment, b2ij. ∼ N(0, σ2b2 is the random effect associated with φ2i, Φ3i is logarithm of the rate constant for the i-th treatment, tk is the time and εijk is the error.
Table 1.
Mean number of Enneothrips flavens (nymphs and adults) found in closed Arachis hypogaea Buds in 3 treatments: Control A. Hypogaea plants (without Chrysoperla externa), on A. Hypogaea plants that received C. Externa eggs, and on A. Hypogaea plants that received C. externa Larvae.
To test for treatment differences, two submodels, namely M2 and M3, were fitted to the data. In M2, the T2 (C. externa eggs) and T3 (C. externa larvae) treatments were grouped; and in M3, the linear predictor is given by
that is, no treatment effect was assumed. The models were compared using likelihood-ratio tests (Verbeke & Molenberghs 2000).
Model M2 did not differ statistically from model M1, and so treatments T2 (C. externa eggs) and T3 (C. externa first-instar larvae) did not differ statistically (p = 0.05), see Table 2. Also, model M3 fit the data poorly compared to model M2 (Table 2). Therefore, the control treatment differed from the T2 and T3 group (p = 0.05).
The parameter Φ3 estimate for treatments T2 and T3 was not significant (F1,171 = 0.10, p = 0.75), so two submodels were fitted to the data: model M 4 with a linear predictor given by
that is, parameter Φ3 is the same for all treatments; and model M5, with the linear predictor given by
Table 2.
Likelihood-ratio tests for nested models M1, M2 AND M3.
Table 3.
Likelihood-ratio tests for nested models M2, M4 and M5.
that is, parameter Φ3 is set to zero. The likelihood-ratio tests (Table 3) showed that the fit from model M4 did not differ from that of model M2; however, the fit of model M5 was significantly different (Table 2). Therefore, model M4 fit the data as well as model M2 and could be used as a final model.
Table 4 shows the parameter estimates and associated standard errors for model M4, which can be written as
Results and Discussion
On day zero, each control plant had a mean of 7.75 thrips, the plants that subsequently received C. externa eggs had 11.85 thrips/plant, and the group that subsequently received lace wing larvae had 13.90 thrips/plant (Table 1). The passage of time did not influence the population of thrips in the control, but there were significant differences in the effects of C. externa releases over time on thrips densities at 0 to 4, 9 and 15 days post-release (Figs. 1 and 2).
Table 4.
Parameter estimates (standard errors) FOR MODEL M4 (Φ1 is the asymptote, Φ2 is a scaling parameter, Φ3 is logarithm of the rate constant and σ282 is the variance of the random effect associated with Φ2)
These results suggest that C. externa requires some time after being released in order to show significant impacts on the pest population. On the sampling dates, the mean number of E. flavens thrips was significantly reduced on plants that had received C. externa when compared with the control plants. These results provided evidence for the potential of C. externa as a biological control agent of E. flavens, under specific conditions on potted peanuts in a greenhouse. The statistical modeling confirmed that the thrips population decreased in the presence of C. externa, as shown in Fig. 1. This result is easily observed by comparing the 2 trends, the constant line describing the E. flavens population in the control and the curves describing the E. flavens populations under the influence of C. externa that had been released either as eggs or larvae (Figs. 1 and 2).
In the absence of predators, it is expected that the mean number of thrips in closed peanut buds will increase, as seen with M. persicae aphids on eggplants (Hassan 1977). In the current study, the number of E. flavens thrips in the control remained stable, while in the other treatments the number was reduced. On day 15, C. externa third-instars started to pupate, and in response the thrips population increased slightly. It is important to use appropriate intervals between predator releases, in order to prevent the pest from persisting when the C. externa larvae are in a post-feeding period. The appropriate release intervals for the control of M. persicae by C. carnea have been estimated from 2 to 5 weeks (Hassan 1977). However, for E. flavens these release intervals would vary from 9 to 15 days, based on the results of the current study.
Fig. 1.
Biological control of Enneothrips flavens with Chrysoperla externa, showing daily trends in each of 20 replicates. The control consisted of Enneothrips flavens thrips-infested peanut plants that received no Chrysoperla externa individuals. In treatment T2 the thrips-infested peanut plants received 4 C. externa eggs/plant, and in treatment T3 the thrips-infested peanut plants received 3 newly hatched C. externa first-instar larvae/plant.
A few studies involving the genus Chrysoperla and thrips have been designed to investigate preferences between different prey. In a recent study, Shrestha & Enkegaard (2013) analyzed the prey choice by 3rd-instar C. carnea on the western flower thrips Frankliniella occidentalis and the lettuce aphid Nasonovia ribisnigri (Mosley) (Aphididae) in the laboratory, by using different prey ratios. The results of the study suggest a slight preference of C. carnea for aphids compared to thrips. However, the results were also significantly influenced by the predator-prey ratios; and at some ratios, no preference was observed (Shrestha & Enkegaard 2013). Although this result indicated an apparent weak interaction between C. carnea and thrips, survey results have shown that members of the genus Chrysoperla are frequently present on plants of different species containing thrips (Bettiol et al. 2004; Mann et al. 2010; Saeidi & Adam 2011), encouraging studies to evaluate the probable interaction dynamics between these species. Unfortunately, the lack of studies investigating possible interactions between populations of C. externa and E. flavens make any specific comment about the interaction strength between them impossible. To our knowledge, studies examining the biological control of E. flavens are not common, and ours is a pioneer study on the use of C. externa for this purpose.
Fig. 2.
Reduction of Enneothrips flavens on peanut plants by Chrysoperla externa. The average (± SE) numbers of thrips per peanut plant were fitted the curve using model M4. The control consisted of Enneothrips flavens thrips-infested peanut plants that received no Chrysoperla externa individuals. In treatment T2 the thrips-infested peanut plants received 4 C. externa eggs/ plant, and in treatment T3 the thrips-infested peanut plants received 3 newly hatched C. externa first-instar larvae/plant.
Acknowledgments
CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) provided financial support for this study, and COPLANA (Cooperativa Agroindustrial, Jaboticabal-SP, Brazil) supplied us with peanut seeds. Dr. José Carlos Barbosa (FCAV/UNESP) helped with data analysis and Dr. Renata C. Monteiro (USP/ ESALQ) identified the thrips. Dr. Sérgio de Freitas (in memoriam) gave us support during the study and identified specimens of green lacewings. We thank Janet Reid for revising the English, and FAPESP for the financial support.
