Introduction
The tomato borer, Tuta absoluta (Meyrick) (Lepidoptera: Gelechiidae), a devastating pest of tomato from South America (Miranda et al., Reference Miranda, Picanco, Zanuncio and Guedes1998) was first detected in Europe in Spain at the end of 2006. It spread quickly throughout the Mediterranean Basin (Desneux et al., Reference Desneux, Wajnberg, Wyckhuys, Burgio, Arpaia, Narváez-Vasquez, González-Cabrera, Catalán Ruescas, Tabone, Frandon, Pizzol, Poncet, Cabello and Urbaneja2010), ravaging tomato crops. Insecticidal control of T. absoluta is problematic due, in part, to pesticide resistance (Siqueira et al., Reference Siqueira, Guedes and Picanco2000; Lietti et al., Reference Lietti, Botto and Alzogaray2005; Haddi et al., Reference Haddi, Berger, Bielza, Cifuentes, Field, Gorman, Rapisarda, Williamson and Bass2012; Campos et al., Reference Campos, Silva, Silva, Silva and Siqueira2014), its effects on non-target organisms (Biondi et al., Reference Biondi, Mommaerts, Smagghe, Viñuela, Zappalà and Desneux2012a , Reference Biondi, Desneux, Siscaro and Zappalà b ; Reference Biondi, Zappalà, Stark and Desneux2013a ), market and governmental residue tolerance requirements (COUNCIL OF THE EUROPEAN COMMUNITIES, 2005), and environmental and human health concerns (Pimentel, Reference Pimentel2005). Therefore, interest in biological control of this and other pests is increasing.
Different indigenous natural enemies to Europe have been found attacking T. absoluta in Europe, which include the eulophid Necremnus artynes (Walker) (Hymenoptera: Eulophidae), (Ferracini et al., Reference Ferracini, Ingegno, Navone, Ferrari, Mosti, Tavella and Alma2012; Urbaneja et al., Reference Urbaneja, Gonzalez-Cabrera, Arno and Gabarra2012; Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2013; Zappalá et al., Reference Zappalá, Biondi, Alma, Al-Jboory, Arnó, Bayram, Chailleux, El-Arnaouty, Gerling, Guenaoui, Shaltiel-Harpaz, Siscaro, Stavrinides, Tavella, Vercher Aznar, Urbaneja and Desneux2013; Abbes et al., Reference Abbes, Biondi, Zappalà and Chermiti2014; Chailleux et al., Reference Chailleux, Desneux, Arnó and Gabarra2014a ; Gebiola et al., Reference Gebiola, Bernardo, Ribes and Gibson2015). Nevertheless, early reports of N. artynes parasitizing T. absoluta are now thought, as it is the case of Calvo et al. (Reference Calvo, Soriano, Bolckmans and Belda2013), to refer Necremnus tutae Ribes & Bernardo (Hymenoptera: Eulophidae), which has been found to be the most abundant and widespread species within the N. artynes group in Europe (Gebiola et al., Reference Gebiola, Bernardo, Ribes and Gibson2015). N. tutae is an ectoparasitoid, which parasitizes second to fourth instars of T. absoluta larvae and inflicts extra mortality by host-killing, i.e. host-feeding and host-stinging (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2013). Additionally, it exhibits a higher intrinsic rate of increase than T. absoluta when reared on tomato (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2013), altogether indicating potential to control this pest. Integrating the use of such agent alone or combined with other control agents into the present biological control-based integrated pest management programmes (IPM) could provide growers with more options to control T. absoluta. However, critical parameters such as rate, timing, and methods or frequency of release under greenhouse conditions needed to implement N. tutae for augmentative biological control in protected tomato crops are lacking.
The current standard biological control-based management of tomato pests in the Mediterranean area is based on augmentation of the mirid bug, Nesidiocoris tenuis Reuter (Heteroptera: Miridae) (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2012a ; Urbaneja et al., Reference Urbaneja, Gonzalez-Cabrera, Arno and Gabarra2012; Biondi et al., Reference Biondi, Chailleux, Lambion, Han, Zappalà and Desneux2013b ; Zappalá et al., Reference Zappalá, Biondi, Alma, Al-Jboory, Arnó, Bayram, Chailleux, El-Arnaouty, Gerling, Guenaoui, Shaltiel-Harpaz, Siscaro, Stavrinides, Tavella, Vercher Aznar, Urbaneja and Desneux2013). This predator commonly appears naturally in tomato and other agricultural crops as well as uncultivated vegetation in the Mediterranean region and the Canary Islands (Malausa & Henao, Reference Malausa and Henao1988; Goula & Alomar, Reference Goula and Alomar1994; Tavella & Goula, Reference Tavella and Goula2001). It is known as an effective natural enemy of whiteflies (Sánchez & Lacasa, Reference Sánchez and Lacasa2008; Calvo et al., Reference Calvo, Bolckmans, Stansly and Urbaneja2009) and has been used for whitefly control in tomato crops since 2002, and more recently against T. absoluta (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2012a ; Urbaneja et al., Reference Urbaneja, Gonzalez-Cabrera, Arno and Gabarra2012). Two release methods for adult N. tenuis are commonly used for augmentative biological control: (1) post-plant application: released 3 or 4 weeks after planting (Calvo & Urbaneja, Reference Calvo and Urbaneja2004; Calvo et al., Reference Calvo, Bolckmans, Stansly and Urbaneja2009), or (2) pre-plant application: release on tomato seedlings at the plant nursery prior to transplanting (Calvo et al., Reference Calvo, Lorente, Stansly and Belda2012b , Reference Calvo, Bolckmans and Belda c ). In both cases, Ephestia kuehniella eggs Zeller (Lepidoptera: Pyralidae) are provided as a supplemental food during the first weeks after the release (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2012a ). Pre-plant application of N. tenuis has proven to be very effective for whitefly and T. absoluta control, and it is now widely implemented (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2012a ). Implementation of N. tutae for biological control of T. absoluta into the existing programme for commercial tomato greenhouses would be justified if it is proven cost-effective. Nevertheless, multispecies-based programmes can lead to different interactions (Straub et al., Reference Straub, Finke and Snyder2008), which are expected to benefit biological control if the species belong to different functional groups i.e. species, which do not share a resource/habitat and/or seasonal occurrence (Northfield et al., Reference Northfield, Crowder, Jabbour, Snyder, Ohgushi, Schmitz and Holt2012). N. tutae and N. tenuis attack different stages of T. absoluta (Urbaneja et al., Reference Urbaneja, Montón and Mollá2009; Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2013), but they could interact negatively through: (1) Kleptoparasitism, the predator feeds on paralyzed host larvae or hostfed larvae by the parasitoid, what induces to ectoparasitoid larval mortality (Chailleux et al., Reference Chailleux, Wajnberg, Zhou, Amiens-Desneux and Desneux2014b ); (2) ‘Unidirectional’ intraguild predation, the omnivorous predator feeds on parasitoid larvae directly; (3) Competition, either the predator or the parasitoid reduces future prey or host for the other.
For all of this, we conducted the present study whose overall objective was to determine whether the inclusion of N. tutae in the existing IPM programme for tomato would result in better T. absoluta control, as well as providing general guidelines for the practical use of N. tutae in tomato greenhouses. This was done in two subsequent experiments of which the first reported here (N. tutae-alone releases) was aimed at optimizing release rates for N. tutae needed to provide good control of T. absoluta under a worst case scenario of rapid immigration of the pest into a tomato greenhouse. The second experiment focused on evaluating the potential for improving augmentative biological control of T. absoluta by N. tenuis with the addition of N. tutae (N. tutae–N. tenuis joint releases) based on the information obtained from the first experiment.
Material and methods
Greenhouse
The experiments were conducted in a multi-tunnel greenhouse located in Vicar (Almeria, Andalusia, Spain), in which 40 walk-in cages were constructed to accommodate the plants and maintain treatments. Twelve and 20 of these cages were used for the first (N. tutae-alone releases) and second (N. tutae-N. tenuis joint releases) experiments, respectively. Walk-in cages were 5 × 3.5 × 4 m3 (L × W × H) i.e. 17.5 m2 with walls and ceiling constructed of ‘anti-thrips’ polyethylene screening with 220 × 331 µm2 interstices and supported by heavy wires. Floors were covered with woven 2-mm-thick polyethylene ground cloth and access to each cage was through a zippered doorway. The greenhouse was equipped with a Climatec™ system (Novedades Agrícolas, Murcia, Spain) for temperature and relative humidity (RH) control. Temperature and relative humidity were monitored in four randomly selected walk-in cages with HOBO H8 RH/Temp Loggers (Onset Computer, Bourne, MA, USA). Mean weekly temperature ranged from 19.3 ± 0.96 to 25.6 ± 2.14°C during the first experiment and 23.4 ± 1.54 to 29.1 ± 1.34°C during the second experiment. Mean weekly RH fluctuated from 63.9 ± 2.11 to 72.1 ± 1.97 and 56.4 ± 2.41 to 68.3 ± 2.78% during the two experiments, respectively.
Pests, control agents, and supplemental food
T. absoluta adults used to infest the tomato plants, were collected from a colony maintained on tomato and originally obtained from field collections in several locations within the Province of Murcia (Spain; 37°59′10″N, 1°7′49″W). For each release, pest adults belonged to the same cohort to assure homogeneity in age. N. tutae specimens used in the assay were reared in the facilities of Koppert Biological Systems located in Aguilas (Murcia, Spain) on tomato plants using T. absoluta as host. Rearing was initiated with more than 200 adults emerging from tomato-leaf samples infested only with T. absoluta larvae and collected within the Region de Murcia (Spain), from March to June 2015. For all releases during the experiments, 3-day old or less N. tutae adult were used. N. tenuis was provided in bottles containing 500 adults (Nesibug™; Koppert Biological Systems, Berkel en Rodenrijs, The Netherlands). Eggs of E. kuehniella used as supplemental food during the experiment were supplied by Koppert Biological Systems in bottles containing 10 g of eggs (ENTOFOOD™, Koppert Biological Systems, The Netherlands).
Experimental design and procedure
Experiment 1: N. tutae-alone releases
Seeds of tomato, Solanum lycopersicum L. (Solanaceae) cv. ‘Razymo’ (Rijk Zwaan, De Lier, The Netherlands) were planted into 5.4 cm2 peat moss root cubes. Seedlings with five fully-expanded leaves were transplanted (2 September 2010) into the above-mentioned walk-in cages in 25 l coco peat fiber bags. Twenty-four seedlings were transplanted per walk-in cage, although a typical plant density for tomato cultivation of 2 plants m−2 (12 m2 cage−1) was considered to estimate the number of N. tutae adults to be released into cages. Crop cultivation techniques typical for greenhouse tomato cultivation were followed: plants were trained by the main stem to a black polyethylene string tied to a stainless steel overhead wire, secondary shoots were removed and water and fertilizers were supplied as required through a drip irrigation system.
Three treatments were compared in a completely randomized block design with four replicates, in which each of the 12 walk-in cages used constituted a plot, and each block (replicate) consisted in three adjacent cages. Plants from all cages were infested by releasing one T. absoluta couple (male × female) per plant every week for 3 weeks since the transplanting date. The treatments were: (1) T. absoluta: T. absoluta only; (2) 1 N. tutae m−2: T. absoluta + 1 N. tutae m−2 (12 adults per walk-in cage released weekly for 7 weeks beginning 2 weeks after planting); and (3) 2 N. tutae m−2: T. absoluta + 2 N. tutae m−2 (24 adults per walk-in cage released weekly for 7 weeks beginning 2 weeks after planting). The release schedule for T. absoluta was meant to simulate gradual but heavy immigration of the pest into the greenhouse (Calvo et al., Reference Calvo, Lorente, Stansly and Belda2012b ). N. tutae neither parasitizes nor feeds on first instar T. absoluta larvae (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2013), and thus N. tutae releases began 2 weeks after planting coinciding with first availability of second instar larvae of T. absoluta. Tuta absoluta and parasitoid adults were cooled briefly in a cold room at 8°C for counting before being released into designated walk-in cages at a sex ratio of 1:1.
Experiment 2: N. tutae–N. tenuis joint releases
Five treatments were compared in a completely randomized block design with four replicates. Five adjacent walk-in cages constituted a block (replicate), and again each of the 20 walk-in cages used during the experiment constituted a plot, into which T. absoluta was released as above. Treatments were: (1) T. absoluta: T. absoluta only; (2) N. tenuis pre-plant: T. absoluta + one N. tenuis per two plants released in transplant trays 5 days before planting; (3) N. tenuis pre-plant + N. tutae: T. absoluta + N. tenuis released as in treatment 2 + 1 N. tutae m−2 (12 adults per walk-in cage released weekly for 5 weeks beginning 2 weeks after planting); (4) N. tenuis post-plant: T. absoluta + one N. tenuis per two plants released into walk-in cage the day of planting and (5) N. tenuis post-plant + N. tutae: T. absoluta + N. tenuis released as in treatment 4 + 1 N. tutae m−2 released as in treatment 3. Timing and rate for N. tenuis releases were established in accordance with Calvo et al. (Reference Calvo, Lorente, Stansly and Belda2012b , Reference Calvo, Bolckmans and Belda c ) and for N. tutae in accordance with results observed during the ‘N. tutae-alone releases’ experiment.
For pre-plant inoculation, groups of 24 tomato seedlings at the four-leaf stage grown as above were moved into ‘inoculation’ cages (1 × 1 × 1.5 m3). Twelve N. tenuis adults were then released into the ‘inoculation’ cages at a sex ratio of 1:1 after being cooled briefly in a cold room at 8°C for counting. Four paper strips (3 × 1 cm2) with ca. 0.01 g eggs of E. kuehniella glued to one side had been placed inside each inoculating cage to serve as a food source for the mirids. Plants were maintained inside the inoculation cages for 5 days at 25°C, 75% RH and 16:8 (L:D) photoperiod, after which adult N. tenuis were removed and the 24 seedlings transplanted into walk-in cages in coco peat fiber bags as above on 2 June 2011. Plants for cages designated for the remaining treatments were maintained under the same conditions during these 5 days, after which they were also transplanted. For post-plant inoculation adult N. tenuis were first counted as above before being released into designated walk-in cages at a sex ratio of 1:1. In all cages with N. tenuis, eggs of E. kuehniella were sprinkled weekly on all plants at a rate of 0.04 g per walk-in cage (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2012a , Reference Calvo, Lorente, Stansly and Belda b ), since the transplanting date and for 4 weeks thereafter as a necessary supplement to insufficient T. absoluta eggs and larvae for N. tenuis nymphs to reach maturity (Urbaneja et al., Reference Urbaneja, Tapia and Stansly2005). Procedures for T. absoluta and N. tutae release and plant management were the same as those described for the ‘N. tutae-alone releases’ experiment.
Sampling
Five randomly selected plants in each walk-in cage were monitored weekly for 11 and 9 weeks after transplanting during the ‘N. tutae-alone releases’ and ‘N. tutae-N. tenuis joint releases’ experiments, respectively, beginning 1 week after the first T. absoluta release. T. absoluta eggs and larvae were counted on five leaves selected from the upper-mid third of each of the five selected plants (Calvo et al., Reference Calvo, Lorente, Stansly and Belda2012b ). Additionally, ten leaves were selected from each of the five selected plants and mined area by T. absoluta rated visually as 0, 1, 2, 3, 4, or 5 where 0 was no damage, 1 = 1–25%, 2 = 26–50%, 3 = 51–75%, 4 = 76–99%, and 5 = 100% of the leaf surface damaged, respectively. Additionally, leaflets were collected weekly from each walk-in cage at all plant strata providing ca. 30 second-fourth instar larvae of T. absoluta per walk-in cage. Leaflets were packed in a separate plastic container and labeled by cage. Mines of T. absoluta were opened in the laboratory and inspected using a 40× stereoscopic microscope and classified as parasitized (T. absoluta larvae with N. tutae eggs or larvae), dead (dead larva with no parasitoid eggs, which included host-killing) or alive. Nymphs and adults of N. tenuis were counted on three leaves from the upper third of each of the above-mentioned five selected plants (Calvo et al., Reference Calvo, Bolckmans, Stansly and Urbaneja2009; Arnó et al., Reference Arnó, Castañé, Riudavets and Gabarra2010). Leaves were turned carefully to count first N. tenuis adults and then nymphs. Finally, fruits from all plants were collected at the end of both experiments, counted, and classified as damaged or not by T. absoluta.
Analysis
Incidence of parasitism on T. absoluta was expressed as the number of parasitized larvae by N. tutae observed per cage divided by the number of live and parasitized larvae per cage. Treatment effects on T. absoluta (both experiments) and N. tenuis (second experiment) were analyzed using linear mixed effects models (α = 0.05), with time (weeks after planting) as random factor nested in blocks to correct for pseudoreplication due to repeated measures (Crawley, Reference Crawley2002). Treatments were compared, contingent on a significant model, through model simplification by combining treatments (Crawley, Reference Crawley2002). Differences among treatments in numbers of N. tenuis per leaf at each sampling event and proportion of T. absoluta-damaged fruits at the end of the experiments were evaluated using a two-way analysis of variance and Tukey's test for mean separation (α = 0.05). Numbers of N. tenuis per leaf were log(x + 1) transformed, whereas percentages of affected area and damaged fruits by T. absoluta were arcsin √(x) transformed prior to analysis to stabilize error variance although untransformed values are given in the text.
Results
Experiment 1: N. tutae-alone releases
Eggs and larvae of T. absoluta
More T. absoluta eggs were recorded in cages with the pest only than in cages with parasitoid release (fig. 1a; table 1), with fewest eggs at the higher parasitoid release. Dynamics of T. absoluta larvae were similar to those recorded for eggs during almost the entire experiment, though only the higher rate of N. tutae significantly reduced larval host (fig. 1b; table 1). At the end of the experiment, plants in the absence of N. tutae collapsed due to pest damage and no longer supported larval feeding, whereas plants receiving the lower rate of the parasitoid were still suitable to host T. absoluta larvae, thus resulting in a similar abundance of pest larvae compared with the control at the end of the experiment.
Mean (±SE) T. absoluta eggs (a) and larvae (b) per leaf per week in the ‘N. tutae alone releases’ experiment in each treatment: (1) T. absoluta; (2) 1 N. tutae m−2; (3) 2 N. tutae m−2. First T. absoluta adults were released the week of planting (week 0) and first N. tutae adults 2 weeks later. Evaluations started 1 week after the first T. absoluta release. Treatments (legends) with the same letter were not significantly different (GLMM, P > 0.05).
Pair-wise comparison between treatments in the N. tutae rate Experiment.
Treatments: (1) T. absoluta: T. absoluta only; (2) 1 N. tutae m−2: T. absoluta + 1 N. tutae m−2 released weekly for 7 weeks beginning 2 weeks after planting; (3) 2 N. tutae m−2: T. absoluta + N. tutae m−2 released weekly for 7 weeks beginning 2 weeks after planting. Total mortality included dead larvae by host-killing and parasitism. ** indicates differences between treatments were significant (GLMM, α < 0.05).
Plant and fruit injury by T. absoluta
Leaf area damaged by T. absoluta increased continuously in the absence of N. tutae, reaching ca. 95% at the end of the experiment (fig. 2a). Parasitoid releases significantly reduced plant damage, especially the higher release rate (table 1). More T. absoluta-damaged fruits were found on untreated plants compared with those receiving N. tutae, especially at the higher rate (F 2,6 = 83.2; P < 0.001; fig. 2b). Nevertheless, nearly 60 and 15% of the fruits were still damaged by T. absoluta from plants receiving 1 and 2 N. tutae m−2, respectively.
Percentage (±SE) of mined leaf area by T. absoluta per week (a) and mean percentage (±SE) of fruits affected by T. absoluta (b) in the ‘N. tutae alone releases’ experiment in each treatment: (1) T. absoluta; (2) 1 N. tutae m−2; (3) 2 N. tutae m−2. First T. absoluta adults were released the week of planting (week 0) and first N. tutae adults 2 weeks later. Evaluations of mined leaf area started 1 week after the first T. absoluta release and fruit damage was assessed at the end of the experiment when all fruits were collected. Treatments (legends) (a) and columns (b) with the same letter were not significantly different: (a) GLMM, P > 0.05; (b) Tukey, P > 0.05.
Incidence of parasitism and mortality
Mortality of T. absoluta larvae (excluding parasitism) was lowest and close to zero in the absence of N. tutae, in contrast to plants receiving the parasitoid, presumably due to host-killing. Consequently, greatest mortality was seen with the higher release rate of N. tutae (fig. 3). Incidence of parasitism followed a similar pattern (F 1,39 = 14.392; P = 0.001; fig. 3). Thus, the combined effects of host-killing and parasitism resulted in a greater total mortality of T. absoluta larvae in cages receiving the higher rate of the parasitoid (table 1).
Overall incidence (±SE) of parasitism and mortality (dead larvae) by N. tutae on T. absoluta larvae from leaves samples collected in the ‘N. tutae alone releases’ experiment in each treatment: (1) T. absoluta; (2) 1 N. tutae m−2; (3) 2 N. tutae m−2. Columns with the same small (mortality) or capital (parasitism) letter were not significantly different (Tukey, P > 0.05).
Experiment 2: N. tutae-N. tenuis joint releases
Nesidiocoris tenuis
More N. tenuis per leaf were observed on plants receiving the predator pre- vs. post-planting during most weeks of the experiment (Week 1: F 3,9 = 14.703; P < 0.001; Week 2: F 3,9 = 6.599; P = 0.012; Week 4: F 3,9 = 4.241; P = 0.006; Week 5: F 3,9 = 4.167; P = 0.042; Week 6: F 3,9 = 6.608; P = 0.012; Week 7: F 3,9 = 5.271; P = 0.023; fig. 4), and consequently abundance of N. tenuis over all 9 weeks was greatest on plants receiving the predator before planting, intermediate when the predator was released alone after planting and lowest in cages receiving the predator after planting with supplementary releases of N. tutae (table 2; fig. 4).
Mean (±SE) number of N. tenuis per leaf per week during the ‘Necremnus tutae-N. tenuis joint releases’ experiment in each treatment receiving the predator: (1) T. absoluta; (2) N. tenuis pre-plant; (3) N. tenuis pre-plant + N. tutae; (4) N. tenuis post-plant; (5) N. tenuis post-plant + N. tutae. N. tenuis was released 5 days before or the day of planting in pre-plant or post-plant treatments, respectively. First T. absoluta adults were release the day of planting (Week 0) and evaluations started 1 week later. Treatments with the same letter on the right were not significantly different (GLMM, P > 0.05), whereas treatments with the same letter on each week were not significantly different in that week (Tukey, P > 0.05).
Pair-wise comparison between treatments in the N. tutae + N. tenuis Experiment.
Treatments: (1) Tuta absoluta: T. absoluta only; (2) N. tenuis post-plant: T. absoluta + one N. tenuis per two plants released into walk-in cage the day of planting; (3) N. tenuis post-plant + N. tutae: T. absoluta + N. tenuis released as in treatment 2 and 1 N. tutae m−2 released weekly for 5 weeks beginning 2 weeks after planting; (4) N. tenuis pre-plant: T. absoluta + one N. tenuis per two plants released in transplant trays 5 days before planting; (5) N. tenuis pre-plant + N. tutae: T. absoluta + N. tenuis released as in treatment 4 + 1 N. tutae m−2 released as in treatment 3.** indicates differences between treatments were significant (GLMM, α < 0.05). Tab: T. absoluta; Nt: N. tenuis; Nect: N. tutae.
Eggs and galleries of T. absoluta
Most eggs of T. absoluta were found on plants with T. absoluta only (table 2), except for the last week when plants collapsed due to pest attack and no T. absoluta eggs were found (fig. 5a). Release of N. tenuis reduced T. absoluta eggs numbers, especially when released before planting (table 2). Egg suppression was improved under both release scenarios with no difference between them when combined with N. tutae (table 2). Pre-plant release of N. tenuis plus N. tutae was the most effective in reducing T. absoluta larvae (fig. 5b; table 2) and the combination of N. tutae and post-plant release of the N. tenuis was as effective as pre-plant release of the predator alone. Post-plant application of N. tenuis alone reduced larval numbers significantly compared with no release, but was the least effective treatment among those receiving beneficial insects.
Mean (±SE) T. absoluta eggs (a) and larvae (b) per leaf per week in the ‘Necremnus tutae-Nesidiocoris tenuis joint releases’ experiment in each treatment: (1) T. absoluta; (2) N. tenuis pre-plant; (3) N. tenuis pre-plant + N. tutae; (4) N. tenuis post-plant; (5) N. tenuis post-plant + N. tutae. First T. absoluta adults were released the week of planting (week 0) and first N. tutae adults 2 weeks later. N. tenuis was released 5 days before or the day of planting in pre-plant or post-plant treatments, respectively, and evaluations started 1 week after the first T. absoluta release. Treatments (legends) with the same letter were not significantly different (GLMM, P > 0.05).
Plant and fruit injury by T. absoluta
Mined leaf area increased most rapidly on plants receiving only T. absoluta (fig. 6a). Release of N. tenuis after planting reduced damage, but was less effective than either the combination of post-planting releases of N. tenuis plus N. tutae or pre-plant releases of N. tenuis alone, with similar results from these latter two treatments. The combination of pre-plant releases of N. tenuis with supplementary releases of N. tutae was most effective in reducing plant feeding by T. absoluta. Most fruits were damaged on plants receiving T. absoluta only, with intermediate damage in those receiving only the predator after planting and least damage with the remaining treatments, which were not significantly different from each other (F 4,11 = 4.997; P = 0.014, fig. 6b).
Percentage (±SE) of mined leaf area by T. absoluta per week (a) and mean percentage (±SE) of fruits affected by T. absoluta (b) in the ‘Necremnus tutae-N. tenuis joint releases’ experiment in each treatment: (1) T. absoluta; (2) N. tenuis pre-plant; (3) N. tenuis pre-plant + N. tutae; (4) N. tenuis post-plant; (5) N. tenuis post-plant + N. tutae. First T. absoluta adults were released the week of planting (week 0) and first N. tutae adults 2 weeks later. N. tenuis was released 5 days before or the day of planting in pre-plant or post-plant treatments, respectively. Evaluations of mined leaf area started 1 week after the first T. absoluta release and fruit damage was assessed at the end of the experiment when all the fruits were collected. Treatments (legends) (a) and columns (b) with the same letter were not significantly different: (a) GLMM, P > 0.05; (b) Tukey, P > 0.05.
Incidence of parasitism and mortality
Mortality of T. absoluta larvae was mainly observed on plants receiving N. tutae, with no significant effects of N. tenuis (fig. 7, table 2). These findings reflect low natural mortality of T. absoluta larvae and little larval predation by N tenuis, whereas the parasitoid killed T. absoluta larvae through both host-feeding and parasitism. Pre-plant release of N. tenuis reduced overall incidence of parasitism by N. tutae (F 1,56 = 5.969; P = 0.021), but larval mortality (excluding parasitism) was no different between the two treatments receiving N. tutae: F 1,56 = 0.007; P = 0.932).
Overall incidence (±SE) of parasitism and mortality (dead larvae) by N. tutae on T. absoluta larvae from leaves samples collected in the ‘N. tutae and Nesidiocoris tenuis’ experiment in each treatment: (1) T. absoluta; (2) N. tenuis pre-plant; (3) N. tenuis pre-plant + N. tutae; (4) N. tenuis post-plant; (5) N. tenuis post-plant + N. tutae. Columns with the same small (mortality) or capital (parasitism) letter were not significantly different (Tukey, P > 0.05).
Discussion
Parasitism and host-killing following releases of N. tutae over seven consecutive weeks resulted in decreasing T. absoluta larval populations and consequently reduced plant and fruit injury, especially at the higher release rate (2 wasps m−2 week−1). Dead hostfed larvae on plants receiving N. tutae exceeded parasitism rates either in the absence or presence of N. tenuis (figs 3 and 7 respectively), which correlates with earlier reports (Ferracini et al., Reference Ferracini, Ingegno, Navone, Ferrari, Mosti, Tavella and Alma2012; Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2013) that concurred with this finding. While possibly advantageous for pest control, a high host-killing rate is a source of inefficiency for mass rearing. This is particularly important in the current T. absoluta-tomato plant based-rearing system for N. tutae, which provides the parasitoid with a limited number of suitable hosts for parasitization and which is reduced more by host-feeding. This increases costs per insect and consequently reduces cost-effectiveness. In our study, two N. tutae m−2 per week during 7 weeks provided more reduction of T. absoluta damage, but it was insufficient to reduce plant and fruit damage to an acceptable level (Desneux et al., Reference Desneux, Wajnberg, Wyckhuys, Burgio, Arpaia, Narváez-Vasquez, González-Cabrera, Catalán Ruescas, Tabone, Frandon, Pizzol, Poncet, Cabello and Urbaneja2010). Thus, we conclude that N. tutae would probably require assistance from N. tenuis to provide adequate control of T. absoluta in commercial tomato greenhouses.
Given that N. tenuis can be released either after (post-plant application) or before (pre-plant application) transplanting (Calvo & Urbaneja, Reference Calvo and Urbaneja2004; Calvo et al., Reference Calvo, Bolckmans, Stansly and Urbaneja2009, Reference Calvo, Soriano, Bolckmans and Belda2012a , Reference Calvo, Lorente, Stansly and Belda b , Reference Calvo, Bolckmans and Belda c ), both release methods were evaluated in the second experiment, with and without supplementary releases of N. tutae. Our results confirmed the superiority of the pre- over the after-plant application of N. tenuis both in number of mirids produced during the critical early phase of the crop cycle and ultimately in more efficient reduction of plant and fruit damage by T. absoluta. Addition of N. tutae to the system increased control of T. absoluta with both N. tenuis release methods. Results were comparable in cages receiving the predator after planting and N. tutae to cages with N. tenuis released before planting only and best when the pre-plant application of the predator was combined with N. tutae, which was the only tactic being able to keep defoliation levels under action thresholds (Desneux et al., Reference Desneux, Wajnberg, Wyckhuys, Burgio, Arpaia, Narváez-Vasquez, González-Cabrera, Catalán Ruescas, Tabone, Frandon, Pizzol, Poncet, Cabello and Urbaneja2010). Nevertheless, except in cages treated with N. tenuis after planting only, excess of action threshold did not result in fruit damage. N. tutae exerted an additive effect on control when combined with N. tenuis regardless of the release method, killing ca. 60% of second-third instar larvae, whereas almost no larval mortality was seen with the predator alone. N. tutae and N. tenuis would belong to different functional groups (Northfield et al., Reference Northfield, Crowder, Jabbour, Snyder, Ohgushi, Schmitz and Holt2012), and as expected, their combination provided better control than N. tenuis alone released by either method. With both natural enemies present, all immature stages of T. absoluta except pupae are subject to attack. Such complementarity could be particularly important at the critical beginning of the crop cycle. Nevertheless, substantial improvement would be needed to justify adding another biological control agent into the system due to pre-plant release of N. tenuis alone was sufficient to prevent fruit damage under our experimental conditions, obviating the necessity of N. tutae. Differently, Calvo et al. (Reference Calvo, Lorente, Stansly and Belda2012b ) found that the combination of the egg parasitoid Trichograma achaeae Nagaraja & Nagarkatti (Hymenoptera: Trichogrammatidae) and/or Bacillus thuringiensis Berliner var. kurstaki (Bt) did not increase control of T. absoluta over pre-plant release of N. tenuis alone. Although Bt has been demonstrated effective against T. absoluta either alone or in combination with natural enemies (González-Cabrera et al., Reference González-Cabrera, Mollá, Montón and Urbaneja2011; Mollá et al., Reference Mollá, González-Cabrera and Urbaneja2011), it acts on T. absoluta larvae primarily when they exit their galleries. Contrarily, N. tutae is a specialist parasitoid that can reach T. absoluta larvae inside the plant tissue (and at some extent also outside the plant tissue with high pest densities; FJ Calvo, Personal observation). This provides a better opportunity for N. tutae to attack and kill T. absoluta compared with Bt, which has little contribution when few T. absoluta eggs escape a well-established N. tenuis population (Calvo et al., Reference Calvo, Lorente, Stansly and Belda2012b ).
Likewise, T. achaeae, which naturally occurs in the Mediterranean region (Zappalá et al., Reference Zappalá, Biondi, Alma, Al-Jboory, Arnó, Bayram, Chailleux, El-Arnaouty, Gerling, Guenaoui, Shaltiel-Harpaz, Siscaro, Stavrinides, Tavella, Vercher Aznar, Urbaneja and Desneux2013), did not improve control of T. absoluta when N. tenuis was released before planting in nurseries i.e. when the predator was already well established before arrival of T. absoluta or T. achaea releases began (Calvo et al., Reference Calvo, Lorente, Stansly and Belda2012b ). Macrolophus pygmaeus (Rambur) (Heteroptera: Miridae), which is biologically comparable with N. tenuis, was also found to reduce effectiveness of T. achaeae against T. absoluta, but not at a greenhouse level (Chailleux et al., Reference Chailleux, Bearez, Pizzol, Amiens-Desneux, Ramirez-Romero and Desneux2013, Reference Chailleux, Wajnberg, Zhou, Amiens-Desneux and Desneux2014b ). Contrarily, Desneux et al. (Reference Desneux, Wajnberg, Wyckhuys, Burgio, Arpaia, Narváez-Vasquez, González-Cabrera, Catalán Ruescas, Tabone, Frandon, Pizzol, Poncet, Cabello and Urbaneja2010) found that supplementary releases of T. achaeae following release of N. tenuis after planting improved control of T. absoluta. Nevertheless, attempts to use different Trichogramma species against T. absoluta in Europe (Zappalá et al., Reference Zappalá, Biondi, Alma, Al-Jboory, Arnó, Bayram, Chailleux, El-Arnaouty, Gerling, Guenaoui, Shaltiel-Harpaz, Siscaro, Stavrinides, Tavella, Vercher Aznar, Urbaneja and Desneux2013) have provided little contribution to the control of the pest, and thus required periodical inundative releases and/or combination with other control agents, primarily mirid predators (Chailleux et al., Reference Chailleux, Desneux, Seguret, Do Thi Khanh, Maignet and Tabone2012, Reference Chailleux, Desneux, Arnó and Gabarra2014a , Reference Chailleux, Wajnberg, Zhou, Amiens-Desneux and Desneux b ). Rates, timing, methods, and frequency of natural enemy release as well as synchronization between prey and predator, abiotic factors (humidity, photoperiod, temperature, etc.), and pesticide use can affect control capacity of a natural enemy (Collier & Van Steenwyk, Reference Collier and van Steenwyk2004; Stiling & Cornelissen, Reference Stiling and Cornelissen2005; Crowder, Reference Crowder2006; Desneux et al., Reference Desneux, Decourtye and Delpuech2007) and thus could help to explain differences among above-mentioned results. Additionally, N. tenuis exhibits better life-history traits than M. pygmaeus when fed on T. absoluta eggs (Mollá et al., Reference Mollá, Biondi, Alonso-Valiente and Urbaneja2014), suggesting that N. tenuis attacking T. absoluta would leave even less room for T. achaeae than M. pygmaeus.
In a short-term interaction study, Chailleux et al., (Reference Chailleux, Desneux, Arnó and Gabarra2014a ) observed a strong negative effect of kleptoparasitism on Stenomesius japonicus Ashamed (Hymenoptera: Eulophidae), another ectoparasitoid of T. absoluta larvae, when combined with M. pygmaeus. Similarly, N. tutae as S. japonicus parasitizes and kills T. absoluta larvae by host-killing (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2013; Chailleux et al., Reference Chailleux, Wajnberg, Zhou, Amiens-Desneux and Desneux2014b ), which reduces future prey for the predator. Nevertheless, N. tutae uses oldest T. absoluta larval stages for parasitization (Calvo et al., Reference Calvo, Soriano, Bolckmans and Belda2013), whereas predation of N. tenuis is reduced on these stages (Urbaneja et al., Reference Urbaneja, Montón and Mollá2009). The predator can also reduce availability of future hosts for the parasitoid (predation) as well as can feed directly on parasitoid larvae (intraguild predation). Consequently, excluding kleptoparasitism of which we have no estimate, a putative reduction of future hosts by unidirectional omnivorous intraguild predation, where N. tenuis could feed directly on N. tutae ectoparasitoid larvae, or resource competition between both species were probably the most important interactions between the predator and the parasitoid during our study.
In summary, N. tutae alone did not reduce plant or fruit damage by T. absoluta under an acceptable level, whereas releasing N. tutae on plants that had been inoculated with N. tenuis provided adequate control. Nevertheless, this combination just slightly improved pest control in leaves over the pre-plant release of N. tenuis. Additionally, pre-plant release of N. tenuis also provides excellent control of whiteflies (Calvo et al., Reference Calvo, Bolckmans and Belda2012c ) and thus, under most situations, this method would be the most cost-effective option for an IPM program in Mediterranean tomato greenhouses, since strategies involving fewer natural enemy species are often simpler and cheaper. Thus, while our study demonstrates potential for integrating N. tutae into biologically-based pest management systems for tomato, the additional cost of such addition has yet to be justified. An efficient rearing method for N. tutae might open the door for future inclusion of N. tutae within an integrated biological control system. Meanwhile, conservation biological control based on the manipulation of the ubiquitous N. tutae could be a feasible sanitary measure against T. absoluta in many tomato areas of Europe.
Acknowledgements
The authors thank Marco Gebiola (CNR – Instituto per la Protezione Sostenibile delle Plante, Portici, Italy) for identifying morphologically and genetically the species of the parasitoid used in these studies and three anonymous reviewers for providing valuable remarks.
