Flora surrounding rice fields as a source of alternative prey for coccinellids feeding on the pests of rice

Coccinellids are effective predators and a key component of the predator guild in rice ecosystems. In order to enhance their effi cacy, a study was undertaken to assess the seasonal movement of coccinellids into rice fi elds and the role of the surrounding fl ora on their colonization. The seasonal abundance of coccinellids and their prey was recorded on the rice crop and the surrounding fl ora at fortnightly intervals from 2012 to 2015. Coccinellid prey range was assessed using PAGE electrophoresis. The herbivorous insects associated with weeds were Aphis gossypii Glover, Aphis craccivora (Koch), Cicadulina bipunctata (Melichar), Schizaphis graminum (Rondani), Sitobion sp., Thaia oryzivora Ghauri and Zygina maculifrons Matsumura. Of the species of coccinellids recorded in rice fi elds, Harmonia octomaculata (Fabricius), Micraspis discolor (F.), Propylea dissecta (Mulsant), Coccinella transversalis Fabricius, Cheilomenes sexmaculata (Fabricius), Scymnus nubilus Mulsant and Brumoides suturalis (Fabricius) were also recorded on weeds. The esterase profi les indicated that the leafhoppers and aphids on the weeds were the prey of the coccinellids before they colonized the rice fi elds. The coccinellids recorded on the weeds showed bands corresponding to the insects present on the weeds. Beetles collected from rice fi elds had different bands, some of which corresponded to the green leafhopper (GLH) Nephotettix virescens Distant, the brown planthopper (BPH) Nilaparvata lugens Stal and white backed planthopper (WBPH) Sogatella furcifera Hovarth infesting rice. In addition, some bands corresponded to hoppers and aphids that were present on the surrounding fl ora. The results indicate the importance of surrounding fl ora in the conservation and colonization of rice fi elds by coccinellids.


INTRODUCTION
Generalist predators play an important role in regulating insect pests in agricultural ecosystems (Joon & Seung, 2001).Although their ecological role in pest management was underestimated in the past, recently, the importance of generalist predators has been reconsidered (Symondson et al., 2002).In particular, the development of molecular techniques for detecting the presence of prey in the guts of the predators has enhanced our understanding of the trophic links in agricultural ecosystems (Pompanon et al., 2012).By manipulating the abundance and assemblage of generalist predators it is possible to signifi cantly affect the abundance of pests (Symondson et al., 2002).For example, a recent meta-analysis of small-scale fi eld studies reveal that an increase in plant diversity, realized by a minimal control of weeds, fosters the activity of generalist predators in suppressing herbivorous pests (Dassou & Tixier, 2016).Generalist predators can thus be a key component of conservation biological control.Simple techniques for conservation of indigenous natural enemy diversity can abundance was averaged for the fi ve locations in order to reveal the overall seasonal trend.

Electrophoretic assessment of the prey of the coccinellids
A polyacrylamide gel electrophoresis (native PAGE) method, modifi ed from Murray & Solomon (1978) and Giller (1982Giller ( , 1984Giller ( and 1986) ) was used to characterize banding patterns for different species of prey and predator.Esterases were studied because they are ubiquitous enzymes in organisms and as they have species specifi c patterns they can be used to discriminate between the different species of prey (Murray & Solomon, 1978).
Two coccinellid species, Harmonia octomaculata (Fabricius) and Micraspis discolor (F.) and three hopper species, the green leaf hopper (GLH) Nephotettix virescens Distant, the brown plant hopper (BPH) Nilaparvata lugens Stal and the white backed planthopper (WBPH) Sogatella furcifera Hovarth, were used to confi rm the unique banding patterns for coccinellids and their prey.Adults of each species were released individually into 100 × 41 mm polystyrene Petri dishes with a 4-cm diameter ventilation hole in the lid covered with a 0.05 mm fabric mesh (SPL Life Sciences Inc., Korea).Five third-instar nymphs of each of the three species of planthopper from glasshouse cultures were placed in the Petri dishes.There were fi ve replicates of each combination of predator and prey, totalling 15 adults per taxa.Beetles were removed for analysis after they were observed to fed on one hopper.In order to assess the feeding behaviour of coccinellids collected from fi eld, the insects were analyzed in the following categories: Adult beetles of H. octomaculata, M. discolor, Propylea dissecta (Mulsant), Coccinella transversalis Fabricius, Cheilomenes sexmaculata (Fabricius), Scymnus nubilus Mulsant and Brumoides suturalis (Fabricius) were collected in the fi eld between 8:00 and 9:00, placed in a refrigerator below 4°C within 1 h of collecting and then quickly analyzed in order to preserve their prey content.
Adult coccinellids of all the species listed above were previously starved for 24 h so that their guts did not contain any remnants of prey.This allowed the identifi cation of bands specifi c for each predator.
The signature banding pattern for esterase activity was identifi ed using the Bio-Rad Mini-PROTEAN-3 apparatus following the procedure of Murray & Solomon (1978).Individual predators and prey were macerated in small glass wells using a glass rod.The sample buffer consisted of 6-10 μl aliquots of a 0.1 M Tris/ HCl (pH 7.0) or a 0.15 M phosphate pH 8.2 buffer, with 0.2% Triton X-100 and 10% sucrose.The samples were later centrifuged for 10 min at 6500 rpm and the supernatant transferred using a micropipette to the sample holders at the top of the gel.5-10 μl of supernatant from each sample was loaded onto the gels, and a constant current of 90 V was applied to the gels for 2.5-3 h, or until the molecular weight standard reached the bottom of the gel.After electrophoresis, the gel was stained briefl y in sodium phosphate buffer pH 7.4 containing 0.1% fast blue RR and 0.05% α-napthyl acetate for 30 min, and then photographed.The bands recorded for the fi eld-collected coccinellids were screened against the bands of each individual prey, in order to detect the presence of any prey remnants in the predator.standing of terrestrial food webs and the alternative food items that enhance the survival of these predators during periods when their main prey is scarce.
The natural food of coccinellids can be ascertained in several ways (Harwood & Obrycki, 2005;Weber & Lundgren, 2009;Hodek & Evans, 2012).Most studies are based on the dissection of their gut contents.Other techniques such as radio-labelling of prey, microscopic examination of faecal samples, electrophoresis, stable isotope analysis and ELISA using antibodies have been used in fi eld studies with varying degrees of success (reviewed by Michaud & Harwood, 2012).Recent advances in molecular biology have resulted in the development of group-specifi c or species-specifi c DNA primers, which can be used in standard or multiplex PCR to identify the prey in a predators' gut (Michaud & Harwood, 2012).While this method is the most accurate for detecting a given type of prey it is expensive when screening a large number of predators for several types of prey.In addition, the development of an appropriate primer is rather time-consuming.In fact, primer design based on sequenced individuals, evaluation of prey specifi city using a set of non-target organisms, primer sensitivity and laboratory feeding trials to estimate the DNA half-detectability time should be done before undertaking a fi eld study (Harwood & Obrycki, 2005;Hodek & Evans, 2012).Considering the small size and cryptic nature of their prey, an analysis of the prey-specifi c enzymes in the gut of generalist predators, such as coccinellids, could be used to more rapidly identify their prey and defi ne their feeding patterns.In this study, we used an electrophoretic approach rather than PCR-based methods.Our approach proved to be useful in a fi rst assessment of the predators' food.In particular, when the prey and predator enzyme banding patterns are distinct, electrophoresis can be a cheap and quick tool for the detection of prey items in fi eld-caught predators (Solomon et al., 1996).It has already been shown that electrophoretic techniques can be used to identify the prey of aquatic and terrestrial invertebrate predators (Lister et al., 1987;Dicke & De Jong, 1988;Schulz & Yurista, 1995;Corey et al., 1998).In order to determine the importance of alternative prey in conservation biological control we used an electrophoretic approach to detect the predation by coccinellids on essential and alternative prey occurring in rice fi elds and surrounding areas.

Spatial and temporal distribution of coccinellids
Field and laboratory studies were carried out on the experimental farm of the ICAR-Indian Institute of Rice Research, Hyderabad, India for three years (2012)(2013)(2014)(2015).In this region rice is grown during the wet season (July to October) and dry season (January to April).This study, therefore, continued throughout each year in order to include the two cropping seasons and the non-crop period in between.
The coccinellids were sampled every fortnight by sweep netting and visual counts in rice fi elds and on fi eld bunds at fortnightly intervals at fi ve different locations.Sampling was done in the morning up to 10:00.Ten sweeps were made in each rice fi eld in plots of 100 m × 25 m and 40 m of fi eld bund.Total coccinellid The ranking procedure suggested by Hobson (1974) and adapted for studying the gut content of fi sh (Eya et al., 2011) was used based on the presence or absence of the prey items in the predators' gut as follows: where RI = Ranking Index of each prey item, A = number of coccinellids per taxa containing each prey item, B = total number of coccinellids per taxa with stomach contents, C = percentage of each prey item, which can be computed as: No.of each prey detected Total of all prey items combined

Spatial and temporal distribution of coccinellids
Around 30 weeds, including grasses sedges and broad leaved plants, were surveyed on fi eld bunds and a total of 1620 coccinellids were collected from twelve species of plants over a period of three years.In total, nine species of coccinellids were recorded on weeds and on rice: B. suturalis, C. sexmaculata, C. transversalis, H. octomaculata, Illeis cincta (Fabricius), M. discolor, P. dissecta, Scymnus latemaculatus Motschulsky and S. nubilus Mulsant.All of them were recorded on weeds, while only seven were also recorded on rice.Illeis cincta was recorded on Acmella uliginosa where it was feeding on fungal spores.Direct observations confi rmed feeding on insects or pollen on twelve species of weeds (Table 1).Feeding activity was also confi rmed by the feeding studies conducted in the laboratory.In all the three years, H. octomaculata, P. dissecta, C. transversalis, C. sexmaculata, S. latemaculatus and B. suturalis were recorded frequently on weeds during the early part of the season and on rice in both the wet and dry seasons (Fig. 1).Conversely, M. discolor was not recorded initially on weeds in the wet season, which is the main season for the rice crop.It was detected on rice during its tillering stage and then moved to weeds when the rice crop reached the harvesting stage (Fig. 1).In the dry season, however, a few were recorded on weeds that had moved from the rice crop to weeds.The phytophagous insects associated with weeds included A. gossypii, A. craccivora, Schizaphis graminum and Sitobion sp.(all Hemiptera: Aphididae); C. bipunctata, T. oryzivora, Z. maculifrons (Hemiptera: Cicadellidae) and all stages of Altica cyanea (Weber) (Coleoptera: Chrysomelidae).All species of coccinellids were recorded on weeds when the rice crop neared the harvesting stage both in December-January following the wet season and in April-May following the dry season.The numbers decreased over time, but coccinellids remained on the weeds until the start of the next rice crop.The presence of coccinellid eggs on the surrounding fl ora increased greatly as the rice approached the harvesting stage.Many adults of M. discolor and H. octomaculata were recorded on Acmella uliginosa (Sw.)Cass., Ageratum conyzoides, Caesularia axillaris Roxb., L. and Eclipta alba (L.) Hassk.(Family: Asteraceae) and Echinocloa colona (L.), (Family: Poaceae), when the rice crop was near the harvesting stage and the hopper population had also started diminishing in the fi eld.The adults were observed mostly feeding on pollen from fl owers and

Electrophoretic assessment of the prey of the coccinellids
Specifi c unique bands in the esterase banding patterns for the different predators and their prey were used to identify the different predators and prey (Figs 2 and 3).The comparison of the esterase banding pattern of H. octomaculata collected from rice fi elds (Fig. 2: Lanes 11-15) and weeds (Fig. 2: Lanes 6-10) with that of prey and starved coccinellids revealed that all the adults had additional bands that were not present in starved adults.These bands corresponded to C. bipunctata and A. craccivora from the leguminous creeper, M. atropurpureum, S. graminum collected from grass, C. barbata Sw and the three ricehoppers, mainly GLH and a few BPH and WBPH, which reveals their possible role as prey of H. octomaculata.Similarly, a few fi eld collected M. discolor individuals showed bands that corresponded with the three species of ricehoppers.In addition, we also recorded other bands that did not correspond with the prey studied indicating that this coccinellid fed on other insects or possibly plant tissues.Coccinella transversalis exhibited bands that correspond to A. gossypii collected from Tridax procumbens L., A. craccivora from leguminous weeds and S. graminum from grasses, indicating that this predator mainly feeds on aphids (Fig. 2).C. sexmaculata had bands corresponding to only A. craccivora found on leguminous weeds where all stages of this predator were recorded.The results clearly indicate that leafhoppers and aphids infesting weeds were the main or alternative prey on which the coccinellids fed before moving to rice fi elds.
The ranking index of the coccinellids recorded in rice fi elds with regard to the prey recorded on weeds and in the rice fi elds was 64% for H. octomaculata, which showed bands corresponding to all the ricehoppers, N. lugens, S. furcifera and N. virescens, and T. oryzivora, C. bipunctata and Z. maculifrons on weeds, and the aphids A. craccivora, S. graminum and A. gossypii on weeds.The green leafhopper N. virescens was the preferred prey of this coccinellid with a ranking index of 31.25 (Table 2A, B), followed by the brown planthopper N. lugens and T. oryzivora (Ranking index of 12.50).The higher preference for the rice pests, GLH and BPH, and all the prey on rice and weeds indicate that H. octomaculata might be a suitable candidate for conservation biological control programmes.There was prey in only the guts of 22% of the M. discolor adults, the most abundant coccinellid in the rice fi elds.A greater ranking index (27.27)was recorded for A. gossypii than for all three ricehoppers.Sixty to eighty % of the guts of C. transversalis and C. sexmaculata contained some prey.C. transversalis showed higher preference for A. craccivora, A. gossypii and S. graminum developing on weeds, while C. sexmaculata only exhibited bands corresponding to A. craccivora.Brumoides suturalis also had only consumed A. craccivora.On the other hand, P. dissecta showed equal preference for the T. oryzivora, Z. maculifrons and S. graminum infesting weeds.Scymnus spp.were aphidophagous as only the two aphids A. craccivora and S. graminum were detected in their guts.

DISCUSSION
Surrounding vegetation and patches of plants seem to have a direct effect on the density of coccinellids by infl uencing their immigration and emigration patterns (Grez & Prado, 2000).The uncultivated patches of vegetation give stability to the agro-ecosystem by providing refuges, alternative prey, nectar and pollen for natural enemies, thus indirectly contributing to crop pest management (Landis et al., 2000;Sutherland et al., 2001).In order to utilize these reservoirs for conservation biological control it is necessary to understand the population dynamics of natural enemies such as coccinellids and their movement into crops from a particular resource or habitat (Gurr et al., 2000).In  this study the movement of coccinellids from rice to weeds was especially marked just before the rice harvest, which also corresponds with a decline in the hopper populations both in the wet and dry seasons (Win et al., 2011).Villegas et al. (2013) report similar movements of coccinellids from thistles in fi eld margins into alfalfa fi elds, which is synchronised with the changes in the availability of prey on the crop and weeds.The coccinellid population moving from weeds caused a reduction in the aphid population on alfalfa during November.Among the coccinellids included in this study, H. octomaculata has the widest prey spectrum on both the rice crop and weeds.In this respect, it is reasonable to use this species in conservation biological control programmes of planthopper pests of rice.
Feeding behaviour of predators can be assessed to a certain extent by direct observations of predation in the fi eld, but is unlikely to identify all the possible interactions between predators and their possible prey spectrum (Harwood & Obrycki, 2005).Predation remains one of the most diffi cult ecological processes to study.A commonly used approach is the analysis of extracts of homogenized predators (or just their guts) using PAGE gels to record enzyme activity, especially that of esterases.A similar approach was used to assess predation by spiders in the fi eld (Amalin et al., 2000).However, Amalin et al. (2000) only include in their analysis the band patterns recorded for spiders fed with particular prey items in a laboratory.In another study, the predation of Coccus viridis in coffee plantations by the foliage dwelling spiders, Oxyopes sp., Dieta virens (Thorell), Olios milleti (Pocock, 1901) and Telomonia dimidiata (Simon, 1899), was detected using esterase banding patterns (Senthil Kumar & Regupathy, 2009).Similarly, Traugott (2003) use an electrophoretic approach to analyse the prey spectrum of larvae and adults of Cantharis spp.collected in an open fi eld system.The combination of our laboratory and fi eld results clearly revealed that coccinellids feed on prey present on the rice crop and surrounding fl ora.The changes in their seasonal abundance showed that the populations on the crop and weeds were associated with seasonal changes in the spatial availability of prey, with weeds acting as a good reservoir for coccinellids during the lean season when prey is less abundant on rice.Selective conservation of the surrounding fl ora can enhance the fi tness of natural enemies and lead to more effective pest management (Gurr et al., 2010;Shanker et al., 2013b;Horgan et al., 2016).Simple techniques for conserving non-crop fl ora will enhance the presence of native natural enemies and help create healthy and sustainable ecosystems and a possible reduction in the use of insecticides.

Fig. 1 .
Fig. 1.Seasonal abundance of the species of coccinellids recorded in rice fi elds and on the surrounding fl ora.Abundances were recorded at fortnightly intervals with each point representing the mean of fi ve replicates collected over a period of three years.

Table 1 .
Visual observations of feeding by coccinellids on insects associated with weeds growing on rice bunds.