Efficacy of Common Carp and Nile Tilapia as biocontrol agents of rice insect pests in the Philippines

Efficacy of Common Carp and Nile Tilapia as biocontrol agents of rice insect pests in the

Philippines

M. Halwarta, J.A. Litsingerb*, A.T. Barrionc, M.C. Virayd and G. Kaulee

aFisheries and Aquaculture Department, Aquaculture Service, Food and Agriculture Organization, Via delle Terme di Caracalla00153, Rome 06 57051, Italy; b1365 Jacobs Place, Dixon, CA 95620, USA; cPhilippine Rice Research Institute (PhilRice)Maligaya, Science City of Munoz, Nueva Ecija 3119, Philippines; dFreshwater Aquaculture Center, Central Luzon State University,Science City of Munoz, Nueva Ecija 3120, Philippines; eInstitute of Landscape Planning and Ecology, University of Stuttgart,Keplerstrasse 11, 70174 Stuttgart, Germany

(Received 30 December 2011; final version received 13 June 2012)

Replicated field plot trials were carried out to evaluate the role of fish as biocontrol agents of insect pests of rice inthe Philippines. Evidence from suction sampling, examination of fish gut contents, and rice plant damage assessmentstrongly suggest that Common Carp Cyprinus carpio (L.) and Nile Tilapia Oreochromis niloticus (L.) do significantlyreduce the densities of stemboring moths and chironomid midges, while the abundance levels of other arthropods arenot significantly affected. This impact by fish often was more important in the wet season when heavy rains and windlikely caused the arthropods to become dislodged from the plant and fall onto the water. Predation by fish wasparticularly evident during the first 45 days of the crop during the wet season when adult chironomid midgeabundance was high. The fish gut content analysis revealed that all insect guilds are represented in the diet of both ofthe fish species regardless of whether the life-cycle of the arthropod prey is aquatic, semi-terrestrial, or terrestrial.Thus, fish in rice fields are part of the assemblage of natural enemies of rice pests. While fish alone cannot completelysuppress rice pest populations below the economic threshold, their significant contribution should not be overlooked.

Keywords: biological control; insect pests; insect predators; insect parasitoids; rice-fish culture

Introduction

The practice of rearing fish in rice fields is at least a2000-year-old tradition in large parts of Asia (Bray1986; Fernando 1993; Cai et al. 1995). As a result ofthe Green Revolution in the 1960s, long-duration,traditional rice cultivars were replaced by semi-dwarfsof shorter duration, and insecticides were seen asindispensable for achieving high yields (Litsinger2008); these activities led to a rapid decline ofcultured fish (Moulton 1973; Koesoemadinata 1980;Li 1988). As fish are highly sensitive to pesticides, thewide-scale adoption of pesticides by farmers in rice-growing communities is often cited as one of themajor constraints upon the popularization of rice-fishfarming (Koesomadinata 1980; Cagauan and Arce1992). Irrigation water often flows from field to field;thus, even farmers who do not use pesticides areaffected. However, beginning in the 1980s severalfactors including the use of pest-resistant rice culti-vars, the realization of the importance of naturalenemies in suppressing and regulating pest popula-tions, and integrated pest management, all mitigatedagainst the need for insecticides (Heong and Schoenly1998). Thus the door was opened for a rebirth of rice-fish farming.

Rice-fish farming is often reported to be a mutuallysupporting system. It has been observed that fish growbetter in rice fields than in ponds (Ardiwinata 1957),while a rice crop benefits from the presence of fish interms of a more diverse source of nutrients along withreduced rice pest incidence (Coche 1967; Litsinger1993; Halwart 1994). A wide diversity of rice pests hasbeen reported to be food for fish: insects, snails, weeds,and even sclerotia of plant-pathogenic fungi such as thecausal agent of sheath blight (Halwart 1994). Thestudies that provided these data, however, often werenot rigorously carried out, particularly regarding aproper experimental design and appropriate statisticalanalyses of results (Halwart 1994); thus, the conclu-sions drawn were often based on anecdotes or theywere surveys of adjacent fields, each having differentmanagement practices.

Despite the fact that rice and fish are the mainstaysof local diets, there is not a tradition of fish culture inrice fields in the Philippines. This is due mainly to alack institutional support and the willingness of farm-ers to collectively raise fish, as well as the need toprotect against theft and to refrain from pesticide usein the watershed. Rainfed wetland rice farmers have along history of capturing aquatic wildlife in rice fields

*Corresponding author. Email: [email protected]

International Journal of Pest ManagementVol. 58, No. 4, October–December 2012, 331–346

ISSN 0967-0874 print/ISSN 1366-5863 online

� 2012 Taylor & Francis

http://dx.doi.org/10.1080/09670874.2012.705918

http://www.tandfonline.com

(especially soon after harvest when the fields are drai-ned) as a source of protein – for example, fish, frogs,clams, snails, and shrimps – which enter rice fields fromrivers, canals, and land run-off after storms (Halwart2006). The diversity of fauna in rice fields is comparableto that found in marshlands. This practice is differentfrom the deliberate stocking of fish in rice fields, wherethe yields obtained are low. Nowadays, there arenational institutions which culture and provide finger-lings to farmers for the purpose of rice-fish farming.

The food of the two most popularly cultured fish,Common Carp Cyprinus carpio L. and Nile TilapiaOreochromis niloticus (L.), often consists of immatureand mature insects, grass seeds, and assorted aquaticflora (Fernando 1993, 1996). Carp is a benthos feederwhile tilapia prefers to take prey from the water column.The feeding habit of Nile Tilapia is characterized aspredominantly phytoplanktivory with a preference forblue-green algae. Invertebrates are found at very lowdensities in fish guts (Harbott 1982), but their role as adietary component is incompletely known. Bowen (1982)stated that the inclusion of invertebrates in the diet maybe an important variable in the feeding strategy oftilapias. More quantitative data are needed to describecircumstances as to how prey is selected. Thus, acollaborative project was organized to overcome thesedeficiencies by the Freshwater Aquaculture Center ofCentral Luzon State University (FAC/CLSU), theInternational Rice Research Institute (IRRI), and theInstitute of Landscape Planning and Ecology of theUniversity of Stuttgart in Germany with support fromthe World Fish Center. The purpose of the study was toexpand on the previous studies, to document thearthropod food items of CommonCarp andNile Tilapia,as well as provide more evidence of the benefit of fish inrice pest control. The expected outcome, if favourable,would promote rice-fish culture and further reduce theneed for pesticides, and also slow the destruction ofwetlands for expansion of aquaculture or rice culture.

Common Carp and Nile Tilapia are both present inthe Philippines and are important to aquaculture inAsia. Both are adapted to shallow ponding, aredependent on detrital aggregates for food, and canadjust their feeding preferences (Spataru 1978). Thesetwo fish species were compared to a non-fish control inexperimental plot trials over four seasons on the FAC/CLSU campus under natural and artificial infestationwhere fish ponds were constructed. Artificial infesta-tion took place with yellow stemborer Scirpophagaincertulas and rice leaffolder Cnaphalocrocis medinalis(both are pyralid moths). Various arthropod samplingmethods were employed to measure moth and otherarthropod densities over each season. Timing of thetrials reflected nearby farms that practise double-cropped transplanted rice culture. As the aim was tocharacterize fish predation of various arthropod guildsin typical rice fields, modern pest-resistant rice cultivarswere used.

Materials and methods

Screens were placed to exclude indigenous carnivorousfish from the permanent experimental fields maintainedat the research station. A refuge trench or pond wasconstructed in one part of each field 0.5 m deep thatoccupied about one-tenth of the 200-m2 field plots.After 2 weeks, water was maintained for the fish, at adepth of 10 cm. Deep ponding controls weeds; phyto-plankton oxygenates the water. There were two ricecrops per year and trials were conducted on four crops.

Field trials

Three treatments were set out in rice field plots in arandomized complete block design with six replicateseach season. All plots were maintained at 10 cm waterdepth using water from a well. Earthen bunds delineatedthe plots. Water inlets and outlets to individual plotswere covered by wooden frames with wire mesh screensto keep stocked fish in and prevent wild predatory fishfrom entering. Rice varieties used were IR64, a 112-dayvariety in the two crops of 1991, and IR42, a 138-dayvariety in the wet season (WS) of 1992, and IR72, avariety with 110-day maturity in the dry season (DS) of1992. These varieties are resistant to brown planthopperNilaparvata lugens and green leafhopper Nephotettixvirescens and are moderately resistant to stemborers.Rice was transplanted with 5 to 8 seedlings per hill, withhills at a 20 cm6 20 cm distance in both directions.Inorganic 90-45-45 kg NPK/ha was applied in two splitswith N and PK basal and a second N application at 40days after transplanting (DT). No pesticides were usedand fields were hand weeded. Arthropods collected fromthe field were identified and counted in the laboratory bytechnicians trained in the Taxonomy Laboratory ofIRRI Entomology Department by A.T. Barrion.

Fish

Prior to introducing fish, the field was seined and thewater electrified biweekly (Regis et al. 1981; Zalewski1985) to remove all indigenous fish. The fish fry weresupplied by the Philippine Bureau of Fisheries and Aqua-tic Resources, the Tanay Carp Hatchery, and FAC’s ownstock and were introduced into the field plots, 7–16 DT.Individual initial fish weight was estimated from a 20%sample of the stocked fish one day before stocking. In1991 the introduced carp and tilapia weighed an averageof 19.9 g and 24.7 g/fish, respectively, in the DS, and 9.0 gand 2.6 g/fish in the WS. In 1992 DS, carp and tilapiaweighed 7.7 g and 16.4 g/fish, respectively, and 1.5 g and4.8 g/fish in the WS. Fish were stocked at 10,000 fish/ha,except in the 1991 WS with only 5000 fish/ha.

Vacuum sampling of arthropods

A motor driven, portable D-Vac1 suction machine(EcoBlower D-Vac Model 122 http://www.rinconvitova.

332 M. Halwart et al.

http://www.rinconvitova.com/d-vac.htm

com/d-vac.htm) was used to remove arthropods fromrice foliage. Arthropods were vacuumed from rice foliagebiweekly between 7 a.m. and 10 a.m. during fourconsecutive seasons. A conically shaped Mylar1 plasticcylinder was placed over each randomly selected hill; 20hills were sampled per replication. Arthropods weretransferred from a plastic bag to a cyanide killing jar andpreserved in vials of 70% ethanol for later identificationin the laboratory using stereo-microscopes. Arthropodswere grouped into guilds: rice pests, predators, para-sitoids, detritivores, and ‘‘indifferents’’. Detritivores alsoincluded plankton feeders and were composed of mainlychironomid midges, while indifferents included non-ricefeeding herbivores as well as transients/tourists that didnot fit into the other categories. Some insect species wereplaced into more than one guild. For example, corixidsare predators, detritivores, and ‘‘pests’’ that can feed onroots of rice seedlings. The ephydrid flyNotiphila is botha detritivore and a herbivore of grasses, and wasclassified as an indifferent. Chironomids are knowndetritivores and plankton feeders, and like corixids feedon roots of rice seedlings. In some countries chironomidlarvae that feed on roots of rice seedlings cause loss ofstand in directly seeded rice (Clement et al. 1977).Feeding injury from corixids and chironomics isconsidered to be sub-economic in the Philippines(Barrion and Litsinger 1984).

Sticky traps

Sticky traps were set out in the three treatments withand without fish only in the 1992 WS crop on IR42.Transparent Mylar1 sheets were cut into 0.25-mdiameter circles with an opening cut in the centre tofit around a rice hill. Pieces of Styrofoam1 wereattached to the bottom to make the disk float. Theupper surface was coated with Tanglefoot1 (http://www.contech-inc.com), a sticky substance designed toimmobilize arthropods. Sampling was done biweeklyfrom 22–90 DT. Traps were placed around threerandomly selected rice hills per plot, and on eachsampling occasion remained in the field for 12 h, oneset during the day, 6 a.m. to 6 p.m., and the other set atnight, 6 p.m. to 6 a.m. After each sampling period thesticky traps were taken to the laboratory for arthropodidentification and counting. Specimens were cleaned inkerosene as needed. In most cases identification was tothe family level only.

Fish gut analysis

The method for examining fish gut contents to notefood preferences was based on that developed byHyslop (1980). Fish (298 in the DS and 268 in the WS)were sampled six times per season at 12–16-dayintervals from 36–106 DT in the DS and 31–99 DTin the WS. Sampling occurred between 8 a.m. and 10a.m. by seining or electrofishing. It was considered

adequate to sample both Nile Tilapia and CommonCarp in the morning hours. Sampling of fish in ricefields is difficult and sometimes seining had to berepeated several times until a sufficient number of fishwas caught. Seining causes stress to fish in the ricefields and should not be performed during the hours ofpeak sunlight. On each sampling date a total of 5 fishper plot were removed and replaced (in exceptionalcases, even repeated sampling was unsuccessful andfewer fish were taken as indicated). Fish wereimmediately killed and placed in a deep freezer in thelab where they were dissected the same day. Recordswere kept on individual fish which were weighed andmeasured for length. Unlike Chapman and Fernando(1994), who identified organisms in tilapia stomachsand carp foreguts, we examined the whole digestivetract, thereby alleviating the confounding effects ofvariability in both gut fullness and evacuation rates.For both fish species the entire digestive tract wasexcised and preserved in a 10% sugar–formaldehydesolution and sealed in plastic bags.

Control of leaffolders

Gathering of data on the role of fish as predators of riceleaffolders was carried out over four rice crops. The1991 DS crop was infested artificially and the remainingthree crops were naturally infested. Artificial infestationinvolved releasing 500 greenhouse-reared Cnaphalocro-cis medinalis adults in each of two plots at 45 DT.Damaged leaves were assessed at 65 and 85 DT. Threecrops where natural infestation was measured had sixreplicates (1991 and 1992WS and 1992 DS) whereas the1991 DS crop was replicated three times. Naturalinfestation was mainly by C. medinalis and Marasmiaspp. (Barrion et al. 1991). Leaffolder damage wasassessed visually as percentage of damaged leaves on 20hills per plot taken by stratified random sampling 10days before harvest. Damage during the ripening stageis considered to be the most severe in relation to yieldloss when the rate of photosynthesis decreases duringgrain filling (Heong 1990).

Control of stemborers

Stemborer damage was monitored for whiteheads(severed panicles) 10 days before harvest in both 1991crops and the 1992 WS crop in six replications of 20hills per crop under natural infestation. Percentagewhiteheads was calculated from counts of all paniclesper hill. In 1992 DS, rice hills were artificially infestedwith yellow stemborer. Two egg masses were attachedto three randomly selected hills per plot at 23, 29, 36,56, and 64 DT in three replications following themethod of Bandong and Litsinger (2005). The same 15hills were infested on each date. Egg masses, obtainedfrom an insectary, were uniform in size (ca. 50 eggs)and all eggs were unparasitized (final selection was

International Journal of Pest Management 333

http://www.rinconvitova.com/d-vac.htm

http://www.contech-inc.com

http://www.contech-inc.com

made at the blackhead stage at which the moth larva’shead capsule becomes visible). At each of the 15infestation sites, whiteheads were counted 10 daysbefore harvest from a 1-m2 area (25 hills).

Results

D-Vac sampling

Suction sampling carried out over four seasons pro-duced a list of 139 species (44 parasitoids, 36 predators,24 detritivores, 23 indifferents, 12 rice pests). There were92 families, predators topping the list with 30 followedby parasitoids (22), detritivores (18), indifferents (15),and rice pests (7). Of the total number of arthropodscollected, the most abundant were detritivores (42%)dominated by chironomids, predators (18%), rice pests(18%), parasitoids (14%), and indifferents (8%). Aqua-tics barely outnumbered terrestrials in terms of numbersof individuals, viz. 53% to 47%.

The abundance of arthropods collected by D-Vacsuction in the four crops (no./hill) over all guilds didnot show significant differences between the fish andnon-fish treatments in 1991 and 1992 DS (Table 1).However, in the 1991 WS the arthropod density was22% less with Common Carp than rice alone, with NileTilapia being statistically intermediate at 14% less. Inthe 1992 WS the degree of arthropod reduction by eachfish species was significant and similar. Carp reducedarthropod density by 10% while tilapia depresseddensities by 9%. In both WS crops the dominantarthropods were chironomids.

Detritivores suffered a 40% reduction with bothfish species in the 1991 WS over the 29–85 DTsampling period (Table 1). The main detritivores wereadult chironomids (Chironomus spp. and Cryptochir-onomus spp.) as well as adults of the chloropid shiningflower fly Mepachymerus (¼Steleocerellus) ensiferwhich also is a combined detritivore and predator.

The rice pest guild was dominated by the rice whorlmaggot Hydrellia philippina in the vegetative stage ofthe 1992 WS crop, where carp produced a 46%reduction at 31 DT but tilapia did not (Table 1). Otherrice pests recorded later in the crop were greenleafhoppers, Nephotettix virescens, and N. nigropictus,as well as zig-zag Recilia dorsalis and white Cofanaspectra leafhoppers. Two planthoppers were present:the whitebacked Sogatella furcifera and brown Nila-parvata lugens. Noted in addition were several rice seedbugs (alydids and pentatomids) and adult moths ofScirpophaga incertulas, C. medinalis/Marasmia sp., aswell as leaf-feeding larvae of Noctuidae.

The predator guild was dominated by the mirid egg-nymph predator Cyrtorhinus lividipennis, the veliidnymph-adult predator Microvelia douglasi atrolineata,spiders of the genera Atypena and Tetragnatha, and thechloropid Steleocerellus ensifer and the phalacrid beetleStilbus sp. Notably, the parasitoid and indifferent guildsdid not show any significant difference from the presenceof fish compared to no fish on any sampling date inthe four crops. Parasitoids were mainly mymarids (e.g.Gonatocerus sp.) which attack leafhopper eggs, as well asbraconids which parasitize leaffolder and stemborerlarvae.

The species of arthropods collected by D-Vac werevery similar to those reported by Schoenly et al. (2010)from a nearby location in Nueva Ecija province. Themain pest was whorl maggot in the early season, andlater on, planthoppers and leafhoppers dominated.Sampling began for the 1991 DS crop during thereproductive stage (50 DT) where beneficials generallyoutnumbered pests with and without fish culture, asituation which accelerated toward harvest (Figure1A). Initially, pests without fish predation outnum-bered those with fish by about 2 per hill versus 1 perhill, respectively, but after 66 DT, the trend reversed.From 66 DT the number of pests in fish culture

Table 1. Arthropods collected as an average of time series sampling in four rice crops by D-Vac suction in fields with andwithout fish culture. Total arthropod abundance is presented for each of four crops and in a comparison of pest and detritivoreguilds, Munoz, Nueva Ecija, Philippines, 1991 and 1992.

No. per 20 hills1

Total arthropods2Detritivores3 Pests4

1991 DS IR64 1991WS IR64 1992 WS IR42 1992 WS IR72 1991 WS IR64 1992 WS IR72Treatment (52–112 DT) (29–85 DT) (22–106 DT) (31–99 DT) (29–85 DT) (31 DT)

Rice 183.1a 266.4 a 255.0 a 196.3 a 79.4 a 34.5 aRice-carp 196.1a 206.7 b 234.9 a 177.0 b 47.3 b 18.8 bRice-tilapia 202.8 a 229.7 ab 268.6 a 179.1 b 47.4 b 58.3 a

ns P � 0.05 ns P � 0.05 P � 0.01 P � 0.01

1A ‘hill’ is 3–4 seedlings transplanted in a clump. Single rice hills were suctioned over six sampling dates per crop between the indicated crop agesmeasured as days after transplanting (DT).

In each column, means followed by different letters are significantly different (P 5 0.05) by LSD test.2Total of pest, predator, parasitoid, detritivore, and indifferent guilds. The indifferent guild was composed of non-rice herbivores and transients3Mainly chironomids but also Mepachymerus ensifer (Chloropidae) and Stilbus sp. (Phalacridae).4Mainly Hydrellia philippina whorl maggot.


averaged 1–1.5 per hill, whereas without fish thishovered around 1 per hill. Differences were notsignificant, however. On the following crop, also withIR64, pests (predominantly whorl maggot) outnum-bered natural enemies, viz. 3.5–5 versus 2.5–3.5 per hillfrom 29–43 DT (Figure 1B). By 57 DT during the cropperiod when hoppers were abundant, beneficialsdominated (4 vs. 2–2.5/hill) with or without fish.Subsequently, pest abundance steadily declined to 1per hill at 85 DT, whereas beneficials declined to 2.5–3per hill. The effect of fish was not noted, as both pestand beneficial curves with and without fish stronglymirrored one another.

In the third crop (1992 DS), with the more pestsusceptible and longer maturing IR42 (Figure 1C),numbers of pests and their natural enemies rose steeplyat 20–65 DT from 0.5–1 per hill to 6–7 per hill with fishand 8–9 per hill without fish. Arthropods in all fourtreatments rapidly declined to 2–3 per hill at 78 DT,and thereafter natural enemies with and without fishrose again to 8–10 per hill, while pests remainedsuppressed at 3–4 per hill. In the 1992 WS crop withthe more resistant IR72, there was no effect of fish

culture on either pests or beneficials whose curvesmirrored each other very closely (Figure 1D). Pestsdeclined steadily thereafter from an initial 2 per hill at31 DT to less than 1 per hill. Beneficials averaged 3–5-fold higher densities than pests throughout the crop.

Sticky traps

Sticky traps ensnared arthropods that either hadbecome detached from the foliage or had attemptedto land on water. As there was no significant differencein densities collected in respect of the presence orabsence of fish, the data were pooled. Greatestcollections over the whole season (mean no./hill) werefirst of all pests (43) followed by predators (29) anddetritivores (19). Parasitoids (8) and indifferents (8)made up the fewest collected (Figure 2). The absolutenumber of arthropods collected declined over thegrowing season from 240 per hill at 22 DT to 42 perhill at 90 DT. The number of pests, however, changedmost over the growing season, declining relativelymore than the other guilds, starting at 155 per hill at 22DT and reaching 22 per hill at 50 DT. This period

Figure 1. Arthropod densities collected by D-Vac suction machine over four rice crops with and without fish culture. On eachsampling date 20 hills were sampled and the arthropods were divided into pest and natural enemy guilds, double rice crops of1991 and 1992, Munoz, Philippines.


was dominated by whorl maggot followed mostly byplant- and leafhoppers, ranging from 10 to 5 per hill onthe last sampling date. Other pests collected were molecrickets (Gryllotalpidae), stemborer and leaffoldermoths, and rice seed bugs (Alydidae and Pentatomidae).Predators declined slowly but steadily in abundancethroughout the crop, ranging from 32 per hill at 22 DTto 15 per hill at 90 DT. There was a wide array ofpredator types dominated by spiders (Araneidae, Lyco-sidae, Tetragnathidae), but included aquatic (Belosto-matidae, Gerridae, Hydrometridae, Hydrophilidae,Mesoveliidae, Notonectidae, Veliidae), and arboreal(Anthocoridae, Carabidae, Coccinellidae [Micraspiscrocea], Formicidae, Gryllidae, Miridae [Cyrtorhinuslividipenis], Staphylinidae, and Tettigoniidae) insects.

Detritivores, mainly chironomids, were prevalentduring the whole crop, and also Ephydridae, Corixidae,Ephemeroptera, and Tipulidae. Their abundanceschanged over time, peaking at 35 DT with 31 per hilland then declining slowly to 10–12 per hill on the finalthree sampling dates. Parasitoid wasps peaked on thefirst sampling date at 18 per hill and thendeclined to 3perhill at 90 DT. These were predominantly Braconidae,Ichneumonidae, Mymaridae, and Scelionidae, but alsoincludedBethylidae,Diapriidae, Elasmidae, Encyrtidae,Eulophidae, Pteromalidae, and Trichogrammatidae.Indifferents were low in number throughout the crop,ranging from 4 to 12 per hill, and included Acari(Oribatellidae), Chrysomelidae, and Phlaeothripidae.

Gut contents

A variety of food items was identified in dissected guts,but the bulk was mainly detrital aggregate, which refers

to a mixture of plant debris (i.e. organic detritus),zooplankton, and phytoplankton. Plant material wasoccasionally found within the guts of tilapia, whileweed seeds were often encountered in the guts of carp.Animal food items that could be recognized includedmolluscs, small fingerlings, as well as members of 23families of arthropods. Only 43% of carp and 13% oftilapia contained arthropods in their guts in the DS,whereas 79% and 75%, respectively, did in the WS(Table 2). At most, 10 arthropods were recovered perfish (Figure 3), but the average was 4.6 from carp and1.6 from tilapia. Carp consumed an average of 2.4arthropods/fish in the DS and 6.8/fish in the WS.Tilapia consumed 0.5/fish in the DS and 2.7/fish in theWS. Significantly more arthropods were consumed inthe wet than dry season for both fish species (283versus 949 arthropods per 100 fish) and carp consumedsignificantly more arthropods than tilapia (318 versus914) (Table 2). The likely reason more arthropods wereconsumed in the WS is that there were frequentmonsoon storms which dislodged them from thefoliage onto the water. Slapping the foliage is acommonly used sampling method, as the insects areeasily dislodged and end up on the water surface.

A taxonomically wide range of arthropod groupswas encountered in the digestive tracts of both fishspecies; seven of the eight dominant families wereaquatic, so naturally in harm’s way to fish (Table 3).The exception was formicid ants, which maintain nestsin the rice bunds, but as the fields are continuouslyflooded they stay on the rice foliage preying onarthropod eggs and larvae. They travel from plant toplant and can be blown off the plants by high winds.Ephydrids included larvae and pupae from the

Figure 2. Mean abundance of arthropod guilds collected on sticky traps in time series sampling from the vegetative stage to theripening stage, Munoz, Nueva Ecija, Philippines, 1992 wet season.


detritivore relatives of whorl maggot. Adult ephydridswere mostly whorl maggots. Interestingly, chironomidsmade up 60% of the intake of carp with 32% beingpredators, whereas the opposite was the case for tilapiawhere 59% of the arthropods were predators and only32% were chironomids. However, as carp consumethree times as many arthropods per fish as tilapia, theirtoll also on predators is probably similar.

Some 7–8% of the arthropods from both fishspecies were insect pests, which were mostly whorlmaggots. Most of the predators were dytiscids,formicids, notonectids, and corixids. Eight of the 11families of predators in fish guts were aquatic.Parasitoids and indifferents made up the smallestfraction of collected arthropods (� 0.4%). These wereadults that had probably been blown onto the water.

There was no significant (P4 0.05, n¼ 566) corre-lation between the length of fish or their weight andnumber of arthropods recovered from their guts, eitherby season or fish species. The range in weight for carp

was 3.5–150 g and for tilapia 7.9–82.9 g, which reflectsthe periodic sampling over the entire season.

Control of stemborers and leaffolders

Significantly less stemborer damage was recorded intwo of the four seasons (Table 4). In 1991 DS, bothfish species registered significantly less (61–69%)stemborer whitehead damage, namely 2.2–2.7% com-pared to 7% without fish. In the 1992 WS there was ahigher infestation level with 18.1% whiteheads in thenon-fish control and 12.7% with carp and 15.1% withtilapia. This translated into 30% less damage withcarp and 17% less with tilapia. These were naturalinfestations. In the 1992 DS, even with high artificialinfestation levels, there was no significant differencebetween any of the treatments. It is also notable thatin each of the four seasons, the plots with carp hadlower, but not always significant, infestation levelsthan tilapia.

Table 2. Number of arthropods identified in the digestive tracts of common carp and Nile Tilapia, Munoz, Nueva Ecija,Philippines, 1992.

Arthropod abundance (no./100 fish)1% Fish witharthropods

Average (no. arthropods/fish)Fish Dry season (DS) Wet season (WS) DS WS

Common carp2 235 + 93 b 679 + 136 a 43.2 79.4 4.6Nile Tilapia3 48 + 36 c 270 + 87 b 13.3 75.0 1.6Total 283 949Mean 28.3 77.2

1Means in both columns and rows followed by different letters are significantly different (P � 0.05) by LSD test.2120 fish in the DS and 178 fish in the WS.3120 fish in the DS and 148 fish in the WS.

Figure 3. Frequency of occurrence of arthropods recovered from the digestive tracts of carp (CC) and tilapia (ON) cultured inIR64 rice fields, Munoz, Nueva Ecija, Philippines, 1991 wet (WS) and dry (DS) seasons.


Leaffolder damage was not affected by the presenceof fish and there was no significant difference betweenany of the treatments in all four crops. The artificially

infested 1991 DS crop registered the highest infestationlevel of 11.0–16.7% damaged leaves, whereas both1992 crops had very low injury levels (0–0.4%damaged leaves). The 1991 WS crop was intermediate,ranging from 1.9% to 2.4% damaged leaves. Averagesfor the four crops were 4.9% damaged leaves withoutfish, and 4.4% for carp and 3.4% for tilapia, but thesevalues were not significantly different.

Discussion

Fish food selection

Detrital aggregate is the major food of both theCommon Carp and Nile Tilapia; it is derived from therice-field bottom by carp and periphyton by tilapia(Chapman and Fernando 1994). In California rice fields,chironomid larvae are also the single most importantarthropod found in Common Carp (Washino andHokama 1967). Halwart (1994) concluded that chirono-mids were the dominant food of Common Carp becauseof their benthic feeding behaviour, during which mud isfiltered and insect larvae, oligochaete worms, weed seeds,and other digestible matter are retained. Chironomidlarvae live in the mud and form silken cases which mightprotect the larvae against aquatic predatators . Thelarvae mostly feed on decaying organic matter thataccumulates at the interface of the soil surface andponded water. Fish predation on chironomids affects notonly the larvae, but also the adults during emergencefrom the pupa, alighting on the water surface whenovipositing, or accidentally falling into the water. InIndonesian rice fields, Ardiwinata (1957) noted thatCommon Carp mainly feed on chironomids andoligochaetes. Halwart (1994) stressed the importance ofmacrofauna (molluscs, chironomids, mayflies, caddisflies, and dytiscids) and macroflora (especially Echino-chloa spp. grass seed) for Common Carp and noted adietary difference compared to pond-reared carp, whichpreferred copepods and cladocerans (zooplankton).Meien (1940) reported that carp in Russian and Italianrice fields consumed large numbers of insect larvae(mainly chironomids and Anopheles spp.), algae, macro-phytes and large amounts of grassy weed seeds.

Nile Tilapia, however, feed in the water columnwhere aquatic predatory arthropods forage for prey.From gut analysis in this study, arthropods from 25families were found to be food items of carp andtilapia, and carp consumed three times as manyarthropods per fish than did tilapia. Most of thearthropod intake by tilapia was invertebrate predators,while that of carp was chironomids. Seven of the eightarthropod families consumed by tilapia were aquaticwith only ants being terrestrial; therefore it isunsurprising that many ended up in the gut of tilapia.Thus, the difference in feeding behaviour of the twotypes of fish explains their differing food selection.Chironomids usually peak early in the season in ricefields (Simpson et al. 1994). Our study corroborated

Table 3. Identification of the arthropods found in thedissected guts of common carp and tilapia, Munoz, NuevaEcija, Philippines, 1992 DS and WS.

Arthropods collected(%)1

Carp Tilapia

Guild Taxa Guild Taxa Guild Taxa

Pest 6.8 7.6Ephydridae2 6.7 7.6Other3 0.1 0.01

Predator 28.4 48.7Dytiscidae4 13.6 16.2Formicidae 8.5 14.9Notonectidae5 3.9 13.9Coenagrionidae 1.9 1.6Veliidae6 0.2 1.9Other7 0.3 0.6

Parasitoid8 0.2 0.2 0.4Detritivore 64.7 43.3

Chironomidae 59.6 31.8Corixidae9 2.9 10.4Ephydridae10 2.2 1.1

Indifferent11 0.01 0.02

1Gut dissection of 178 carps and 148 tilapias,2Hydrellia philippina,3Pyralidae larvae (probably leaffolder), Cicadellidae (Recilia dorsa-lis),4Lacophilus sharpi,5Anisops sp.6Microvelia douglasi,7Belostomatidae, Carabidae, Coccinellidae, Gerridae, Hydrometri-dae, Hydrophilidae (Helochares sp.),8Braconidae, Mymaridae, Phoridae (Megaselia sp.), Scelionidae(Telenomus sp.),9Micronecta sp.10Notiphila, Psilopa.11Cryptostigmata (Oribatei), Cucujid.

Table 4. Stemborer damage recorded in four crops with andwithout fish culture,Munoz, Nueva Ecija, Philippines, 1991–92.

Stemborer whiteheads(%)1

Treatment 1991 DS 1991 WS1992DS2

1992WS Average

Rice 7.0 a 2.2 a 17.5 a 18.1 a 11.2 aRice-carp 2.2 b 1.8 a 14.7 a 12.7 b 7.9 bRice-tilapia 2.7 b 2.2 a 17.2 a 15.1 b 9.3 ab

1In each column, means followed by a different letter are significantlydifferent (P � 0.05) by LSD test. Plots were sampled 10 days beforeharvest.

There were six replications in each of three seasons under naturalinfestation where 20 hills were sampled per replicate2Artificial infestation was carried out in three replications where twoegg masses were attached to 3 randomly selected hills per plot on fiveinfestation dates(23, 29, 36, 56, and 64 DT) giving a total of 15infested hills per replicate. 25 hills were sampled per replicate forwhiteheads at 126 DT.


this fact, based on sticky trap data (Figure 2).Chironomids are also an important food source ofnatural enemies, particularly for spiders early in theseason (Settle et al. 1996). It is at this time that bothpests and beneficials are building up in numbers and ifthe latter are to be able to contain pests during the cropthey need an early start.

Chapman and Fernando (1994) also found thatarthropods make up only a small proportion of the dietof both carp and tilapia, in agreement with our results.Arthropod rice pests, with the exception of whorlmaggot, made up only a small fraction of the diet ofcarp and tilapia but provided vital protein. Based onthe data from Table 2, carp averaged 4.6 while tilapiaaveraged 1.6 arthropods per fish in their guts. Thistotal for carp is within the same range reported by Wanet al. (1995), who found in Shan-Gao, China, thatCommon Carp weighing 150 g ate 1.3 g of insects/day(a black widow spider weighs about 1 g) (http://animals.nationalgeographic.com/animals/bugs/black-widow-spider/?source¼A-to-Z).

Mosquitoes were not found in fish guts in ourstudy, as populations appeared to be low and only afew adults of two genera were found in D-Vacsampling. The consumption of mosquito larvae hasbeen reported in China for both the Common Carpand Nile Tilapia (Pao 1981). Fish have been used formosquito control in rice fields in China for many years(Wu 1995). There is even a report that predation by fishon mosquitoes dramatically reduced malaria incidencein Guangxi, Zhuang (Wu et al. 1995).

Between 75% and 79% of fish had arthropods intheir gut in WS compared to only 13–43% in DS. Whatcould be the reason for this? Phytoplankton is theprimary source of energy in the aquatic arthropod foodcycle which is limited by solar radiation. Following thisreasoning, then, the DS should give rise to the higherarthropod counts. Comparing abundance over the fourseasons in the D-Vac samplings, the 1992 DS had thehighest mean density, but the 1991 WS had the secondlargest density. The 1992 DS was sown to the very pestsusceptible IR42. But comparing the two seasons in1991, where the same variety was grown (IR64), the WSaveraged 21% more. Munoz lies at the edge of the largeirrigation system in C Luzon and is bordered byextensive rainfed rice areas. The larger rice area in theWS presumably contributed to the higher arthropoddensities from dispersal. Monsoon winds also causedmore arthropods to fall into the floodwater providingfood for fish. Thus the larger rice area of the WS plus themonsoon winds were probably responsible for the higherarthropod numbers in the digestive tracts of the fish.

Which rice insect pests are more vulnerable to fishpredation?

The main aim of this study was to describe andquantify the role of fish as biocontrol agents of rice

pests. Up to 10 arthropod species were found in theguts of individual fishes with only 7–8% being riceinsect pests. The low incidence of insect pests isattributable to their rarity in rice-field pond water.Fish gut content analysis revealed that prey werepredominantly aquatic or neustonic (at or near watersurface) invertebrates with only modest numbers ofterrestrial groups. Pests monitored by D-Vac werepredominantly whorl maggot during the vegetativestage which were replaced by non-economically im-portant levels of leaf- and planthoppers until cropsenescence, a period when arthropod numbers de-clined. D-Vac is best suited for sampling less-mobilespecies living on the rice plant and water surface.Although disturbance of arthropods can be minimizedby experienced staff such as was achieved in this study,highly mobile insects such as odonatans (dragonfliesand damselflies) and orthopterans (grasshoppers andcrickets) normally escaped capture. Aquatic speciesliving in the water column such as odonatan nymphs,water beetles, and water bugs were collected, albeit inlower numbers than if the water column had beendirectly sampled. Previous field studies have shownthat purely aquatic predators interact little with riceinsect pests in the Philippines, except when theyinadvertently fall into, or alight on, the water. Otherpests are also recorded as being preyed upon by aquaticarthropods (Hemiptera, Coleoptera, Odonata), andcan be noted in the food webs of leaf- and planthop-pers as well as stemborers, but these occur infrequently(Jahn et al. 2007; Kandibare et al. 2007).

Our study has shown that rice pests that live in thewater are the most vulnerable to fish. Aquaticchironomids and corixids feed on the roots of riceseedlings, but they are never sufficiently abundant inthe Philippines to be counted as economically im-portant pests. Of all the insects pests of rice, the semi-aquatic rice caseworm Nymphula depunctalis (Lepidop-tera: Pyralidae) is probably the most vulnerable to fishpredation in the Philippines, as its larvae float on thewater surface within cases made of rolled rice leaves(Litsinger et al. 1994a, 1994b). Floating larvaeaccumulate at the tail end of irrigation systems or atthe basins of a toposequence where they can causelarge losses. Larvae protrude from their cases whilenavigating to new plants, so alerting fish to theirpresence. Females oviposit underneath floating leavesmaking themselves as well as eggs vulnerable. Thus therice caseworm is exposed for most of its life-cycle. InVietnam, Vromant et al. (1998) found that a poly-culture (Thai Silver Barb Barbonymus gonionotus,Common Carp and tilapia) was able to reduce thenumber of rice caseworm larvae 493% and quelled anoutbreak.

Several species of aquatic water weevils feed on riceroots. In Hubei, Echinocnemus squamous adults werefound in dissected fish with more found in fish 47 cmlong (Wu 1995). Two other studies in northern China


http://animals.nationalgeographic.com/animals/bugs/black-widow-spider/?source=A-to-Z




also noted that fish consumed water weevil adults.Early in the season, eggs are laid next to rice roots,making adults highly vulnerable to predation. Adultsalso could have been blown by wind into the flood-water, or fish found them when they emerged fromunderground pupal cells (Luo 1995; Pan et al. 1995).

Rice whorl maggot is a minor pest of vegetative-stage rice and occurs throughout the Philippines. Itthrives in fields with standing water, but no stage livesin water. Its attraction to flooded fields may be apredator avoidance strategy (Jahn et al. 2007). Femalesof this ephydrid fly lay single eggs, which adhere tightlyto the foliage, and the larvae feed within the developingleaf buds. Pupation takes place at the leaf bases, thusthe adult is the most vulnerable stage to fish predation,as flies rest on plants near the water or when they alighton the surface. Immature stages could, however,become accessible to fish if the paddy water rose as aresult of heavy rainfall or irrigation. Ephydrid larvaeand puparia were found in fish guts, but most probablywere from related species. Adults of several ephydridsof the genera Notiphila and Psilopa were prevalent, butthey are not considered to be pests, although they dofeed on rice and weeds (Barrion and Litsinger 1986).There were two species of Notiphila (N. similis and N.latigenis) whose larvae feed mainly on detritus. All ofthe ephydrids were commonly found in sticky trapsand fish guts as well as in D-Vac samplings. CommonCarp in particular may be an effective predator as itreduced ephydrid adult densities by 46% (see Table 1).As the whorl maggot is an early colonizer of rice fields,fish may not be of great importance as predators, asthey are generally introduced into the fields severalweeks after transplanting. Fish predation may play amore important role in areas of staggered plantings bysuppressing the overall whorl maggot population. InBangladesh, fish in rice fields significantly reduced thenumber of dipteran insects sampled by a vacuumsuction device from an average of 6.6/m2 to 1.0/m2

(Frei et al. 2007).In the literature, plant- and leafhoppers are the

most commonly mentioned insect pests whose popula-tions have been reduced by fish. Most of the studieswhere the greatest effects were found have been inChina, with one study from each of Vietnam andIndonesia. Fourteen reports found in the literature givedata which averaged 60%+ 22% suppression (range16%–98%) (Xiao 1992; Halwart 1994; Hendarsih et al.1994; Tuan 1994; MacKay 1995; Weimin 2010), andfour studies from China on green leafhoppers reported49%+ 32% suppression (range 30%–88%) (Xu andGuo 1992; Halwart 1994; Xiao-Fan 1995; Weimin2010). These reports were based on field counts ofhoppers comparing fields with fish to those withoutfrom the same area. In Hubei, fish guts were analysedand leaf- and planthoppers were found in CommonCarp and Silver Carp Hypophthalmichthys molitrix.Wu (1995) claimed that the presence of fish did away

with the need to apply insecticide. Li et al. (1995)observed that carp even jump from the water surface toreach planthoppers on nearby tillers.

Leafhoppers and planthoppers are vulnerable dur-ing most of their life-cycle. Carp which consume tillerswould therefore consume hopper eggs deposited inside.Brown planthopper (Nilaparvata lugens) is particularlyvulnerable, as it is alone among the main hopper pestsin preferring to reside on tillers near to the watersurface. On the other hand, white-backed planthopperand green, white, and zig-zag leafhoppers reside morein the upper canopy. If the water level were to risesuddenly, brown planthoppers would be particularlyvulnerable. Carp and tilapia are known to collide withrice plants during feeding, which can cause nymphsand adults to detach and fall in the floodwater(Sinhababu and Majumdar 1981; Chapman andFernando 1994). Even without fish collisions or astrong wind, studies have shown that hoppers readilybecome dislodged from the tillers even in protectedenvironments such as greenhouses. In a greenhousestudy, Nakasuji and Dyck (1984) found that two-thirdsof brown planthopper adults and late-instar nymphsfell onto the water. Almazon and Heong (1992) foundthat dislodgement rates were density dependent, andthat dislodgement mostly occurred during the mon-soon WS due to high winds. Another reason fordetachment is to escape terrestrial predators whilefeeding. The surface tension of water prevents hoppersfrom sinking, but the ripples caused by the impact onthe water surface alerts nearby fish

In the D-Vac study there was no significantdifference between leafhoppers and planthoppers inany of the four seasons due to the presence of fish.Leafhoppers and planthoppers colonize a young cropand build up over two to three generations to becomemore important as pests in the later crop growth stages.The only significant reduction of pests occurred at 31DT, when hopper numbers were only modest and theobserved differences were due to whorl maggot.

Yellow stemborer larvae possess aquatic character-istics in so far as they can develop from larva to adultin completely submerged rice plants (Catling and Islam1995). Tillers are sealed off and made water-tight bythe larvae secreting several layers of silk over theentrance hole. Neonate larvae undertake dispersalbehaviour soon after emergence by secreting silkthreads from which larvae dangle in the wind to reachnew plants (Shiraki 1917). This occurs early in themorning when the wind speed is typically low, butduring this process the larvae may fall to onto thewater. They can float due to their waxy cuticle and lowbody mass, and they move by wiggling to reachanother plant, which would alert the fish. Older larvaealso disperse; like rice caseworm larvae, they can makea case from a cut rice leaf (which naturally rolls into acylinder when cut) and float on the water surface,eventually reaching a new plant. Here again stemborer


larvae are potential prey of fish. This dispersal stagecan last around two to eight days (Shiraki 1917). Adultmoths cling to rice plants during the day and becomeactive at night to avoid predators such as Odonata orbirds. Winds, however, can cause them to becomedislodged and fall to the water’s surface. Disturbedmoths fly away when one walks through the fields. Iffish can disturb them, then higher predation of mothsmay occur. Shiraki (1917) reported that after a stormmany dead moths were seen floating on the water.Apparently moths are not preferred food for fish, sinceadults or other evidence such as parts of wings werenot found in fish guts.

Thirteen trials were found in the literature, twelvefrom China and one in the Philippines, where stem-borers averaged 49%+ 24% suppression (range 11%–100%) (Xiao 1992; Xu and Guo 1992; Halwart 1994;MacKay 1995; Halwart and Gupta 2004). Various carpwere involved in all trials, with Common Carp in eight,but also polycultures with Grass Carp Ctenopharyngo-don idella, Silver Carp, Japanese Crucian CarpCarassius carassius, and/or tilapia in six of the cases.Among fish species, the lowest degree of control wasfrom Common Carp. Grass Carp eat the lower leaves,a behaviour which would disturb moths (Yu et al.1995). As rice stemborers are the most importantchronic insect pests of rice in Asia, their predation byfish that reduced damage 17–29% is of high populationdynamic significance.

In the current study, lower stemborer damage levelswere recorded in two of the four seasons and larvaewere observed in the guts of both Common Carp andNile Tilapia, indicating that predation had occurred(see Table 3). It would be expected that the effect of fishpredation would be greater in the WS’s due to therainfall and wind washing larvae and moths into thewater, but in this study equally significant predationoccurred in both seasons.

Rice leaffolders are also chronic pests in all parts ofAsia and some studies have shown that they canbecome vulnerable prey for fish. The average of fivetrials (four in China and one in Indonesia involvingCommon Carp in all but also with tilapia in Indonesia)suppression averaged 49%+ 13% (range 30%–64%).It is interesting to note that unidentified pyralid larvaewere found in fish guts, which most likely wereleaffolder larvae blown from plants by the wind. Inthe our study, only the artificially infested plotsregistered significant damage (11–17% damagedleaves), whereas in the other trials, damage rangedfrom 0 to 2%. Leaffolders are normally under heavypressure from a wide range of natural enemies (Barrionet al. 1991). Egg predation and larval parasitismgenerally are high. Like stemborers, the adult mothsremain in the foliage during the day and become activeat night. They lay eggs on the foliage and do not comeinto contact with the floodwater. Larvae do notdisperse by making cases as do stemborers, but move

directly from plant to plant. They typically occupy thetop half of the foliage. Observations in field studieshave been made of carp colliding with plants, probablyinadvertently during their feeding, and this action hasbeen found to startle leaffolder moths into flight, whichare then preyed upon by Odonata and birds (Litsinger1993).

Three reports support our findings that there wasno significant control of leaffolders by fish. Studies inVietnam (Vromant et al. 2003), Bangladesh (Frei et al.2007), and Malaysia (Teo 2006) found no effect of fishon leaffolders. On the other hand, three surveys donein Zhejiang Province, China revealed a large increasein leaffolders in association with fish (Yu et al. 1995).In Xiaoshan, there was a 6.5-fold greater quantity ofleaffolder-damaged leaves in association with fish in1986 and 1.9-fold in the following year. In Shangyuthere was a 57% increase in larvae with Grass Carpalone, but only a 30% increase in a polyculture ofGrass Carp, Common Carp and tilapia. Reasons givenas to why leaffolder was more abundant in fish culturewere that the benthos-feeding carp churn the soil,aerating it, and their faeces add to organic fertilizer,making the leaves greener, and the more humidmicroclimate from the deeper water favours oviposi-tion and larval survival (Yu et al. 1995). On the otherhand, the same authors claimed that control of sheathblight disease was from the lower humidity due toremoval of excess tillers and diseased leaves by the fishwhich opened the canopy to let in more light and wind.The most probable explanation is the higher level of N,which is well known to increase fecundity andsurvivorship (Litsinger et al. 2011).

Weimin (2010) also recorded suppression of thegreen-semilooper Naranga aenescens (80% predation)and the skipper butterfly Parnara guttata (50% preda-tion) due to fish. Rice seed bugs (Leptocorisa spp.) arepests of developing grain and were detected in the D-Vacsampling, but their numbers were small and no signi-ficant difference was noted. Wind could knock them intothe water, but it remains to be seen whether fish wouldeat them, as they secrete a defensive chemical whendisturbed, which may be repellent to fish. Teo (2006)reported a large infestation of rice bugs in Malaysia, butthe presence of fish did not affect their densities.

Why fish are not always effective predators

Several studies have failed to show insect pestsuppression in rice-fish cultures (Vromant et al. 2002;Teo 2006; Frei et al. 2007). Our study also found thatfish did not significantly reduce densities of hoppersand leaffolder, whereas in other studies, suppressionwas high. Lack of control in some studies may havebeen due to several possible causes:

(1) The pest population was not sufficiently highfor us to be able to measure differences.


This probably explains why hoppers were notsuppressed by fish in the current study. Themost likely reason for the low pest populationwas that insect-resistant rice cultivars were usedin our trials. The three cultivars used werehighly resistant to the green leafhopper andbrown planthopper, although IR42 was onlymoderately resistant. IR cultivars, however, arenot resistant to either whitebacked planthopperor white or zig-zag leafhoppers. However,because insecticides were not used, high densi-ties of beneficials were allowed to build up,particularly in a double rice-crop system(Litsinger 2008). Only sub-economic densitiesof 510 hoppers/hill were recorded from thefour field trials where beneficials mostly domi-nated (see Figure 1). Seven reports from Chinaand Thailand, where control was noted, aspreviously reported, averaged 38 planthoppersper hill, which is above the 20 per hill actionthreshold used to trigger insecticide application(Way et al. 1991). Hopper populations arehigher in China and Thailand than in thePhilippines due to a lack of genetic resistancein common rice cultivars, higher use of N, singlerice crop systems, and migration of planthop-pers from south to north into crops with lowlevels of beneficials (Litsinger 2008). The longrice-free fallow in single crop systems reducesbeneficials more than pests, while double cropsystems favour both beneficials and pests. Thus,insecticide usage and genetic resistance dictatethe outcome in both systems locally. Leaffoldersalso behave much like hoppers, in that popula-tions are stimulated by both N fertilizer andinsecticide resurgence (Bandong and Litsinger1986; de Kraker et al. 2000). Hopper predationby fish occurs more under higher pest densitieson a crop as detachment is density dependent(Almazon and Heong 1992; Hendarsih et al.1994). Weather can also play a similar role inlowering infestation level. Frei et al. (2007)during the winter months noted low arthropoddensities in Bangladesh where cool temperaturescurtailed population development.

(2) Sampling pests too late in the season. It is nowonder that Teo (2006) did not find evidence offish predation, as sampling for arthropods wasdone only once and just before harvest whenthe crop was in senescence. Few insect pests areactive when the crop is senescing. Whiteheadscan be readily seen at this stage, but leaffolderdamaged leaves may have withered away, andcertainly early-season insect pests would leaveno evidence. Pests should be sampled periodi-cally to detect peaks of injury or density.

(3) Grass Carp versus Common Carp. Within carp,Grass Carp aggressively consume rice tillers

disturbing plants; this flushes out moths and, itis believed, leads to more predation by Odonataand birds. Stems also contain eggs of hoppersand other insect pests that would be consumed.Therefore Yu et al. (1995) believed that GrassCarp is superior to Common Carp in reducingdensities of insect pests. But most reports fromChina where both fish are prevalent show thatCommon Carp exert greater control of riceinsect pests (Mackay 1995).

(4) Selection of larger-bodied fish should result inhigher predation rates. Yu et al. (1995) con-cluded that large, more mature Common Carp,Grass Carp, and Nile Tilapia were better thanyounger fish as biocontrol agents. However, asignificant difference was found in the currentstudy between fish size and the number ofarthropods found in fish gut analyses, with carpweighing 3.5–150 g and tilapia weighing 7.9–82.9 g per fish.

(5) A higher stocking rate should lead to greaterpredation. Stocking rates in reports from Chinaranged from 3000 to 15,000/ha for mature fishand 20,000–225,000/ha for fingerlings (Mackay1995). In our study, however, there was nodifference between the two rates tested, viz. 5000and 10,000 fish/ha. In fact more predation wasnoted in the 1991WSwith a stocking rate of 5000fish/ha. This was one of the crops where fishexerted significant suppression of arthropods aswell as detritivores/predators (see Table 1), andall other crops were stocked at 10,000/ha.

(6) Polyculture – by employing different species offish, each specializing in a different feeding modeor part of the aquatic habitat to feed – shouldoffer improved predation over single species.The current trial showed that tilapia and carpfeed in different parts of rice field profiles, andperhaps predation would be greater if they weremixed. The scant evidence of trials wherecomparisons have been made between polycul-ture and single species does not seem to bear thisout. Yu et al. (1995) found polyculture did notincrease predation of planthoppers over singlespecies, whereas polyculture increased leaffolderonly 30% whereas grass carp alone increased itmore (57%), but the differences may not havebeen statistically significant.

(7) Poor design of a trench system could minimizepredation, if fish would spend most of theirtime in the trench rather than in the field. Thecurrent study had one 0.5-m deep trenchrunning along the length of the field. Halwartet al. (1996) showed that the fish spent moretime in the field than in the trench, thus thedesign aided predation opportunity. In most ofthe literature the design of the refuge system isnot described.


(8) Pest behaviour is a significant factor in predis-position to fish predation. Stemborers, leaf-folders, and rice-seed bugs are terrestrial as theyfeed on aerial portions of the plant and onlyoccasionally will fall from the plant and land onthe water, so they are less affected by fish.Stemborers are more predisposed than leaf-folders to fish predation owing to the aquaticbehaviour of dispersing larvae. Leaffolderlarvae, because they reside in shelters of foldedleaves, are less likely to detach and fall onto thewater. This is probably why artificially infestinghigh numbers of moths had no effect. Veliids,gerids, lycosids, ephydrids, and other predatorsthat inhabit or frequent the water surface aremore in harm’s way (Frei et al. 2007). Aquaticpredatory arthropods that live in the watercolumn such as water beetles (dytiscids, hydro-philids), water bugs (notonectids, hydrome-trids, corixids), and culicids (mosquitoes), aswell as chironomids that frequent the sedimentare most available to fish and were conse-quently most impacted by fish.

The role of fish in the arthropod food web

The results of this study support those of other studieswhich showed that rice fields have a rich and diversearthropod food web where beneficials generally dom-inate over pests (Schoenly et al. 1998). Thus sustain-ability of rice integrated pest management in thePhilippines is based, to a large degree, on naturalenemy conservation, particularly in double rice-crop-ping systems. Normally, there is little need for farmersto apply insecticides owing to the action of beneficials(Ooi and Shepard 1994; Way and Heong 1994; Heongand Escalada 1999). Farmers have their own actionthresholds which are much lower than those ofresearchers (Bandong et al. 2002). Consequently themajority of Filipino farmers still spray, despite the factthat most applications are uneconomical, thus there isthe constant potential for insecticide resurgence (Lit-singer et al. 2005).

Fish can play a role in the arthropod food web aspredators. Phytoplankton, the basis for fish food,decreases in quantity as the crop grows. Invertebratesincreasingly play a key role as food, as the seasonprogresses. More natural enemies than pests are trulyaquatic, thus beneficials come into contact with fishmore than do pests. Fish feed indiscriminately oninvertebrates irrespective of guild but take a higher tollon aquatic predatory arthropods (coenagrionid dam-selflies, belostomatids, hydrometrids, notonectids, cor-ixids) than those that inhabit the neuston zone andleast among terrestrial than aquatic, neustonic or semi-aquatic beneficials. This is borne out by the fact that49% of the arthropods recovered from tilapia guts and28% from carp were predators, whereas only 9% were

pests. The most common semi-aquatic predators areveliids, gerrids, as well as wolf spiders and mycro-phantid spiders, which run across the water surface insearch of prey or make webs just above the waterline.

Interestingly, no spiders were recovered from fishguts. Semi-aquatic wolf spiders related to Paradosapseudoannulata dive into the water after prey (Bleck-mann and Lotz 1987). Such spiders have evolvedbehaviours that aid them to avoid being prey them-selves to fish. This evasive behaviour may also occurwith rice field Araneae. In an aquarium it was notedthat spiders were readily consumed by fish, but in suchan environment there was no shelter. Xu and Guo(1992) in Fujian, China found that spiders increasedthree-fold in abundance in an azollaþ fish (polycultureof grass, common, and crucian carp) compared to ricemonoculture without fish or azolla. The presence of theazolla floating fern provided a refuge from predation,above the water level. In rice fields spiders can seekshelter in between rice tillers and on weeds.

Our results indicate that fish prey more heavily onnatural enemies than pests do. Does this upset thebalance of this important group of pest deterrents?Gupta et al. (1998) in Bangladesh found that predator(lady beetles, spiders, and damselflies) densities were 5–48% higher in rice-fish farms than rice-only farms up to10–12 weeks after transplanting, but nearer to harvestthe converse was observed. However, pest infestationwas 40–167% higher in rice-only farms during all stagesof rice growth. From Figure 1 we note from D-Vaccollections that there were times when pests out-numbered predators, as in 1991 WS and briefly in1992 DS, but most of the time the opposite was true.

Fortunately, the natural enemy guild includes agreater number that reside in the crop canopy inassociation with most insect pests. It is generallybelieved that predators are more important thanparasitoids in suppressing rice pests (Ooi and Shepard1994). Notably, relatively few parasitoids were repre-sented in the D-Vac collections (14% of the arthro-pods), sticky traps (8%), and gut analyses (5 1%).Whorl maggot has few documented parasitoids, butegg parasitoids play a key role against leaf- andplanthoppers as well as stemborers. Larval parasitoidsare important against leaffolders, more so than againstthe more protected stemborers. As a food for fish inrice fields, parasitoids are of minor importance.Natural enemies tend to be low in density at thebeginning of the season, but then increase to levelsabove those of pests during the last half of the ricecrop. This does not always occur, however (see Figure1). Because each natural enemy usually will kill morethan one individual, it is not necessary that beneficialsoutnumber pests in order to contain their densities. It isestimated, for example, that lycosid spiders kill anaverage of five hoppers per day (Kenmore et al. 1984).But arthropod predators prey on one another, evenwithin members of the same species. More often than


not (Figure 1), comparisons with and without fishshowed greater similarity than difference, thus it isconcluded there was no significant effect of fishselectively diminishing beneficials over pests. Weexpect that at times there will be brief upsets in thedensities of beneficials, particularly if farmers useinsecticides. This is one argument in support of farmersusing preventive means of pest control, such as geneticresistance or a long dry fallow between rice crops, tocomplement natural enemies. We saw that in associa-tion with fish during four crops under naturalinfestation, only stemborers reached an action thresh-old in the 1992 WS. The food chain is so rich thatother, more terrestrial predator groups fill the gap.

Fish replace aquatic invertebrates as predators.They make useful predators, as they exist at a constantdensity 0.5–1/m2 that declines only slightly over aseason, whereas beneficials tend to fluctuate dynami-cally over the season (see Figure 1). The flora andfauna of a rice field together have formed a stable agro-ecosystem for many millennia, and fish have been apart of it. The introduction of large numbers of exoticfish did not appear to upset the pest–natural-enemybalance. Fish play a role as predators and anyadditional link to the food chain strengthens integratedpest management.

Insecticides may be unnecessary or even detrimen-tal to insect pest control in the Philippines, sincenatural enemies typically keep pests at manageablelevels and varieties possess genetic resistance. The mainbenefit of rice-fish culture, therefore, is to give farmersan additional reason not to use insecticides and thusspare beneficials (Rothuis et al. 1998). Research hasshown that farmers stocking fish tolerate a higheraction threshold due to the value of the fish (Waibel1992). Fish may play a larger role in pest suppressionin countries where rice varieties do not have geneticresistance against green leafhoppers and brownplanthopper or where, due to long fallows, the naturalenemy densities are low. Because of the value of fish,farmers manage their rice crops better. Better agro-nomic management has been shown to increase thecompensatory ability of the crop to tolerate pestdamage (Litsinger et al. 2011). Increased crop com-pensation is another benefit of rice-fish culture.

Acknowledgements

The following are gratefully acknowledged for their supportand cooperation: the Freshwater Aquaculture Center (FAC)of the Central Luzon State University in Munoz, theInternational Rice Research Institute (IRRI) in Los Banos,and the World Fish Center. We are particularly appreciativeof R.C. Sevilleja and his predecessor R.C. Arce of FAC forbeing the host for the study. Also, thanks go to P. Flor-Blanco, farm manager at FAC. We thank V.R. Carangal whoco-ordinated the Asian Rice Farming Systems Network forproject administration at IRRI. G.T. Chua of IRRI’s Manilaoffice also provided valuable assistance. Support was alsorendered by the Biometrics Unit of IRRI through V.A.

Samaranayke, J. Mojica, and V. Bartolome. This study wasperformed as a PhD thesis by the senior author in theInstitute of Landscape Planning and Ecology, University ofStuttgart, Germany and we thank the support of K. Henle.The members of the research team are also acknowledged fortheir contribution to the study including Fe Ibanez, E.V.Circa and M. Macasiray. We acknowledge the participationof U. Borlinghaus, M. Borlinghaus, and T. Becker of theUniversity of Hohenheim, Germany.

References

Almazon MLP, Heong KL. 1992. Fall-off rates of Nilparvatalugens (Stal) and efficiency of the predator Limnogonusfossarum (F.). Int Rice Res Notes. 17:17.

Ardiwinata RO. 1957. Fish culture on paddy fields inIndonesia. Proc Indo-Pacif Fish Counc. 7:119–154.

Bandong JP, Litsinger JA. 1986. Egg predators of riceleaffolder and their susceptibility to insecticide. Int RiceRes Newslett. 11(3):21.

Bandong JP, Canapi BL, dela Cruz CG, Litsinger JA. 2002.Insecticide decision protocols: A case study of untrainedFilipino rice farmers. Crop Protect. 21:803–816.

Bandong JP, Litsinger JA. 2005. Rice crop stage suscept-ibility to the rice yellow stemborer Scirpophaga incertulas(Walker) (Lepidoptera: Pyralidae). Int J Pest Manage.51:37–43.

Barrion AT, Litsinger JA. 1984. Chironomid, corixid, andostracod pests of irrigated rice seedling roots. Int RiceRes Newslett. 9(6):19.

Barrion, AT, Litsinger JA. 1986. Ephydrid flies (Diptera:Ephydridae) of rice in the Philippines. Int Rice ResNewslett. 11(4):29–30.

Barrion AT, Litsinger JA, Medina EB, Aguda RM, BandongJP, Pantua PC, Viajante VD, dela Cruz CG, Vega CR,Soriano JS. 1991. The rice Cnaphalocrocis and Marasmia(Lepidoptera: Pyralidae) leaffolder complex in thePhilippines: taxonomy, bionomics and control. PhilippEntomol. 8:987–1074.

Bleckmann H, Lotz T. 1987.The vertebrate-catching beha-viour of the fishing spider Dolomedes triton (Araneae,Pisauridae). Anim Behav. 35:641–651.

Bowen SH. 1982. Feeding, digestion and growth: quantitativeconsiderations. In: Pullin RSV, Lowe-McConnell RH,editors. The biology and culture of tilapias. Manila(Philippines): International Center for Living AquaticResources Management (ICLARM). p. 141–156.

Bray F. 1986. The rice economies. Technology and develop-ment in Asian societies. Oxford (UK): Blackwell.

Cagauan AG, Arce RG. 1992. Overview of pesticide use inrice-fish farming in Southeast Asia. In: dela Cruz CR,Lightfoot C, Costa-Pierce BA, Carangal VR, BimbaoMP, editors. Rice–fish research and development in Asia.International Center for Living Aquatic ResourcesManagement (ICLARM) conference proceedings No.24. p. 217–233.

Cai R, Ni D, Wang J. 1995. Rice-fish culture in China: thepast, present, and future. In: Mackay KT, editor. Rice-fish culture in China. Ottawa (Canada): InternationalDevelopment Research Centre (IDRC). p. 3–14.

Catling HD, Islam Z. 1995. Pests of deepwater rice and theirmanagement. Integ Pest Manage Rev. 4:193–229.

Chapman G, Fernando CH. 1994. The diets and relatedaspects of feeding of Nile tilapia (Oreochromis niloticusL.) and common carp (Cyprinus carpio L.) in lowland ricefields in northeast Thailand. Aquaculture 123:281–307.

Clement SL, Grigarick AA, Way MO. 1977. The coloniza-tion of California rice paddies by chironomid midges. JAppl Ecol. 14:379–389.


Coche AG. 1967. Fish culture in rice fields a world-widesynthesis. Hydrobiol 30:1–44.

de Kraker J, van Huis A, Heong KL, van Lenteren JC,Rabbinge R. 1999. Population dynamics of rice leaf-folders (Lepidoptera: Pyralidae) and their natural ene-mies in irrigated rice in the Philippines. Bull EntomolRes. 89:411–421.

de Kraker J, Rabbinge R, Huis A van, Lenteren JC van,Heong KL. 2000. Impact of nitrogenous fertilization onthe population dynamics and natural control of riceleaffolders (Lep.: Pyralidae). Int J Pest Manage. 46:225–235.

Fernando CH. 1993. Rice field ecology and fish culture: anoverview. Hydrobiologica. 259:91–113.

Fernando CH. 1996. Ecology of rice fields and its bearing onfisheries and fish culture. In: De Silva SS, editor.Perspectives of Asian fisheries. Manila (Philippines):Asian Fisheries Society. p. 217–237.

Frei M, Khan MAM, Razzak MA, Hossain MM, Dewan S,Becker K. 2007. Effects of a mixed culture of commoncarp, Cyprinus carpio L., and Nile tilapia, Oreochromisniloticus (L.), on terrestrial arthropod population,benthic fauna, and weed biomass in rice fields inBangladesh. Biol Cont. 41:207–213.

Gupta MV, Sollows JD, Abdul Mazid M, Rahman A,Hussain MG, Dey MM. 1998. Integrating aquaculturewith rice farming in Bangladesh: feasibility and economicviability, its adoption and impact. International Centerfor Living Aquatic Resources Management ICLARMTech Rep. 55:1–90.

Halwart M. 1994. Fish as biocontrol agents in rice: thepotential of common carp Cyprinus carpio (L.) and Niletilapia Oreochromis niloticus (L.). Trop Agroecol. 8:1–169. Stuttgart (Germany): Margraf Verlag / Institute ofLandscape Planning and Ecology, University ofStuttgart.

Halwart M, Borlinghaus M, Kaule G. 1996.Activity patternof fish in rice field. Aquaculture. 145:159–170.

Halwart M. 2006. Biodiversity and nutrition in rice-based aquatic ecosystems. J Food Comp Anal. 19:747–751.

Halwart M. Gupta MV. 2004. Culture of fish in rice fields.Penang (Malaysia): World Fish Center/FAO. 83 p.

Harbott BJ. 1982. Studies on the feeding activity ofSarotherodon niloticus (L.) in Lake Turkana. In: HopsonAF, editor. Lake Turkana: a report on the findings of theLake Turkana project, 1972–1975. London: OverseasDevelopment Administration. p. 1357–1368.

Hendarsih S, Suriapermana S, Fagi A, Manwan I. 1994.Potential of fish in rice-fish culture as a biological controlagent of rice pests. In: dela Cruz CR, editor. Role of fishin enhancing ricefield ecology and in integrated pestmanagement. International Center for Living AquaticResources Management (ICLARM) conference proceed-ings No. 43. p. 32–33.

Heong KL. 1990. Feeding rates of the rice leaffolder,Cnaphalocrocis medinalis (Lepidoptera: Pyralidae), ondifferent plant stages. J Agric Entomol. 7:81–90.

Heong KL, Schoenly KG. 1998. Impact of insecticides onherbivore-natural enemy communities in tropical riceecosystems. In: Haskell PT, McEwen P, editors. Ecotox-icology: pesticides and beneficial organisms. London:Chapman and Hall. p. 381–403.

Heong KL, Escalada MM. 1999. Quantifying rice farmers’pest management decisions: beliefs and subjective normsin stem borer control. Crop Prot. 18: 315–322.

Hyslop EJ. 1980. Stomach contents analysis: a review ofmethods and their application. J. Fish Biol. 17: 411–429.

Jahn GC, Litsinger JA, Chen Y, Barrion AT. 2007. Integratedpest management of rice: ecological concepts. In: Koul O,Cuperus GW, editors. Ecologically based integrated pestmanagement. Wallingford (UK): CABI. p. 315–366.

Kandibare M, Raguraman S, Mahadevan NR. 2007.Diversity and community structure of aquatic arthropodsin an irrigated rice ecosystem of Tamil Nadu, India.Asian J Plant Sci. 6:741–748.

Kenmore PE, Carino FO, Perez CA, Dyck VA, GutierrezAP. 1984. Population regulation of the rice brownplanthopper (Nilaparvata lugens Stal) within rice fieldsin the Philippines. J Plant Prot Trop. 1:19–37.

Koesoemadinata S. 1980. Pesticides as major constraints tointegrated agriculture–aquaculture farming systems. In:Pullin RSV, Shehadeh Z, editors. Integrated agriculture–aquaculture farming systems. Manila (Philippines): Inter-national Center for Living Aquatic Resources Manage-ment (ICLARM). p. 45–51.

Li K. 1988. Rice-fish culture in China: a review. Aquaculture.71: 173–186.

Li X, Wu H, Zhang Y. 1995. Economic and ecologicalbenefits of rice-fish culture. In: Mackay KT, editor. Rice-fish culture in China. Ottawa (Canada): InternationalDevelopment Research Centre (IDRC). p. 129–138.

Litsinger JA. 1993. A farming systems approach to insectpest management for upland and lowland rice farmersin tropical Asia. In: Altieri MA, editor. Crop protectionstrategies for subsistence farmers. Westview Studies inInsect Biology. Boulder (CO): Westview Press. p. 45–101.

Litsinger JA. 2008. Areawide rice insect pest management: aperspective of experiences in Asia. In: Koul O, Cuperus G,Elliott N, editors. Areawide pest management: theory andimplementation. Wallingford (UK): CABI. p. 351–440.

Litsinger JA, Chantaraprapha N, Bandong JP, Barrion AT.1994a. Natural enemies of the rice caseworm Nymphuladepunctalis (Guenee) (Lepidoptera: Pyralidae). Insect SciAppl. 15:261–268.

Litsinger JA, Bandong JP, Chantaraprapha N. 1994b. Massrearing, larval behaviour, and effects of plant age on therice caseworm Nymphula depunctalis (Guenee) (Lepidop-tera: Pyralidae). Crop Prot. 13:494–502.

Litsinger JA, Bandong JP, Canapi BL, dela Cruz CG, PantuaPC, Alviola III AL, Batay-an E. 2005. Evaluation ofaction thresholds against chronic insect pests of rice inthe Philippines: I. Less frequently occurring pests andoverall assessment. Int J Pest Manage. 51:45–61.

Litsinger JA, Bandong J P, Canapi BL. 2011. Effect ofmultiple infestations from insect pests and other stresseson irrigated rice in the Philippines: I. Damage functions.Int J Pest Manage. 57:93–116.

Luo G. 1985. Rice-fish culture in ricefield ditchponds. In:Mackay KT, editor. Rice-fish culture in China. Ottawa,Canada: International Development Research Centre(IDRC). p. 147–152.

Mackay KT. 1995. Rice-fish culture in China. Ottawa(Canada): International Development Research Centre(IDRC), 265 p.

Meien VA. 1940. Fish culture in rice fields. Moscow, Russia:Food Industry Publishers. Machackova B, translator;Fernando CH, editor.

Moulton TP. 1973. More rice and less fish: some problemswith the ‘‘Green Revolution’’. Austral Nat Hist. 6:322–327.

Nakasuji F, Dyck VA. 1984. Evaluation of the role ofMicrovelia douglasi atrolieata (Bergroth) (Heteroptera:Veliidae) as predator of the brown planthopper Nilapar-vata lugens (Stal) (Homoptera: Delphacidae). Res. PopulEcol. 26:134–148.


Pan S, Huang Z, Zheng J. 1995. Ecological mechanisms forincreasing rice and fish production. In: Mackay KT,editor. Rice-fish culture in China. Ottawa (Canada):International Development Research Centre (IDRC). p.195–200.

Ooi PAC, Shepard BM. 1994. Predator and parasitoids ofrice insect pests. In: Heinrichs EA, editor. Biology andmanagement of rice insects. New Delhi (India): WileyEastern. p. 585–612.

Pao LL. 1981. The present status of biocontrol of mosquitoesin China. In: Laird M, editor. Biocontrol of medical andveterinary pests. Proceedings of the 16th internationalCongress of Entomology, Kyoto, Japan. New York:Praeger. p. 54–77.

Regis J, Pattee E, Lebreton JD. 1981. A new method forevaluating the efficiency of electric fishing. Arch Hydro-biol Beih. 93:68–82.

Rothuis AJ, Nam CQ, Richter CJJ, Ollevier F. 1998.Polyculture of silver barb, Puntius gonionotus (Bleeker),Nile tilapia, Oreochromis niloticus (L.), and commoncarp, Cyprinus carpio L., in Vietnamese rice fields: fishproduction parameters. Aquaculture Res. 29:661–668.

Schoenly KG, Justo HD, Barrion AT, Harris MK, BottrellDG. 1998. Analysis of invertebrate biodiversity in aPhilippine farmer’s irrigated rice field. Environ Entomol.27:1125–1136.

Schoenly KG, Cohen JE, Heong KL, Litsinger JA, BarrionAT, Arida GS. 2010. Fallowing did not disrupt inverte-brate fauna in Philippine low-pesticide irrigated ricefields. J Appl Ecol 47:593–602.

Settle WH, Ariawan H, Astuti ET, Cahyana W, Hakim AL,Hindaya-na, Lestari AS, Pajarningsih. 1996. Managingtropical rice pests through conservation of generalistnatural enemies and alternate prey. Ecology. 77:1975–1988.

Shiraki T. 1917. Paddy borer (Schoenobius incertellusWalker). Taihoku Agricultural Experiment Station,Formosa. 256 p.

Simpson I, Roger PA, Oficial B, Grant IF. 1994. Effects offertilizer and pesticide management on floodwaterecology of a wetland ricefield: II. Dynamics of micro-crustaceans and dipteran larvae. Biol Fertil Soils. 17:138–146.

Sinhababu DP, Majumdar N. 1981. Evidence of feeding onbrown planthopper, Nilaparvata lugens (Stal) by commoncarp, Cyprinus carpio var. communis L. J Inland Fish SocIndia. 13:17–21.

Spataru P. 1978. Food and feeding habits of Tilapia zillili(Gervais) (Cichlidae) in Lake Kinneret (Israel). Aqua-culture. 14:327–338.

Teo SS. 2006. Evaluation of different species of fish forbiological control of golden apple snail Pomaceacanaliculata (Lamarck) in rice. Crop Prot. 25:1004–1012.

Tuan NA. 1994. Culturing fish in ricefields as a means ofpreventing insect infestations: preliminary observations.In: dela Cruz CR, editor. Role of fish in enhancingricefield ecology and in integrated pest management.International Center for Living Aquatic ResourcesManagement (ICLARM) conference proceedings No.43. p. 31 (abstract).

Vromant N, Rothuis AJ, Cuc NTT, Ollevier F. 1998. Theeffect of fish on the abundance of the rice casewormNymphula depunctalis (Guenee) (Lepidoptera: Pyralidae)in direct seeded, concurrent rice-fish field. Biocont SciTechnol. 8: 539–546.

Vromant N, Nhan DK, Chau NTH, Olleviera F. 2002. Canfish control planthopper and leafhopper populations inintensive rice culture? Biocont Sci Technol. 12:695–703.

Vromant N, Nhan DK, Chau NTH, Olleviera F. 2003. Effectof stocked fish on rice leaffolder Cnaphalocrocis medinalisand rice caseworm Nymphula depunctalis populations inintensive rice culture. Biocont Sci Technol 13:285–297.

Waibel H. 1992. Comparative economics of pesticide use inrice and rice–fish farming. In: dela Cruz CR, Lightfoot C,Costa-Pierce BA, Carangal VR, Bimbao MP, editors.Rice–fish research and development in Asia. Interna-tional Center for Living Aquatic Resources Management(ICLARM) conference proceedings No. 24. p. 245–254.

Wan Q, Li K, Li P, Gu H, Zhou X. 1995. Developing rice-fishcultures in shallow waters of lakes. In: Mackay KT, editor.Rice-fish culture in China. Ottawa (Canada): InternationalDevelopment Research Centre (IDRC). p. 67–74.

Washino RK, Hokama Y. 1967. Preliminary report on thefeeding pattern of two species of fish in a rice fieldhabitat. Proc Pap Annu Conf Calif Mosq Cont Assoc.35:84–87.

Way MJ, Heong KL. 1994. The role of biodiversity in thedynamics and management of insect pests of tropicalirrigated rice: a review. Bull Entomol Res. 84:567–587.

Way MO, Grigarick AA, Litsinger JA, Palis F, Pingali PL.1991. Economic thresholds and injury levels for insectpests of rice. In: Heinrichs EA, Miller TA, editors. Riceinsects: management strategies. New York: Springer-Verlag. p. 67–105.

Weimin M. 2010. Recent developments in rice-fish culture inChina: a holistic approach for livelihood improvement inrural areas. In: De Silva SS, Davy FB, editors. Successstories of Asian aquaculture. Ottawa (Canada): Springerand International Development Research Centre(IDRC). p. 15–40.

Wu L. 1995. Methods of rice-fish culture and their ecologicalefficiency. In: Mackay KT, editor. Rice-fish culture inChina. Ottawa (Canada): International DevelopmentResearch Centre (IDRC). p. 91–96.

Wu N, Liao G, Lou Y, Zhong G. 1995. The role of fish incontrollingmosquitoes in rice fields. In:MackayKT, editor.Rice-fish culture in China. Ottawa, Canada: InternationalDevelopment Research Centre (IDRC). p. 213–216.

Xiao F. 1995. Fish culture in ricefields: rice fish symbiosis. In:Mackay KT, editor. Rice-fish culture in China. Ottawa(Canada): International Development Research Centre(IDRC). p. 183–194.

Xiao QY. 1992. Role of fish in pest control in rice farming.In: dela Cruz CR, Lightfoot C, Costa-Pierce BA,Carangal VR, Bimbao MP, editors. Rice-fish researchand development in Asia. International Center for LivingAquatic Resources Management (ICLARM) conferenceproceedings No. 24. p. 235–243.

Xu Y, Guo Y. 1992. Rice-fish farming systems research inChina. In: dela Cruz CR, Lightfoot C, Costa-Pierce BA,Carangal VR, Bimbao MP, editors. Rice-fish researchand development in Asia. International Center for LivingAquatic Resources Management (ICLARM) conferenceproceedings No. 24. p. 315–323.

Yu SY, Wu WS, Wei HF, Ke DA, Xu JR, Wu QZ. 1995.Ability of fish to control rice diseases, pests, and weeds.In: Mackay KT, editor. Rice-fish culture in China.Ottawa (Canada): International Development ResearchCentre (IDRC). p. 223–228.

Zalewski M. 1985. The estimate of fish density and biomassin rivers on the basis of relationships between specimensize and efficiency of electrofishing. Fish Res. 3:147–155.


Efficacy of Common Carp and Nile Tilapia as biocontrol agents of rice insect pests in the Philippines

Documents