BIOLOGICAL AND MICROBIAL CONTROL

Insecticides Suppress Natural Enemies and Increase PestDamage in Cabbage

RICCARDO BOMMARCO,1,2 FREDDY MIRANDA,3 HELENA BYLUND,1

AND CHRISTER BJORKMAN1

J. Econ. Entomol. 104(3): 782Ð791 (2011); DOI: 10.1603/EC10444

ABSTRACT Intensive use of pesticides is common and increasing despite a growing and historicallywell documented awareness of the costs and hazards. The beneÞts from pesticides of increased yieldsfrom sufÞcient pest control may be outweighed by developed resistance in pests and killing ofbeneÞcial natural enemies. Other negative effects are human health problems and lower pricesbecause of consumersÕ desire to buy organic products. Few studies have examined these trade-offs inthe Þeld. Here, we demonstrate that Nicaraguan cabbage (Brassica spp.) farmers may suffer eco-nomically by using insecticides as they get more damage by the main pest diamondback moth, Plutellaxylostella (L.) (Lepidoptera: Plutellidae), at the same time as they spend economic resources oninsecticides. Replicated similarly sized cabbage Þelds cultivated in a standardized manner were eithertreated with insecticides according common practice or not treated with insecticides over two seasons.Fields treated with insecticides suffered, compared with nontreated Þelds, equal or, at least in someperiods of the seasons, higher diamondback moth pest attacks. These Þelds also had increased leafdamage on the harvested cabbage heads. Weight and size of the heads were not affected. The farmersreceived the same price on the local market irrespective of insecticide use. Rates of parasitizeddiamondback moth were consistently lower in the treated Þelds. Negative effects of using insecticidesagainst diamondback moth were found for the density of parasitoids and generalist predatory wasps,and tended to affect spiders negatively. The observed increased leaf damages in insecticide-treatedÞelds may be a combined consequence of insecticide resistance in the pest, and of lower predationand parasitization rates from naturally occurring predators that are suppressed by the insecticideapplications. The results indicate biological control as a viable and economic alternative pest man-agement strategy, something that may be particularly relevant for the production of cash crops intropical countries where insecticide use is heavy and possibly increasing.

KEY WORDS Nicaragua, Brassicae oleracea, Plutella xylostella, Diadegma insulare, Polybia sp.

Insecticides are generally considered as a key input toobtain high agricultural productivity, but their use isnot unequivocally positive. The number of observedcases of pest species that are resistant against insec-ticides has steadily increased since the large-scale in-troduction of insecticides that formed part of theGreen Revolution (Whalon et al. 2008). In parallel,insecticides have repeatedly been demonstrated to belethal or sublethal to beneÞcial arthropods that arenatural enemies to crop pests (Croft 1990, Desneauxet al. 2007). Early studies show that eliminating naturalenemies with insecticides often lead to more severepest attacks as the natural pest control exerted byenemies is reduced (Pickett and Patterson 1953, Huf-faker and Kenneth 1956, Ripper 1956, Dempster 1968).Based on these Þndings, an expectation is that chem-ical treatment against an insecticide resistant pest, via

the elimination of natural enemy control, more fre-quently will increase pest attacks, leading to decreasednet economic gain to the farmer and perhaps evenyield loss. Despite a steadily increasing use of insec-ticides worldwide, there is with a few exceptions (Fur-long et al. 2004, 2008) a lack of up to date experimentalevidence clearly monitoring and testing this in majoragricultural crops.

A related problem are secondary pests that occurwhen insecticides eliminate nontarget parasitic andpredatory natural enemies so that previously low andeconomically unimportant populations of herbivorousinsects are released from natural control and becomecrop pests. In some cases, emergence of secondarypests in combination with the development of insec-ticide resistance in the primary pest, has led to thecollapse of entire cropping systems, such as for cotton,Gossypium hirsutum L., in Central America (Metcalf1980). Despite these catastrophes, not enough efforthas been devoted to research on the impact of insec-ticides on naturally occurring beneÞcial organisms:The few cases that exist clearly demonstrate the efÞ-

1 Department of Ecology, Swedish University of Agricultural Sci-ences, Box 7044, SE-75007 Uppsala, Sweden.

2 Corresponding author, e-mail: riccardo.bommarco@slu.se.3 Universidad Nacional Agraria, Box 453, Managua, Nicaragua.

0022-0493/11/0782Ð0791$04.00/0 � 2011 Entomological Society of America

ciency of conservation biological control and a trade-off between biological and chemical control of pestinsects (Settle et al. 1996; Heong and Schoenly 1998;Furlong et al. 2004, 2008; KÞr 2004).

Pesticides, however, remain the main approach tomanaging agricultural pests worldwide. Pesticide useis, in fact, increasing in response to intensiÞed foodproduction mobilized to meet an increasing demand(Tilman et al. 2001). Primary reliance on natural en-emies in integrated pest management (IPM) schemeshas in reality not been implemented to its full extent(Zalucki et al. 2009). The situation is perhaps partic-ularly alarming in developing countries where farmersand their environment are exposed to increasingamounts of pesticides, to a large extent including ac-tive ingredients that have been banned or restricted inindustrialized countries due to their recognized tox-icity and persistence (Ekstrom and Ekbom 2011). Es-timates of actual beneÞts from insecticide use in termsof the farmerÕs economy and agricultural productivityare lacking in developing countries. An importantshortcoming in many cropping regimes is that thegains of the ecosystem services provided by naturallyoccurring beneÞcial arthropods through pest controlremain largely unexplored or neglected (Gardiner etal. 2009).

An example of this is the diamondback moth, Plu-tella xylostella (L.) (Lepidoptera: Plutellidae), a seri-ous pest insect on cruciferous crops worldwide (Sar-fraz et al. 2005, Grzywacz et al. 2010). Diamondbackmoth in cabbage (Brassica spp.) is a suitable candidateto study how insecticide applications affect an insec-ticide-resistant pest, predator, and parasitoid popula-tions, crop damage, and yields. Although Furlong et al.(2004, 2008) have done some ground-breaking re-search showing the negative effects of insecticide useon diamondback moth control via natural enemiesthere is still more to learn. Diamondback moth wasone of the very Þrst species reported to be resistant toDDT and is now known to be resistant against manyinsecticides (Ovalle and Cave 1989, Talekar and Shel-ton 1993, Perez et al. 2000). Despite considerableresearch, not enough is known about the role of nativenatural enemies in the regulation of diamondbackmoth populations. Important exceptions are certainparasitoid species, in particular Diadegma spp. (Hy-menoptera: Ichneumonidae), which also have beenreared and released as biological control agents (Ooi1992, Poelking 1992). Other natural enemies of dia-mondback moth, such as spiders, predatory beetles,and bugs, have been identiÞed as effective controlagents against other pests (Murdoch et al. 1985, Sy-mondson et al. 2002), but their potential role in thecontrol of diamondback moth has yet to be quantiÞedto a higher extent (but see Sarfraz et al. 2005, Mirandaet al. 2011).

To assess the efÞcacy of insecticides to control di-amondback moth in cabbage, and also to obtain anindication of the impacts on and beneÞts of naturalenemies, we examined effects of insecticide treat-ments on time series of diamondback moth, its pred-ators and parasitoids, diamondback moth parasitism

rates, crop damage, and yield for two crop cycles inNicaraguan head cabbage. Entire similarly cultivatedand sized Þelds on separate farms were either treatedwith insecticides according to common practice or nottreated with insecticides against diamondback moth.We assessed how the use of insecticides affected 1)seasonal predator and pest population development,2) parasitism rates of diamondback moth, 3) cropdamage measured as feeding damage on the cabbagehead at harvest, and 4) the Þnal pest managementcosts, crop yield and sale price for each farmer.

Materials and Methods

Study Area. Field experiments were performed inthe Tisey-Estanzuela area (12� 59�208� N, 86� 21� 973�W) in Nicaragua at 1,348Ð1,483-m elevation, near thecity of Estelõ. Farms in the area are usually family farmsand the main crops grown are potatoes (Solanumspp.), cabbage, maize (Zea mays L.), and beans(Phaseolus spp.). The Þeld study took place during twoperiods: from June to September 2006, correspondingto the growing period called Primera, and from Sep-tember to December 2006, corresponding to the grow-ing period called Postrera. Both growing periods occurduring the rainy season.Field Experiment. During Primera, 2006, we se-

lected six farms, each with a cabbage Þeld that was nottreated with insecticides, and six farms, each with acabbage Þeld that was treated with insecticides. Werepeated theexperiment inPostrera, 2006, inÞveÞeldsthat were not treated with insecticides, and Þvetreated Þelds, all on separate farms. No single Þeld wasused in two seasons. All experimental Þelds were sim-ilarly sized (mean, 0.21 ha; maximum, 0.30 ha; mini-mum, 0.09 ha; Tables 1 and 2). The treated andnontreated Þelds were spatially separated and inter-spersed with a mean distance between Þelds of 2.8 km(maximum, 8.3 km; minimum, 0.1 km).

A common practice to control diamondback moth,the main pest insect on cabbage in Nicaragua, is toapply insecticides, sometimes as frequently as twiceper week (personal communications with farmers).Farms using insecticides in our study sprayed theexperimental Þelds several times according to com-mon practice. The farmers decided themselves howmany times to spray and what substance to use tocontrol the diamondback moth, and Þelds weretreated six to 10 times each crop cycle with variousformulations (Tables 1, 2). The farms not using insec-ticides had stopped using insecticides in 2005 exceptfor one farm (farm 3) that had not treated with in-secticides since 2000. To control weeds, all farms madeone application of the herbicides Gramoxone or Para-quat (3 liters/ha).

Head cabbage of the variety capitata (Izalco) wassown in trays with small holes in a netted rearing areafor farmers not treating with insecticides and in aseedbed for farmers by using insecticides. This differ-ence in rearing regime did not affect the seedlings orthe plants in the Þeld in any detectable way after visualobservation. The seedlings were, as is common prac-

June 2011 BOMMARCO ET AL.: INSECTICIDES INCREASE DAMAGE IN CABBAGE 783

tice, transplanted by hand at a row spacing of 0.6 and0.5 m between plants. All Þelds received inorganicfertilization equivalent to 182 kg/ha (formula NÐPÐK,12Ð30Ð10) 8 d after transplanting, and an additional 95kg/ha nitrogen fertilizer (N, 45) applied 30 d aftertransplanting.Densities of Diamondback Moth and Natural Ene-mies. Densities per plant of natural enemies, and ofeggs, larvae, pupae and adults of diamondback moth,

were assessed by weekly visual counts. Five samplingpoints, each including 10 cabbage plants, were in-spected in each Þeld. Each sampling point was a 2- by2-m area including two to three rows of cabbage fromwhich the 10 plants were randomly selected. Differentsampling points were selected each week. We re-corded the most abundant natural enemies on theplants, which were spiders, and Polybia sp. (Hyme-noptera: Vespidae), a eusocial predatory wasp that ßy

Table 1. Primera growing season information on farms and insecticide applications

Farmno.

Active ingredientNo.

applicationsField size

(ha)Spray cost

Sum spraycosts

Price/head Crop value

1 0.30 0.142 1,4182 0.20 0.148 9833 0.25 0.17 1,4174 0.25 0.198 1,6545 0.15 0.085 4256 0.09 0.113 3407 Betacyßuthin � thiacloprid 2 0.30 73.1

B. thuringensis 2 43.1Deltametrin 2 33.6Spinosad 2 70.4 220.3 0.113 1,134

8 B. thuringensis 2 0.15 43.1Deltametrin 4 67.2Methamidophos � deltametrin 2 44.1Methamidophos � sdimetil-fosforamidotioato 2 33.6 188.1 0 0

9 B. thuringensis 4 0.18 86.2Deltametrin 1 16.8Methamidophos 2 33.6 136.7 0.142 851

10 B. thuringensis 5 0.25 107.8Deltametrin 1 16.8Spinosad 3 105.6 230.2 0.142 1,187

11 Fipronil 3 0.15 105.6Indoxacarb 2 70.4Oxamyl 1 25.2 201.2 0.156 780

12 Indoxacarb 7 0.15 246.4Methamidophos � sdimetil-fosforamidotioato 1 16.8 263.2 0.057 285

All costs and prices are in U.S. dollars. Spray costs include workers payment of $11.56 per application. Crop value is calculated from a cropdensity of one head per 0.3 m2.

Table 2. Postrera growing season information on farms and insecticide applications

Farmno.

Active ingredientNo.

applicationsField size

(ha)Spray cost

Sum Spraycosts

Price/head Crop value

1 0.25 0.112 9332 0.26 0.084 7283 0.25 0.112 9334 0.2 0.112 7465 0.14 0.14 653

13 Betacyßuthin � thiacloprid 1 0.3 36.6Endosulfan 1 16.8Indoxacarb 2 70.4Spinosad 2 70.4 194.2 0.14 1400

8 Deltametrin 4 0.15 67.2Methamidophos � sdimetil-fosforamidotioato 4 67.2B. thuringensis 2 43.1 177.6 0.056 280

9 Deltametrin 2 0.18 33.6Endosulfan 1 16.8B. thuringensis 2 43.1Methamidophos 1 16.8Spinosad 2 70.4 180.8 0.112 672

10 Endosulfan 1 0.25 16.8Spinosad 7 246.4 263.2 0.14 1167

11 Fipronil 6 0.15 211.2B. thuringensis 2 43.1Indoxacarb 1 35.2 289.5 0.112 560

All costs and prices are in U.S. dollars. Spray costs include workers payment of $11.56 per application. Crop value is calculated from a cropdensity of one head per 0.3 m2.

784 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 3

from its nest into the Þeld to forage. Parasitoid pupaeand adults were also recorded. In this region, dia-mondback moth is mainly parasitized by Diadegmainsulare (Cresson) (Hymenoptera: Ichneumonidae)(Cordero and Cave 1992).Parasitism Rates. To estimate the rate of parasitism

on diamondback moth, 30Ð50 third- to fourth-instarlarvae were collected every second week from eachÞeld. The collected larvae were reared individuallyuntil pupation in plastic cups (4.5 cm in diameter, 4 cmin height) in the laboratory, with slices of fresh cab-bage leaves as food. The degree of parasitism wasdetermined based on the number of parasitized andunparasitized diamondback moth larvae. Larvae werethen reared until eclosion. Emerging adult parasitoidswere determined to species following Azidah et al.(2000).Crop Damage, Yield, Costs, and Sales Price. Ten

randomly chosen cabbage heads were collected fromeach experimental Þeld within 1 wk before harvest. Todetermine the damage level, 10 wrapper leaves, i.e.,the outermost leaves, were inspected on each head.The damage rating was based on the measure devel-oped by Chalfant and Brett (1965), where the leafdamage on each wrapper leaf was classiÞed into oneof six classes based on percentage of leaf damage: 0, 1,5, 10, 25, or �50%. The yield was estimated by mea-suring the weight (kilograms), diameter (centime-ters), and height (centimeters) of the 10 cabbageheads. The balance between costs for insecticidetreatments and income from the cabbage yield wascalculated for each Þeld (Tables 1 and 2). Cost for theinsecticide is calculated from market price per hectarein 2006 and the surface area of each Þeld. Farmersoften employ workers to apply the pesticides. Workercosts for one insecticide application for a 0.2-ha Þeldwas estimated to be US$11.56 based on four hours ofwork for two persons, one supplying water and oneapplying the insecticide. The sales price for each Þeldwas the price reported by each farmer when sellingthe cabbage heads to retailers in Estelõ.Statistics. We used linear repeated measures in

mixed effects models to test the importance of insec-ticide regimeon thenumberperplantofdiamondbackmoth eggs, larvae, pupa, and adults over each season.We further tested the effect of insecticides on para-sitism rates, development of number of adults andpupae of the parasitoids, and on number of spiders andPolybiawasps per plant over each growing season. Allpopulation densities (N) were log-transformed[log10(N� 0.02)], where 0.02 was the lowest possibleobserved value in a Þeld in a week, i.e., one observa-tion on 50 plants.

Sampling occasion and insecticide treatment, andthe interaction between these were assumed as Þxedfactors in the mixed model. Because the consecutivemeasurements over time are likely to covary within aÞeld, we analyzed data with a repeated measures ap-proach with Þeld as the measurement unit. Using Ai-kaikeÕs information criteria (AIC), we tested, for eachdependent variable separately, which covariancestructure to assume, given the data. We tested com-

pound symmetric, Þrst-order autoregressive, and un-structured covariance structures, and chose the onethat gave the best (lowest) AIC value in each case(Littell et al. 2006). Effects of insecticide treatmentson crop damage, yield measures, and sale price wereanalyzed with mixed models with season � treat-ment � farm as random factor to consider the sam-pling within farm. The average percentage of damageper head in a Þeld was arcsine and square root trans-formed before the statistical analysis. Proc Mixed inSAS 9.2 for Linux (SAS Institute, Cary, NC) was usedfor all statistical analyses.

Results

Densities of Diamondback Moth. Due to el Ninoclimatic events, the weather was unusually dry in 2006,with a total precipitation of 208 mm in Primera and 238mm in Postrera. Dry weather usually enhances dia-mondback moth attacks as population growth rateincreases with shortened generation time in the warmweather and the insects are not disturbed by rainfall.

A signiÞcant interaction week � treatment for di-amondback moth pupae in Primera and diamondbackmoth adults in Postrera shows that these were moreabundant at least some weeks during the season (Fig.1; Table 3). In general however there was no greatdifferences in abundance in Þelds treated with insec-ticides compared with nontreated Þelds (Fig. 1; Table3). Diamondback moth larvae per plant, which prob-ably is the measure that best estimates diamondbackmoth abundance in a Þeld, surpassed an economicthreshold of 0.3Ð0.5 individual per plant (Kirby andSlosser 1984, Diaz et al. 1999) in the Primera season inboth treated (mean � 1.6 diamondback moth larvaeper plant, SE � 0.32, N� 54) and untreated (mean �0.82, SE � 0.14, N� 54) Þelds. In the Postrera season,the average attack was lower and closer to the eco-nomic thresholds both in treated (mean � 0.68, SE �0.11,N� 45), and untreated Þelds (mean � 0.48, SE �0.04,N� 45). In no case did we Þnd signiÞcantly lowerdiamondback moth densities in treated Þelds, indicat-ing that Þelds repeatedly treated with insecticides didnot result in a greater pest suppression compared withÞelds under natural control (Fig. 1; Table 3).ParasitismRates.All emergent parasitoids wereD.insulare. The rate of parasitism on diamondbackmoth was higher on farms not using insecticidesthan on farms using insecticides in both seasons(Fig. 2; Table 4).Natural Enemies. Natural enemy density was gen-

erally lower in Þelds treated with insecticides in bothseasons although not always signiÞcantly so. In severalcases, the natural enemy populations started off atsimilar abundances that then diverged in the differenttreatments as shown by a signiÞcant week � treatmentinteraction (Fig. 3; Table 4). The difference betweentreatments was apparent for parasitism rates in bothseasons and for Diadegman adults and pupae and forPolybiawasps in the Primera season (Fig. 3; Table 4).Crop Damage, Yield, and Sales Price. Leaf dam-

age was higher in Þelds treated with insecticides in

June 2011 BOMMARCO ET AL.: INSECTICIDES INCREASE DAMAGE IN CABBAGE 785

Postrera (F1, 90 � 8.0; P � 0.006), but despite ap-parent differences in average damage levels (Fig. 4)not signiÞcantly so in Primera (F1, 108 � 2.7; P �0.10). The weight (mean [SE]) of a cabbage headdid not differ between treated (Primera, 1.93 [0.24]kg; Postrera, 2.18 [0.26] kg) and untreated Þelds(Primera, 1.68 [0.20]; Postrera, 1.70[0.14]) (P �0.11). Head diameter or height did not differ be-

tween treatments in either season (P� 0.14). Farm-ers received the same price per cabbage head irre-spective of insecticide use in either season (P �0.20). Farmers using insecticides spent between 14and 92% (mean � 35%) of their crop value forinsecticides, excluding one farmer whose cropfailed entirely (calculations are based on informa-tion in Tables 1 and 2).

Fig. 1. Mean number of diamondback moth per plant throughout the season in cabbage Þelds using insecticides and inÞelds with no use of insecticides in Nicaragua. Data from the two growth seasons Primera and Postrera are presented. Errorbars show SEs.

Table 3. Results for effects of insecticide applications on the life stages of diamondback moth in head cabbage fields in two croppingseasons: Primera and Postrera, 2006

Dependentvariable

Primera Postrera

Effect df F PCovariancestructure

df F PCovariancestructure

Egg Treatment 1, 10 0.4 0.57 un 1, 8 0.0 0.91 csWk 8, 10 17.0 �0.0001 8, 64 1.2 0.31Wk � treatment 8, 10 2.5 0.086 8, 64 1.4 0.22

Larvae Treatment 1, 10 1.7 0.23 un 1, 8 0.1 0.80 ar(1)Wk 8, 10 19.5 �0.0001 8, 64 3.6 0.0018Wk � treatment 8, 10 1.8 0.20 8, 64 0.2 0.99

Pupa Treatment 1, 10 1.9 0.20 un 1, 8 0.2 0.66 csWk 8, 10 8.1 0.0017 8, 64 4.0 0.0006Wk � treatment 8, 10 3.3 0.042 8, 64 0.6 0.79

Adult Treatment 1, 10 0.0 0.84 ar(1) 1, 8 0.3 0.61 csWk 8, 80 7.9 �0.0001 8, 64 5.1 �0.0001Wk � treatment 8, 80 0.5 0.83 8, 64 2.6 0.014

Note that the calculation of degrees of freedom may vary depending on the choice of covariance structure. Covariance structures areunstructured (un), compound symmetry (cs), or Þrst order autoregression ar(1).

786 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 3

Discussion

We found that diamondback moth densities werenot affected or became somewhat higher in Þeldstreated with insecticides. The rates of parasitized di-amondback moth were consistently higher in un-treated Þelds in both seasons. In addition, the densitiesof two types of generalist predators, spiders and apredatory wasp, were generally lower in sprayed com-pared with unsprayed Þelds, although not always sig-niÞcantly so. This pattern was most obvious in the Þrst(Primera) crop cycle. The crop suffered more damageon the wrapper leafs at harvest in sprayed comparedwith unsprayed Þelds, but crop yields, measured ascabbage head size and weight, did not differ due toinsecticide applications. Finally, differences in dam-age levels were not reßected in the sales price for thefarmer that, on average, received the same price perhead whether they had treated their Þeld or not. Thus,farmers that did not use insecticides made a largerproÞt as they did not invest in insecticides.

Diamondback moth populations have repeatedlybeen found to be resistant against most insecticides

used against them (Ovalle and Cave 1989, Talekar andShelton 1993, Grzywacz et al. 2010). In a study, thatincluded diamondback moth populations from theTisey area, Nicaraguan diamondback moth popula-tions demonstrated 40Ð200-fold higher LD50 valueswith Þve common insecticide types compared with acontrol population (Perez et al. 2000). Diamondbackmoth resistance has even been found to commercialformulations of Bacillus thuriengiensis ssp. kurstaki(Perez and Shelton 1997). The seemingly unaffectedpopulation levels of diamondback moth in Þelds thatwere treated with insecticides up to 10 times in a cropcycle, are consistent with the well documented highlevels of resistance in diamondback moth to pyre-throids and other insecticides. Less well documentedis possible resistance against more modern insecti-cides, such as spinosad and the neonicotinoid thiaclo-prid, that were used by farmers in the current study.

In insecticide-treated Þelds, we observed declinesin abundances of natural enemies and consistentlylower parasitism rates. This indicates that these com-mon natural enemies are vulnerable nontarget organ-

Fig. 2. Levels of parasitism by D. insulare on DBM in cabbage Þelds by using insecticides and in Þelds with no use ofinsecticides after transplanting of cabbage seedlings into the Þelds. Results from the two growth seasons, Primera and Postrera,are presented. Error bars show SEs.

Table 4. Results for effects of insecticide applications on natural enemy densities per plant and parasitism rates by D. insulare ondiamondback moth in head cabbage fields in two cropping seasons: Primera and Postrera, 2006

Dependentvariable

Primera Postrera

Effect df F PCovariancestructure

df F PCovariancestructure

Parasitism rate Treatment 1, 10 9.1 0.013 cs 1, 8 21.8 0.0016 csWk 3, 30 0.2 0.89 3, 24 3.1 0.048Wk � treatment 3, 30 4.1 0.015 3, 24 3.3 0.037

Adult Diadegma Treatment 1, 10 10.0 0.01 un 1, 8 4.8 0.060 csWk 8, 10 77.2 �0.0001 8, 64 4.6 0.0002Wk � treatment 8, 10 40.5 �0.0001 8, 64 1.2 0.30

Pupa Diadegma Treatment 1, 10 3.3 0.10 un 1, 8 1.7 0.23 csWeek 8, 10 15.7 �0.0001 8, 64 8.5 �0.0001Wk � treatment 8, 10 5.0 0.01 8, 64 1.5 0.16

Spiders Treatment 1, 10 3.7 0.082 un 1, 8 1.5 0.26 cswk 8, 10 6.0 0.0055 8, 64 4.9 �0.0001Wk � treatment 8, 10 2.6 0.082 8, 64 1.0 0.42

Polybia Treatment 1, 10 13.3 0.0045 cs 1, 8 1.3 0.29 csWeek 8, 80 0.9 0.49 8, 64 1.4 0.22Week � treatment 8, 80 1.1 0.36 8, 64 0.7 0.73

Note that the calculation of degrees of freedom may vary depending on the choice of covariance structure. Covariance structures areunstructured (un), or compound symmetry (cs).

June 2011 BOMMARCO ET AL.: INSECTICIDES INCREASE DAMAGE IN CABBAGE 787

isms to the insecticide applications. Negative effectson nontarget natural enemies may, together with in-secticide resistance in the pest, explain the observedunaffected or higher pest levels, often lower naturalenemy abundances, and, ultimately, increased leafdamage in Þelds treated with insecticides. Severalstudies also indicate that natural enemies are oftenmore sensitive to insecticides and evolve insecticideresistance slower compared with their prey (Hill andFoster 2000, Xu et al. 2001). The farmers in the currentstudy used both modern biological and chemical in-secticides with various modes of action, as well asolder broad-spectrum insecticides. In industrializedcountries the latter, that are often highly toxic tohumans, have to large extent been replaced by pyre-throids for controlling insect pests. However, also py-rethroids, neonicotinoids, and broad-spectrum con-trol agents with biological origin such as spinosad,have been shown to be highly toxic to parasitoids incabbage (Hill and Foster 2000, Xu et al. 2001, Corderoet al. 2007) and for other beneÞcial arthropods (Croft1990, Desneaux et al. 2007).

Generalist natural enemies have been found to beefÞcient to suppress pests in managed ecosystems(Murdoch et al. 1985, Symondson et al. 2002) or canprovide an important complement to the control ex-erted by specialized natural enemies (Snyder and Ives2003, Dwyer et al. 2004). The declines in generalistnatural enemy abundances that we observed in thecurrent study may well have enhanced diamondbackmoth attacks in sprayed Þelds. The actual role andeffectiveness of generalist predators for controllingdiamondback moth population is, however, not as wellstudied as the parasitoids, notablyD. insulare that hasbeen shown able to control diamondback moth pop-ulations in cabbage Þelds (Talekar and Shelton 1993,Sarfraz et al. 2005, Grzywacz et al. 2010). Our resultstogether with those of others (Furlong et al. 2004,Sarfraz et al. 2005) indicate that the role of predatorsas biological control agents in cabbage deserves moreattention. Active management of parasitoids and pred-ators in IPM schemes with natural control as primarystrategy may provide an economically viable alterna-tive to the current insecticide based pest control prac-

Fig. 3. Density (log10) of natural enemies per head cabbage plant over two crop cycles (Primera and Postrera), of spiders,Polybia sp., pupae, and adults of the parasitoid D. insulare. Error bars show SEs.

788 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 3

tice. For this, management schemes need to be de-veloped based on a thorough understanding ofdiamondback moth population control by multiplepredatory arthropods that underpin the ecosystemservice of biological pest control in diverse agroeco-systems (Vandermeer et al. 2010).

Insecticides were applied six to 10 times each seasonand more applications per crop cycle are not unusualin Nicaragua. Pest management costs are therefore ofsubstantial importance for the farmerÕs economy andvaried between 14 and 92% of the crop value. Thisinvestment was, however, not returned to the farmersas the market price was similar between treatments.Also, farmers without insecticides did not receive ahigher sales price despite lower crop damage levels.This is probably a result of a crude pricing systemwhere the buyer estimates diamondback moth dam-age and head size in discussion with the farmer. Thereis, furthermore, no organized market for organic ag-ricultural products in Nicaragua. Excessive use of pes-ticides also has created obstacles for the export ofCentral American agricultural products for more prof-itable markets, such as the United States and Canada,that is hindered due to excessive pesticide residues(Nicholls and Altieri 1997). These factors are crucialto consider for the farmerÕs economy and should beincluded when calculating the economic beneÞt ofinvesting in insecticides and for developing futureefÞcient pest management schemes.

There is, especially in developing countries, still aserious lack of up to date studies that quantify theactual agronomic and economic beneÞts of pesticides,weighed against their potential drawbacks. Pesticidescan, for instance, have serious consequences for entireecosystems (Pimentel and Edwards 1982), general

biodiversity and the services it provides (Gibbs et al.2009, Geiger et al. 2010). Health risks due to pesticideexposure in Central America and elsewhere are in-creasingly highlighted but remain poorly documentedworldwide (Nicholls and Altieri 1997, Kesavachan-dran et al. 2009). A recent study shows that pesticideexposure is a public health concern in Nicaragua withan estimated 50,000Ð80,000 cases of acute pesticidepoisoning in a year (Corriols et al. 2010).

The main idea with IPM when it was conceived, wasthat natural control alternatives should be imple-mented as the main control strategy with chemicalsonly as a backup option designed to interfere as littleas possible with nonchemical options (Kogan 1998).This idea needs to be actualized as excessive use ofpesticide has evidently continued. Resistance man-agement and IPM are not implemented in Nicaraguaand probably not in many other developing countries.The results from the current study indicate that atransition to IPM based on biological pest control ismotivated not only by health and environmental rea-sons but also by considering farmerÕs economy.

Acknowledgments

We thank Adriana Diaz Valery for Þeld assistance. Weare grateful to the participating farmers for their helpfulcollaboration and Barbara Ekbom for constructive com-ments on the manuscript. This study was Þnancially sup-ported by The Swedish International Development Co-operation Agency/Department for Research Cooperation(SIDA/SAREC). R.B. was, in addition, supported by theresearch program Multifunctional Agriculture: HarnessingBiodiversity for Sustaining Agricultural Production andEcosystem Services (SAPES) funded by the Swedish Re-search Council for Environment, Agricultural Sciencesand Spatial planning (FORMAS).

References Cited

Azidah, A. A., M. G. Fitton, and D.L.J. Quicke. 2000. Iden-tiÞcation of the Diadegma species (Hymenoptera: Ich-neumonidae, Campopleginae) attacking the diamond-back moth, Plutella xylostella (Lepidoptera: Plutellidae).Bull. Entomol. Res. 90: 375Ð389.

Chalfant, R. B., and C. Brett. 1965. Cabbage Loopers andimported cabbageworms; feeding damage and control oncabbage in western North Carolina. J. Econ. Entomol. 58:28Ð35.

Cordero, J., and R. D. Cave. 1992. Natural enemies of Plu-tella xylostella (Lep, Plutellidae) on crucifers in Hondu-ras. Entomophaga 37: 397Ð407.

Cordero, R. J, J. R. Bloomquist, and T. P. Kuhar. 2007. Sus-ceptibility of two diamondback moth parasitoids, Dia-degma insulare (Cresson) (Hymenoptera; Ichneu-monidae) and Oomyzus sokolowskii (Kurdjumov)(Hymenoptera; Eulophidae), to selected commercial in-secticides. Biol. Control 42: 48Ð54.

Corriols, M., J. Marın, J. Berroteran, L. M. Lozano, and I.Lundberg. 2010. Incidence of acute pesticide posion-ings in Nicaragua: a public health concern. Occup. En-viron. Med. 66: 205Ð210.

Croft, B. A. 1990. Arthropod biological agents and pesti-cides. Wiley, New York.

Fig. 4. Average percentage leaf damage by diamondback moth on cabbage in Þelds with no insecticide (graybars) spraying and in Þelds with use of insecticides ac-cording to common practice (white bars), in two crops(Primera and Postrera seasons) of cabbage. Error barsshow SEs.

June 2011 BOMMARCO ET AL.: INSECTICIDES INCREASE DAMAGE IN CABBAGE 789

Dempster, J. P. 1968. The control of Pieris rapaewith DDT.II. Survival of the young stages of Pieris after spraying.J. Appl. Ecol. 5: 451Ð462.

Desneux, N., A. Decourtye, and J. M. Delpuech. 2007. Thesublethal effect of pesticides on beneÞcial arthropods.Annu. Rev. Entomol. 52: 81Ð106.

Diaz, J., F. Guharay, F. Miranda, J. Molina, M. Zamora, andR. Zeledon. 1999. Manejo integrado de plagas en el cul-tivo de repollo. Manual Tecnico No 38 Managua, Nica-ragua.

Dwyer, G., J. Dushoff, and S. H. Yee. 2004. The combinedeffects of pathogens and predators on insect outbreaks.Nature 430: 341Ð345.

Ekstrom, G., and B. Ekbom. 2011. Pest control in agro-ecosystems: an ecological approach. Crit. Rev. Plant Sci.30: 74Ð94.

Furlong, M. J., S. Zu-Hua, L. Yin-Quan, G. Shi-Jian, L. Yao-Bin, L. Shu-Sheng, andM. P. Zalucki. 2004. Experimen-tal analysis of the inßuence of pest management practiceon the efÞcacy of an endemic arthropod natural enemycomplex of the diamond back moth. J. Econ. Entomol. 97:1814Ð1827.

Furlong,M. J., K.H. Ju, P.W. Su, J. K.Chol, R.C. Il, andM.P.Zalucki. 2008. Integration of endemic natural enemiesand Bacillus thuringiensis to manage insect pests of Bras-sica crops in North Korea. Agric. Ecosyst. Environ. 125:223Ð238.

Gardiner,M.M., D. A. Landis, C. Gratton, C. D.Difonzo,M.O’Neal, J. M. Chacon, M. T. Wayo, N. P. Schmidt, E. E.Mueller, and G. E. Heimpel. 2009. Landscape diversityenhances biological control of an introduced crop pest inthe north-central USA. Ecol. Appl. 19: 143Ð154.

Geiger, F., J. Bengtsson, F. Berendse, W. W. Weisser, M.Emmerson, M. B. Morales, P. Ceryngier, J. Liira, T.Tscharntke, C.Winqvist, et al. 2010. Persistent negativeeffects of pesticides on biodiversity and biological controlpotential on European farmland. Basic Appl. Ecol. 11:97Ð105.

Gibbs, K. E., R. L. Mackey, and D. J. Currie. 2009. Humanland use, agriculture, pesticides and losses of imperiledspecies. Biodiversity and Distributions. 15: 242Ð253.

Grzywacz, D., A. Rossbach, A. Rauf, D. A. Russell, R. Srini-vasan, andA.M. Shelton. 2010. Current control methodsfor diamondback moth and otherBrassica insect pests andthe prospects for improved management with lepidopter-an-resistant BT vegetable brassicas in Asia and Africa.Crop Prot. 29: 68Ð79.

Heong, K. L., and K. G. Schoenly. 1998. Impact of insecti-cide on herbivore-natural enemy communities in tropicalrice ecosystems, pp 381Ð403. In P. T. Haskell and P.McEwen (eds.), Ecotoxicology: pesticides and beneÞcialorganisms. Chapman & Hall, London, United Kingdom.

Hill, T. A., and R. E. Foster. 2000. Effect of insecticides onthe Diamondback moth (Lepidoptera: Plutellidae) andits parasitoid Diadegma insulare (Hymenoptera: Ichneu-monidae). J. Econ. Entomol. 93: 763Ð768.

Huffaker, C. B., and C. E. Kenneth. 1956. Experimentalstudies on predation: predation and cyclamen mite pop-ulations on strawberries in California. Hilgardia 26: 191Ð222.

Kesavachandran, C. N., M. Fareed, M. K. Pathak, V. Bihari,N. Mathur, and A. K. Srivastava. 2009. Adverse healtheffects of pesticides in agrarian populations of developingcountries. Rev. Environ. Contam. Toxicol. 200: 33Ð52.

Kfir,R. 2004. Effect of parasitoid elimination on populationsof diamondback moth in cabbage, pp. 197Ð206. In N. M.Endersby and P. M. Ridland (eds.), The Management ofDiamondback Moth and Other Crucifer Pests. Proceed-

ings of the Fourth International Workshop, 26Ð29 No-vember 2001, Melbourne, Victoria, Australia.

Kirby, R. D., and J. E. Slosser. 1984. Composite economicthreshold for three lepidopterous pests of cabbage. J.Econ. Entomol. 77: 725Ð733.

Kogan, M. 1998. Integrated pest management: historicalperspectives and contemporary developments. Annu.Rev. Entomol. 43: 243Ð270.

Littell, R. C., G. A. Milliken, W. W. Stroup, and R. D.Wolfinger. 2006. SAS system for mixed models, 2nd ed.SAS Institute, Cary, NC.

Metcalf, L. R. 1980. Changing role of insecticides in cropprotection. Annu. Rev. Entomol. 25: 219Ð256.

Miranda, F., H. Bylund, L. Gronberg, L. Larsson, and C.Bjorkman. 2011. Population density and killing capacityby predators of eggs and larvae of the diamondback mothin Nicaragua. Environ. Entomol. 40: 333Ð341.

Murdoch, W. W., J. Chesson, and P. L. Chesson. 1985. Bi-ological control in theory and practice. Am. Nat. 125:344Ð366.

Nicholls, C. I., and M. A. Altieri. 1997. Conventional agri-cultural development models and the persistence of thepesticide treadmill in Latin America. International J. Sus-tain. Dev. World Ecol. 4: 93Ð111.

Ooi, P.A.C. 1992. Role of parasitoids in managing diamond-back moth in the Cameron Highlands, Malaysia, pp. 255Ð262. InN. S. Talekar (ed.), Diamondback Moth and OtherCrucifer Pests, Proceedings of the Second InternationalWorkshop, Tainan, Taiwan, 10Ð14 December 1990, AsianVegetable Research and Development Centre, Shanhua,Taiwan, AVRDC Publication 92-368.

Ovalle, G. O., and R. D. Cave. 1989. Determinacion de re-sistencia de Plutella xylostella L. (Lepidoptera: Plutelli-dae) a insecticidas comunes en Honduras. Ceiba 30: 119Ð128.

Perez, C. J., P. Alvarado, C. Narvaez, F. Miranda, L. Her-nandez, H. Vanegas, A. Hruska, and A. M. Shelton. 2000.Assessment of insecticide resistance in Þve insect pestsattacking Þeld and vegetable crops in Nicaragua. J. Econ.Entomol. 93: 1779Ð1787.

Perez, J. C., and A. M. Shelton. 1997. Resistance of Plutellaxylostella (Lepidoptera: Plutellidae) to Bacillus thurin-giensis Berliner in Central America. J. Econ. Entomol. 90:87Ð93.

Pickett, A. D., and N. A. Patterson. 1953. The inßuence ofspray programs on the fauna of apple orchards in NovaScotia IV. A review. Can. Entomol. 85: 472Ð498.

Pimentel, D., and C. A. Edwards. 1982. Pesticides and eco-systems. Bioscience 32: 595Ð600.

Poelking, A. 1992. Diamondback moth in the Philippinesand its control with Diadegma semiclausum, pp. 271Ð278.In N. S. Talekar (ed.), Diamondback Moth and OtherCrucifer Pests, Proceedings of the Second InternationalWorkshop, 10Ð14 December 1990, Tainan, Taiwan.

Ripper, W. E. 1956. Effect of pesticides on balance of ar-thropod populations. Annu. Rev. Entomol. 1: 403Ð438.

Sarfraz,M.,A.B.Keddie, andL.M.Dosdall. 2005. Biologicalcontrol of the diamondback moth, Plutella xylostella: areview. Biocontrol Sci. Technol. 15: 763Ð789.

Settle, W. H., H. Ariawan, E. Tri Astuti, W. Cahyana, L. A.Harin, D. Hindayana, S. A. Lestari, and Pajarningsih.1996. Managing tropical rice pests through conservationof generalist natural enemies and alternative prey. Ecol.Appl. 77: 1975Ð1988.

Snyder, W. E., and A. R. Ives. 2003. Interactions betweenspecialist and generalist natural enemies: parasitoids,predators, and pea aphid biocontrol. Ecology 84: 91Ð107.

790 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 3

Symondson, W.O.C., K. D. Sunderland, and M. H. Green-stone. 2002. Can generalist predators be effective bio-logical agent. Annu. Rev. Entomol. 47: 561Ð594.

Talekar, N. S., and A. M. Shelton. 1993. Biology, ecology,and management of the diamondback moth. Annu. Rev.Entomol. 38: 275Ð301.

Tilman, D., J. Fargione, B.Wolff, C. D’Antonio, A. Dobson, R.Howarth, D. Schindler, W. H. Schlesinger, D. Simberloff,and D. Swackhamer. 2001. Forecasting agriculturallydriven global environmental change. Science 292: 281Ð284.

Vandermeer, J., I. Perfecto, andS.Philpott. 2010. Ecologicalcomplexity and pest control in organic coffee production:

uncovering an autonomous ecosystem service. Biosci-ence 60: 527Ð537.

Whalon, M. E., D. Mota-Sanchez, and R. M. Hollingworth(eds.). 2008. Global pesticide resistance in arthropods.CABI, Oxfordshire, United Kingdom.

Xu, J., A. M. Shelton, and X. Cheng. 2001. Variation in sus-ceptibility of Diadegma insulare (Hymenoptera: Ichneu-monidae) to permethrin. J. Econ. Entomol. 94: 541Ð546.

Zalucki, M. P., D. Adamson, and M. J. Furlong. 2009. ThefutureofIPM:whitherorwither?Aust.J.Entomol.48:85Ð96.

Received 10 December 2010; accepted 8 April 2011.

June 2011 BOMMARCO ET AL.: INSECTICIDES INCREASE DAMAGE IN CABBAGE 791