The impacts of metals and metalloids on insect behavior


*Correspondence: E-mail:


In toxicology studies, the use of death as an endpoint often fails to capture the effects a pollutant has on disruptions of ecosystem services by changing an animal’s behavior. Many toxicants can cause population extinctions of insect species at concentrations well below the EC25, EC50, or EC90 concentrations traditionally reported from short-term bioassays. A surprising number of species cannot detect metal and metalloid contamination, and do not always avoid food with significant metal concentrations. This frequently leads to modified ingestion, locomotor, and reproductive behaviors. For example, some species show a tendency to increase locomotor behaviors to escape from locations with elevated metal pollution, whereas other insects greatly decrease all movements unrelated to feeding. Still others exhibit behaviors resulting in increased susceptibility to predation, including a positive phototaxis causing immatures to move to exposed positions. For purposes of reproduction, the inability to avoid even moderately polluted sites when ovipositing can lead to egg loss and reduced fitness of offspring. Ultimately, impaired behaviors result in a general reduction in population sizes and species diversity at contaminated sites, the exceptions being those species tolerating contamination that become dominant. Regardless, ecosystem services, such as herbivory, detritus reduction, or food production for higher trophic levels, are disrupted. This review evaluates the effects of metal and metalloid pollution on insect behaviors in both terrestrial and aquatic systems reported in a diverse literature scattered across many scientific disciplines. Behaviors are grouped by ingestion, taxis, and oviposition. We conclude that understanding how insect behavior is modified is necessary to assess the full scope and importance of metal and metalloid contamination.


Despite the importance of insects in most ecosystems, and the worldwide pollution of systems by heavy metals, surprisingly little information is available on the effects of metal and metalloid pollution on the behaviors of insects. Stark & Banks (2003) reviewed the available literature on toxicant effects on insects and determined that 95% of the reports used the lethal concentration, LC50 (50% of insects die), or simple mortality as a toxicological endpoint. Although short-term assays are efficient and allow for comparisons between compounds and insect species, they may not always provide a clear ecological picture on the potential effects of contaminants. For example, toxicants can affect populations enough to cause extinction at levels well below the LC values reported in the literature (Bechmann, 1994). Because insect behaviors are key contributors to the ecology of insect interactions with other plant and animal species, as well as with their abiotic environments, these behaviors are critical to the stability and diversity of ecosystems (Fisher, 1998). Thus, our review focuses on ecologically important behaviors related to ingestion, taxis, and reproduction as affected by natural and anthropogenic sources of a widespread class of pollutants: the heavy metals and metalloids.

Although there are many natural sources of elevated concentrations of metals (Boyd, 2004), anthropogenic activities such as mining, smelting, and industrial use have created both localized and regional pollution problems in nearly every country in the world (Nriagu, 1996). In some cases the pollution has been extensive enough to lead to environmental disasters and ecosystem shutdown (Hopkin, 1989; Sainz et al., 2004). Insects may be exposed by direct contact with dissolved elements in aquatic systems, via contact with contaminated soils, through airborne pollution or atmospheric deposition, and through herbivory on plants that have sequestered these materials. Predators and parasites are also exposed when feeding on insects that contain elevated concentrations of these elements (Vickerman & Trumble, 2003). A few of the more common metal and metalloid pollutants that are discussed in this paper are briefly characterized below. A detailed description of every metal is beyond the scope of this paper, instead a few key references have been included.

Zinc (Zn) is a common metal in the earth’s crust, averaging about 75 mg kg−1 soil (Emsley, 2001). For animals, Zn is an essential element; however, levels of 100–250 mg day−1 can cause significant health effects (OhioEPA, 2002). Zinc is commonly used in manufacturing of paints, dyes, wood preservatives, and rubber (Emsley, 2001). Zinc compounds found at industrial sites, at mines and nearby watersheds, and in sludge spread on agricultural fields include zinc chloride, zinc oxide, zinc sulfate, and zinc sulfide. In field conditions, hyperaccumulator plants may accumulate in excess of 12 000 mg Zn kg−1 dry weight (Deng et al., 2006).

Copper (Cu), which has been mined throughout the world for thousands of years, has many industrial uses. Widespread pollution has resulted from mining and smelting, brake dust from automobiles, uses as a marine antifoulant, the spreading of sewage sludge on agricultural lands, and the application of Cu as a fungicide in agriculture (Hutchinson & Whitby, 1974). Soil levels can exceed 2 890 mg Cu kg−1, whereas concentrations in Cu-accumulating plants have been reported above 1 300 mg kg−1 (Fernandes & Henriques, 1991). Because of the ubiquitous use in both developing and developed countries, and a long history of smelters releasing Cu as an air pollutant, Cu pollution occurs on nearly all continents and in most countries (Nriagu, 1996). As a result, insects frequently interact with elevated concentrations of Cu either through atmospheric deposition or by uptake and sequestration by plants.

Dissolved inorganic selenium (Se), in the form of sodium selenate, is sequestered by many plants. This material is available directly to aquatic insects, or may be modified into a form bioavailable to herbivores which can be either inorganic (sodium selenite) or organic (selenomethionine and selenocystine). For most plants the total Se concentration rarely exceeds 50 mg g−1, but some hyperacumulators may have total Se levels exceeding 5 000 mg g−1 (Galeas et al., 2007). Selenium is a common pollutant in most Pacific Rim countries, and this metalloid is a major pollutant in the western United States where large deposits are leached by rainfall and irrigation practices (McNeal & Balistrieri, 1989). In water collection sites without outlets, such as the Kesterson Reservoir in central California, concentrations can exceed 1 400 μg l−1 (Wu, 2004). However, most available studies examine the effects of Se at much lower concentrations. Selenium can also reach high concentrations in vegetation found near coal burning power plants and some industrial sites (McNeal & Balistrieri, 1989; Huggins et al., 2007).

Arsenic (As), commonly found as arsenate, is an important pollutant of groundwater that is often used for drinking and irrigation. Arsenic contamination has become a significant problem in Southeast Asia, where concentrations in well water may exceed 3 000 μg As l−1, and levels in soils can exceed 30 μg g−1 (Berg et al., 2007). Arsenic contamination results from natural and anthropogenic disturbance of rock, resulting in oxidation and release of inorganic forms of As (arsenate [As(V)] and arsenite [As(III)]), which are available to plants. These form arsenobetaine and arseno-sugars (among other compounds), which may be complexed with phytochelatins, which are important in heavy metal detoxification (Meharg & Hartley-Whitaker, 2002). Some plants reduce arsenate to arsenite, and then further transform it into several methylated forms (Zaman & Pardini, 1996). Thus, insects are likely exposed to a range of As species. Extreme As concentrations in surface waters, soils, and plants have been reported to result from mining effluent (particularly at gold mines) (Eisler, 2004) and as a result of burning coal high in As (Huggins et al., 2007). At industrial sites, As levels can reach up to 38 000 μg l−1 in water that is available to plants (Cappuyns et al., 2002). Arsenic tolerant vegetation can sequester concentrations of 500 to nearly 3 500 μg g−1 (Porter & Peterson, 1975). However, most reports on crops describe concentrations that are 75 μg g−1 or less (see references in Meharg & Hartley-Whitaker, 2002).

Another widespread pollutant is cadmium (Cd), with contamination resulting from the application of sludge or urban composts, pesticides, fertilizers, emissions from waste incinerators, waste water irrigation, and residues from metalliferous mining and metal smelting (McGrath et al., 2001). Though unlikely to affect plant growth, Cd negatively influences enzymatic systems of cells in higher organisms as a result of transfer up the food chain (Sanità di Toppi & Gabbrielli, 1999) and has a very long soil residency time. The US Environmental Protection Agency (USEPA) considers Cd a priority toxic pollutant, with an acute exposure limit in freshwater of 2.0 mg l−1 for up to 10 days and a chronic exposure limit of 0.25 mg l−1 (USEPA, 2001). In soils intended for agricultural use, acceptable limits range from 1 to 8 mg Cd kg−1 soil dry weight, depending on pH (Environment Agency, 2002).

Although understanding the individual effects of metals and metalloids is important, most metals occur in combination, and joint effects must be evaluated. Yang (1994) reviewed the literature on the toxicology of metals to all classes of organisms and determined that >95% of all journal articles reported the effects of individual compounds or elements. In combination, effects may not simply be additive, but possibly potentiating or antagonistic. For example, the joint toxicity of mercury and Se to an insect detritivore, the phorid fly Megasilia scalaris Loew, was strongly potentiating, with just 5% of the LC50’s of the two elements combined producing significantly increased developmental time and significantly greater mortality than the LC50 of either element alone (Jensen et al., 2006). Where available, we have included the literature that provides information on the joint effects of metals on insect behaviors. However, we recognize that in some cases the concentrations/mixtures will be difficult to replicate exactly (particularly in field studies).


ISI Web of Knowledge databases were searched using terminology including metals, metalloids, behavior, insects, and pollution. To be considered, papers were required to meet the following criteria: report quantifiable data regarding insect behavioral responses to metal pollutants, include a control or reference concentration, and include statistical analyses comparing test data. A substantial proportion of papers (>50%) did not include control or reference concentrations, or described behaviors without including data or analyses to verify that behaviors changed as a result of the contaminant. Once suitable papers were obtained, their references were examined, and ISI Web of Knowledge searches conducted for any recent papers citing those already obtained. A total of 75 papers meeting the above criteria were found. Papers from both terrestrial and aquatic systems were considered, evaluating the behavioral responses of first instars through adults. Behaviors were further divided into three categories: ingestion, taxis (locomotion), and reproductive (oviposition) behaviors. See Tables 1 and 2 for a summary of metals, species, and behavioral outcomes for these categories.

Table 1.   Summary of contaminants and the resultant behavioral outcome observed for insect species in terrestrial habitats
Metal/metalloidFormSpeciesBehavioral outcomeReference
  1. Only individual metals/metalloids are considered for behavioral outcomes because mixtures may lead to synergistic or antagonistic interactions otherwise unaccounted for in the behavioral response. Positive and negative outcomes correspond to stimulation or suppression of the particular behavior as a result of metal presence, respectively, and ‘no effect’ means the organism was unaffected at the experimental conditions.

  2. *Indicates the behavioral outcome was only observed at high concentrations and the organism was unaffected at lower concentrations.

  3. **Indicates that there was an initial positive response to low concentrations vs. controls that then became negative as concentrations increased. In choice assays, aversion results in a negative behavioral outcome. See text for the measured concentrations.

A. Ingestion behavior
 ArsenicNa2HAsO4Schistocerca americanaNegativeRathinasabapathi et al., 2007
 CadmiumCdChorthippus spec.NegativeMigula & Binkowska, 1993
Lochmaea capreaeNo effectRokytová et al., 2004
Neochetina bruchiNegativeJamil et al., 1989a,b
Neochetina eichhorniaeNo effectKay & Haller, 1986
CdCl2Frankliniella occidentalisNegativeJiang et al., 2005
Cd(NO3)2Agasicles hygrophilaNegativeQuimby et al., 1979
Bactra verutanaNo effectQuimby et al., 1979
Folsomia candidaNegativeFountain & Hopkin, 2001
CdSO4·2.67H2ODrosophila melanogasterNegativeBahadorani & Hilliker, 2009
 CopperCuFolsomia candidaNegativeFilser & Hölscher, 1997
Neochetina eichhorniaeNo effectKay & Haller, 1986
CuCl2·3Cu(OH)2Folsomia manolacheiPositiveFilser et al., 2000
Folsomia quadrioculataPositiveFilser et al., 2000
Isotomurus palustrisNo effectFilser et al., 2000
Onychiurus armatusNegativeFilser et al., 2000
Cu(NO3)2Folsomia candidaNegativeFountain & Hopkin, 2001
CuSO4Drosophila melanogasterNegativeBahadorani & Hilliker, 2009
Leptinotarsa decemlineataNegativeEl-Bassiouny, 1991
Mamestra brassicaeNegative*El-Bassiouny, 1991
Mamestra oleraceaNegative*El-Bassiouny, 1991
Pieris brassicaeNegativeEl-Bassiouny, 1991
Pieris napiNegativeEl-Bassiouny, 1991
 IronFeOrchesella cinctaNegativeNottrot et al., 1987
FeSO4·7H2ODrosophila melanogasterNegativeBahadorani & Hilliker, 2009
Fe2(SO4)3·xH2ODrosophila melanogasterNegativeBahadorani & Hilliker, 2009
 LeadPbChorthippus spec.NegativeMigula & Binkowska, 1993
Folsomia candidaNegative*Fountain & Hopkin, 2001
Lochmaea capreaeNo effectRokytová et al., 2004
Neochetina eichhorniaeNo effectKay & Haller, 1986
Pb(NO3)2Leptinotarsa decemlineataNegativeKwartirnikov et al., 1999
Orchesella cinctaNo effectvan Capelleveen et al., 1986
 ManganeseMnLochmaea capreaeNegativeRokytová et al., 2004
Orchesella cinctaNo effectNottrot et al., 1987
 NickelNiAphididaeNo effectBoyd & Martens, 1999; Jhee et al., 2005
Delia radicumNegativeJhee et al., 2005
Evergestis rimosalisNegativeJhee et al., 2005
Melanoplus femurrubrumNegativeJhee et al., 2005
Pieris rapaeNegativeMartens & Boyd, 1994
 SeleniumSePlutella xylostella G88NegativeFreeman et al., 2006
Plutella xylostella StanleyiNo effectFreeman et al., 2006
Pieris rapaeNegativeFreeman et al., 2006
Na2SeO3Spodoptera exiguaNegativeVickerman & Trumble, 1999
Na2SeO4Acheta domesticaNegativeFreeman et al., 2007
Myzus persicaeNegativeHanson et al., 2004
Spodoptera exiguaNegativeVickerman & Trumble, 1999; Vickerman et al., 2002b
Seleno-cystineSpodoptera exiguaNo effectVickerman & Trumble, 1999
Seleno-methionineSpodoptera exiguaNo effectVickerman & Trumble, 1999
 ZincZnLochmaea capreaeNegativeRokytová et al., 2004
Neochetina bruchiNegativeJamil et al., 1989a,b
ZnCl2Drosophila melanogasterNegativeBahadorani & Hilliker, 2009
Zn(NO3)2Folsomia candidaNegativeFountain & Hopkin, 2001
ZnSO4Heliothis virescensNegativeSell and Bodznick, 1971
Ostrinia nubialisNegativeGahukar, 1975
Pieris brassicaeNegativePollard & Baker, 1997
Schistocerca gregariaNegativePollard & Baker, 1997
Schistocerca gregariaNegativeBehmer et al., 2005
Spodoptera littoralisNegativeSalama & El-Sharaby, 1972
B. Taxis behavior
3 Cu(OH)2
Folsomia manolacheiNegativeFilser & Hölscher, 1997
Mesophorura macrochaetaNegativeFilser & Hölscher, 1997
Other CollembolaNo effectFilser & Hölscher, 1997; Filser et al., 2000
Pseudosinella albaNegativeFilser et al., 2000
Pterostichus cupreusNegativeBayley et al., 1995
CuSO4Drosophila melanogasterNegativeBahadorani & Hilliker, 2009
 IronFeSO4·7H2ODrosophila melanogasterPositiveBahadorani & Hilliker, 2009
 ZincZnCl2Drosophila melanogasterNo effectBahadorani & Hilliker, 2009
C. Oviposition
 CadmiumCdSO4·2.67H2ODrosophila melanogasterNegativeBahadorani & Hilliker, 2009
 ChromiumCr(VI)Megaselia scalarisNo effectTrumble & Jensen, 2004
 CopperCuSO4Drosophila melanogasterNegativeBahadorani & Hilliker, 2009
 IronFeSO4·7H2ODrosophila melanogasterNegative**Bahadorani & Hilliker, 2009
Fe2(SO4)3·xH2ODrosophila melanogasterNegative**Bahadorani & Hilliker, 2009
 SeleniumSePlutella xylostella G88NegativeFreeman et al., 2006
Plutella xylostella StanleyiNo effectFreeman et al., 2006
Pieris rapaeNegativeFreeman et al., 2006
Na2SeO4Spodoptera exiguaPositiveVickerman et al., 2002b
Spodoptera exiguaNo effectVickerman et al., 2002a
 ZincZnCl2Drosophila melanogasterNegative**Bahadorani & Hilliker, 2009
Table 2.   Summary of contaminants and the resultant behavioral outcome observed for insect species in aquatic habitats
MetalFormSpeciesBehavioral outcomeReference
  1. Only individual metals/metalloids are considered for behavioral outcomes because mixtures may lead to synergistic or antagonistic interactions otherwise unaccounted for in the behavioral response. Positive and negative outcomes correspond to stimulation or suppression of the particular behavior as a result of metal presence, respectively, and ‘no effect’ means the organism was unaffected at the experimental conditions. In choice assays, aversion results in a negative behavioral outcome. See text for the measured concentrations.

A. Ingestion behavior
 CopperCuSO4·5H2OParatanytarsus parthenogeneticusNegativeHatakeyama & Yasuno, 1981
 CadmiumCdHydropsyche slossonaeNegativeTessier et al., 2000
CdCl2Baetis tricaudatusNegativeIrving et al., 2003; Riddell et al., 2005
Glyptotendipes pallensNegativeHeinis et al., 1990
Kogotus nonusNegativeRiddell et al., 2005
 SeleniumNa2SeO4Culex quinquefasciatusNegativeJensen, 2006
Sympetrum corruptumPositiveJensen, 2006
 ZincZnSO4·7H2OHydropsyche betteniNegativeBalch et al., 2000
B. Taxis behavior
 AluminumAlCl3ChironomidaePositiveBernard et al., 1990
EphemeropteraPositiveBernard et al., 1990; Bernard, 1985
PlecopteraNo effectBernard et al., 1990; Bernard, 1985
Simulium spp.NegativeBernard, 1985
Simulium spp.No effectBernard et al., 1990
TrichopteraPositiveBernard et al., 1990
Al2(SO4)3Baetis rhodaniPositiveOrmerod et al., 1987
Dicranota spp.PositiveOrmerod et al., 1987
Dixa puberulaPositiveOrmerod et al., 1987
Elmis aeneaNo effectOrmerod et al., 1987
Ephemera danicaPositiveHerrmann & Andersson, 1986
Ephemerella ignitaPositiveOrmerod et al., 1987
Heptagenia fuscogriseaPositiveHerrmann & Andersson, 1986
Heptagenia sulphureaPositiveHerrmann & Andersson, 1986
Leuctra spp.No effectOrmerod et al., 1987
Protonemura meyeriNegativeOrmerod et al., 1987
SimuliidaeNegativeOrmerod et al., 1987
 CadmiumCdChironomus salinariusNo effectHare & Shooner, 1995
Procladius spp.No effectHare & Shooner, 1995
Sergentia coracinaNo effectHare & Shooner, 1995
CdCl2Baetis tricaudatusNegativeRiddell et al., 2005
Glyptotendipes pallensNegativeHeinis et al., 1990
Hexagenia limbataNo effectGosselin & Hare, 2004
Hexagenia rigidaPositiveOdin et al., 1995
3CdSO4·8H2OBaetis rhodaniNegativeGerhardt, 1990
HydropsychidaePositiveVuori, 1994
Leptophlebia marginataNo effectGerhardt, 1990
 CopperCuAdenophlebia auriculataPositiveGerhardt & Palmer, 1998
Chimarra spp.NegativeClements et al., 1989
Cinygmula spp.No effectStitt et al., 2006
Hydropsyche morosaNegativeClements et al., 1989
Paragnetina mediaNo effectClements et al., 1989
CuCl2Hydropsyche angustipennisNegativevan der Geest et al., 1999
CuSO4Baetis spp.PositiveLeland, 1985
ChironomidaePositiveLeland, 1985
Chironomus ripariusNo effectDornfeld et al., 2009
Cleptelmis addendaNo effectLeland, 1985
Lepidostoma spp.PositiveLeland, 1985
Optioservus divergensPositiveLeland, 1985
Paraleptophlebia pallipesPositiveLeland, 1985
Simulium spp.No effectLeland, 1985
Symphitopsyche oslariNegativeLeland, 1985
 IronFeSO4·7H2OLeptophlebia marginataNegativeGerhardt, 1992, 1994
 LeadPbCl2Leptophlebia marginataNegativeGerhardt, 1994
 MercuryCH3HgClHexagenia rigidaPositiveOdin et al., 1995
C. Oviposition
 CadmiumCdCl2·2½H2OChironomus ripariusNegativeWilliams et al., 1987
 CopperCuSO4·5H2OChironomus ripariusNo effectDornfeld et al., 2009
 MercuryCH3ClHgCulex quinquefasciatusNo effectJensen et al., 2007
 SeleniumNa2SeO4Culex quinquefasciatusNo effectJensen et al., 2007


Terrestrial systems

Ingestion behavior  The majority of research concerning feeding behaviors of terrestrial insects has investigated the effects of metals on feeding preference when individuals were exposed to various concentrations and combinations of metals. For herbivorous insects, this has focused mostly on antifeedant properties of metals on agricultural pests. Zinc, Cu, Ni, Se, and As have been evaluated individually, whereas several other studies examined combinations of metals.

Zinc sulfate has been extensively studied, particularly in experiments involving first-instar lepidopteran pests. Zinc salts are known to cause toxicity in insects and their abilities as feeding deterrents for agricultural pests were quantified. Ostrinia nubilalis (Hübner) (Lepidoptera: Crambidae) responded to Zn sulfate treatments in choice experiments by avoiding meridic diet containing concentrations ≥0.1% ZnSO4 (Gahukar, 1975). Although variability in individual responses was high and not always significant, aversion increased with increasing ZnSO4 in the diet. These results are consistent with the results obtained by Sell & Bodznick (1971) for Heliothis virescens Fabricius (Lepidoptera: Noctuidae), which was deterred by concentrations ≥0.2% ZnSO4, and Spodoptera littoralis Boisduval (Lepidoptera: Noctuidae), which was deterred by concentrations ≥0.1 m ZnSO4 (Salama & El-Sharaby, 1972). Similarly, Pollard & Baker (1997) demonstrated significant preference of Schistocerca gregaria (Forskål) (Orthoptera: Acrididae) and Pieris brassicae (L.) (Lepidoptera: Pieridae) for low Zn treatments over high Zn treatments in choice experiments. Behmer et al. (2005) came to the same conclusion in choice experiments, and further showed that S. gregaria learned associatively to avoid Zn-treated foods. Behavioral responses of insect predators and parasitoids to zinc sulfate-contaminated prey have not been reported.

Gustatory perception assays with adult Drosophila melanogaster Meigen (Diptera: Drosophilidae) showed that they also prefer low Zn treatments to high Zn treatments (Bahadorani & Hilliker, 2009). These authors observed the same pattern for Fe (II) and Fe (III), while adults preferred control diets to Cd- and Cu-contaminated diets significantly more often. Like adults, larvae also avoided feeding on high concentrations of heavy metals (Bahadorani & Hilliker, 2009).

Copper, as copper sulfate (CuSO4), has also been investigated as a feeding deterrent against agricultural pests. Mixing CuSO4 with lime creates a Bordeaux mixture effective as a fungicide ( El-Bassiouny (1991) investigated possible feeding deterrent properties for several lepidopteran species. These responded to CuSO4 feeding deterrents with mixed responses, depending on species. Oligophagous species [P. brassicae and Pieris napi (L.)] were deterred at lower concentrations (0.05–0.1 m CuSO4), whereas polyphagous species (Mamestra brassicae L. and Mamestra oleracea L.) (Lepidoptera: Noctuidae) were only inhibited by higher concentrations (0.2 m CuSO4) (El-Bassiouny, 1991). Pieris brassicae took shorter meals before feeding ceased, and experienced an increase in palpation frequency.

Some research has focused on hyperaccumulating plants and documented that high levels of metals in plant tissues may serve to deter herbivory. For example, herbivorous insects preferred Streptanthus polygaloides Gray (Brassicaceae) grown in low nickel (Ni) soils (15.6–76.5 μg g−1) vs. high Ni soils (1 820–7 960 μg g−1) (Jhee et al., 2005). These included the folivores Melanoplus femurrubrum (De Geer) (Orthoptera: Acrididae), Evergestis rimosalis Guenée (Lepidoptera: Pyralidae), and the rhizovore Delia radicum L. (Diptera: Anthomyiidae). Pieris rapae also preferred unamended to treated plants (180 and 7 400 mg Ni kg−1 soil, respectively) (Martens & Boyd, 1994). The feeding behaviors of aphids and other vascular feeding insects were not altered by Ni accumulation in plants (Boyd & Martens, 1999; Jhee et al., 2005).

The lepidopteran, Spodoptera exigua Hübner, exposed to different forms of Se were deterred from feeding by inorganic Se compounds (sodium selenate and sodium selenite) at LC30 values and greater for first and third instars (Vickerman & Trumble, 1999). By contrast, this same study revealed that organic Se compounds did not serve as feeding deterrents for third-instar S. exigua, though first instars preferred controls to these compounds 50–75% of the time. Sodium selenate accumulated by Brassica juncea (L.) Czern. (Brassicaceae) also effectively prevented Acheta domestica (L.) (Orthoptera: Gryllidae) feeding in choice experiments (Freeman et al., 2007) with 5× as many crickets preferring controls to treated leaves (546 ± 38 μg Se g−1 dry weight). At 10 mg kg−1 dry leaf weight, B. juncea with incorporated sodium selenate also successfully deterred Myzus persicae (Sulzer) (Hemiptera: Aphididae) feeding and prevented colonization (Hanson et al., 2004). When fed alfalfa with incorporated Se, first-instar S. exigua were unable to distinguish between low and high (2.88 ± 0.52 vs. 305.81 ± 52.14 μg g−1 plant dry weight) concentrations of Se compared with controls (Vickerman et al., 2002b). Fourth instars did not differentiate between low Se and controls, but avoided high Se plants. Alternatively, various polyphagous acridid grasshoppers chose low-Se Stanleya pinnata (Pursh) Britton (Brassicaceae) significantly more often than high-concentration alternatives (1 vs. 230 μg g−1 dry weight) in choice experiments (Freeman et al., 2007). A recently discovered biotype of a lepidopteran, Plutella xylostella Stanleyi (Lepidoptera: Plutellidae), was shown to withstand accumulated concentrations of 2 000 μg Se g−1 dry weight on S. pinnata and larvae showed no preference for low or high (47 vs. 792 μg g−1 dry weight) Se-treated plants in choice experiments, as opposed to P. xylostella G88 and P. rapae which avoided higher concentrations (Freeman et al., 2006).

Although there are many studies reporting the effects of As on insects, relatively few report behavioral impacts. In terms of ingestion behaviors, only a single paper was found. Rathinasabapathi et al. (2007) reported avoidance by Schistocerca americana (Drury) of lettuce contaminated with As when given a choice with low-As treated plants (46.14 ± 22 vs. 2.3 ± 0.2 mg kg−1). They showed adult S. americana took taste bites before rejecting highly contaminated lettuce, indicating As is detected through gustation.

In polluted areas, metals often exist as simple or complex mixtures. Migula & Binkowska (1993) investigated the ability of populations of Chorthippus spp. (Orthoptera: Acrididae) from heavily and weakly polluted sites to distinguish between Cd, lead (Pb), and Cd + Pb exposed diets. They found that grasshoppers locally adapted in weakly polluted sites did not have the ability to distinguish between leaves with different metal concentrations, whereas those from heavily polluted sites reduced their consumption rate with increasing Cd and Pb concentrations. This may indicate learned avoidance behavior in Chorthippus populations living in taxing environments. In a different experiment examining the effects of Cd alone, Frankliniella occidentalis (Pergande) (Thysanoptera: Thripidae) also experienced a significant decrease in feeding, as measured by the ‘leaf feeding damage index’ for treatment concentrations ranging from 0 to 300 mg kg−1 in Thlaspi caerulescens J. & C. Presl (Brassicaceae) varieties (Jiang et al., 2005).

A subset of research investigating metal impacts on ingestion behaviors examined plant biocontrol agents, with varying results. First-instar Bactra verutana Zeller (Lepidoptera: Tortricidae) exposed to purple nutsedge for up to 4 weeks were unaffected by Cd concentrations up to 18 μg g−1 (Quimby et al., 1979). Agasicles hygrophila Selman & Vogt (Coleoptera: Chrysomelidae) exposed to 8.7 μg Cd g−1 alligatorweed showed an inability to distinguish between Cd contaminated and uncontaminated plants, though they did experience feeding depression when fed on Cd-contaminated leaves in choice experiments (Quimby et al., 1979).

Neochetina bruchi Hustache (Coleoptera: Curculionidae) is used in the control of water hyacinth, an emergent, metal-accumulating aquatic plant, and spends its life on the leaf surface. When separately exposed to Cd and Zn, Jamil et al. (1989a,b) found a significant decrease in the number of water hyacinth feeding lesions, reflecting a decrease in feeding activity with increasing exposure concentration for both metals. There was no significant difference between numbers of feeding lesions found in plants accumulating up to 89.5 and 165 μg Zn/100 g dry weight; however, lesions were significantly fewer when N. bruchi were fed on plants accumulating 232 μg Zn/100 g dry weight (Jamil et al., 1989a,b). Cd exposure accumulating to levels of 3.78, 6.20, and 66.70 μg/100 g dry weight showed the same pattern of feeding depression (Jamil et al., 1989a), with no effect of the lower concentrations on number of feeding lesions. This finding supports results by Quimby et al. (1979).

A different species of water hyacinth beetle, Neochetina eichhorniae Warner, had conflicting behavioral outcomes in the presence of Cd when compared with N. bruchi (Kay & Haller, 1986). Water hyacinth with 8.00 and 17.20 μg Cd g−1 leaves did not experience decreased feeding activity of N. eichhorniae when compared with controls. Neochetina eichhorniae feeding activity when exposed to 21.62 and 44.77 μg g−1 Cu and 5.89 and 9.84 μg g−1 Pb was also not significantly different from controls. Kay & Haller (1986) exposed beetles to contaminated water hyacinth for 10 days, vs. Jamil et al. (1989a,b) who exposed beetles for 7 days, and did not report feeding depression at any point during their assays. This implies that N. eichhorniae is more tolerant of metals uptake by water hyacinth than N. bruchi.

Other studies on the antifeedant effects of metals on beetles showed a consistent decrease in feeding activity as a result of dietary exposure. Third instars of Leptinotarsa decemlineata (Say) (Coleoptera: Chrysomelidae) exposed to CuSO4 (El-Bassiouny, 1991) and Pb(NO3)2 (Kwartirnikov et al., 1999) significantly decreased feeding activity compared with controls, which were more pronounced for increasing concentrations of each antifeedant. Adult L. decemlineata exposed to Pb(NO3)2 also showed decreased feeding activity, though this result was less pronounced than for larvae (Kwartirnikov et al., 1999).

Complex preference experiments by Rokytová et al. (2004) revealed that adult Lochmaea capreae L. (Coleoptera: Chrysomelidae) did not alter feeding activity on birch leaves dipped in Cd (2–250 μg ml−1) and Pb (4–500 μg ml−1). For manganese (Mn) and Zn, feeding activity significantly decreased between low and high concentrations (100–500 and 10 000 μg Mn ml−1, and 80–400 and 8 000 μg Zn ml−1, respectively). They also avoided high concentrations of Mn and Zn more often than all concentrations of Cd and Pb. Another chrysomelid, Melasoma lapponica L. (Coleoptera: Chrysomelidae) showed a preference for very high (273.3 mg Ni kg−1 and 95.4 mg Cu kg−1) and very low (27.7 mg Ni kg−1 and 16.9 mg Cu kg−1) concentrations of Cu and Ni in willow foliage along a distance gradient from a smelter (Zvereva & Kozlov, 1996). For this species, some feeding was necessary on undamaged leaves before rejection, though damaged leaves with metal exposures were identified and rejected before feeding, possibly due to an increased release of deterrent substances (Zvereva & Kozlov, 1996).

Finally, a handful of studies have examined the effects of metals on soil-dwelling invertebrate feeding behavior. Orchesella cincta (L.) (Collembola: Entomobryidae) showed no significant preference for green algae diet contaminated with Pb(NO3)2 up to 1 600 μg g−1, and the authors concluded that avoidance was not necessary due to efficient excretion mechanisms already in place (van Capelleveen et al., 1986). Fountain & Hopkin (2001) reported significant avoidance of Pb-contaminated diet at 2 170 μg g−1 for Folsomia candida Willem (Collembola: Isotomidae), but no significant avoidance of diet contaminated with 406 μg g−1 Pb, consistent with van Capelleveen et al. (1986). Orchesella cincta was similarly unaffected by Mn in the diet up to 9.2 ± 0.3 μmol g−1 dry mass, though iron (Fe) caused a significant decrease in feeding activity, especially at higher concentrations (Nottrot et al., 1987). In a field study, Sphaeridia pumilis Krausbauer (Collembola: Sminthuridae), Parisotoma notabilis (Schäffer) (Collembola: Isotomidae), and Mesaphorura macrochaeta (Rusek) (Collembola: Onychiuridae) gut contents reflected preferential avoidance of the organic horizon where the majority of Cd, Pb, and Zn were concentrated (Gillet & Ponge, 2003).

Copper-contaminated diet significantly deterred F. candida at 1 500 μg g−1 dry weight in the laboratory (Filser & Hölscher, 1997), and Onychiurus armatus (Tullberg) (Collembola: Onychiuridae) with a 13.5% Cu solution-soaked diet (Filser et al., 2000). By contrast, Isotomurus palustris (Müller) (Collembola: Isotomidae) fed on diet with Cu contamination as often as on uncontaminated diet, and F. quadrioculata and F. manolachei preferred Cu-contaminated diet (Filser et al., 2000). Folsomia candida significantly avoided Cu-contaminated yeast at concentrations exceeding 10 μg g−1 (Fountain & Hopkin, 2001). Folsomia candida also avoided yeast contaminated with Cd at concentrations exceeding 28 μg g−1, and always preferred controls to Zn-contaminated diet (Fountain & Hopkin, 2001).

Taxis behavior  Taxis is an oriented movement in response to a directional stimulus or a stimulus gradient. All of the available research investigating locomotory effects of pollution has focused on carabids and collembolans, with the exception of one study investigating pupation-site preference in D. melanogaster. Unfortunately, the studies appear to be contradictory, making deduction of patterns impossible. Bayley et al. (1995) found that Pterostichus cupreus L. (Coleoptera: Carabidae) larvae exposed to 500 μg Cu g−1 in soil and diet experienced severely impaired locomotion as adults. This resulted in decreased prey capture success as adults, despite the absence of antifeedant properties in the larval diet (Bayley et al., 1995). Though collembolans were shown to avoid Cu-contaminated diets, they tended to not avoid Cu-contaminated soils, the exceptions being M. macrochaeta and Folsomia manolachei Bagnall (Filser & Hölscher, 1997) and Pseudosinella alba (Packard) (Collembola: Entomobryidae) (Filser et al., 2000).

By contrast, Lock et al. (2001) sampled seven sites within the vicinity of an abandoned Pb-Zn mine and found no significant relationship between relative activity of carabids and measureable metal concentrations in soils. Activity was measured using pitfall traps in combination with diversity sampling to determine whether certain species were more active given particular soil metal concentrations. The lack of significance despite a gradient in total Pb, Cd, and Cu concentrations was attributed to these metals not being bioavailable to the predatory carabids (Lock et al., 2001).

Parisotoma notabilis exposed to a gradient of Cd, Pb, and Zn pollution in the field followed the distribution of its weakly polluted food source by shifting position in the soil from surface to deeper horizons, and thus avoided changing its feeding habits (Gillet & Ponge, 2003). Using methodology identical to that of Lock et al. (2001, 2003) determined that a location with similar contaminants as that of Gillet & Ponge (2003) showed no significant relationship between activity of collembolans and metal concentrations in soils. This difference in collembolan activity may have been caused by the high concentrations of Cd, Pb, and Zn reported in Gillet & Ponge (2003), compared with those in Lock et al. (2001). Other studies examining mixtures of metals in contaminated soils revealed that F. candida consistently avoided heavily contaminated soils, though there was high variability in individual response (Natal da Luz et al., 2004). Reporting an overall response of a conglomerate soil insect fauna, Gongalsky et al. (2009) found consistent avoidance of heavily contaminated soils. Collectively, the broadly contradictory results from the available research suggest there are many environmental factors that may influence movement. Thus, substantial opportunities exist for additional research on this topic.

In choice trials, Bahadorani & Hilliker (2009) found no significant difference in pupation-site preference in late-instar D. melanogaster when presented with normal food and food with a concentration of 70 mmol Zn l−1. Larvae did, however, significantly prefer normal food to Cu-contaminated food (20 mmol l−1). They also significantly preferred pupating in Fe (II)-contaminated food (70 mmol l−1) to non-contaminated food.

Oviposition  To date, only five publications reported the effects of metals on ovipositional response in terrestrial insects. When given a choice between concentrations of ovipositional substrate exposed to hexavalent chromium (Cr VI) (50, 500, 1 000 μg g−1), Trumble & Jensen (2004) found that females of Megaselia scalaris Loew (Diptera: Phoridae) did not discriminate between control, low, and high concentrations. This occurred despite the observation that the highest level was toxic to the larvae. In a different study on a dipteran, Bahadorani & Hilliker (2009) reported that mated female D. melanogaster significantly decreased egg laying at relatively high concentrations of heavy metals (2 mmol l−1 Cd, 10 mmol l−1 Cu, 40 mmol l−1 Fe (II), 20 mmol l−1 Fe (III), and 30 mmol l−1 Zn). Interestingly, for both Fe and Zn, oviposition increased significantly relative to controls (0 mmol l−1 for each) at lower concentrations. This indicates that the female not only senses metals in the environment, but also knows which concentrations will maximize the fitness of her offspring.

Two studies were available that examined oviposition by S. exigua in response to Se. Females preferred to oviposit on low concentrations of Se-treated alfalfa (2.88 ± 0.52 μg Se g−1 dry weight) over controls (Vickerman et al., 2002b). However, adult females were unable to distinguish between low and high (305.81 ± 52.14 μg Se g−1 dry weight) concentrations of Se for oviposition, despite the fact that the high level was toxic. In a second study examining oviposition on Atriplex spp. plants which accumulate Se, S. exigua did not distinguish between plants that contained concentrations of Se that were toxic to their offspring (Vickerman et al., 2002a). For both M. scalaris and S. exigua, the inability to distinguish between lethal concentrations of Cr VI and Se, respectively, puts eggs and larvae at risk at exposed oviposition sites.

Finally, the newly discovered P. xylostella Stanleyi biotype does not differentiate between high and low concentrations of Se in S. pinnata when ovipositing. This is in direct contrast to a different ecotype, P. xylostella G88, and P. rapae, which avoided ovipositing on highly Se-contaminated plants (Freeman et al., 2006).

Aquatic systems

Ingestion behavior  As was the case with CuSO4 and oligophagous terrestrial herbivores, Hatakeyama & Yasuno (1981) found that aquatic systems with Cu available in concentrations from 0.01 to 0.64 mg Cu l−1 resulted in reduced food uptake in first-instar Paratanytarsus parthenogeneticus Freeman (Diptera: Chironomidae), as measured by the area of deposited feces. Lowered egestion rates were similarly observed for Chironomus riparius Meigen (Diptera: Chironomidae) exposed to Zn, Cd, and Fe contaminated sediments (Leppänen et al., 1998). However, this was only the case at one of their treatment locations (2 356.4 μg Cd g−1 sediment, 38 mg Zn g−1, and 17 mg Fe g−1). One of the reference locations with minimal contamination also contained chironomids with decreased egestion (7.2 μg Cd g−1 sediment, 0.9 mg Zn g−1, and 42.4 mg Fe g−1) when compared with the other reference location (9.0 μg Cd g−1 sediment, 1.1 mg Zn g−1, and 17.4 mg Fe g−1). Because of this discrepancy, the authors were unable to conclude whether metal-contaminated sediments had a significant impact on feeding rates of C. riparius (Leppänen et al., 1998). A separate field study examining post-exposure feeding depression of C. riparius revealed no significant difference between reference, low contamination, and highly contaminated field locations (Faria et al., 2006). Contaminated sites contained a mixture of metals, including As, Cd, Cr, Cu, Fe, Pb, Mn, Ni, and Zn. Post-exposure feeding depression was measured as the amount of algae consumed in an uncontaminated environment after exposure for 6 days, to evaluate the effect of acute pollution pulses on long term population viability of C. riparius (Faria et al., 2006).

Irving et al. (2003) reported significantly decreased feeding rates for Baetis tricaudatus Dodds (Ephemeroptera: Baetidae) exposed to dietary Cd, though they did not preferentially avoid Cd-contaminated diatom mats. Feeding inhibition was apparent after 8 days for both Cd treatments (10 and 84 μg Cd g−1 diatoms). Cd concentrations of 0.5–1.0 mg l−1 were further shown to disrupt filter feeding behaviors in fourth-instar Glyptotendipes pallens (Meigen) (Diptera: Chironomidae) and result in increased defecation rates, possibly in an attempt to regulate Cd uptake (Heinis et al., 1990).

Only two studies evaluated the impacts of heavy metal contamination on predator hunting behaviors. In a complex factorial design, B. tricaudatus, Kogotus nonus (Needham & Claassen) (Plecoptera: Perlodidae), and two fish species were placed in a mesocosm and the effects of dietary and waterborne Cd contamination evaluated (Riddell et al., 2005). Because K. nonus are predators, only waterborne Cd contamination (0.5 and 5 μg Cd l−1) was relevant to their behaviors in this experiment. At both concentrations, locomotory activities were significantly reduced, resulting in impaired foraging abilities. Only two of nine attacks on prey were successful, and both occurred at the 0.5 μg l−1 treatment, so further extrapolation about effects on predation behaviors could not be made (Riddell et al., 2005). Jensen (2006) evaluated the effects of Se and/or methyl-mercury (MeHg) on consumption rates of Sympetrum corruptum (Hagen) (Odonata: Libellulidae) when fed Culex quinquefasciatus Say (Diptera: Culicidae). He found that S. corruptum in Se treatment solutions consumed significantly more mosquito larvae per day than controls; however, predators eating prey contaminated with Se + MeHg consumed significantly fewer prey per day. Predators’ consuming more in Se-treated water with non-treated prey was attributed to the mosquito larvae experiencing a reduction in avoidance behavior. Predators’ consuming less in the Se + MeHg treatments was attributed to treatments making the prey unpalatable or suppressing the predator’s appetite (Jensen, 2006).

Several studies have evaluated the impacts of metal exposure on the construction of capture nets by Hydropsyche spp. (Trichoptera: Hydropsychidae). These nets are used to capture drifting plant and animal materials; therefore, construction anomalies have the potential to negatively impact the efficiency with which nymphs are able to recover food items. Fifth-instar H. betteni adapted to either polluted or unpolluted environments were collected from the field and exposed to waterborne (5.4 or 10.7 mg l−1) and dietary (113 μg g−1) Zn (Balch et al., 2000). After 5 and 7 weeks, nymphs from unpolluted and polluted populations, respectively, exhibited significantly looser net structures with such large openings between net strands that capture efficiency was negligible. Interestingly, larvae exposed to waterborne Zn at 42.1 and 21.7 mg l−1 showed no significant difference in net spinning capabilities (Balch et al., 2000). Third and fourth-instar H. slossonae exposed to Cd also exhibited an increase in net anomalies (Tessier et al., 2000). After 5 days, approximately 60% of nymphs exposed to 43.3 and 21.4 μg Cd l−1 experienced strand crossover anomalies, with 100% of nymphs showing anomalies after 10 and 20 days, respectively. After 10 days, approximately 60% of nymphs exposed to 11.6 μg Cd l−1 experienced strand crossover anomalies, with 100% of nymphs showing anomalies after 15 days. The lowest concentrations tested (1.2 and 0.37 μg Cd l−1) exhibited approximately 50% of nymphs with net anomalies apparent after 15 and 20 days, respectively. Background anomalies in control treatments were found to be approximately 20% (Tessier et al., 2000).

By contrast, Petersen & Petersen (1983) pooled data for Hydropsyche spp. net anomalies after determining that the number of anomalies was independent of species. Specifically, they evaluated the possibility of increased strand crossover frequencies which can result in smaller mesh openings with less uniform strand arrangements. They found that net strand crossover frequency at heavy metal-contaminated field locations was not significantly different from control sites. Metals present at different locations varied and though exact concentrations were not reported, exposure levels were always sublethal. However, the lack of information on metal concentrations in the Petersen & Petersen (1983) report makes direct comparisons with the studies by Balch et al. (2000) and Tessier et al. (2000) impossible. This highlights the critical need for reporting detailed information on metal concentrations which are necessary for documenting potential patterns in insect responses.

Taxis behavior  Drift was the most commonly measured response of invertebrates in aquatic systems to pollution, and is defined as an organism detaching from the substrate and swimming or floating downstream. This behavior is used to escape localized pollution, with the tradeoff being greater exposure to predation. It is easily measured in the field just downstream of point sources of pollution and allows for a community-wide assessment of the impacts of contaminants on downstream movements of insects.

For example, aluminum (Al) (0.95 mg l−1) caused chironomids to enter the drift column 4–8 times more frequently after exposure for 6 h than controls (Bernard et al., 1990). In this same study, Ephemeroptera and Trichoptera spp. entered the drift column 10–15× and 3–5× more frequently, respectively, while Simulium spp. (Diptera: Simuliidae) and Plecoptera spp. experienced no significant increase in drift throughout the 12-h study duration. When exposed to Al with varying pH values, Plecoptera nymphs failed to respond by increasing drift when dissolved Al concentration was increased from 31.0 to 40.2 μg l−1, and the pH decreased from 7.0 to 5.9 (Bernard, 1985). This may be due to a lack of sensory capacity to detect Al increases within the range tested, or because the chemical was not physiologically stressful at the tested concentration. Trichoptera and Simulium spp. both experienced a delayed response at the above concentrations, possibly due to a disruption of physiological processes. Ephemeroptera, the most sensitive order evaluated, showed an immediate response to the Al influx, possibly because they are able to detect this ion in their environment through chemoreceptors (Bernard, 1985). Chironomids became sensitive to Al only after pH had decreased in this experiment. In a different study on pH impacts on Al toxic response, Ormerod et al. (1987) examined a wider range of taxa. Both Leuctra spp. (Plecoptera: Leuctridae) and Elmis aenea (Müller) (Coleoptera: Elmidae) did not change drift patterns in response to an acid and Al pulse. Ephemerella ignita (Poda) (Ephemeroptera: Ephemerellidae) showed significantly increased drift during the episode, and Baetis rhodani Pictet (Ephemeroptera: Baetidae) drift density increased 8.4× relative to the control. Simuliidae drift increased 3.6×, and Protonemura meyeri Pictet (Plecoptera: Nemouridae) increased 1.6× during treatment and remained high the following day. Both Dixa puberula Loew (Diptera: Dixidae) and Dicranota spp. (Diptera: Pediciidae) showed significantly increased drift, but only during the episode (Ormerod et al., 1987).

However, not all metals produce a significant drift response. There was no significant difference in drift rate of Cinygmula spp. (Ephemeroptera: Heptageniidae) at concentrations of 78 and 229 μg l−1 Cu (Stitt et al., 2006), and rates of drift at 3 μg l−1 Cu were not significantly different from the control. After dosing experimental stream channels with Cu at three concentrations (2.5, 7, and 15 μg l−1), Leland (1985) found that Cleptelmis addenda (Fall) (Coleoptera: Elmidae) showed no significant increase in drift compared with controls, whereas drift rate of Optioservus divergens (Leconte) (Coleoptera: Elmidae) increased slightly at 2.5 and 15 μg Cu l−1. Paraleptophlebia pallipes (Eaton) (Ephemeroptera: Leptophlebiidae), Baetis spp., and Lepidostoma spp. (Trichoptera: Lepidostomatidae) increased drift at 7 and 15 μg Cu l−1, whereas Symphitopsyche oslari Banks (Trichoptera: Hydropsychidae) decreases drift at these concentrations. Drift of chironomids decreased slightly at 7 μg Cu l−1, and increased at 15 μg Cu l−1, whereas Simulium spp. drift was unaffected by Cu (Leland, 1985).

One study documented a decreasing response as pollution levels increased. Riddell et al. (2005) demonstrated that B. tricaudatus drift decreased with increasing concentrations of Cd. However, they used a recirculating stream system, so drift behavior did not allow insects to escape to a lower concentration. They suggested that continued exposure could have reduced locomotory behavior, or that an increased energy demand due to contaminant acclimation or detoxification could have reduced the energy available for relocation.

When exposed to multiple metals (2.2 μg Cd l−1, 24 μg Cu l−1, and 200 μg Zn l−1), various aquatic insects (Coleoptera: Elmidae; Diptera: Chironomidae; Ephemeroptera, Plecoptera, and Trichoptera: Hydropsychidae) from populations locally evolved in uncontaminated streams experienced significant increases in drift (Clements, 1999). Because all of the organisms collected in drift nets were still alive, the authors were able to conclude that drift was a behavioral response to avoid heavy metals in the water. In choice experiments with contaminated sediments, first-instar C. riparius exposed to 2 mg Cu l−1 (Dornfeld et al., 2009) and fourth-instar C. salinarius, Sergentia coracina (Zetterstedt), (Diptera: Chironomidae) and Procladius spp. (Diptera: Chironomidae) exposed to 0.15 μg Cd l−1 (Hare & Shooner, 1995) were unable to distinguish between treated and control sediments. In choice experiments using sediments collected from five treatments and two reference locations, C. tentans preferred control over treatment sediments for the most highly contaminated locations only (774 mg l−1 Cd, 11 134 mg l−1 Zn, 1 393 mg l−1 Cr; 964 mg l−1 Cd, 16 397 mg l−1 Zn, 2 129 mg l−1 Cr; and 1 029 mg l−1 Cd, 17 262 mg l−1 Zn, 1 640 mg l−1 Cr) (Wentsel et al., 1977).

Evaluations of heavy metal impacts on other locomotory behaviors are also common. Cd concentrations ranging from 2.5 to 10 mg l−1 were shown to significantly increase time spent in inactive states for G. pallens (Heinis et al., 1990). Exposure to levels as low as 0.02 mg Cd l−1 were shown to reduce locomotion in B. rhodani, whereas Leptophlebia marginata (L.) (Ephemeroptera: Leptophlebiidae), exposed to 0.2 mg Cd l−1, showed no difference in locomotory activities compared with controls (Gerhardt, 1990). Based on these results, Gerhardt (1990) concluded that locomotion was a good parameter to measure in cases of suspected subacute chemical stress. He reached the same conclusion when Fe-exposed L. marginata decreased motility in proportion to the concentration of dissolved Fe (10, 20, and 50 mg l−1) (Gerhardt, 1992). At much lower levels of contamination, female Hexagenia limbata (Serville) (Ephemeroptera: Ephemeridae) exposed to 18.9 μg Cd g−1 in sediment and 5.8 μg Cd l−1 in water experienced no discernable effect on burrowing activities (Gosselin & Hare, 2004). Similarly, concentrations of 0.05–0.1 mg Cd l−1 had relatively minor impacts on activity of G. pallens (Heinis et al., 1990).

Exposure at concentrations of 0.05–0.296 mg Cu l−1 led to increased escape behavior by Adenophlebia auriculata Eaton (Ephemeroptera: Leptophlebiidae): mayflies searched for stones away from areas of Cu input (Gerhardt & Palmer, 1998). There were more ventilation and abdominal undulations observed at these concentrations as well, possibly in an attempt to rid Cu ions bound in gill membranes. Finally, A. auriculata was more prone to climbing on top of rocks instead of maintaining negative phototactic behaviors observed in controls. By contrast, L. marginata showed a decrease in escape behavior correlated to increasing exposure time and Fe (10–500 mg l−1) and Pb (0.1–5.0 mg l−1) concentrations (Gerhardt, 1994). Odin et al. (1995) reported increased bioturbation activity of Hexagenia rigida McDunnough (Ephemeroptera: Epheme-ridae), as measured by turbidity in the water column, when nymphs were exposed to sediment concentrations up to 10 mg Cd kg−1 and MeHg concentrations up to 2.98 mg kg−1. Interestingly, nymphs were unaffected when the exposure route was water only (Odin et al., 1995).

For Hydropsyche angustipennis (Curtis) (Trichoptera: Hydropsychidae), there was a significant decrease in ventilation, or abdominal undulatory movements, at 20 μg Cu l−1, resulting in a proportional increase in other locomotory behaviors and inactivity (van der Geest et al., 1999). Concentrations ranging from 100 to 600 μg Cu l−1 caused H. angustipennis to spend very little time ventilating, and individuals were, for the most part, inactive. When exposed to waterborne Al at 2.0 mg l−1, nymphs of Heptagenia fuscogrisea (Retzius), Heptagenia sulphurea (Müller) (Ephemeroptera: Heptageniidae), and Ephemera danica Müller (Ephemeroptera: Ephemeridae) showed significant increases in respiration (Herrmann & Andersson, 1986).

There are a few studies on combinations of metals and industrial effluents inhibiting locomotory behaviors of insects. Unfortunately, because of the variability, complexity, and unknown constituents of the industrial effluents, analysis of potential synergistic and antagonistic interactions with metals and other non-metals are not possible. However, a few of these studies are included here and will allow the reader to access this literature. Nymphs of H. angustipennis exposed to effluent downstream of an industrial area exhibited decreased ventilation, but other locomotory activities increased in frequency (Gerhardt, 1996). Exposed H. pellucidula also increased activity relative to controls (Macedo-Sousa et al., 2008). Choroterpes picteti (Eaton) (Ephemeroptera: Leptophlebiidae) experienced an initial increase in locomotion, but by the end of the assay, individuals in the control treatment were more active than those exposed to the acid mine drainage (AMD) treatment (Macedo-Sousa et al., 2008). In a similar study, C. picteti in AMD treatments increased locomotion in response to heavy metals and lower pH, which had the potential to increase nocturnal drift behavior (Gerhardt et al., 2005).

Inter- and intraspecific behaviors have not been frequently reported, particularly with relation to impacts of metals. When exposed to 0.5 and 5 μg l−1 dissolved Cd in water and diet, B. tricaudatus were more vulnerable to predators as a result of decreased predator avoidance behaviors (Riddell et al., 2005). When crop and stomach contents from Paragnetina media (Walker) (Plecoptera: Perlidae) were assessed in Cu-exposed vs. control individuals, Clements et al. (1989) found an increased amount of Hydropsyche morosa Hagen and Chimarra spp. (Trichoptera: Philopotamidae) remains, indicating increased susceptibility to predation in 5.5 μg Cu l−1 contaminated waters for these species. However, various species of caddisflies exposed to a wide range of Pb, Zn, and Cd concentrations simultaneously in the field showed no difference from controls in amount of time taken to emerge from their cases after a predatory threat (Lefcort et al., 2000). In a factorial experiment, Kiffney (1996) showed that metal-contaminated water (0.7 μg Cd l−1, 6 μg Cu l−1, and 50.3 μg Zn l−1) had no impact on B. tricaudatus and Rhithrogena hageni Eaton (Ephemeroptera: Heptageniidae) predator avoidance behaviors, while it increased predation risk to Hydropsyche spp. Prostoia besametsa (Ricker) (Plecoptera: Nemouridae) predation decreased in metals treatments compared with controls (Kiffney, 1996). Therefore, it would appear that the impact of metals on predator avoidance behavior is species dependent.

In an experiment investigating the impacts of Cd on competition behavior between conspecifics of hydropsychid larvae for optimal foraging habitat, Vuori (1994) found that exposed intruders performed shorter and less fierce attacks when paired with control residents. Exposed residents were surprisingly active during attacks against exposed intruders and fiercer than control residents, though the fights were still shorter. Attacks between control residents and intruders were longer than those involving exposed individuals. Vuori (1994) speculated this Cd-induced behavioral change might have been due to individuals weighing the personal risk involved in combat to the energy that had already been expended in spinning silk to construct a net. Though the animal may be poisoned from Cd exposure, energetically a prime territory is worth defending when a net has already been spun. By contrast, an invading caddisfly may abandon a fight to construct a net in less suitable territory, particularly if they are outmatched.

Oviposition  As was the case for terrestrial reproductive behaviors, only a few papers analyzed ovipositional responses of aquatic insects to metal-contaminated environments. Williams et al. (1987) found that when given a choice between water contaminated with different levels of Cd, C. riparius adult females distinguished between control and low concentrations (0, 0.3, and 30 mg l−1) vs. high concentrations (100 and 300 mg l−1). Significantly fewer eggs were laid in Petri dishes with high concentrations of Cd vs. dishes with low or control concentrations. Despite this, female aversion was only sensitive enough to avoid concentrations acutely toxic to eggs; although not toxic to eggs, these concentrations are acutely toxic to first instars (Williams et al., 1987). In another preference study, Dornfeld et al. (2009) found C. riparius females unable to distinguish between control media and treatment media with 1.3 mg Cu l−1 when ovipositing. Though egg hatchability was significantly reduced in Cu treatments, Cu did not affect larval survival (Dornfeld et al., 2009).

Similarly, C. quinquefasciatus did not discriminate between water contaminated with sodium selenate (30 mg l−1), MeHg chloride (7 mg l−1), or a mixture of sodium selenate and MeHg chloride (at the above concentrations) (Jensen et al., 2007). The authors concluded that females were either unable to detect these compounds at the tested concentrations, or did not prefer unpolluted to polluted water when ovipositing.


A summary of these studies and the observed behavioral outcome for a particular contaminant can be found in Tables 1 and 2. These tables allow for a generic analysis of the broader impacts of metal and metalloid pollution in both terrestrial and aquatic systems, but are not comprehensive and not meant to serve as a quantitative meta-analysis. This qualitative classification is meant to offer the reader a quick summary of the published literature. Further, the designation of positive and negative outcomes does not necessarily confer a fitness advantage or disadvantage, and in some cases a positive behavioral outcome may have negative fitness impacts. For any given study, the outcomes may be dependent on the instar and concentrations tested, as well as exposure routes.

Over 95% of studies (53 of 55) on terrestrial ingestion behaviors reported either no effect or negative impacts as a result of individual pollutant exposure for the various behaviors quantified (Table 1A). In these studies there was some degree of repellency or feeding inhibition. Only 3.6% reported positive effects, equating to a stimulation of feeding behavior as a result of the metal being present. The substantial majority of these studies investigated impacts on lepidopteran pests, Collembola, and Orthoptera; absent or underrepresented orders merit future research. For aquatic taxis behaviors, 40% of studies (18 of 45) reported a positive behavioral stimulation of insects to some form of pollution, with the rest reporting suppressed behaviors or no effects.

Although many insect species were capable of distinguishing contaminated from uncontaminated locations, a surprising number of species evidently cannot detect the presence of metal and metalloid contamination. For purposes of reproduction, an inability to avoid heavily polluted sites would lead to loss of eggs and reduced fitness. Although some species showed a tendency to increase locomotory behaviors to escape from locations with elevated metal pollution, other species remained and greatly decreased all movements unrelated to feeding. Still other species exhibited behaviors that would result in increased predation, including positive phototaxis that caused immatures to move to exposed positions. Ultimately, for some insects these behaviors result in reduced species fitness at contaminated sites, a general reduction in population sizes as well as species diversity, and a trend toward preponderance of those species that can tolerate pollution.

Due to the paucity of information regarding terrestrial taxis and reproductive behaviors, and aquatic ingestion and reproductive behaviors, further conclusions cannot be drawn about patterns of insect response to metals and metalloids. Analyses by feeding guild, environment (terrestrial vs. aquatic), and systematic classification did not provide evidence for a single dominant response. Additionally, the total number of species that have been investigated is relatively small. Patterns may not become evident until more research is published, particularly as many responses appear to be species-specific. A large number of papers also do not include a comparison with behaviors at uncontaminated sites, or document concentrations of the key pollutants. Although still valuable, these cannot be used as reliable evidence for behavioral changes that occur in response to metals and metalloids.

Because of the extent of the problem with metal and metalloid pollution worldwide, there is considerable opportunity for additional research. Knowledge of the effects of these pollutants at the bottom of the food web will be critical to understanding the true impact of metal contamination and to the potential reconstruction of damaged ecosystems.


We would like to thank C. Butler, R. Cardé, B. Carson, G. Kund, K. Hladun, and W. Walton whose comments and suggestions helped improve earlier versions of the manuscript.