The immigration of spiders (Araneida) into a new polder

Abstract. (1) The immigration of spiders into the Lauwerszeepolder (constructed in 1969) was studied during four years in four areas. The spider fauna was sampled weekly and 65 000 specimens of eighty-four species were caught using strip traps, window traps, simple pitfall traps and fences.
(2) About sixty species from all kinds of habitats were caught rarely; nineteen species were caught in numbers that suggest that they had established populations in the study areas.
(3) The four successful pioneer species, plus two marsh-dwelling species, were most abundant during 1969 and 1970. The 1971 catch was dominated by species from saline habitats; they became less abundant in 1972 when species from non-saline habitats increased. Only halotolerant species established populations.
(4) No evidence was found that dispersing spiders select certain areas. The differences between the spider fauna of the differennt study areas arise from degrees of success in colonization by immigrants. Pioneer species are least influenced by the abiotic environment.
(5) A greater aeronautic dispersal power is found in species inhabiting unstable habitats than in species from stable habitats.
(6) Aeronautic activity in adult linyphiids (s.l.) is not restricted to a special season, but is related to definite phases of the phenology.
(7) Male and female erigonids are equally active in aeronautic behaviour. In erigonids and linyphiids intense aeronautic activity coincides with great ground activity; on the ground males are more active than females.     

Do seasonal changes in numbers of aerially dispersing spiders reflect population density on the ground or variation in ballooning motivation?

Groups of linyphiid spiders (Erigone spp.) (Araneae, Linyphiidae), collected at intervals from arable land, were tested in laboratory bioassays to determine the proportion of individuals that exhibited ballooning behaviour on each field sampling occasion. There was no significant variation in the proportions of spiders in each test group ballooning in the laboratory over a year. Investigations of ground density and aerial dispersal, in a grass field and a winter-wheat field, confirmed that peaks in numbers of spiders observed ballooning in the field generally coincided with population peaks in the summer and autumn. Significant correlations between ground populations and aerial catches were found for total spiders, immature spiders, and Bathyphantes gracilis (Blackwall) in both fields, adult spiders in the grass field, Erigone spp. in the grass, and Meioneta rurestris (C.L. Koch) in the wheat. The other groups analysed, Lepthyphantes tenuis (Blackwall) in both fields, Erigone spp. in wheat, and M. rurestris in grass, showed similar but non-significant trends. No significant difference was found between overall ground-to-air ratios for males compared to females, but adult spiders were more likely to balloon than immatures.     

Edge or dispersal effects – Their relative importance on arthropod densities on small islands

Dispersal behaviour and edge effects are two potential factors determining population densities, and both effects are likely to vary with patch size. However, the relative importance of these two effects may be hard to separate because they may produce similar patterns. Here, we separate these two effects on population densities of seven groups of arthropods on small islands. To separate dispersal behaviour and edge effects, we use the fact that the slope of the density–area relationships (DAR-slope) should change with the absolute rates of dispersal, as would occur in response to island isolation, whereas the edge effect is expected not to be dependent on island isolation. For lycosid spiders, parasitic wasps and possibly herbivorous Homoptera DAR-slopes changed between isolated and non-isolated islands, suggesting dispersal behaviour to be relatively more important for explaining variation in their densities. Other arthropods (ants and Collembola), typically those with a predicted low dispersal among islands, showed similar DAR-slopes between isolated and non-isolated islands consistent with dominant edge effects. For two groups (web spiders and Diptera) the results were inconclusive. We conclude that both migratory processes and edge effects should be considered in the evaluation of patch size and isolation on density–area relationships.     

Characterization of restricted area searching behavior following consumption of prey and non-prey food in a cursorial spider, Hibana futilis

Cursorial spiders are important predators of crop pests in a variety of agricultural systems. Their survivorship, growth, and fecundity can be enhanced by the consumption of extra‐floral nectar. We recently showed that Hibana futilis (Banks) (Araneae: Anyphaenidae) engages in restricted area search following contact with nectar, is stimulated by nectar aroma, and can learn to recognize novel aroma cues. Studies have shown that H. futilis is also responsive to solvent extracts of the eggs and scales of the corn earworm, Helicoverpa zea Boddie (Lepidoptera: Noctuidae), one of its primary prey insects in cotton. The arrestment behavior of cursorial spiders following consumption of prey and non‐prey food has not been characterized. In the present study, the responses of spiders were measured following consumption of prey (H. zea eggs) or non‐prey (droplets of dilute honey) food items and compared with individuals tested without food items. The food items were presented to the spiders in test arenas constructed from the top of an inverted glass Petri dish cover. A combination of real time and recorded observations were made via a video camera attached to a computer. The behaviors and movement patterns of individual spiders were analyzed with behavioral tracking software. Significant differences in the behaviors and motion paths of spiders tested in the different treatments were observed. Hibana futilis displayed significantly more dispersal behavior on a blank test arena, than on test arenas supplied with honey droplets or moth eggs. Likewise, spiders tested on the blank arena crawled faster and their motion paths were significantly less tortuous than those of spiders tested in the arenas with honey or moth eggs. Following consumption of both the honey droplets and moth eggs, spiders showed elevated levels of restricted area search and lowered levels of dispersal behavior. The analysis showed that these spiders could crawl rapidly for extended distances. Behaviors such as restricted area search and learned recognition of food‐based stimuli would facilitate efficient location of the food resources needed to maintain their high activity levels.     

Subsocial spiders in space and time: A fine scale approach to the dynamics of dispersal

Less than 0.2% of all spider species live in close associations with conspecifics. Among these, subsocial spiders show characteristics of both solitary spiders (e.g., individuals disperse for breeding) and social spiders (e.g., prolonged cooperative behaviours at least prior to independent reproduction). Dispersing individuals build small webs, usually with one inhabitant, whereas colonies are large webs with plant debris and harbouring multiple females. We studied the spatiotemporal dynamics of dispersal in the subsocial spider Anelosimus baeza. We followed the occupancy of all colonies and dispersal webs over the breeding season by mapping the number and sex of spiders with respect to their location in three dimensions. We studied the settlement patterns of new webs and fluctuation in web occupancy through movement between occupied and abandoned webs of colonies and dispersal webs. The occupancy of webs was highly dynamic with changes occurring at small time scales. The similarity in the patterns of web occupancy by females among dispersal webs was partially explained by their spatial and their temporal proximity. Our results suggest that dispersal webs may be used by spiders as a temporary refuge by both sexes during the breeding season. Patterns described here suggest new approaches to dispersal studies in group living spiders.     

Behavioural and molecular evidence for selective immigration and group regulation in the social huntsman spider,Delena cancerides

Movement among social groups interacts with the costs and benefits of group‐living in complex ways. Unlike most other social spiders, the social huntsman spider, Delena cancerides, appears to enter foreign colonies, discriminates kin from non‐kin, and has very limited dispersal options because their bark retreats are rare, making this species an interesting model organism with which to examine the role of inter‐colony movement on group‐living. We examined movement among field colonies of D. cancerides in three ways: (1) by tracking the dispersal and immigration of marked spiders into foreign colonies; (2) by recording resident spiders' behaviour toward introduced immigrants; and (3) by inferring intra‐colony relatedness and immigration patterns through allozyme electrophoresis. Of the marked spiders, only young juveniles moved into neighbouring colonies, whereas subadults and adults did not. Introduced juveniles were tolerated in foreign colonies, whereas introduced adult males and subadults were usually attacked by the resident adult female, unless she had similar sized subadult/adult offspring of her own. Allozyme profiles from unmanipulated field colonies showed that 47% of sampled colonies contained at least one immigrant and that average within colony relatedness was below 0.5. These data align with previous research on the costs and benefits of group‐living for D. cancerides, suggesting that spiders actively seek and regulate group membership based on interests of both the immigrant and the colony. © 2012 The Linnean Society of London, Biological Journal of the Linnean Society, 2012, ?? , ??–??.     

Habitat selection in a large orb‐weaving spider: vegetational complexity determines site selection and distribution

1. The distribution of the large orb‐weaving spider Argiope trifasciata in old field habitats of North America and the habitat selection process this species used was studied for 2 years. 2. Because web spiders have limited dispersal abilities and an energetically costly prey capture device, they do not have the ability to sample potential foraging sites. Structural complexity of the vegetation to which the web must be attached is relatively easy to assess. The hypothesis that the structural complexity is a primary factor in determining initial web site selection was tested both by relating the natural distribution of the spiders across habitats to vegetational complexity and by manipulating the complexity of the habitats in a series of experiments. 3. Argiope trifasciata was not distributed evenly among three old field vegetation types. Habitat complexity was related to spider density in both years although no measure of insect activity, prey capture, or prey consumption was correlated with spider distribution. 4. Three experimental manipulations were conducted to test the impact of habitat structure on spider establishment: (1) the amount of natural vegetation was reduced, (2) structures were added to a simple habitat, and (3) the complexity of the structures added was varied. In each case, spiders were introduced and establishment of webs was monitored. In all manipulations, spider establishment was related to the complexity of the substrate available. 5. These results are important for understanding the cues that influence foraging site selection and therefore provide insight into the distribution of species with limited dispersal abilities and high site investment requirements.     

AN EXPERIMENTAL EVALUATION OF SELECTION ON COLOR MORPHS OF THE POLYMORPHIC SPIDER ENOPLOGNATHA OVATA (ARANEAE: THERIDIIDAE)

The annual theridiid spider Enoplognatha ovata exhibits a genetically based color polymorphism of red and nonred phenotypes. We evaluated fitness differences between red and nonred spiders by manipulating morph frequencies in a population in which red morphs were rare (≤5%). Broods from red females were introduced to open experimental plots from which natural aggregations of spiders had been repeatedly removed. Control plots in which spiders were removed but not replaced were used to estimate spider immigration from surrounding vegetation into experimental plots. Morph frequencies observed in experimental plots one year following the manipulation were adjusted by immigration estimates and tested against frequencies predicted with the hypothesis of no selection. We found no evidence of selection against red morphs: female morph frequencies in experimental plots did not differ significantly from expected frequencies assuming no selection; female frequencies did not change significantly between subadult and adult stages; and red and nonred spiders exhibited similar fecundities. We conclude that 1) selection on E. ovata color morphs is not likely to be detected easily within a single population because of the swamping effect of dispersal and 2) local patterns of morph-frequency variation may arise more from dispersal and drift than from selection on the color phenotypes.     

Wolf spider feeding strategies: optimality of prey consumption in Pardosa hortensis

Feeding behaviour of the wolf spider Pardosa hortensis Thorell (Araneae, Lycosidae) was studied in the laboratory. Characteristics of feeding were measured while prey availability was increased and the results were compared with the predictions of three models: the marginal value theorem (MVT), gut limitation theory (GLT) and the digestion rate limitation model (DRL). As a result of more frequent encounters with prey, the wolf spiders were able to modify their feeding behaviour so that their net energy intake rate increased substantially. Handling time decreased by 30%, and consumption rate increased by 40%. Partial consumption of prey did not occur until the spiders became nearly satiated. This indicated that spiders did not reach the optimum predicted by MVT. The most plausible mechanism for the increased efficiency was prey-stimulated digestive enzyme production as suggested in DRL. The predictions of GLT were not applicable for most of the feeding session, though gut satiation had an influence on the final stages of feeding. P. hortensis seemed to apply a responsive but cautious strategy: (i) spiders improved feeding efficiency on entering the higher quality habitat, but (ii) feeding times appeared to be sub-optimal and (iii) spiders were also willing to continue feeding when, as they approached satiation, the previously high efficiency could not be maintained. Such feeding behaviour optimizes long-term energy intake when food is scarce and unpredictable, which corresponds well with the known degree of natural food limitation of these animals.     

Increased propensity for aerial dispersal in disturbed habitats due to intraspecific variation and species turnover

Animal dispersal depends on multiple factors, such as habitat features and life‐history traits of the species. We studied the propensity for ballooning dispersal in spiders under standardized laboratory conditions. The 1269 tested individuals belonged to 124 species and originated from 16 sites with wide variation in habitat type. Spiders from disturbed habitats ballooned 5.5 times more than spiders from stable habitats. In Meioneta rurestris , for which we had enough data for a single‐species analysis, individuals were most dispersive if they originated from highly disturbed habitats. While the data for the other species were not sufficient for single‐species analyses, a hierarchical model that included the data simultaneously on all species suggested that dispersal propensity generally increases within species with the level of habitat disturbance. Dispersal probability showed a trend to increase with niche width, but the higher commonness of species with wide niches provides an alternative explanation for this pattern. As the prevalence of especially dispersive species was highest in disturbed habitats, variation in dispersal propensity was influenced by both inter‐ and intraspecific factors. We conclude that the positive correlation between niche width and dispersal propensity enables generalist species to utilize highly disturbed habitats, whereas the persistence of specialist species with restricted dispersal ability requires the conservation of stable habitats.     

The Effect of Feeding History on Prey Capture Behaviour in the Orbweb Spider Argiope keyserlingi Karsch (Araneae: Araneidae)

The prey capture behaviour of the orb-web spider Argiope keyserlingi Karsch was examined experimentally by subjecting spiders to two different feeding regimes (food deprived and food satiated) and three types of prey: Drosophila, blowflies (Lucilia cuprina) and bees (Apis mellifera). The attack behaviour of the spiders was influenced by both their foraging history and the type of prey. Food deprived spiders attacked Drosophila and bees more frequently than food satiated spiders, and food satiated spiders travelled more slowly to any of the prey types than food deprived spiders. Furthermore, Drosophila were never wrapped in silk but only grasped with the chelicerae, whereas both blowflies and bees were always wrapped. This provides experimental confirmation that feeding history affects the decision of orb-web spiders to accept or reject any given prey.     

Cannibalism and kin recognition in Delena cancerides (Araneae: Sparassidae), a social huntsman spider

Social behaviour in spiders is rare: of the 39 000 species of spiders known, only 23 are considered to be cooperatively social. Delena cancerides is a social species of the huntsman spider that is endemic to Australia. This species is virtually unique among social spiders, having evolved social behaviour in the absence of a snare web. It is thought that this form of social behaviour in D. cancerides has evolved via the sub-social route, that is, the extension of an ancestrally occurring period of maternal care and the delayed dispersal of juveniles. Most social spiders show no aggression towards non-kin conspecifics, prompting suggestions that spiders cannot recognize kin; however, D. cancerides individuals are highly aggressive towards conspecifics introduced from outside their own colony. In order to determine whether selective aggression in D. cancerides has its basis in kin recognition, tolerance behaviour was assessed in the context of kinship and size. We observed that, in general, juveniles preferred to starve than engage in cannibalism of any conspecifics, related or not. However, where cannibalism did occur, non-kin were preferentially eaten, indicating that this species is clearly capable of kin recognition. Size thresholds were also established, below which juveniles are tolerated by adults and above which aggressive interactions leading to death occur. We conclude that kin recognition and juvenile dispersal explain the uncharacteristically high levels of genetic polymorphism in this species.     

Wasp Attacks and Spider Defence in the Orb Weaving Species <Emphasis Type="Italic">Zygiella x-notata</Emphasis>

Predator–prey relationships are generally based on arm-race. Wasps and spiders are both predators, which could be potential prey for each other. The orb weaver spider Zygiella x-notata is sometimes a prey for the wasp Vespula germanica. We observed the wasp hunting behaviour under natural conditions, and we tested the influence of the spider’s behaviour on the wasp attack success. Wasps were active predators during the reproductive period of the spider. Results showed that wasps located more easily male spiders than females particularly when they were engaged in mate guarding. Female location depended on the presence of a web, but also of prey or prey remains in the web. On the other hand, their location depend neither on the characteristics and the position of the retreat in the environment nor on the size of the web. After location, males were more often captured than females whatever their behaviour (mate guarding or not). Presence of prey remains or prey in the web did not increase the risk for the spider to be captured. There was also no influence of the retreat’s characteristics or of its position in the habitat on the risk for the spider to be captured; but wasp successful attacks were less numerous when silk was present around the entrance of the retreat or when the spider was completely inside. As prey and prey remains favoured location of spiders by the wasps, we tested spider web cleaning behaviour as a response to wasp predatory pressure. By throwing small polystyrene pellets in the webs, we observed that more 80% of the spiders rejected the pellets in less than one minute. Our data indicated that wasps were significant predators of Z. x-notata and wasp attack could have been a selective pressure that had favoured spider defensive behaviours such as web cleaning.     

Natal Dispersal Patterns of a Subsocial Spider Anelosimus cf. jucundus (Theridiidae)

Species that alternate periods of solitary and social living may provide clues to the conditions that favor sociality. Social spiders probably originated from subsocial‐like ancestors, species in which siblings remain together for part of their life cycle but disperse prior to mating. Exploring the factors that lead to dispersal in subsocial species, but allow the development of large multigenerational colonies in social species, may provide insight into this transition. We studied the natal dispersal patterns of a subsocial spider, Anelosimus cf. jucundus, in Southeastern Arizona. In this population, spiders disperse from their natal nests in their penultimate and antepenultimate instars over a 3‐mo period. We tracked the natal dispersal of marked spiders at sites with clustered vs. isolated nests. We found that most spiders initially dispersed less than 5 m from their natal nests. Males and females, and spiders in patches with different densities of nests, dispersed similar distances. The fact that both sexes in a group dispersed, the lack of a sex difference in dispersal distance, and the relatively short distances dispersed are consistent with the hypothesis that natal dispersal results from resource competition within the natal nest, rather than inbreeding avoidance in competition for mates. Additionally, an increase in the average distance dispersed with time and with the number of spiders leaving a nest suggests that competition for nest sites in the vicinity of the natal nest may affect dispersal distances. The similar distances dispersed in patches with isolated vs. clustered nests, in contrast, suggest that competition among dispersers from different nests may not affect dispersal distances.     

Inbreeding depresses short and long distance dispersal in three congeneric spiders

Dispersal is one of the most important precopulatory inbreeding avoidance mechanisms and subject to landscape related selection pressures. In small populations, inbreeding within and between populations may strongly affect population dynamics if it reduces fitness and gene‐flow. While inbreeding avoidance is generally considered to be a key evolutionary driver of dispersal, potential effects of inbreeding on the dispersal process, are poorly known. Here, I document how inbreeding within a population, so by mating among relatives, affects the survivorship and the dispersal behaviour of three congeneric spider Erigone species (Araneae: Linyphiidae) that differ in habitat preference and regional rarity. The three species were chosen as a model because they allow the assessment of both long and short distance dispersal motivation (respectively ballooning and rappelling) under laboratory conditions. Inbreeding reduced both long and short distance dispersal modes in the three congeneric species. Because survival was depressed after inbreeding, with a tendency of reduced survival loss in the rare and highly stenotopic species, energetic constraints are likely to be the underlying mechanism. Inbreeding consequently depresses silk‐related dispersal in three related spiders. This may induce an inbreeding depression vortex with important consequences for range expansion and metapopulation dynamics of aerially dispersing species from highly fragmented landscapes.     

Aerial dispersal ability does not drive spider success in a crop landscape

1. Although spiders can colonise ecosystems by air, dispersal capabilities differ among spider species. Web‐building spiders are thought to balloon at higher rates than hunting spiders. Spider success in agricultural systems may also depend on habitat preferences. Few studies have examined the success of aerially dispersing spiders in crop systems, and information about the dispersal capabilities of spiders in putative source habitats is limited. 2. Spiders were monitored in the air and on the foliage of vineyards and adjacent oak woodland in order to compare the aerial spider faunas between these disparate habitats and to determine whether highly dispersive species contributed disproportionately to the spider community in vineyards. 3. The results show that most aerially dispersing spiders in both habitats were web‐building dwarf spiders, Erigone spp. (Linyphiidae), although hunting spiders were also well represented in the air, especially in oak woodland. Most woodland spiders in the air appeared to be residents of oak woodland and probably dispersed only short distances. 4. Conversely, only a subset of the aerial spider fauna established in vineyards in high numbers. Spiders that dominated the aerial fauna were under‐represented on vineyard foliage, whereas several hunting spiders dispersed aerially at low rates but dominated vineyard spider composition. 5. These results suggest that aerial dispersal ability may allow spiders to reach crop systems, but that establishment depends on habitat preferences and/or competitive ability.     

Water as an Essential Resource: Orb Web Spiders Cannot Balance Their Water Budget by Prey Alone

Water is essential for all living organisms because it acts as a major solvent and reaction medium. Terrestrial animals may lose water through evaporation and excretion and consequently have evolved strategies to balance their water budget by either minimising losses or by gaining water. The major pathway to gain water is via food intake, although many animals additionally drink free water. Spiders acquire substantial amounts of water by ingesting enzymatically liquefied prey. However, this may not account for the water needs of some species. We tested whether drinking is essential for orb web spiders of the genus Argiope by experimentally manipulating the diet (flies or crickets) and water supply (no water or a daily shower) to females and then measuring their subsequent drinking behaviour. Individuals of Argiope trifasciata, which are typically found in dry habitats, increased their body mass when fed crickets but not when fed flies. However, spiders deprived of water subsequently ingested significantly more water than spiders that received water every day, regardless of their feeding regime. This pattern was replicated in Argiope aetherea, which is found in the tropics and perhaps less likely to be water deprived in natural populations. Our results reveal that drinking allows these spiders to realise their water balance independent from the nutritional status. We suggest that the spiders may need to drink fresh water to process ingested nutrients.     

The influence of food supply on foraging behaviour in a desert spider

We tested the alternative hypotheses that foraging effort will increase (energy maximizer model) or decrease (due to increased costs or risks) when food supply increased, using a Namib desert burrowing spider, Seothyra henscheli (Eresidae), which feeds mainly on ants. The web of S. henscheli has a simple geometrical configuration, comprising a horizontal mat on the sand surface, with a variable number of lobes lined with sticky silk. The sticky silk is renewed daily after being covered by wind-blown sand. In a field experiment, we supplemented the spiders' natural prey with one ant on each day that spiders had active webs and determined the response to an increase in prey. We compared the foraging activity and web geometry of prey-supplemented spiders to non-supplemented controls. We compared the same parameters in fooddeprived and supplemented spiders in captivity. The results support the costs of foraging hypothesis. Supplemented spiders reduced their foraging activity and web dimensions. They moulted at least once and grew rapidly, more than doubling their mass in 6 weeks. By contrast, food-deprived spiders increased foraging effort by enlarging the diameter of the capture web. We suggest that digestive constraints prevented supplemented spiders from fully utilizing the available prey. By reducing foraging activities on the surface, spiders in a prey-rich habitat can reduce the risk of predation. However, early maturation resulting from a higher growth rate provides no advantage to S. henscheli owing to the fact that the timing of mating and dispersal are fixed by climatic factors (wind and temperature). Instead, large female body size will increase fitness by increasing the investiment in young during the period of extended maternal care.     

Survival and growth in groups of a subsocial spider (Stegodyphus lineatus)

Summary In spiders, known as potentially cannibalistic, mutual tolerance is one important requirement for group life. Using the subsocial spiderStegodyphus lineatus which possibly resembles the ancestors of the social species, the effects of competition were investigated in the laboratory. When dispersal was prevented, spiderlings were capable of living in groups. The intensity of competition for food among spiders in groups was varied experimentally by varying group size or the relative size differences of individuals. Body mass and mortality were compared in the different experiments. Prey availability, the size of the spiders and initial body size differences among group members all influenced the survival probability and growth of the spiders. Spiders of equal size tolerated each other with a higher probability than spiders of different sizes. Feeding in groups was always disadvantagous even for the largest spiders.     

From lynx spiders to cotton: Behaviourally mediated predator effects over four trophic levels

Food web studies often examine density and behaviourally mediated effects of predators on herbivores, but are less likely to assess the plant targeted by the herbivore. We conducted a study that incorporated four trophic levels examining the effect of two generalist predators (damsel bugs, Nabis kinbergii Reuter; and lynx spiders, Oxyopes molarius L. Koch) on damage to cotton bolls caused by green mirids (Creontiades dilutus (Stål)). First we tested whether lynx spiders and damsel bugs could control mirid numbers and cotton boll damage in field cages. We found that in cages containing mirids and only lynx spiders, lynx spiders reduced both mirid numbers and boll damage. However, in cages which contained mirids and both predators (lynx spiders and damsel bugs) only mirid numbers were reduced. To explain the negative effect of damsel bugs on boll damage, we examined the interactions between lynx spiders, damsel bugs and mirids. We found that lynx spiders were better mirid predators than damsel bugs, and that lynx spiders attacked damsel bugs, but not vice versa. Behaviourally, mirids responded to increasing predator pressure regardless of whether the predators were lynx spiders or damsel bugs. However, damsel bugs seemed to alter the behaviour of lynx spiders because in their presence, a higher proportion of lynx spiders moved to the top of the plant, towards the damsel bugs but away from the bolls found lower on the plant. These results suggest that the most likely explanation for the increase in boll damage in the presence of damsel bugs was that lynx spiders moved to the top of the plant in the presence of damsel bugs, which then exposed the bolls lower down on the plant to mirid attack. This work emphasizes the importance of behaviourally mediated effects in food webs extending over four trophic levels.