Exploration of the acarine fauna on coconut palm in Brazil with emphasis on Aceria guerreronis (Acari: Eriophyidae) and its natural enemies

Coconut is an important crop in tropical and subtropical regions. Among the mites that infest coconut palms, Aceria guerreronis Keifer is economically the most important. We conducted surveys throughout the coconut growing areas of Brazil. Samples were taken from attached coconuts, leaflets, fallen coconuts and inflorescences of coconut palms in 112 localities aiming to determine the occurrence and the distribution of phytophagous mites, particularly A. guerreronis, and associated natural enemies. Aceria guerreronis was the most abundant phytophagous mite followed by Steneotarsonemus concavuscutum Lofego & Gondim Jr. and Steneotarsonemus furcatus De Leon (Tarsonemidae). Infestation by A. guerreronis was recorded in 87% of the visited localities. About 81% of all predatory mites belonged to the family Phytoseiidae, mainly represented by Neoseiulus paspalivorus De Leon, Neoseiulus baraki Athias-Henriot and Amblyseius largoensis Muma; 12% were Ascidae, mainly Proctolaelaps bickleyi Bram, Proctolaelaps sp nov and Lasioseius subterraneus Chant. Neoseiulus paspalivorus and N. baraki were the most abundant predators on attached coconuts. Ascidae were predominant on fallen coconuts, while A. largoensis was predominant on leaflets; no mites were found on branches of inflorescences. Leaflets harboured higher mite diversity than the attached coconuts. Mite diversity was the highest in the state Pará and on palms surrounded by seasonal forests and Amazonian rain-forests. Neoseiulus paspalivorus, N. baraki and P. bickleyi were identified as the most promising predators of A. guerreronis. Analyses of the influence of climatic factors revealed that dry ambient conditions favour the establishment of A. guerreronis. Neoseiulus paspalivorus and N. baraki have differing climatic requirements; the former being more abundant in warm and dry areas, the latter prevailing in moderately tempered and humid areas. We discuss the significance of our findings for natural and biological control of A. guerreronis.     

Population dynamics of Aceria guerreronis (Acari: Eriophyidae) and other mites associated with coconut fruits in Una, state of Bahia, northeastern Brazil

Aceria guerreronis Keifer is a major coconut pest in the Americas, Africa and some Asian countries, and occurs in high population levels in northeastern Brazil. The determination of the climatic conditions that favorably affect its population growth and the prevalence and abundance of predatory mites can promote more efficient control practices. Our objective was to evaluate the pattern of occurrence and population dynamics of A. guerreronis, their associated predators and other mites during a 2?year period in a hybrid coconut plantation in the municipality of Una, state of Bahia, northeastern Brazil. Monthly samples of fruits were taken from June 2008 to May 2010 for qualitative and quantitative evaluation of mites. Aceria guerreronis represented 99.9?% of the mites. An average density of 1,117 mites per fruit and a maximum of 23,596 mites per fruit indicated that the level of infestation can be high in Bahia. Bdella ueckermanni Hernandes, Daud and Feres was the most abundant and frequent predator. Population increase of A. guerreronis was directly related to the temperature rise and inversely related to both the increase of air relative humidity and rainfall. The highest population densities occurred from November to March. The largest A. guerreronis populations occurred in fruits with 32 and 48?% of damaged surface. The relationship between prevailing wind direction and incidence of A. guerreronis could not be corroborated or refuted.     

Olfactory response of predatory mites to vegetative and reproductive parts of coconut palm infested by <Emphasis Type="Italic">Aceria guerreronis</Emphasis>

The phytophagous mite Aceria guerreronis Keifer is an important pest of coconut worldwide. A promising method of control for this pest is the use of predatory mites. Neoseiulus baraki (Athias-Henriot) and Proctolaelaps bickleyi Bram are predatory mites found in association with A. guerreronis in the field. To understand how these predators respond to olfactory cues from A. guerreronis and its host plant, the foraging behavior of the predatory mites was investigated in a Y-tube olfactometer and on T-shaped arenas. The predators were subjected to choose in an olfactometer: (1) isolated parts (leaflet, spikelet or fruit) of infested coconut plant or clean air stream; (2) isolated parts of non-infested or infested coconut plant; and (3) two different plant parts previously shown to be attractive. Using T-shaped arenas the predators were offered all possible binary combinations of discs of coconut fruit epidermis infested with A. guerreronis, non-infested discs or coconut pollen. The results showed that both predators were preferred (the volatile cues from) the infested plant parts over clean air. When subjected to odours from different infested or non-infested plant parts, predators preferred the infested parts. Among the infested plant parts, the spikelets induced the greatest attraction to predators. On the arenas, both predators preferred discs of coconut fruits infested with A. guerreronis over every other alternative. The results show that both predators are able to locate A. guerreronis by olfactory stimuli. Foraging strategies and implications for biological control are discussed.     

Intraguild predation and cannibalism between the predatory mites Neoseiulus neobaraki and N. paspalivorus, natural enemies of the coconut mite Aceria guerreronis

Neoseiulus neobaraki and N. paspalivorus are amongst the most common phytoseiid predators of coconut mite, Aceria guerreronis, found in the spatial niche beneath coconut fruit bracts. Both predators may occur on the same coconut palms in Benin and Tanzania and are therefore likely to interact with each other. Here, we assessed cannibalism and intraguild predation (IGP) of the two predators in the absence and presence of their primary prey A. guerreronis. In the absence of the shared extraguild prey, A. guerreronis, N. neobaraki killed 19 larvae of N. paspalivorus per day and produced 0.36?eggs/female/day, while the latter species killed only 7 larvae of the former and produced 0.35?eggs/female/day. Presence of A. guerreronis only slightly decreased IGP by N. neobaraki but strongly decreased IGP by N. paspalivorus, which consumed 4-7 times less IG prey than N. neobaraki. Resulting predator offspring to IG prey ratios were, however, 4-5 times higher in N. paspalivorus than N. neobaraki. Overall, provision of A. guerreronis increased oviposition in both species. In the cannibalism tests, in the absence of A. guerreronis, N. neobaraki and N. paspalivorus consumed 1.8 and 1.2 conspecific larvae and produced almost no eggs. In the presence of abundant herbivorous prey, cannibalism dramatically decreased but oviposition increased in both N. neobaraki and N. paspalivorus. In summary, we conclude that (1) N. neobaraki is a much stronger intraguild predator than N. paspalivorus, (2) cannibalism is very limited in both species, and (3) both IGP and cannibalism are reduced in the presence of the common herbivorous prey with the exception of IGP by N. neobaraki, which remained at high levels despite presence of herbivorous prey. We discuss the implications of cannibalism and IGP on the population dynamics of A. guerreronis and the predators in view of their geographic and within-palm distribution patterns.     

Compatibility of Neoseiulus paspalivorus and Proctolaelaps bickleyi, candidate biocontrol agents of the coconut mite Aceria guerreronis: spatial niche use and intraguild predation

The eriophyid mite Aceria guerreronis occurs in most coconut growing regions of the world and causes enormous damage to coconut fruits. The concealed environment of the fruit perianth under which the mite resides renders its control extremely difficult. Recent studies suggest that biological control could mitigate the problems caused by this pest. Neoseiulus paspalivorus and Proctolaelaps bickleyi are two of the most frequently found predatory mites associated with A. guerreronis on coconut fruits. Regarding biological control, the former has an advantage in invading the tight areas under the coconut fruit perianth while the latter is more voracious on the pest mites and has a higher reproductive capacity. Based on the idea of the combined use/release of both predators on coconut fruits, we studied their compatibility in spatial niche use and intraguild predation (IGP). Spatial niche use on coconut fruits was examined on artificial arenas mimicking the area under the coconut fruit perianth and the open fruit surface. Both N. paspalivorus and P. bickleyi preferentially resided and oviposited inside the tight artificial chamber. Oviposition rate of P. bickleyi and residence time of N. paspalivorus inside the chamber were reduced in the presence of a conspecific female. Residence of N. paspalivorus inside the chamber was also influenced by the presence of P. bickleyi. Both N. paspalivorus and P. bickleyi preyed upon each other with relatively moderate IGP rates of adult females on larvae but neither species yielded nutritional benefits from IGP in terms of adult survival and oviposition. We discuss the relevance of our findings for a hypothetic combined use of both predators in biological control of A. guerreronis.     

Distribution patterns of coconut mite, Aceria guerreronis, and its predator Neoseiulus aff. paspalivorus in coconut palms

Distribution patterns and numerical variability of the coconut mite Aceria guerreronis Keifer (Acari: Eriophyidae) and its predator Neoseiulus aff. paspalivorus DeLeon (Phytoseiidae) on the nuts of 3- to 7-month-old bunches of coconut palms were studied at two sites in Sri Lanka. At the two sites, coconut mites were present on 88 and 75% of the nuts but no more than three-quarters of those nuts showed damage symptoms. N. aff. paspalivorus was found more on mature nuts than on immature nuts. Spatial and temporal distribution of coconut mites and predatory mites differed significantly. The mean number of coconut mites per nut increased until 5-month-old bunches and declined thereafter. The densities of predatory mites followed a similar trend but peaked 1 month later. Variability in the numbers of mites among palms and bunches of the same age was great, but was relatively low on 6-month-old bunches. The results indicate that assessment of infestation levels by damage symptoms alone is not reliable. Sampling of coconut and/or predatory mite numbers could be improved by using several nuts of 6-month-old bunches. The effect of predatory mites on coconut mites over time suggests that N. aff. paspalivorus could be a prospective biological control agent of A. guerreronis.     

Status of Aceria guerreronis Keifer (Acari: Eriophyidae) as a Pest of Coconut in the State of Sao Paulo, Southeastern Brazil

The coconut mite, Aceria guerreronis Keifer, is one of the main pests of coconut palms (Cocos nucifera) in northeastern Brazil. The objective of this study was to evaluate the levels of the coconut mite and other mites on coconut palms in the state of S?o Paulo and to estimate the possible role of predatory mites in the control of this pest. The effect of cultivated genotypes and sampling dates on the mite populations was also estimated. We sampled attached fruits, leaflets, inflorescences, and fallen fruits. The coconut mite was the main phytophagous mite found on attached and fallen fruits, with average densities of 110.0 and 20.5 mites per fruit, respectively. The prevalent predatory mites on attached and fallen fruits were Proctolaelaps bulbosus Moraes, Reis & Gondim Jr. and Proctolaelaps bickleyi (Bram), both Melicharidae. On leaflets, the tenuipalpids Brevipalpus phoenicis (Geijsks) and Tenuipalpus coyacus De Leon and the tetranychid Oligonychus modestus (Banks) were the predominant phytophagous mites. On both leaflets and inflorescences, the predominant predatory mites belonged to the Phytoseiidae. Neoseiulus baraki (Athias-Henriot) and Neoseiulus paspalivorus (De Leon), predators widely associated with the coconut mite in northeastern Brazil and several other countries, were not found. The low densities of the coconut mite in S?o Paulo could be related to prevailing climatic conditions, scarcity of coconut plantations (hampering the dispersion of the coconut mite between fields), and to the fact that some of the genotypes cultivated in the region are unfavorable for its development.     

The coconut mite, Aceria guerreronis, in Benin and Tanzania: occurrence, damage and associated acarine fauna

The coconut mite Aceria guerreronis (Eriophyidae) is considered the most important pest of coconut fruits in Africa; however, quantitative knowledge about its distribution and abundance is lacking. We conducted four diagnostic surveys—three in Southern Benin and one along the coast of Tanzania—to determine the distribution of A. guerreronis and the severity of its damage to coconut fruits, as well as the diversity and abundance of other associated mites and potential natural enemies. Aceria guerreronis was found in all visited plantations with the percentage of damaged fruits varying considerably among plantations—67–85% in Benin and 43–81% in Tanzania. Overall, 30–40% of the fruit surfaces were damaged by A. guerreronis. Damage severity increased with fruit age and negatively affected fruit weight of 7- to 12-months-old fruits. Aceria guerreronis was by far the most abundant mite on coconut fruits but its abundance depended on fruit age. The highest densities of A. guerreronis were observed on 3- to 4-months-old fruits. Neocypholaelaps sp. (Ameroseiidae) was the most abundant mite on inflorescences. Three species of predatory mites (Phytoseiidae)—Neoseiulus baraki, N. neobaraki and N. paspalivorus—were the most commonly found predatory mites beneath the coconut bracts in association with A. guerreronis. Neoseiulus neobaraki was the prevailing predator in Tanzania while N. paspalivorus was the most frequent predator in Benin. Other mites found beneath the bracts were the herbivore Steneotarsonemus furcatus (Tarsonemidae) and the detritivore and fungivore Tyrophagus putrescentiae (Acaridae).     

Limitations of <Emphasis Type="Italic">Neoseiulus baraki</Emphasis> and <Emphasis Type="Italic">Proctolaelaps bickleyi</Emphasis> as control agents of <Emphasis Type="Italic">Aceria guerreronis</Emphasis>

Several predatory mites have been found in association with the coconut mite, Aceria guerreronis Keifer, in northeast Brazil. However, the latter still causes damage to coconut in that region. The objectives of this work were to compare the frequencies of occurrence of Neoseiulus (Phytoseiidae) and Proctolaelaps (Melicharidae) species on standing and aborted coconuts in coastal Pernambuco State, northeast Brazil and to analyze their possible limitations as control agents of the coconut mite, based on evaluations of the restrictions they may have to access the microhabitat inhabited by the pest and their functional and reproductive responses to increasing densities of the latter. Neoseiulus baraki (Athias-Henriot) was found mostly on standing coconuts whereas Proctolaelaps bickleyi (Bram) was found mostly on aborted coconuts. Measurements of the entrance to the microhabitat occupied by the coconut mite, between the bracts and the subjacent fruit surface, showed that this different pattern of predator prevalence could be related to predator sizes, although other environmental factors could not be disregarded. Progressively higher predation rate of N. baraki was observed up to an experimental density that corresponded to 1,200 coconut mites per fruit, which is close to the average number determined in northeast Brazil, reducing slightly afterwards. Predation rate of P. bickleyi reduced consistently but slightly with increasing prey densities, but in absolute values, rates were always much higher than determined for N. baraki. The excessively high killing capacity of P. bickleyi, probably related to its high feeding requirement, may be detrimental in terms of stability. In fact, such high requirement for food suggests that P. bickleyi might not have a strong relation with the coconut mite and that the latter may not be its main food source under natural conditions. It is concluded that body sizes of both predators and the exceedingly high feeding requirement of P. bickleyi may limit their performance as control agents of the coconut mite.     

Neoseiulus californicus (McGregor) and Neoseiulus fallacis (Garman): larval responsesto prey and humidity, nymphal feeding drive and nymphal predation on phytoseiid eggs

Are Neoseiulus californicus (McGregor) and Neoseiulus fallacis (Garman) both specialized predators of spider mites? As part of a series of studies made to answer this question, responses by larvae were assessed in treatments either with or without prey (Tetranychus urticae Koch) and with either moderate or high relative humidities (a factor that may distinguish between these two morphologically similar species). Neoseiulus fallacis larvae had more feeding, ambulatory activity and jerking (intra– or interspecific communication) in all treatments than N. californicus. The percent egg hatch was less and development took longer for N. fallacis larvae than for N. californicus larvae without prey at moderate humidity levels. The nymphs of both mites showed similar drives to feed 1–2>h after being held 12–48 h without food and then given eggs of T. urticae. Neoseiulus californicus nymphs fed more on the eggs of either phytoseiid and thereafter became adults than did N. fallacis nymphs. These data indicate that N. californicus may be a less specialized predator of spider mites than is N. fallacis.     

Two predatory mite species as potential control agents of broad mites

The broad mite is a key pest of various crops worldwide, including chili pepper (Capsicum frutescens), where it is controlled with chemical pesticides. Two phytoseiid predators (Amblyseius herbicolus and Neoseiulus barkeri) and a blattisociid mite species (Lasioseius floridensis) occur in association with the broad mite in Brazil. Predation of broad mites and oviposition rates of A. herbicolus were higher than those of N. barkeri and lowest for L. floridensis. On intact plants, both phytoseiids controlled broad mite populations seven days after their release. In a separate experiment, the predators controlled broad mites on flowering plants during 15 days. After two months, plants with predators produced heavier fruits than plants without predators. Concluding, these two phytoseiid species can control broad mites on chili peppers plants at different densities and over time, reducing production losses, and should be evaluated under field conditions.     

Influence of species of host plants on the predation of thrips by Neoseiulus cucumeris, Iphiseius degenerans and Orius laevigatus

A leaf-disc bioassay was used to compare the predation levels of two species of predatory mites (Neoseiulus cucumeris (Oudemans) and Iphiseius degenerans (Berlese)) and a predatory bug (Orius laevigatus (Fieber)), on the thrips Frankliniella occidentalis (Pergande) and Heliothrips haemorrhoidalis (Bouché), feeding on a range of susceptible plant species from twelve plant families. The predatory bug, O. laevigatus, reduced the number of thrips to a greater extent than the predatory mites and all three predators showed greater levels of predation on F. occidentalis than on H. haemorrhoidalis. The level of predation caused by each predator varied among the species of plants; the variation was greater on the plant hosts of H. haemorrhoidalis than of F. occidentalis.     

Morphological variation and reproductive incompatibility of three coconut-mite-associated populations of predatory mites identified as Neoseiulus paspalivorus (Acari: Phytoseiidae)

Predatory mites identified as Neoseiulus paspalivorus DeLeon (Phytoseiidae) have been considered as agents for classical biological control of the coconut mite, Aceria guerreronis Keifer (Eriophyidae), in Africa and elsewhere. Preliminary identification of geographically distinct populations as belonging to the same species (N. paspalivorus) was based on their morphological similarity. However, laboratory studies recently conducted have shown large differences in feeding behaviors and biological characteristics among individuals collected from three geographic origins: Brazil (South America), Benin and Ghana (West Africa). As morphologically similar specimens do not necessarily belong to the same species, we evaluated under laboratory conditions, reproductive compatibility between the specimens from three geographic locations to ascertain their conspecificity. Morphological measurements were also made to determine whether there is a means of discriminating between them. Inter-population crosses showed complete reproductive isolation between the three geographic populations, but interpopulation discontinuities in morphometric characters were absent. These results indicate that the tested specimens are distinct biological entities despite morphological similarity. Further molecular genetic studies are therefore proposed, including screening for endosymbionts and assessment of genetic differentiation, to determine the cause of reproductive incompatibility and to clarify the taxonomic relationship between those populations.     

Functionally different predators break down antipredator defenses of spider mites

Prey that lives with functionally different predators may experience enhanced mortality risk, because of conflicts between the specific defenses against their predators. Because natural communities usually contain combinations of prey and functionally different predators, examining risk enhancement with multiple predators may help to understand prey population dynamics. It is also important in an applied context: risk enhancement with multiple biological control agents could lead to successful suppression of pests. We examined whether risk enhancement occurs in the spider mite Tetranychus kanzawai Kishida (Acari: Tetranychidae) when exposed to two predator species: a generalist ant, Pristomyrmex punctatus Mayr (Hymenoptera: Formicidae), and a specialist predatory mite, Neoseiulus womersleyi Schicha (Acari: Phytoseiidae). We replicated microcosms that consisted of spider mites, ants, and predatory mites. Spider mites avoided generalist ants by staying inside their webs on leaf surfaces. In contrast, spider mites avoided specialist predatory mites that intruded into their webs by exiting the web, which obviously conflicts with the defense against ants. In the presence of both predators, enhanced mortality of spider mites was observed. A conflict occurred between the spider mites’ defenses: they seemed to move out of their webs and be preyed upon by ants. This is the first study to suggest that risk enhancement occurs in web‐spinning spider mites that are exposed to both generalist and specialist predator species, and to provide evidence that ants can have remarkable synergistic effects on the biological control of spider mites using specialist predatory mites.     

Absence of odour-mediated avoidance of heterospecific competitors by the predatory mite Phytoseiulus persimilis

Arthropods use odours associated with the presence of their food, enemies and competitors when searching for patches. Responses to these odours therefore determine the spatial distribution of animals, and are decisive for the occurrence and strength of interactions among species. Therefore, a logical first step in studying food web interactions is the analysis of behaviour of individuals that are searching for patches of food. We followed this approach when studying interactions in an artificial food web occurring on greenhouse cucumber in the Netherlands. In an earlier paper we found that one of the predators of the food web, the predatory mite Phytoseiulus persimilis Athias-Henriot, used to control spider mites, discriminates between odours from plants with spider mites, Tetranychus urticae Koch, and plants with spider mites plus conspecific predators. The odours used for discrimination are produced by adult prey in response to the presence of predators, and probably serve as an alarm pheromone to warn related spider mites. Other predator species may also trigger production of this alarm pheromone, which P. persimilis could use in turn to avoid plants with heterospecific predators. We therefore studied the response of the latter to odours from plants with spider mites and 3 other predator species, i.e. the generalist predatory bug Orius laevigatus (Fieber), the polyphagous thrips Frankliniella occidentalis and the spider-mite predator Neoseiulus californicus (McGregor). Both olfactometer and greenhouse release experiments yielded no evidence that P. persimilis avoids plants with any of the 3 heterospecific predators. This suggests that these predators do not elicit production of alarm pheromones in spider mites, and we argue that this is caused by a lack of coevolutionary history. The consequences of the lack of avoidance of heterospecific predators for interactions in food webs and biological control are discussed.     

The presence of <Emphasis Type="Italic">Tetranychus urticae</Emphasis> (Acari: Tetranychidae) and its predators on plants in the ground cover in commercially treated vineyards

The presence of Tetranychus urticae and its predators on plants in the ground cover in commercially treated vineyards in the Hex River Valley, Western Cape Province, South Africa was investigated. Six Barlinka and six Dauphine vineyards were surveyed. Leaf samples from the plants on the vineyard floor were taken at monthly intervals and microscopically examined for spider mites and their predators. The coverage of the plants was determined using a co-ordinate sampling system. A wide variety of plant species was found on the vineyard floor. Most of the spider mites found on these plants were T. urticae. The phytoseiid mites on the plants were Euseius rubicolus, Neoseiulus californicus and an undescribed Typhlodromus species. The presence of these plants made it possible for T. urticae and phytoseiid mites to occur throughout the year, with a similar pattern in their seasonal activity on the vineyard floor and the vine leaves.     

Spider Mites Avoid Plants with Predators

While searching for food, prey can use cues associated with their predators to select patches with a reduced predation risk. In many cases, odours indicate the presence of both food and predators. Spider mites are known to use odours to locate food and mates, but also to avoid interspecific competitors. We studied the response of the two-spotted spider mite, Tetranychus urticae, to cues associated with the presence of their predators, the phytoseiid Phytoseiulus persimilis. We found that the spider mites strongly avoid plants defended by this predator, but do not avoid plants with another predatory mite, Neoseiulus californicus. Since P. persimilis is commonly used in the greenhouse where our strain of T. urticae was collected and strains of this pest are known to adapt to greenhouse environments, we hypothesize that there has been selection on the pest to recognize its enemy. We further hypothesize that there has been no selection to recognize N. californicus, as it has not been used against two-spotted spider mites in the greenhouse where our spider mites were collected. We discuss the implications of avoidance of predation by spider mites and non-lethal effects of predators for biological control of this pest in greenhouses.     

Residual bioassay to assess the toxicity of Acaricides against Aceria guerreronis (Acari: Eriophyidae) under laboratory conditions

Aceria guerreronis Keifer (Acari: Eriophyidae) is considered a major pest of the coconut (Cocos nucifera L.), and the use of pesticides is the current method to control it. However, no standard toxicological tests exist to select and assess the efficiency of molecules against the coconut mite. The aim of this study was to develop a methodology that allows for the evaluation of the relative toxicity of acaricides to A. guerreronis through rapid laboratory procedures. We confined A. guerreronis on arenas made out of coconut leaflets and tested two application methods: immersing the leaf fragments in acaricides and spraying acaricides on the leaf fragments under a Potter spray tower. In the latter application method, we sprayed leaf fragments both populated with and devoid of mites. We evaluated the comparative toxicity of two populations (Itamaracá and Petrolina, Pernambuco, Brazil) by spraying on leaflets without mites and submitted the mortality data to probit analysis after 24 h of exposure. No difference was observed in the LC50, regardless of whether the leaflets were immersed or sprayed with acaricide (abamectin, chlorfenapyr or fenpyroximate). The toxicity of chlorfenapyr and fenpyroximate did not differ, irrespective of whether it was applied directly to the leaflet or to the mite; however, the toxicity of abamectin was higher when applied directly to the mite. Chlorpyrifos and abamectin toxicities were lower for the Petrolina population than for the Itamaracá population. Immersing and spraying coconut leaflets can be used to assess the mortality of A. guerreronis under laboratory conditions.     

Predators induce interspecific herbivore competition for food in refuge space

Resource competition among herbivorous arthropods has long been viewed as unimportant because herbivore populations are controlled by predators. Although recently resurrected as an organizing force in arthropod communities on plants, there is still general agreement that resource competition among herbivores is reduced by predators. Here we show the reverse: predators induce interspecific resource competi-tion among herbivores. We found that thrips larvae ( Frankliniella occidentalis ) use the web produced and inhabited by the spider mite Tetranychus urticae as a refuge from predation by the phytoseiid mite Neoseiulus cucumeris . Thrips larvae prefer clean plant parts, but move into the web upon perceiving volatile cues associated with thrips and predators. This behaviour leads to lower predation risk, but also to reduced developmental rate and lower production of thrips larvae due to competition with spider mites. In addition, thrips larvae consume spider-mite eggs. Thus, predators induce interspecific competition and intraguild predation among herbivores within refuge space, even when host plants have an overall green appearance.