UvA-DARE (Digital Academic Repository) What lies …K.A. Muñoz Cárdenas, 2017. What lies beneath?...

UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

What lies beneath?Linking litter and canopy food webs to protect ornamental cropsMuñoz Cárdenas, K.A.

Link to publication

Creative Commons License (see https://creativecommons.org/use-remix/cc-licenses):Other

Citation for published version (APA):Muñoz Cárdenas, K. A. (2017). What lies beneath? Linking litter and canopy food webs to protect ornamentalcrops.

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 18 Jul 2020

https://dare.uva.nl/personal/pure/en/publications/what-lies-beneath(fdea6dbe-484b-4041-a8ce-d2711d56b3fe).html

Karen A. Muñoz Cárdenas

What lies beneath?Linking litter and canopy food webs

to protect ornamental crops

What lies beneath?Linking litter and canopy food webs to

protect ornamental crops

K.A. Muñoz Cárdenas, 2017. What lies beneath? Linking litter and canopy food webs to protect

ornamental crops

PhD thesis, University of Amsterdam, The Netherlands

Publication of this thesis was financially supported by Koppert Biological Systems(www.koppert.nl)

ISBN: 978 94 91407 47 5

Cover design: Jan van Arkel (IBED)Thesis lay-out: Jan Bruin (www.bred.nl)

What lies beneath?Linking litter and canopy food webs to

protect ornamental crops

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctoraan de Universiteit van Amsterdamop gezag van de Rector Magnificus

Prof. dr. ir. K.I.J. Maexten overstaan van een door het College voor Promoties ingestelde

commissie, in het openbaar te verdedigen in deAula der Universiteit

op woensdag 21 juni 2017, te 11:00 uur

door

Karen Andrea Muñoz Cárdenas

geboren te Bogotá, Colombia

Promotiecommissie

PromotorProf. Dr. P.C. de Ruiter Universiteit van Amsterdam

CopromotorDr. A.R.M. Janssen Universiteit van Amsterdam

Overige ledenProf. Dr. E.T. Kiers Vrije Universiteit AmsterdamProf. Dr. S.B.J. Menken Universiteit van AmsterdamDr. W.E. Morriën Universiteit van AmsterdamProf. Dr. W.H. van der Putten Wageningen UniversiteitDr. I.M. Smallegange Universiteit van AmsterdamProf. Dr. F.L. Wäckers Lancaster University, UK

Faculteit der Natuurwetenschappen, Wiskunde en Informatica

Contents

6 Author addresses

7 0General introduction

27 1Generalist red velvet mite predator (Balaustium leanderi) performs better on a mixed diet

57 2Supplying high-quality alternative prey in the litter increases control of an above-ground plant pest by a generalist predator

81 3Alternative food for litter-inhabiting predators decreases pest densities and above-ground plant damage

101 4Single and combined predator releases with alternative food increases thrips control in an ornamental crop

123 General discussion132 Author contributions and project funding133 Summary137 Samenvatting141 Curriculum vitae and publications143 Acknowledgements

Author addressesDepartment of Population Biology, Institute for Biodiversity and Ecosystem Dynamics, University ofAmsterdam, Science Park 904, 1098 XH Amsterdam, The Netherlands

Duarte, Marcus V. A.; Ersin, Firdevs; Janssen, Arne; Muñoz-Cárdenas, Karen; Sabelis, Maurice W.

Programa de Biología Aplicada, Facultad de Ciencias Básicas, Universidad Militar Nueva Granada, Km 3 viaCajicá-Zipaquirá, Cajicá, Colombia

Muñoz-Cárdenas, Karen; Cantor, R. Fernando; Rodríguez, C. Daniel

Centro de Biosistemas, Universidad Jorge Tadeo Lozano, Carretera Central del Norte, 3 kms al norte delpuente de La Caro, Chía, Cundinamarca, Colombia

Fuentes, Luz Stella

Department of Plant Protection, Faculty of Agriculture, University of Ege, 35100 Bornova, Izmir, Turkey

Ersin, Firdevs

Biobest Belgium, Ilse Velden 18, 2260 Westerlo, Belgium

Pijnakker, Juliette

Wageningen UR Greenhouse Horticulture, PO Box 20, 2265 ZG Bleiswijk, The Netherlands

Leman, Ada; Messelink, Gerben J.; Pijnakker, Juliette

Koppert Biological Systems, PO Box 155, 2650 AD Berkel en Rodenrijs, The Netherlands

Hoogerbrugge, Hans; van Houten, Yvonne

Department of Agricultural Development, Laboratory of Agricultural Entomology and Zoology, DemocritusUniversity of Thrace, Pantazidou 193, 68 200 Orestiada, Greece

Pappas, Maria L.

Department of Entomology, Federal University of Viçosa, Viçosa, Minas Gerais, Brazil

Duarte, Marcus V. A.

Mitox Consultants/Eurofins, Science Park 406, 1098 XH Amsterdam, The Netherlands

Faraji, Farid

Laboratorio de Acarología, Departamento de Entomología y Acarología, ESALQ, Univer si dade de São Paulo,Piracicaba, Av. Pádua Dias, 11, São Paulo, Brazil

Rueda-Ramirez, Diana

General introduction

This thesis is on biological control, which is the use of live organisms to keep agri-cultural pests, diseases, and weeds at low levels (DeBach 1974). Since ancienttimes, humans have utilized natural enemies to control pests. Thousands of yearsago, ants and spiders were used in China to protect tree crops from herbivore pests.Through history, there have been many successful programs of biological controlworldwide (Van Driesche & Bellows 1996). Only after World War II, the use of chem-ical pesticides replaced natural and biological control (Bale et al. 2008). Thanks to thepressure of consumers for pesticide-free agricultural products, biological control hasagain become a key component for pest control (Calvo et al. 2015). Moreover, grow-ers have actively chosen to use biological control because of the advent of pestsresistant to chemical pesticides (Bolckmans 1999). Nowadays, biological control is aviable alternative to chemical pest control.

Currently, there is a wide range of biological control agents commercially avail-able, such as parasitoids, entomopathogens and predators. Parasitoids are waspsand flies that deposit their eggs in or on their host. In the case of biological control,these hosts are insect pests. The developing parasitoid larva feeds on thehaemolymph of the host, affecting the host’s development and finally causing itsdead (Godfray 1994). Entomopathogens used in biological control are viruses, fungi,bacteria, and protozoa that cause diseases in insects and mites (Federici 1999).These pathogens infect arthropod pests, eventually killing them (Federici 1999). Inthis thesis I focus on arthropod natural enemies. Arthropod predators living on cropplants can be either specialist, attacking one or a few closely related pest species,or generalist, attacking a wider range of pests. In the past, biological pest control hasfocused on the use of one or a few specialist natural enemies for each pest species(Huffaker & Messenger 1976; Hokkanen & Pimentel 1984; van Lenteren & Woets1988; Hoy 1994) because the use of generalist predators was judged to be morerisky as they could also target non-pest species (van Lenteren et al. 2003). Moreover,they were considered inferior enemies to parasitoids because their dynamics wouldnot be synchronized with the pest, and because they usually do not have a highpotential for increase (Murdoch et al. 1985). Nevertheless, generalist predators arewidely used in biological control nowadays (Symondson et al. 2002). These preda-tors have the advantage of being able to feed on food sources different than the tar-get pest. Thus, they are able to establish in the crop before pest invasions(Symondson et al. 2002). The introduction of different natural enemies (either gener-

7

0

alists, specialists, or both) results in artificial communities in the crops, and the mem-bers of those communities can interact in complex ways. Not only do natural ene-mies interact with the target pest, but interactions also occur among different pestspecies and among natural enemy species (Helle & Sabelis 1985; Minks & Harrewijn1989; Sabelis 1992), and these interactions can affect biological control (Sih et al.1985; Janssen et al. 1998; Prasad & Snyder 2004; Evans 2008). A lot of attention hasbeen given to the effects of interactions among various species of natural enemies.Compared to the effect of each enemy species separately, the effect of such interac-tions can be negative, neutral or positive for biological control (Rosenheim et al.1995; Losey & Denno 1998; Rosenheim 1998; Colfer & Rosenheim 2001; Snyder &Ives 2001; Venzon et al. 2001; Cardinale et al. 2003; Snyder & Ives 2003; Finke &Denno 2004; Cakmak et al. 2009).

Interactions among natural enemiesWhen multiple species of arthropod natural enemies are present in a crop, there canbe negative effects on biological control. Especially generalist predators can interferewith other natural enemies not just through competition for prey or food, but alsothrough hyperpredation or intraguild predation (Rosenheim et al. 1995; Rosenheim1998; Snyder & Wise 2002). Intraguild predation takes place when two natural ene-mies that attack the same prey are also involved in predation (Polis & Holt 1992;Ferguson & Stiling 1996). Intraguild predation is common in food webs in crops andits effects can be strong, resulting in low survival of one or both natural enemies, dis-rupting biological control (Rosenheim et al. 1995; Rosenheim & Harmon 2006;Janssen et al. 2007; Vance-Chalcraft et al. 2007; but see Janssen et al. 2006).

The use of multiple predators can also result in positive effects on biological con-trol (Losey & Denno 1998). A predator with a certain foraging behaviour can causebehavioural changes in a particular prey, making it more available to other predatorspecies (Sih et al. 1998). For example, Losey & Denno (1998) show that better bio-logical control of aphids is achieved with two predators instead of either one sepa-rately. The mechanism for this positive effect is that a canopy-dwelling predatorinduced ‘dropping’ behaviour in the aphids, which makes them available to a preda-tor on the soil. Yet other studies show that the effect of the presence of multiple pred-ators on pest control can be neutral, meaning that the multiple predator effects onthe pest population correspond to the sum of the single-species effects (Gillespie &Quiring 1990; Sokol-Hessner & Schmitz 2002; Wiethoff et al. 2004), which would stillresult in better biological control.

Summarizing, the indirect interactions among arthropods in food webs with gen-eralist arthropod predators are complex and need more study. To make adequate useof generalist predators for biological control, it is crucial to study their interactions

8

CHAPTER 0 | GENERAL INTRODUCTION

with all pests and predators present in a crop. In this thesis, I studied the combinedeffect of multiple predator species on the control of one pest species.

Interactions among pest organismsHerbivorous prey might compete among each other for plant resources, but this onlyoccurs at high pest densities, which is undesirable in agriculture because it wouldcertainly result in exceeding the economic damage threshold. For this reason, I willnot further discuss resource competition among herbivores in this thesis. Herbivorescan also affect each other indirectly by inducing plant defence responses: the attackof an herbivore triggers a plant defensive response that does not only affect theinducing herbivore, but also other herbivore species on the same plant (Karban &Carey 1984). Plants have direct defences against herbivores, for example, they mightproduce toxins or anti-feeding compounds, resulting in reduced survival, fecundity ordevelopmental rate of the herbivore (Kessler & Baldwin 2002). The structure and com-position of herbivore communities are often shaped by plant-mediated interactionsbetween herbivores due to induction of plant defences (Karban & Baldwin 1997;Kessler & Halitschke 2007). For example, resistance against leaf miners increaseswhen tomato and cucumber plants are previously attacked by whiteflies (Inbar et al.1999; Zhang et al. 2005). In contrast, susceptibility to aphids increases when tomatoplants were previously attacked by whiteflies (Nombela et al. 2009). Plant defencescan also be down-regulated, the spider mite Tetranychus evansi suppresses plantdefences (Sarmento et al. 2011a), which enhances the mite’s performance, as well asthat of the closely related spider mite Tetranychus urticae (Sarmento et al. 2011b).

Plant defences can also be indirect by affecting the action of the natural enemiesof the herbivores (see Sabelis et al. 1999 for a review). They can do this by providingfood or shelter to the natural enemies, thus arresting the natural enemies on theplant, and increasing their survival or reproduction. Plants are also known to producespecific blends of volatiles upon attack by herbivores, and these volatiles are attrac-tive to natural enemies (Turlings et al. 1990; Dicke & Sabelis 1988) and herbivores canthus also affect each other through the induction of plant volatiles. For example,Shiojiri et al. (2002) show lower parasitism of P. xylostella on cabbage plants co-infested with another herbivore (Pieris rapae) than on plants infested only with P.xylostella, because the parasitoids of P. xylostella are not attracted to the volatileblends produced by plants that are attacked by both herbivores. The opposite effectis recorded on cabbage plants with the parasitic wasp of P. rapae, which attackedmore hosts on plants attacked by both herbivores (Shiojiri et al. 2002).

Herbivores can also affect each other through structural changes of the environ-ment. For example, several species of spider mites cover the leaf area on which theyfeed with a dense web, which is thought to serve as protection against predators

9

GENERAL INTRODUCTION | CHAPTER 0

(Sabelis & Bakker 1992), but this web can also serve as refuge for other herbivores(Pallini et al. 1998; Magalhães et al. 2007), but may exclude yet other herbivores(Sarmento et al. 2011b). Concluding, interactions among herbivores might affect bio-logical pest control in crops.

Different prey species can also affect each other indirectly through a population ofgeneralist predators and this is one of the main themes of this thesis. For example,when a generalist predator feeds on a target pest and on an alternative prey, the alter-native prey can indirectly affect the densities of the pest (Holt 1977). The presence oraddition of alternative food or prey to a population of generalist predators can result inpositive effects on the pest species (apparent mutualism; Holt 1977). For example, thepredator can direct its attacks to the alternative food or prey and thus become satiat-ed, resulting in fewer attacks of the pest (apparent mutualism; Holt 1977), negativelyaffecting biological control. However, after one or a few generations, the increasedavailability of food will boost the densities of the predators, resulting in increased pre-dation of both the target pest and the supplied food (apparent competition; Holt 1977),thus biological control is increased. When populations of natural enemies and pestsshow fluctuations, periods of positive effects of the two prey on each other will alter-nate with periods of negative effects (Abrams & Matsuda 1996). Whereas apparentmutualism is undesired when supplying alternative food, apparent competition is ben-eficial for biological control (van Rijn et al. 2002). The occurrence of apparent compe-tition in biological control systems has been widely studied (Collyer 1964; Karban et al.1994; Hougen-Eitzman & Karban 1995; Müller & Godfray 1997; Liu et al. 2006;Messelink et al. 2008). For example, Karban et al. (1994) shows that releasing an eco-nomically unimportant herbivore together with predatory mites increases the biologi-cal control of Pacific spider mites, a pest of grapevines. Hence, apparent competitionbetween herbivores and between a pest population and alternative food can improvebiological control. In this thesis, I present several examples of this.

An important aspect that has not been incorporated in the theory of apparent com-petition and apparent mutualism is the effect of a mixed predator diet. Depending onthe quality of the prey or food source, the performance of generalist predators can beaffected by feeding on a mixed diet (Uetz et al. 1992; Toft & Wise 1999; Oelbermann& Scheu 2002; Messelink et al. 2008; Marques et al. 2015). For example, the fecun-dity of the predatory mite Homeopronematus anconai (Baker) is significantlyincreased when it feeds on its prey plus pollen (Hessein & Perring 1988). Messelinket al. (2008) found a shorter immature development of the predatory mite Amblyseiusswirskii (Athias-Henriot) when feeding on two prey (thrips and whiteflies) compared topredators that feed on either of the two prey separately. Other studies show that amixed diet can have positive or negative effects on performance of generalist preda-tors, depending on the combinations (Toft & Wise 1999; Oelbermann & Scheu 2002).

10


11


A meta-analysis performed by Lefcheck et al. (2013) shows that animals that feed onmixed diets have higher fitness than animals feeding on single diets.

A question that arises is why a mixed diet affects performance of generalist pred-ators. There are two possible reasons. Firstly, different prey may provide the preda-tors with different essential nutrients. With certain mixed diets, predators can acquirethe nutrients needed to reproduce more or develop faster compared with a singlefood or another mixed diet (Uetz et al. 1992; Lefcheck et al. 2013). Another reason isthat, by mixing different prey or foods, predators can avoid ingesting harmful quan-tities of poisonous or growth-reducing substances from each food or prey separate-ly (Coll & Guerson 2002; Lefcheck et al. 2013). In order to design a food web thatenhances pest control, it is crucial to assess the effects of mixed diets on the per-formance of predators. In this thesis, I studied the effects of mixed diets on the per-formance of generalist predatory mites used for biological control.

Pollen as alternative foodMany species of plants produce pollen that are edible for arthropods, including pred-ators and parasitoids. Thus, plants invoke a kind of apparent competition betweenthe pollen and the pests attacked by these natural enemies. For some natural ene-mies, pollen is a highly nutritious food. It contains proteins, lipids, carbohydrates, andminerals (Lundgren 2009). The presence of edible pollen can result in the persistenceof populations of arthropod predators in the absence of their prey. Currently, thepollen of some plant species, such as Typha spp., are being used in Europe as analternative food for generalist predators in greenhouse crops (Vangansbeke et al.2014; Pijnakker et al. 2016a). A risk of adding pollen to crops is that insect pests suchas thrips can benefit from it (Chitturi et al. 2006; Leman & Messelink 2015;Vangansbeke et al. 2016). However, studies have shown that pollen can help boost-ing the populations of generalist predators, increasing biological control of thrips andherbivorous mites (van Rijn et al. 1999; Nomikou et al. 2002; Duarte et al. 2015). Forexample, van Rijn et al. (2002) added pollen as alternative food for phytoseiid mitesand found increased control of thrips on cucumber plants compared with treatmentsin which pollen was not added.

Contrasting results have also been found. Delisle et al. (2015) show that by addingpollen as alternative food for predatory mites, thrips cause more damage than whenpollen has not been added. However, the duration of the experiment may have beentoo short to observe a positive effect of adding pollen for predators on thrips control(i.e. the authors observed short-term apparent mutualism, but no longer-term appar-ent competition between the pollen and the pest). Although the addition of pollen andother alternative foods to the above-ground parts of a crop often does seem to resultin better pest control, the disadvantage is that these alternative foods may contam-

inate the marketable parts of the crop. This is especially the case for ornamentals.I therefore investigate the possibilities of adding alternative food for predators to thelitter instead of to the plant canopy. This requires investigating the connectionsbetween food webs occurring in the litter and the above-ground plant parts.

Links between above-ground and litter food websOne clear distinction between the food web occurring above-ground and that occur-ring in the litter is the presence of detritivores. These organisms usually do not causedamage to the plants, but have an important role in recycling nutrients and mineralsfrom the litter (Seastedt 1984). Coincidentally, several detritivore species are used asfood for mass production of natural enemies, hence, there is potential for using the lit-ter food web as a source of alternative food for predators. Although the two food webswere often studied in isolation, it is now clear that the above-ground and below-ground communities associated with plants are strongly interlinked (Scheu 2001; vander Putten et al. 2001, 2009; Miyashita et al. 2003; Moore et al. 2003; Wardle et al.2004). For example, some generalist predators can switch from feeding on detritivoresbelow-ground to herbivores in the canopy, thus forming a direct link between the two(Scheu 2001). There are a few examples of biological control systems in which the linkof above-ground and below-ground food webs is investigated (Settle et al. 1996; Halaj& Wise 2002; Birkhofer et al. 2008; Von Berg et al. 2009). Settle et al. (1996) increaseorganic matter in rice fields, thus enhancing populations of decomposers and plank-ton feeders, resulting in a significant increase in the abundance of spiders and insectpredators and a decrease of above-ground herbivores. Contrasting results are report-ed by Birkhofer et al. (2008), who show that an increase in decomposer prey densi-ties in wheat fields results in an increase of pest populations above-ground 1 monthlater. The authors conclude that increasing the densities of decomposers results inshifts in the foraging of beetles and spiders, predators that typically have long gener-ation times, reducing predation of the pest herbivores (reminiscent of apparent mutu-alism; Holt 1977, Abrams & Matsuda 1996). Another study shows that generalist pred-ators feed on organisms from the above-ground and below-ground food webs, hav-ing an effect on above-ground pest densities (Von Berg et al. 2009). Although most ofthese studies show that the below-ground food web and the herbivore food web onthe above-ground plant parts are connected through generalist predators, many ofthem did not study population dynamics of predators and pests and the impact onplant damage and thus on biological control. In this thesis, I therefore investigate inter-actions among and between food webs on above-ground plant parts and the litter andhow these interactions affect biological pest control in a rose crop (BOX 1). Food webinteractions in such a system can be very complex (FIGURE 0.1), and the effect of theseinteractions are likely to affect biological control. Studying these interactions is impor-

12


13


tant for two reasons: first, it can help to design food webs that allow growers toimprove biological control programs, and second, it can help to test ecological theo-ries. In this thesis, I specifically study the effects of apparent competition and appar-ent mutualism and the interactions between above-ground and below-ground foodwebs on pest densities.

FIGURE 0.1. A food web associated with roses in a greenhouse. (1) Plant resources: (1a) rose plants; (1b) pollen,

which serves as food for some pests and predators. (2) Detritivores in the litter can be used as alternative prey

for predators. (3) Pest species: (3a) spider mite eggs; (3b) thrips, which cause damage on above-ground plant

parts and spend part of their life cycle as pre-pupae and pupae in the litter or soil; (3c) whitefly eggs. Notice that

I have depicted eggs of spider mites and whiteflies, which do not feed on plants, but the other stages of these

pest species do, hence the solid arrows. (4) Generalist predators, which connect above-ground and litter or soil

food webs by feeding on pests, and by feeding on herbivores and alternative prey in the litter or soil: (4a) Ambly -

seius swirskii; (4b) Balaustium leanderi; (4c) Cosmolaelaps n. sp.; (4d) Macrocheles robustulus.

14


Box 1: Cut-rose production

Biological control strategies commonly used for vegetable crops do not always work in

ornamentals (Parrella et al. 1999). This is because damage thresholds are much lower for

ornamentals than for vegetable crops: for example, the presence of thrips or thrips dam-

age is unacceptable for cut-flower commercialization. Thus, growers of ornamental

crops often spray chemical pesticides and fungicides regularly and year-round (Parrella

et al. 1999). In order to use biological control as a feasible alternative to chemical pesti-

cides, natural enemies must therefore be efficient in maintaining ornamentals free from

pests.

Cut-rose production is important worldwide, it provides thousands of jobs in many

countries (Parrella et al. 1999). In some countries, like The Netherlands, roses are grown

using a technique developed in Japan, known as arching (Ohkawa & Suematsu 1999).

Growers buy budded stock plants from propagator companies. The substrate used is

rockwool; fertilizers dissolved in water are provided through the irrigation system. Under

these conditions, the rose plants can last for around 10 years (Parrella et al. 1999).

During this time, rose plants produce a large amount of litter (FIGURE 0.2a) For practical

reasons, this is disposed below the gutters, a special kind of benches where the plants

on the rockwool are growing (FIGURE 0.2b). From each rose plant, a few stems are select-

ed to produce marketable roses, the other stems are bent, but remain attached to the

plant (Ohkawa & Suematsu 1999). The bent stems connect the litter layer to the foliage

of the plants (FIGURE 0.2a). In rose crops, the food web within the litter or soil layer has

not been studied before, nor the interactions among the organisms belonging to this

food web, and with the food web on the above-ground plant parts.

FIGURE 0.2. Characteristics of com-

mercial rose crops in The Nether -

lands. (a) Litter in a commercial

rose greenhouse. (b) Diagram of a

commercial rose production sys-

tem.

15


SynopsisThe main research question of this thesis is how interactions between above-groundand below-ground food webs affect biological control. In CHAPTER 1, I investigatedthe effect of a mixed diet of above-ground plant pests on life history traits of a gen-eralist below-ground predator, Balaustium leanderi, which also forages on the above-ground plant parts (BOX 2). Females of B. leanderi lay one batch of eggs in 1 to 2 daysat the end of their lives, exhibiting big-bang reproduction and their life cycle last c. 2months. I recorded life history traits such as reproductive performance, survivorshipand development when fed on mixed diets of three pest species that inhabit above-ground plant parts (BOX 3).

In CHAPTER 2, I further investigated generalist predators that link above-groundand litter food webs. The canopy-dwelling predator A. swirskii mainly inhabits theabove-ground plant parts but makes excursions to the litter layer (BOX 2). I testedwhether the addition of astigmatic mites (BOX 4) as alternative prey in the litter wouldboost predator densities, resulting in better pest control (apparent mutualism orapparent competition, Holt 1977). I addressed the following questions. 1) Do preda-tors move from the canopy to the litter to feed on an alternative prey? 2) Do preda-tors still control the pest even if the quality of the alternative prey in the litter is high-er than the pest? 3) Do pest densities and plant damage decrease when providingalternative prey in the litter?

Subsequently, I investigated the effect of litter-inhabiting predators (BOX 2) present incommercial rose production greenhouses on thrips control (CHAPTER 3). Many pestspecies, including thrips, spend part of their life cycle in the litter or soil (Wahab 2010),so litter-inhabiting predators may contribute to thrips control. I specifically testedwhether supplying food for these predators helps to increase biological control of thrips.

Growers commonly release canopy-dwelling predators to control pests (Branigan1915; DeBach 1974; Gerson & Weintraub 2007), and these may interact with litter-inhabiting predators, which may affect pest control. Thus, in CHAPTER 4 I investigat-ed the combination of species of canopy-dwelling and litter-inhabiting predatorspecies. Combining natural enemies sometimes results in increased biological con-trol compared with single releases of either of the natural enemy species separately(Losey & Denno 1998, 1999; Sih et al. 1998; Casula et al. 2006). For this reason, andbecause pests such as thrips inhabit above-ground plant parts and the litter or soil,attempts have been made to increase thrips control by combining canopy-dwellingpredators and soil-inhabiting predators (Wiethoff et al. 2004; Thoeming & Poehling2006; Manners et al. 2013; Pozzebon et al. 2015). The novelty of my study is that Iinvestigated whether biological control can be enhanced by supplying different alter-native food sources for canopy-dwelling and litter-inhabiting predators (pollen orastigmatic mites).

16


Box 2: Predators

One of the most successful groups of arthropods used to control pests in greenhouses

are predatory mites (Gerson & Weintraub 2007). After being released, they can rapidly es -

ta blish in the crops; they oviposit on different parts of the plants, including the lower stra-

tum of the canopy. This is

important because not all

predatory mites are removed

when harvesting flowers. The

main predatory mite species

studied in this thesis are

Balaus tium leanderi (Haitlin -

ger) (Actinotrichida: Erythrae -

idae), Amblyseius swirskii

(Athi as-Henriot) (Acari: Phyto -

sei idae), Cosmolaelaps n. sp.

(Acari: Laelapidae) and

Macro cheles robustulus (Ber -

lese) (Acari: Macrochelidae).

These predatory mites are

chosen because they either

occur in rose greenhouses or

are released commercially in

rose crops.

Balaustium leanderi (FIGURE

0.3a) is a generalist predator

which is naturally found in

Mexico and Colombia (Getiva

& Acosta 2004; Fuentes et al.

2014). It inhabits the soil or lit-

ter layer in roses and other

crops, where the applications

FIGURE 0.3. Generalist predators in

greenhouse crops. (a) Balaustium

leanderi adult; (b) Amblyseius swir -

skii adult feeding on a thrips larva;

(c) Cosmolaelaps n. sp. adult.

Photographs J. van Arkel.

17


of chemical pesticides are not frequent (Muñoz-Cárdenas et al. 2015). Some species of

the genus Balaustium have been studied in regard to their potential use in biological con-

trol of agricultural pests (Cadogan & Laing 1977; Makol et al. 2012; CHAPTER 1). Due to dif-

ficulties rearing these predatory mites, biological control studies are scarce. However,

studying the potential for pest control of these predatory mites is important because they

are well adapted to crop conditions and they are found attacking different pest species in

America and Europe (Muñoz-Cárdenas et al. 2015).

Amblyseius swirskii (FIGURE 0.3b) is a predator from the Mediterranean region (Swirskii

et al. 1967; Nomikou et al. 2001). This predator is widely used in Europe and North

America to control pests in vegetable and ornamental crops like roses (Pijnakker &

Ramakers 2008; Buitenhuis et al. 2015; Calvo et al. 2015). In vegetable crops, A. swirskii

successfully controls whiteflies and thrips (Nomikou et al. 2002; Messelink et al. 2008;

Nomikou et al. 2010), but it can also control herbivorous mites (van Maanen et al. 2010)

and feeds on other alternative food sources and pest species (Hoogerbrugge et al. 2008;

Calvo et al. 2015; Janssen & Sabelis 2015; Vangansbeke et al. 2016). This predator inhab-

its the plant canopy and moves to the litter or soil to disperse (Buitenhuis et al. 2010) or to

feed on alternative prey (CHAPTER 2).

Cosmolaelaps n. sp. (FIGURE 0.3c) inhabits the litter layer in rose greenhouses in The

Netherlands (CHAPTER 3). Individuals of this genus feed on nematodes, astigmatic mites,

and thrips (de Moraes et al. 2015). A species of this genus found in Brazil, Cosmolaelaps

jaboticabalensis, is considered to have potential to control thrips (Furtado & de Moraes

2015).

Machrocheles robustulus is a generalist predator that also inhabits the soil in green-

houses in The Netherlands. It feeds on astigmatic mites and it is commercially released in

greenhouses to control insect pests like thrips (Messelink & van Holstein-Saj 2008).

Box 3: Target pests

The following pest species were studied in my thesis because they are cosmopolitan

and represent some of the most important pest species in crops in the field and in

greenhouses: the western flower thrips Frankliniella occidentalis (Pergande)

(Thysanoptera: Thripidae) (FIGURE 0.4a); the two-spotted spider mite Tetranychus urticae

Koch (Acari: Tetranychidae) (FIGURE 0.4b) and the greenhouse whitefly, Trialeurodes vapo-

rariorum (Westwood) (Hemiptera: Aleyrodidae) (FIGURE 0.4b).

The two-spotted spider mite is known to attack over 1100 plant species (Migeon &

Dorkeld 2016). It causes leaf decoloration by feeding on clorophyll. Spider mites punc-

18


ture plant cells feeding on their contents (Helle & Sabelis 1985). Spider mite females pro-

duce silk webs, their populations can reach very high densities, killing the plant in a few

weeks (Brødsgaard & Albajes 1999). However, successful biological control of spider

mites can be achieved in many crops by using predatory mites such as Phytoseiulus

spp. (Bravenboer & Dosse 1962; Hussey & Bravenboer 1971). These predatory mites are

specialist predators of spider mites because they can penetrate the spider-mite web,

which other predators cannot (Sabelis & Bakker 1992).

The greenhouse whitefly attacks more than 850 plant species (http://www.cabi.org). It

feeds on the phloem of the plant, causing leaf deformation, discoloration and defoliation.

Owing to the honeydew production by whiteflies, different parts of the plant can become

fully covered by fungi growing on the honeydew, which blocks the light and reduces

photosynthesis. The greenhouse whitefly also causes plant damage by transmitting plant

viruses (Inbar & Gerling 2008). Successful biological control of this whitefly can be

achieved in different crops by releasing the parasitoids Encarsia formosa Gahan,

Eretmocerus mundus Mercet and Eretmocerus eremicus Rose & Zolnerowich and the

predator Macrolophus pygmaeus (Rambur), but recently, phytoseiid mites have been

used more frequently (Speyer

1927; van Lenteren & Woets

1988; Nomikou et al. 2001,

2002; Perdikis & Lykouressis

2002; Pijnakker et al. 2016b).

The western flower thrips

is another important pest,

feeding on plant leaves, flow-

ers, pollen or fruits of more

than 250 plant species

(http://www.cabi.org). It is the

main pest organism studied

in this thesis. They cause

damage to plant parts by

FIGURE 0.4. Main pest species in

greenhouse crops: (a) thrips pupae;

(b) adults of whitefly (left) and spi-

der mite (right). Photographs J. van

Arkel.

19


piercing epidermal and parenchymal cells, and sucking out their contents. They also

transmit viruses (Brødsgaard & Albajes 1999). Damaged plant parts develop characteris-

tic silvery spots or scars. In ornamentals, low densities of thrips can cause aesthetic

damage to flowers, the marketable part of roses, making them unsaleable (Brødsgaard &

Albajes 1999). Although thrips feed on above-ground plant parts, the pre-pupae and

pupae live in the litter or soil (Brødsgaard & Albajes 1999), where they are less affected

by chemical control. Biological control with predators is efficient in some crops, such as

vegetables (Gillespie 1989; Jacobson et al. 2001; Shipp & Wang 2003; Shipp &

Ramakers 2004). Western flower thrips are omnivourous: besides feeding on plant tis-

sue, larvae and adults consume spider mite eggs, predatory mite eggs and whitefly

crawlers (Trichilo & Leigh 1986; Faraji et al. 2001; Janssen et al. 2003; van Maanen et al.

2012). Due to all these special characteristics, its control is still a challenge in crops with

low damage thresholds, especially ornamentals.

Box 4: Alternative prey: Astigmatic mites

Astigmatic mites such as the mould mite Tyrophagus putrescentiae (Schrank) (FIGURE

0.5), the dried fruit mite Carpoglyphus lactis (L.) and the flour mite Acarus siro (L.) can be

used as alternative prey for predators. These species belong to the Acaridae family, they

are cosmopolitan, so they can be used in different crops and in different countries. They

are easy to rear and are used as food for predators in mass-rearing systems (Ramakers

& van Lieburg 1982).

These astigmatic species inhabit the soil, litter, and decaying vegetable material. Spe -

cies of the genus Tyrophagus commonly inhabit soils, old hay, mushrooms, house dust,

and bird nests (Walter et al.

1986; Mullen & OConnor

2009). They feed on yeast,

FIGURE 0.5. Tyrophagus putrescen-

tiae adults and eggs. These mites

can be used as alternative prey for

generalist predators. Photograph J.

van Arkel.

ReferencesAbrams PA, Matsuda H (1996) Positive indirect effects between prey species that share predators. Ecology

77:610-616Bale JS, van Lenteren JC, Bigler F (2008) Biological control and sustainable food production. Phil Trans R

Soc B 363:761-776Barbosa MFC, de Moraes GJ (2016) Potential of astigmatid mites (Acari: Astigmatina) as prey for rearing

edaphic predatory mites of the families Laelapidae and Rhodacaridae (Acari: Mesostigmata). Exp ApplAcarol 69:289-296

Birkhofer K, Wise DH, Scheu S (2008) Subsidy from the detrital food web, but not microhabitat complexity,affects the role of generalist predators in an aboveground herbivore food web. Oikos 117:494-500

Bolckmans KJF (1999) Commercial aspects of biological pest control in greenhouses. In: Integrated pest anddisease management in greenhouse crops. Albajes R, Lodovica Gullino M, van Lenteren J, Elad Y (eds.)Kluwer Academic Publishers, New York, USA, pp. 310-318

Bolckmans KJF, van Houten YM (2006) Mite composition, use thereof, method for rearing the phytoseiidpredatory mite Amblyseius swirskii, rearing system for rearing said phytoseiid mite and methods for bio-logical pest control on a crop. WO Patent WO/2006/057552

Bolckmans KJF, van Houten YM, van Baal AE, Timmer R, Morel DM (2013) Mite composition, carrier, methodfor rearing mites and uses related thereto. WO Patent WO/2013/103295

Branigan EJ (1915) Vedalia vrs. Icerya on pears. Calif State Comm Hort Mo Bull 4:107-108Bravenboer L, Dosse G (1962) Phytoseiulus riegeli Dosse as a predator of harmful mites of the group of

Tetranychus urticae. Entomol Exp Appl 5:291-304Brødsgaard HF, Albajes R (1999) Insects and mite pests, in integrated pest and disease management in

greenhouse crops. In Albajes R, Guillino ML, van Lenteren JC, Elad Y (eds.) Kluwer Academic Press, TheNetherlands, pp. 48-60

Buitenhuis R, Murphy G, Shipp L, Scott-Dupree C (2015) Amblyseius swirskii in greenhouse production sys-tems: a floricultural perspective. Exp Appl Acarol 65:451-464

20


algae, nematodes, and possibly also on injured and dead mites (Walter et al. 1986). The

mold mite T. putrescentiae (FIGURE 0.4b), is considered a pest of stored products with

high lipid and protein content, such as cheeses and grains. It is commonly used as food

source in rearing systems of different predatory mites of the family Phytoseiidae, and of

soil predators of different families, such as Laelapidae (Ramakers 1990; Bolckmans &

van Houten 2006; Fidgett & Stinson 2008; Pochubay & Grieshop 2012; Barbosa & de

Moraes 2016). Carpoglyphus lactis is considered a pest of saccharide-rich stored prod-

ucts such as dried fruits, wine, beer and honey (Hubert et al. 2014). This astigmatic mite

species is used to mass-rear phytoseiid predators such as Amblyseius swirskii (Athias-

Henriot) (Bolckmans & van Houten 2006; Fidgett & Stinson 2008). Acarus siro is also

considered a pest in stored products such as flour, wheat, and cheese (Solomon 1962).

This mite is used to mass-rear phytoseiid predators from the genus Amblyseius and

Neoseiulus, and soil predators from different families such as Macrochelidae and

Laelapidae (Ramakers & van Lieburg 1982; Bolckmans et al. 2013).

Buitenhuis R, Shipp L, Scott-Dupree C (2010) Dispersal of Amblyseius swirskii Athias-Henriot (Acari:Phytoseiidae) on potted greenhouse chrysanthemum. Biol Control 52:110–114

Cadogan BL, Laing JE (1977) A technique for rearing the predaceous mite Balaustium putmani (Acarina:Erythraeidae), with notes on its biology and life history. Can Entomol 109:1535-1544

Cakmak I, Janssen A, Sabelis MW, Baspinar H (2009) Biological control of an acarine pest by single and mul-tiple natural enemies. Biol Control 50:60-65

Calvo JF, Knapp M, van Houten YM, Hoogerbrugge H, Belda JE (2015) Amblyseius swirskii: What made thispredatory mite such a successful biocontrol agent? Exp App Acarol 65:419-433

Cardinale BJ, Harvey CT, Gross K, Ives AR (2003) Biodiversity and biocontrol: emergent impacts of a multi-enemy assemblage on pest suppression and crop yield in an agroecosystem. Ecol Lett 6:857-865

Casula P, Wilby A, Thomas MB (2006) Understanding biodiversity effects on prey in multi–enemy systems.Ecol Lett 9:995-1004

Chitturi A, Riley DG, Houston JP (2006) Effect of pine pollen on settling behavior of Frankliniella occidentalisand Frankliniella fusca (Thysanoptera: Thripidae) on tomato and peanut. Environ Entomol 35:1396-1406

Colfer RG, Rosenheim JA (2001) Predation on immature parasitoids and its impact on aphid suppression.Oecologia 126:292-304

Coll M, Guerson M (2002) Omnivory in terrestrial arthropods: Mixing plant and prey diets. Annu Rev Entomol47:267-297

Collyer E (1964) The effect of an alternative food supply on the relationship between two Typhlodromusspecies and Panonychus ulmi (Koch) (Acarina). Entomol Exp Appl 7:120-124

de Moraes GJ, Venancio R, dos Santos VLV, Paschoal AD (2015) Potential of Ascidae, Blattisociidae andMelicharidae (Acari: Mesostigmata) as biological control agents of pest organisms. In: Carrillo D, deMoraes GJ, Peña JE (eds.) Prospects for biological control of plant feeding mites and other harmfulorganisms. Springer International Publishing, Cham, Switzerland, pp. 77-102

DeBach P (1974) Biological control ecology. In DeBach P (ed.) Biological control by natural enemies.Cambridge University Press, London, pp. 47-88

Delisle JF, Shipp L, Brodeur J (2015) Apple pollen as a supplemental food source for the control of westernflower thrips by two predatory mites, Amblyseius swirskii and Neoseiulus cucumeris (Acari: Phytoseiidae),on potted chrysanthemum. Exp App Acarol 65:495-509

Dicke M, Sabelis MW (1988) How plants obtain predatory mites as bodyguards. Neth J Zool 38: 148-165Duarte MV, Venzon M, Bittencourt MCDS, Rodríguez-Cruz FA, Pallini A, Janssen A (2015) Alternative food

promotes broad mite control on chilli pepper plants. BioControl 60:817-825Evans EW (2008) Multitrophic interactions among plants, aphids, alternate prey and shared natural enemies–

a review. Eur J Entomol 105:369-380Faraji F, Janssen A, Sabelis MW (2001) Predatory mites avoid ovipositing near counterattacking prey. Exp

Appl Acarol 25:613-623Federici BA (1999) A perspective on pathogens as biological control agents for insect pests. In Bellows TS,

Fisher TW (eds.) Handbook of biological control. Academic Press, San Diego, USA, pp. 517-548Ferguson KI, Stiling P (1996) Non additive effects of multiple natural enemies on aphid populations.

Oecologia 108:375-379Fidgett MJ, Stinson CSA (2008) Method for rearing predatory mites. WO Patent WO/2008/015393Finke DL, Denno RF (2004) Predator diversity dampens trophic cascades. Nature 429:407-410Fuentes Quintero LS, Muñoz-Cárdenas K, Combita O, Jimeno E, Getiva de la Hoz JC, Cantor F, Rodriguez

D, Makol J (2014) A re-description of Balaustium leanderi Comb. Nov. (Actinotrichida, Erythraeidae) withfirst report on characteristics of all active instars and taxonomic notes on the genus. The Fla Entomol97:937-951

Furtado G, de Moraes GJ (2015) The potential of free-living Laelapid mites (Mesostigmata: Laelapidae) as bio-logical control agents. In: Carrillo D, de Moraes GJ, Peña JE (eds.) Prospects for biological control of plantfeeding mites and other harmful organisms. Springer International Publishing, Switzerland, pp. 33-75

Gerson U, Weintraub PG (2007) Mites for the control of pests in protected cultivation. Pest Mgt Sci 63:658-676

21


Getiva JC, Acosta A (2004) Taxonomía de ácaros asociados a cultivos de flores. Asocolflores 66:59-68Gillespie DR (1989) Biological control of thrips (Thysanoptera: Thripidae) on greenhouse cucumber by

Amblyseius cucumeris. Entomophaga 34:185-192Gillespie DR, Quiring DMJ (1990) Biological control of fungus gnats, Bradysia spp. (Diptera: Sciaridae), and

western flower thrips Frankliniella occidentalis (Pergande) (Thysanoptera: Thripidae), in greenhousesusing a soil-dwelling predatory mite, Geolaelaps sp. nr. aculeifer (Canestrini) (Acari: Laelapidae). Can Ent122:975-983

Godfray HCJ (1994) Parasitoids: behavioral and evolutionary ecology. Monographs in Behavior and Ecology.Princeton University Press, Princeton, New Jersey, USA

Halaj J, Wise DH (2002) Impact of a detrital subsidy on trophic cascades in a terrestrial grazing food web.Ecology 83:3141-3151

Helle W, Sabelis MW (1985) Spider mites, their biology, natural enemies and control, World crop pests. Vol1B. Elsevier, Amsterdam, The Netherlands

Hessein NA, Perring TM (1988) The importance of alternative foods for the mite Homeopronematus anconai(Acari: Tydeidae). Ann Entomol Soc Am 81:488-492

Hokkanen H, Pimentel D (1984) New approach for selecting biological-control agents. Can Entomol 116:1109-1121

Holt RD (1977) Predation, apparent competition, and the structure of prey communities. Theor Popul Biol12:197-229

Hoogerbrugge H, Van Houten Y, Van Baal E, Bolckmans K (2008) Alternative food sources to enable establis-hment of Amblyseius swirskii (Athias-Henriot) on chrysanthemum without pest presence IOBC/WPRS Bull32:79-82

Hougen-Eitzman D, Karban R (1995) Mechanisms of interspecific competition that result in successful con-trol of Pacific mites following inoculations of Willamette mites on grapevines. Oecologia 103:157-161

Hoy MA (1994) Parasitoids and predators in management of arthropod pests. In: Metcalf RL, Luckmann WH(eds). Introduction to insect pest management. John Wiley & Sons, New York, pp. 129-198

Hubert J, Nesvorna M, Kopeck J, Sagova-Mareckova M, Poltronieri P (2014) Carpoglyphus lactis (Acari:Astigmata) from various dried fruits differed in associated micro-organisms. J Appl Microb 118:470-484

Huffaker CB, Messenger PS (1976) Theory and practice of biological control. Academic Press, New York, USAHussey NW, Bravenboer L (1971) Control of pests in glasshouse culture by the introduction of natural ene-

mies. In: Huffaker CB (ed) Biological Control. Plenum Press, New York, USA, pp. 195-216Inbar M, Doostdar H, Leibee GL, Mayer RT (1999) The role of plant rapidly induced responses in asymmetric

interspecific interactions among insect herbivores. J Chem Ecol 25:1961-1979Inbar M, Gerling D (2008) Plant-mediated interactions between whiteflies, herbivores, and natural enemies.

Annu Rev Entomol 53:431-448Jacobson RJ, Croft P, Fenlon J (2001) Suppressing establishment of Frankliniella occidentalis Pergande

(Thysanoptera: Thripidae) in cucumber crops by prophylactic release of Amblyseius cucumerisOudemans (Acarina: Phytoseiidae). Biocontrol Sci Technol 11:27-34

Janssen A, Montserrat M, HilleRisLambers R, de Roos AM, Pallini A, Sabelis MW (2006) Intraguild predationusually does not disrupt biological control, in: Brodeur J, Boivin G (eds.), Trophic and Guild Interactionsin Biological Control, Progress in Biological Control. Springer, Dordrecht, The Netherlands, pp. 21-44

Janssen A, Pallini A, Venzon M, Sabelis MW (1998) Behaviour and indirect interactions in food webs of plant-inhabiting arthropods. Exp Appl Acarol 22:497-521

Janssen A, Sabelis MW (2015) Alternative food and biological control by generalist predatory mites: the caseof Amblyseius swirskii. Exp Appl Acarol 65:413-418

Janssen A, Sabelis MW, Magalhães S, Montserrat M, van der Hammen T (2007) Habitat structure affectsintraguild predation. Ecology 88:2713-2719

Janssen A, Willemse E, van der Hammen T (2003) Poor host plant quality causes omnivore to consume pred-ator eggs. J Anim Ecol 72:478-483

Karban R, Baldwin IT (1997) Induced responses to herbivory. University of Chicago Press, Chicago, USAKarban R, Carey JR (1984) Induced resistance of cotton seedlings to mites. Science 225:53-54

22


Karban R, Hougen-Eitzman D, English-Loeb G (1994) Predator-mediated apparent competition between her-bivores that feed on grapevines. Oecologia 97:508-511

Kessler A, Baldwin IT (2002) Plant responses to insect herbivory: The emerging molecular analysis. Ann RevPlant Biol 53:299-328

Kessler A, Halitschke R (2007) Specificity and complexity: the impact of herbivore-induced plant responseson arthropod community structure. Curr Opin Plant Biol 10:409-414

Lefcheck JS, Whalen MA, Davenport TM, Stone JP, Duffy JE (2013) Physiological effects of diet mixing onconsumer fitness: a meta-analysis. Ecology 94:565-572

Leman A, Messelink GJ (2015) Supplemental food that supports both predator and pest: A risk for biologicalcontrol? Exp Appl Acarol 65:511-524

Liu C, Yan L, Li H, Wang G (2006) Effects of predator-mediated apparent competition on the populationdynamics of Tetranychus urticae on apples. BioControl 51:453-463

Losey JE, Denno RF (1998) Positive predator–predator interactions: enhanced predation rates and synergis-tic suppression of aphid populations. Ecology 79:2143-2152

Losey JE, Denno RF (1999) Factors facilitating synergistic predation: the central role of synchrony. Ecol Appl9:378-386

Lundgren JG (2009) Relationships of natural enemies and non-prey foods. Springer Publishers, Heidelberg,The Netherlands, pp. 127-139

Magalhães S, van Rijn PCJ, Montserrat M, Pallini A, Sabelis MW (2007) Population dynamics of thrips preyand their mite predators in a refuge. Oecologia 150:557-568

Makol J, Arijs Y, Wäckers F (2012) A new species of Balaustium Von Heyden, 1826 (Acari: Actinotrichida,Erythraeidae) from Spain. Zootaxa 3178:1-21

Manners AG, Dembowski BR, Healey MA (2013) Biological control of western flower thrips, Frankliniella occi-dentalis (Pergande) (Thysanoptera: Thripidae), in gerberas, chrysanthemums and roses. Aust J Entomol52:246-258

Marques RVR, Sarmento A, Lemos F, Pedro-Neto M, Sabelis MW, Venzon M, Pallini A, Janssen A (2015)Active prey mixing as an explanation for polyphagy in predatory arthropods: synergistic dietary effects onegg production despite a behavioural cost. Func Ecol 29:1317-1324

Messelink G, van Holstein-Saj R (2008) Improving thrips control by the soil-dwelling predatory miteMacrocheles robustulus (Berlese). IOBC/WPRS Bull 32:135-138

Messelink GJ, van Maanen R, van Steenpaal SEF, Janssen A (2008) Biological control of thrips and whitefliesby a shared predator: Two pests are better than one. Biol Control 44:372-379

Migeon A, Dorkeld F (2016) Spider Mites Web: a comprehensive database for the Tetranychidae.http://www.montpellier.inra.fr/CBGP/spmweb

Minks AK, Harrewijn P (1989) Aphids, their biology, natural enemies and control, World crop pests. Vol. 2B.Elsevier, Amsterdam, The Netherlands

Miyashita T, Takada M, Shimazaki A (2003) Experimental evidence that aboveground predators are sustainedby underground detritivores. Oikos 103:31-36

Moore JC, McCann KS, Setälä H, de Ruiter PC (2003) Top-down is bottom-up: Does predation in the rhizo -sphere regulate aboveground dynamics? Ecology 84:846-857

Mullen GR, OConnor BM (2009) Mites. In: Mullen GR, Durden L (eds.) Medical and Veterinary Entomology(2nd ed), Elsevier, Amsterdam, The Netherlands, pp. 433-482

Müller CB, Godfray HCJ (1997) Apparent competition between two aphid species. J Anim Ecol 66:57-64Muñoz-Cárdenas K, Fuentes L, Rueda-Ramirez D, Rodríguez D, Cantor F (2015) The Erythraeoidea

(Trombidiformes: Prostigmata) as biological control agents, with special reference to the genusBalaustium. In: Carrillo D, de Moraes G, Peña J (eds.) Prospects for biological control of plant feedingmites and other harmful organisms. Springer International Publishing, Cham, Switzerland, pp. 207-239

Murdoch WW, Chesson J, Chesson PL (1985) Biological control in theory and practice. Am Nat 125:344-66Nombela G, Garzo E, Duque M, Muñiz M (2009) Preinfestations of tomato plants by whiteflies (Bemisia tabaci)

or aphids (Macrosiphum euphorbiae) induce variable resistance or susceptibility responses. Bull EntomolRes 99:182-191

23


Nomikou M, Janssen A, Schraag R, Sabelis MW (2001) Phytoseiid predators as potential biological controlagents for Bemisia tabaci. Exp Appl Acarol 25:271-291

Nomikou M, Janssen A, Schraag R, Sabelis MW (2002) Phytoseiid predators suppress populations of Bemisiatabaci on cucumber plants with alternative food. Exp Appl Acarol 27:57-68

Nomikou M, Sabelis MW, Janssen A (2010) Pollen subsidies promote whitefly control through the numericalresponse of predatory mites. BioControl 55:253-26

Oelbermann K, Scheu S (2002) Effects of prey type and mixed diets on survival, growth and development ofa generalist predator, Pardosa lugubris (Araneae: Lycosidae). Basic Appl Ecol 3:285-291

Ohkawa K, Suematsu M (1999) Arching cultivation techniques for growing cut-roses. Acta Hort 482:47-51Pallini A, Janssen A, Sabelis MW (1998) Predators induce interspecific herbivore competition for food in

refuge space. Ecol Lett 1:171-177Parrella MP, Stengard L, van Lenteren J (1999) Glasshouse environments. In: Bellows TS, Fisher TW Jr. (eds.)

Handbook of Biological Control: Principles and Applications. Academic Press, San Diego, USA, pp. 819-839Perdikis DC, Lykouressis DP (2002) Life table and biological characteristics of Macrolophus pygmaeus when

feeding on Myzus persicae and Trialeurodes vaporariorum. Entomol Exp Appl 102:261-272Pijnakker J, Arijs Y, de Souza A, Wäckers F (2016a) The use of Typha angustifolia (cattail) pollen to establish

the predatory mites Amblyseius swirskii, Iphiseius degenerans, Euseius ovalis and Euseius gallicus inglasshouse crop. IOBC/WPRS Bull 120:47-52

Pijnakker J, Arijs Y, de Souza A, Wäckers F (2016b) The polyphagous mite Euseius gallicus (Kreiter & Tixier):A new predator able to persist in glasshouse roses. IOBC/WPRS Bull 120:45-46

Pijnakker J, Ramakers P (2008) Predatory mites for biocontrol of Western Flower Thrips, Frankliniella occi-dentalis (Pergande), in cut roses. IOBC/WPRS Bull 32:171-174

Pochubay EA, Grieshop MJ (2012). Intraguild predation of Neoseiulus cucumeris by Stratiolaelaps miles andAtheta coriaria in greenhouse open rearing systems. Biol Control 63: 195-200

Polis GA, Holt RD (1992) Intraguild predation: the dynamics of complex trophic interactions. TREE 7:151-54Polis GA, Strong DR (1996) Food web complexity and community dynamics. Am Nat 147:813-846.Pozzebon A, Boaria A, Duso C (2015) Single and combined releases of biological control agents against

canopy- and soil-dwelling stages of Frankliniella occidentalis in cyclamen. BioControl 60:341-350Prasad RP, Snyder WE (2004) Predator interference limits fly egg biological control by a guild of groundac-

tive beetles. Biol Control 31:428-437Ramakers PMJ (1990). Manipulation of phytoseiid thrips predators in the absence of thrips.IOBC/WPRS Bull

13:169-172Ramakers PMJ, van Lieburg MJ (1982) Start of commercial production and introduction of Amblyseius

mckenziei Sch. & Pr. (Acarina: Phytoseiidae) for the control of Thrips tabaci Lind. (Thysanoptera:Thripidae) in glasshouses. Mededelingen van de Faculteit der Landbouwwetenschappen van deRijksuniversiteit te Gent 47:541-545

Rosenheim JA (1998) Higher-order predators and the regulation of insect herbivore populations. Annu RevEntomol 43:421-447

Rosenheim JA, Harmon JP (2006) The influence of intraguild predation on the suppression of a shared preypopulation: an emperical reassessment. In: Brodeur J, Boivin G (eds.) Trophic and guild interactions inbiological control, Springer, Dordrecht, The Netherlands, pp. 1-20

Rosenheim JA, Kaya HK, Ehler LE, Marois JJ, Jaffee BA (1995) Intraguild predation among biological controlagents: theory and evidence. Biol Control 5:303-335

Sabelis MW (1992) Arthropod predators. In: Crawley MJ (ed.) Natural enemies, the population biology ofpredators, parasites and diseases, Blackwell, Oxford, England, pp. 225-264

Sabelis MW, Bakker FM (1992) How predatory mites cope with the web of their Tetranychid prey – a func-tional view on dorsal chaetotaxy in the Phytoseiidae. Exp Appl Acarol 16:203-225

Sabelis MW, Janssen A, Bruin J, Bakker FM, Drukker B, Scutareanu P, van Rijn PCJ (1999) Interactionsbetween arthropod predators and plants: A conspiracy against herbivorous arthropods? In: Bruin J, vander Geest LPS, Sabelis MW (eds.) Ecology and Evolution of the Acari. Kluwer Academic Publishers,Dordrecht, The Netherlands, pp. 207-229

24


Sarmento RA, Lemos F, Bleeker PM, Schuurink RC, Pallini A, Oliveira MGA, Lima ER, Kant M, Sabelis MW,Janssen A (2011a). A herbivore that manipulates plant defence. Ecol Lett 14:229-236

Sarmento RA, Lemos F, Dias CR, Kikuchi WT, Rodrigues JCP, Pallini A, Sabelis MW, Janssen A (2011b) A her-bivorous mite down-regulates plant defence and produces web to exclude competitors. PLoS ONE 6:1-7

Scheu S (2001) Plants and generalist predators as links between the below-ground and above-ground sys-tem. Basic Appl Ecol 2:3-13

Seastedt TR (1984) The role of microarthropods in decomposition and mineralization processes. Annu RevEntomol 29:25-46

Settle WH, Ariawan H, Astuti EI, Cahyana W, Hakim A, Hindayana D, Lestari AS (1996) Managing tropical ricerests through conservation of generalist natural enemies and alternative prey. Ecology 77:1975-1988

Shiojiri K, Takabayashi J, Yano S, Takafuji A (2002) Oviposition preference by herbivores is affected bytritrophic interaction networks. Ecol Lett 5:1-7

Shipp JL, Ramakers PMJ (2004) Biological control of thrips on vegetable crops. In: Heinz KM, Van DriescheRG, Parrella P (eds.) Biocontrol in Protected Culture. Ball Publishing, Batavia, USA, pp. 265-276

Shipp JL, Wang K (2003) Evaluation of Amblyseius cucumeris (Acari: Phytoseiidae) and Orius insidiosus(Hemiptera: Anthocoridae) for control of Frankliniella occidentalis (Thysanoptera: Thripidae) on green-house tomatoes. Biol Control 28:271-281

Sih A, Crowley P, McPeek M, Petranka J, Strohmeier K (1985) Predation, competition, and prey communities:a review of field experiments. Ann Rev Ecol Syst 16:269-311

Sih A, Englund G, Wooster D (1998) Emergent impacts of multiple predators on prey. TREE 13: 350-355Snyder WE, Ives AR (2001) Generalist predators disrupt biological control by a specialist parasitoid. Ecology

82:705-716Snyder WE, Ives AR (2003) Interactions between specialist and generalist natural enemies: Parasitoids, pred-

ators, and pea aphid biocontrol. Ecology 84:91-107Snyder WE, Wise DH (2002) Predator interference and the establishment of generalist predator populations

for biocontrol. Biol Control 15: 283-292Sokol-Hessner L, Schmitz OJ (2002) Aggregate effects of multiple predator species on a shared prey. Ecology

83:2367-2372Solomon ME (1962) Ecology of the flour mite, Acarus siro L. (=Tyropglyphus farinae DeG.). Ann Appl Biol

50:178-184Speyer ER (1927) An important parasite of the greenhouse white-fly (Trialeurodes vaporariorum Westwood).

Bull Entomol Res 17:301-8Swirski E, Amitai S, Dorzia N (1967) Laboratory studies on the feeding, development and oviposition of the

predaceous mites Amblyseius rubini Swirski and Amitai and Amblyseius swirski Athias-Henriot (Acarina:Phytoseiidae) on various kinds of food substances. Israel J Agric Res 17:101-119

Symondson WOC, Sutherland KD, Greenstone MH (2002) Can generalist predators be effective biocontrolagents? Annu Rev Entomol 47:561-594

Thoeming G, Poehling HM (2006) Integrating soil-applied azadirachtin with Amblyseius cucumeris (Acari:Phytoseiidae) and Hypoaspis aculeifer (Acari: Laelapidae) for the management of Frankliniella occidental-is (Thysanoptera: Thripidae). Environ Entomol 35:746-756

Toft S, Wise DH (1999) Growth, development, and survival of a generalist predator fed single- and mixedspecies diets of different quality. Oecologia 119:191-197

Trichilo PJ, Leigh TF (1986) Predation on spider mite eggs by the western flower thrips, Frankliniella occiden-talis (Thysanoptera: Thripidae), an opportunist in a cotton agroecosystem. Environ Entomol 15:821-825

Turlings TCJ, Tumlinson JH, Lewis WJ (1990) Exploitation of herbivore-induced plant odors by host-seekingparasitic wasps. Science 250:1251–1253

Uetz GW, Bischo J, Raver J (1992) Survivorship of wolf spiders (Lycosidae) reared on different diets. JArachnol 20:207-211

van der Putten WH, Bardgett RD, de Ruiter PC, Hol WHG, Meyer KM, Bezemer TM, Bradford MA, ChristensenS, Eppinga MB, Fukami T, Hemerik L, Molofsky J, Schädler M, Scherber C, Strauss SY, Vos M, Wardle DA(2009) Empirical and theoretical challenges in aboveground–belowground ecology. Oecologia 161:1-14

25


van der Putten WH, Vet LEM, Harvey JA, Wäckers FL (2001) Linking above- and belowground multitrophicinteractions of plants, herbivores, pathogens, and their antagonists. Trends Ecol Evol 16:547-554

Van Driesche RG, Bellows Jr TS (1996) Biological Control. Chapman and Hall, New York, USA, pp. 7-19van Lenteren JC, Woets J (1988) Biological and integrated pest control in greenhouses. Ann Rev Entomol

33:239-269van Lenteren JC, Babendreier D, Bigler F, Burgio G, Hokkanen HMT, Kuske S, Loomans AJM, Menzler-

Hokkanen I, van Rijn PCJ, Thomas MB, Tommasini MG, Zeng QQ (2003) Environmental risk assessmentof exotic natural enemies used in inundative biological control. BioControl 48:3-38

van Maanen R, Broufas G, Oveja MF, Sabelis MW, Janssen A (2012) Intraguild predation among plant pests:western flower thrips larvae feed on whitefly crawlers. BioControl 57:533-539

van Maanen R, Vila E, Sabelis MW, Janssen A (2010) Biological control of broad mites (Polyphagotarsonemuslatus) with the generalist predator Amblyseius swirskii. Exp Appl Acarol 52:29-34

van Rijn PCJ, van Houten YM, Sabelis MW (1999) Pollen improves thrips control with predatory mites.IOBC/WPRS Bull 22:209-212

van Rijn PCJ, van Houten YM, Sabelis MW (2002) How plants benefit from providing food to predators evenwhen it is also edible to herbivores. Ecology 83:2664-2679

Vance-Chalcraft HD, Rosenheim JA, Vonesh JR, Osenberg CW, Sih A (2007) The influence of intraguild pre-dation on prey suppression and prey release: A meta-analysis. Ecology 88:2689-2696

Vangansbeke D, Nguyen DT, Audenaert J, Verhoeven R, Gobin B, Tirry L, De Clercq P (2014) Performance ofthe predatory mite Amblydromalus limonicus on factitious foods. BioControl 59:67-77

Vangansbeke D, Nguyen DT, Audenaert J, Verhoeven R, Gobin B, Tirry L, De Clercq P (2016) Supplementalfood for Amblyseius swirskii in the control of thrips: feeding friend or foe? Pest Manag Sci 72:466-473

Venzon M, Janssen A, Sabelis MW (2001) Prey preference, intraguild predation and population dynamics ofan arthropod food web on plants. Exp App Acarol 25:785-808

Von Berg K, Thies C, Tscharntke T, Scheu S (2009) Cereal aphid control by generalist predators in presenceof belowground alternative prey: complementary predation as affected by prey density. Pedobiologia53:41-48

Wahab S (2010) Integrated pest management strategies for sustainability of agriculture. In: Sharma AK,Wahab S, Srivastava R (eds.) Agriculture Diversification: Problems and Perspectives. I.K. InternationalPublishing House Pvt. Ltd. New Delhi, India, pp. 29-43

Walter DE, Hudgens RA, Freckman DW (1986) Consumption of nematodes by fungivorous mites, Tyrophagusspp. (Acarina: Astigmata: Acaridae). Oecología 70:357-361

Wardle DA, Bardgett RD, Klironomos JN, Setala H, van der Putten WH, Wall DH (2004) Ecological linkagesbetween aboveground and belowground biota. Science 304:1625-1633

Wiethoff J, Poehling HM, Meyhöfer R (2004) Combining plant- and soil-dwelling predatory mites to optimizebiological control of thrips. Exp Appl Acarol 34:239-261

Zhang LP, Zhang GY, Zhang YJ, Zhang WJ, Liu Z (2005) Interspecific interactions between Bemisia tabaci(Hem., Aleyrodidae) and Liriomyza sativae (Dipt., Agromyzidae). J Appl Entomol 129:443-446

26


Generalist red velvet mite predator(Balaustium leanderi) performs better ona mixed diet

Karen Muñoz-Cárdenas, Luz Stella Fuentes, R. Fernando Cantor, C. Daniel Rodríguez,

Arne Janssen & Maurice W. Sabelis

Generalist predators have the potential advantage to control more than one pest

and to be more persistent than specialist predators because they can survive on

different foods. Moreover, their population growth rate may be elevated when

offered a mixture of prey species. We studied a generalist predatory mite

Balaustium leanderi that shows promise for biological control of thrips and white-

flies in protected rose cultures in Colombia. Although starting its life in the soil, this

predator makes excursions onto plants where it feeds on various arthropods. We

quantified life history parameters of the predator, offering high densities of three

pest species: first-instar larvae of Frankliniella occidentalis, eggs of Trialeurodes

vaporariorum and Tetranychus urticae, either alone or in combination. The preda-

tors completed their life cycle on each diet. The egg-to-egg period was c. 2

months. All eggs were laid in one batch in 1-2 days, indicating a pronounced

semelparous reproduction pattern. In general, females reproduced earlier and laid

more eggs on mixed diets, and these early reproducers consequently had higher

population growth rates than late reproducers. The best diet in terms of egg-to-

egg period and juvenile survival was the combination of eggs from whiteflies and

spider mites. Spider mite eggs alone and western flower thrips larvae alone were

the worst diets. It remains to be investigated whether mixed diets promote the

population growth rate of Balaustium leanderi sufficiently for biocontrol of whiteflies

and thrips in the presence of alternative prey, such as spider mites, to become

effective.

Experimental and Applied Acarology (2014) 62: 19-32

29

1

IntroductionWhether to use generalist or specialist predators for biological control of crop pests isan important and hotly debated topic (Symondson et al. 2002). Most attention hasbeen paid to specialist biocontrol agents, because their dynamics are tightly linked tothat of the prey and because of the lower risk of side effects on populations of non-target organisms (Murdoch et al. 1984, 1985; Hassell & May 1986; Murdoch 1994).However, generalist predators are increasingly used for biocontrol (Chiverton 1986;Rosenheim et al. 1993; Settle et al. 1996; Chang & Kareiva 1999; Symondson et al.2002; Messelink et al. 2012). This also applies to generalist phytoseiid mites that liveon plants and feed on various arthropod herbivores (Nomikou et al. 2001; Messelinket al. 2008, 2010, 2012), as well as on plant food such as pollen (Nomikou et al. 2002;van Rijn et al. 2002) and on plant fungi, such as spores of mildew (Pozzebon & Duso2008). Compared to specialist natural enemies, the advantage of using generalistpredators is that they can feed on other prey than the target pest, thereby enablingpersistence of their populations at high densities, even in periods of low pest density.A disadvantage of using generalist predators is that they may also attack non-targetprey or even other predators (intraguild predation or higher-order predation;Rosenheim et al. 1995; Rosenheim 1998). Although this could potentially disrupt bio-logical control, this is often not the case for intraguild predation (Janssen et al. 2006),but can occur in the case of higher-order predation (Messelink et al. 2011).

Generalist predators may be present on the plants, but they may also live in thesoil, yet make foraging excursions onto the plants. In this case, their population sizedepends on the community of arthropod decomposers and predators in the soil orlitter (Settle et al. 1996; Scheu 2001), as well as on the community of arthropod her-bivores on the plant canopy. Some species in the genus Balaustium (Acari:Erythraeidae) are good examples of species that occur in the litter and on plants(Putman 1969; Childers & Rock 1981; Welbourn 1983; Welbourn & Jennings 1991).Moreover, these species prey on a range of species on plants; they have beenobserved to feed on fruit-tree red spider mites, Panonychus ulmi (Acari:Tetranychidae) (Putman 1969; Cadogan & Laing 1977), eggs and larvae ofLepidoptera and Diptera, aphids and pollen (Hayes 1985). Muñoz et al. (2009)observed that an undescribed species of Balaustium (C. Welbourn, pers. comm.2008) feeds preferentially on immatures of three plant pests: two-spotted spidermites Tetranychus urticae (Acari: Tetranychidae), western flower thrips Frankliniellaoccidentalis (Thysanoptera: Thripidae) and greenhouse whiteflies, Trialeurodes vapo-rariorum (Hemiptera: Aleyrodidae). The same species was used in this study. Thesepredatory mites are found in the vegetation outside greenhouses of flower cultureson the Bogotá Plateau in Colombia (Torrado et al. 2001) in cold as well as temperatezones (Getiva & Acosta 2004). Provided pesticides are not applied, they can be

30

CHAPTER 1 | GENERALIST PREDATOR PERFORMS BETTER ON A MIXED DIET

abundant and they are considered as candidate predators for control of several pestspecies (Muñoz et al. 2009).

In this article, we assess the potential of Balaustium leanderi (Fuentes et al. 2014)to feed and reproduce on a diet of spider mites, thrips and whiteflies. We measuredprey mortality when exposed to mobile life stages of this predator and we deter-mined the life table parameters on a diet of each of these pest species as well as allpossible mixtures. The mixtures were offered because several generalist predatorymites have been shown to reproduce better on mixtures of prey species than on eachprey species alone (Oelbermann & Scheu 2002; Messelink et al. 2008), which furtherincreases their capacity for pest control (Messelink et al. 2008, 2010). This researchmay pave the way for future research on the efficacy of B. leanderi as a biocontrolagent and on the role of soil-inhabiting prey (e.g. fungivorous mites) to boost theirpopulations and thereby improve pest control.

Materials and methodsPredator and prey culturesPredators and prey were reared under local greenhouse conditions with a daily max-imum and minimum temperature of 31.7 ± 3.5 and 11.1 ± 2 °C (mean ± s.d.). Thewhitefly T. vaporariorum was reared on tomato plants (Solanum lycopersicum),whereas the thrips F. occidentalis was reared on Pelargonium grandiflorum flowers inplastic containers (30 × 16 × 12 cm) with moist paper towels on the bottom to pro-mote humidity. The spider mite T. urticae was reared on bean plants. The tomato andbean plants were 2 weeks old at infestation; P. grandiflorum flowers were boughtweekly from a commercial producer, which sprayed minimal levels of pesticides. Thecultures of whiteflies and spider mites were kept in separate compartments in agreenhouse inside cages to avoid contamination.

Predator rearing units consisted of closed plastic containers (18 cm diameter, 20 cmheight) with an opening (10 cm diameter) covered with a mite-proof steel mesh for ven-tilation. To provide a suitable substrate for juvenile development, the bottom of the con-tainer was covered with a layer of moist peat (c. 3 cm deep). A disc of paper towel (17cm diameter) was placed on top of the layer to maintain moisture. A total of 15 adultpredator individuals were placed in the container and they were provided with plantmaterial infested with all stages of whiteflies, thrips and spider mites twice per week.

To obtain a cohort of B. leanderi eggs, we offered individual young adult females amixture of the three prey species (Muñoz et al. 2009). To this end, B. leanderi adultswere placed individually in an arena consisting of a rose leaflet on top of an invertedPetri dish (3 cm diameter), which was used as a base for the leaflet; the petiole wasinserted in a water-filled plastic tube (6 cm long, 2 cm diameter) to provide water to theleaflet. Mites were kept on the leaflet inside a ring consisting of an opaque PVC tube

31

GENERALIST PREDATOR PERFORMS BETTER ON A MIXED DIET | CHAPTER 1

(1 cm high, 2 cm diameter), with a hole in its wall, closed by a piece of cork underneathwhich the mites preferred to lay eggs. The ring was closed with plastic wrap (Vinipel)(see SUPPLEMENTARY MATERIAL 1.1). After 7-10 days, they had laid at least 50 eggs, theseegg batches were transferred to a Petri dish with wet cotton wool (Cadogan & Laing1977). As soon as the pre-larvae had hatched from the eggs, they were transferredeach to a separate arena (the same as described above but now without a hole). In thisway, batches of 50 individuals at the onset of the pre-larval stage were obtained. Fordetails of culture methods please see SUPPLEMENTARY MATERIAL 1.1.

Developmental stages, survival and reproductionPreliminary observations (K. Muñoz-Cárdenas, pers. obs., 2008) showed that B.lean deri reproduces by thelytoky, as is the case in a related species (B. murorum,Halliday 2005; B. nr. putmani, Hedges et al. 2012). We therefore studied the life cycleof B. leanderi with individuals kept singly in an arena (as described above) through-out their life span.

There were seven treatments, three of which involved a daily supply with one ofthe three prey species, offered in sufficient numbers to prevent prey depletion: 80spider mite eggs, 80 whitefly eggs or 20 first-instar thrips. Predator larvae wereoffered half these amounts because of their lower prey requirements. Each treatmentwas replicated with c. 50 B. leanderi individuals, each in a separate experimentalunit. We also carried out four treatments involving the following prey combinationsbased on a replacement design (i.e. half or a third of the amounts of prey offered inthe monocultures): (1) 40 eggs of T. urticae and of T. vaporariorum, (2) 40 eggs of T.vaporariorum and 10 larvae of F. occidentalis, (3) 40 eggs of T. urticae, 10 larvae of F.occidentalis, and (4) 30 eggs of T. urticae, 30 eggs of T. vaporariorum and six larvaeof F. occidentalis. Each of these treatments was replicated with 15 B. leanderi indi-viduals. Predator larvae again received half the amount of prey. Prey mortality wasmeasured to verify that enough prey was offered to the predatory mites (seeSUPPLEMENTARY MATERIAL 1.2 for results).

Spider mite eggs and thrips larvae were replaced daily with fresh prey from the cul-tures, using a fine paintbrush. Thus, spider-mite eggs were offered free of the silkenweb produced by T. urticae. Whitefly eggs were replaced once per week as follows.Three whitefly females were allowed to lay eggs on rose leaflets in a clip cage in thepresence of males for a period of 3 days. The leaflets with eggs were subsequentlyused as arenas for predatory mites. Excesses of whitefly eggs and honeydew wereremoved with a small wet cotton swab. We registered the number of days of everystage, the day at which every B. leanderi female laid eggs and the day they died.

Effects of diet on life span and pre-oviposition period were analysed using gener-alized linear models (GLM) with a Poisson error distribution. For fecundity, a quasi-

32


33


Poisson error distribution was used to correct for overdispersion. The effect of dieton survivorship was analyzed using a Cox proportional hazards model (Hosmer &Lemeshow 1999; Crawley 2007). To test the effect of diet on the percentage of non-reproductive predatory mites, we performed a GLM with a quasi-binomial error dis-tribution to correct for overdispersion. Treatments were compared using the mult-comp package (Hothorn et al. 2008).

Life table parametersAs shown below, the reproduction of B. leanderi represents a special type of semel-parity (Stearns 1976, 1992; Roff 1992), appropriately referred to as ‘pronounced’semelparity (Hautekeete et al. 2001): all eggs are deposited within a very short timespan (also called ‘big-bang reproduction’; Diamond 1982; Zeineddine & Jansen2009). Under big-bang reproduction and thelytoky, the life history can be summa-rized by the following variables: (1) the egg-to-egg developmental time, which thenequals the generation time T; (2) the number of eggs produced per reproductivefemale at age T (mT); (3) the survival from egg deposition until reproducing adult (lT).The latter variable is the product of the survival until adulthood (lA) and the propor-tion of adults that reproduces (sT). Based on these life table parameters, we calcu-lated the net reproduction rate (R0 = lTmT) and the intrinsic rate of increase (rm), whichfor the special case of big-bang reproduction equals (Carey 1993):

rm = ln(R0)/(T + 1). (1)

Under big bang reproduction, the variables T and mT can be quantified for eachindividual separately. Such an individual-based assessment is obviously not possiblefor the survival probability until becoming a reproducing adult (lT). We therefore pro-visionally used population estimates of lT when estimating the rm for each individual,which enabled us to estimate the mean and the variance of rm.

To explore trade-offs between the net reproduction and the egg-to-egg period, wecarried out a regression analysis of individual R0 against individual T and individualrm against individual T. To help exploring how the trade-off between individual R0 andindividual T influences the relation between individual rm and individual T, we plottedrm based on the overall mean of R0 and T in relation to T (necessarily leading to ahyperbolic relation because R0 is constant).

The effect of diet on intrinsic rate of population increase was analysed with a GLMwith a Gaussian error distribution. To detect trade-offs and possibly other trends inthe intrinsic rate of population increase and the net reproduction, these parameterswere plotted against the egg-to-egg developmental time. All the statistical tests weredone using the software R (2.15.1).

ResultsDevelopmental stagesWe observed seven developmental stages (see SUPPLEMENTARY MATERIAL 1.1), compa-rable to the related species B. putmani (Putman 1969; Cadogan & Laing 1977) andB. hernandezi (Makol et al. 2012): (1) the spherical egg (c. 0.16 mm diameter), initial-ly red but turning dark red later; (2) the red oval prelarva with an orange band (c. 0.2mm wide, c. 0.3 mm long); (3) the orange-red, six-legged larva; (4) the scarlet, ovaland quiescent (legless and sessile) protonymph; (5) the scarlet, oval and eight-legged

34


FIGURE 1.1. The effect of single prey species and mixtures of species on the life span of Balaustium leanderi. Diets

consisted of whitefly eggs (W), thrips larvae (T), spider-mite eggs (S), or combinations of these prey. ((A) Mean (+

s.e.) life span, including the mean duration of the immature phase indicated by the black bars and the adult

longevity indicated by the white bars. ((B) Mean (+ s.e.) preoviposition period. Different letters indicate significant

differences among diets for the total life span (A, white + black bars) and for the pre-oviposition period (B).

35


deutonymph showing three distinctive parallel white dorsal lines; (6) the orange-redtritonymphs (morphologically similar but larger than the protonymph); (7) the bright-red to dark-red adults with four pairs of legs and three white lines on the dorsum (likethe deutonymph) (1 mm wide, 1.5 mm long). We found only females, as in B. muro-rum (Halliday 2005).

The overall life span varied significantly with diet (FIGURE 1.1A; GLM, deviance =67.7, d.f. = 6,73, P<0.0001). The shortest mean life span was found on a mixed dietof eggs of whiteflies and spider mites, the longest when offered thrips larvae. Mostof the variation in overall life span was due to differences in the longevity of the adults(FIGURE 1.1A).

The time from emergence as an adult to the first egg laid (the pre-oviposition peri-od) also varied significantly with the diet offered (FIGURE 1.1B; GLM, deviance =123.7, d.f. = 6,47, P<0.0001). It was shortest on a mixed diet of whiteflies and spidermites and on a diet of spider mites and thrips, and longest on a diet of thrips or spi-der mites alone, thrips alone and the combination of all three prey species.

SurvivalSurvival varied significantly with diet (FIGURE 1.2; Log rank test = 13.36, d.f = 6, P =0.038). The predatory mites fed with thrips alone or in combination with spider mitesshowed significantly more mortality compared to the predatory mites fed with theother diets. In general, most mortality occurred during the development from pre-lar-vae to larvae, which occurs from days 15-30 (FIGURE 1.2).

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80

Time [days]

Cum

ulat

ive

prop

ortio

n su

rviv

ing

until

adu

lt

W

WT

T

STS

WS

STW

aaa

a

bb

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80

Time [days]

0 20 40 60 80

Time [days]

Cum

ulat

ive

prop

ortio

n su

rviv

ing

until

adu

lt

WW

WTWT

TT

STSTSS

WSWS

STWSTW

aaa

a

bb

FIGURE 1.2. Cumulative survivorship of Balaustium leanderi predators on different diets. S refers to spider-mite

eggs, T to thrips larvae and W to whitefly eggs. Different letters next to the survivorship curves indicate significant

differences.

ReproductionOviposition occurred within 1 or 2 days, implying big-bang reproduction. Thus, tocharacterize reproductive effort, it suffices to focus on fecundity per reproductivefemale (FIGURE 1.3A). Fecundity varied significantly with diet (GLM, F6,46 = 3.34, P =0.008). In particular, fecundity on a diet of spider mite eggs was significantly lowerthan that on a diet of whitefly eggs or of thrips larvae.

The percentage of non-reproductive predatory mites (sT) also varied with diet(FIGURE 1.3B; GLM, F6,71 = 2.58, P = 0.026). On a diet of whitefly eggs either with orwithout spider mites, fewer than 10% of the adult predators did not oviposit, whereason the other diets this was 25% or more. The proportion of non-reproductive preda-

36


FIGURE 1.3. The effect of diet on the average (+ s.e.) ((A) total number of eggs laid by reproductive adult females

of Balaustium leanderi and ((B) proportion of non-reproductive B. leanderi females. Different letters indicate signif-

icant differences. S refers to spider-mite eggs, T to thrips larvae and W to whitefly eggs.

37


tors fed on the mixed diet of eggs of whiteflies and spider mites and that fed on a dietof whiteflies alone was significantly lower than in the other treatments (FIGURE 1.3B).

Life table parametersUsing the individual estimates of T and mT and the population estimates of lT, we cal-culated ‘individual’ rm, their mean and standard error for each diet (FIGURE 1.4) andfound significant dietary effects on rm (GLM, F8,47 = 12.88, P<0.0001). Clearly, amixed diet of whitefly and spider mite eggs together gave the highest value of rm anda diet of exclusively thrips larvae or spider mites resulted in the lowest values.

The effect of diet on R0 was not tested separately because it essentially convergesto analyzing the effect on fecundity (lT being a population estimate). We found a non-significant negative correlation between the generation time T and the net reproduc-tive rate R0 = lT mT (FIGURE 1.5A; linear regression, d.f. = 1, 52, R2 = 0.045, P = 0.12).

Regressing rm against T is not very informative: because T was used to calculaterm, it results in a spurious correlation. Instead, it was thought to be more instructiveto see how rm declines with T when rm is calculated following expression (1), usingeach individual developmental time T (points in FIGURE 1.5B) and the overall mean R0

(drawn line in FIGURE 1.5B). Clearly, the data points tend to be above the line for lowvalues of T and below the line for higher values of T. This demonstrates the extent towhich the R0-T relation determines the individual rm.

FIGURE 1.5 illustrates the main points of our life history assessment. First, a diet ofthrips larvae or spider mites is the least profitable food in terms of R0 and rm (the datapoints cluster at the lower right end of the plot). Second, the mixture of eggs of white-flies and spider mites tend to be the most profitable food source, especially whenlooking at rm (all data point cluster at the higher left end of the plot; FIGURE 1.5B).

aab

b

ab

ab

c

ab

0

0.02

0.04

0.06

0.08

S ST STW T W WS WTDiet

Intr

insi

c ra

te o

f in

crea

se (

r m)

aab

b

ab

ab

c

ab

0

0.02

0.04

0.06

0.08

S ST STW T W WS WTDiet

Intr

insi

c ra

te o

f in

crea

se (

r m)

FIGURE 1.4. The effect of different diets on the intrinsic rate of population increase (rm + s.e.) of Balaustium. Different

letters indicate significant differences. S refers to spider-mite eggs, T to thrips larvae and W to whitefly eggs.

DiscussionPronounced semelparityWe show that B. leanderi exhibits big-bang reproduction or ‘pronounced semelpari-ty’ (Stearns 1976, 1992; Roff 1992; Hautekeete et al. 2001): they lay all their eggs in1 or 2 days after a long period of development and die soon thereafter. Big-bangreproduction is widespread among the hard ticks (Acari: Ixodidae), whereasiteroparous reproduction prevails among the soft ticks (Acari: Argasidae) (Sonen -

38


A

35 45 55 65 75Egg-to-egg developmental period

0

10

20

30

40N

et r

epro

duct

ion

(R0)

S ST STW T W WS WT A



0

10

20

30

40N

et r

epro

duct

ion

(R0)

0

10

20

30

40N

et r

epro

duct

ion

(R0)

S ST STW T W WS WT

B

35 45 55 65 75Egg-to-egg developmental time

0

0.02

0.04

0.06

0.08

Intr

insi

c ra

te o

f inc

reas

e (r

m)

B



0

0.02

0.04

0.06

0.08

Intr

insi

c ra

te o

f inc

reas

e (r

m)

0

0.02

0.04

0.06

0.08

Intr

insi

c ra

te o

f inc

reas

e (r

m)

FIGURE 1.5. Relation of ((A) net reproduction (R0) and ((B) the intrinsic rate of population increase (rm) to egg-to-egg

developmental period (T) of Balaustium leanderi on different diets. Panel B only serves for visual inspection and

was not used for a regression analysis because it would yield spurious correlations. The interpretation of this fig-

ure is facilitated by including a line expressing rm as a function of T (see text). The deviation of the data from this

line shows the extent to which the relation between R0 and T (Panel A) affects the value of rm.

39


shine 1991). On a continuous scale from ‘pronounced semelparity’ to ‘pronouncediteroparity’ (Hautekeete et al. 2001), many fast-reproducing mites exhibit semel-parous reproduction; examples are physogastric heterostigmatic mites (Bruce &Wrensch 1990; Kaliszewski et al. 1995), prostigmatic mites such as eriophyoids(Sabelis & Bruin 1996) and tetranychoid mites (Sabelis 1991), and mesostigmaticmites such as phytoseiids (Sabelis & Janssen 1994). However, all these mites requirea few weeks to produce all offspring and do not immediately die after reproducing(Blommers & van Arendonk 1979). Hence, they do not even come close to the phe-nomenon of big-bang reproduction as described here for B. leanderi and elsewherefor B. hernandezi (Makol et al. 2012). Theory on semelparous and iteroparous repro-duction predicts that big-bang reproduction is favoured by natural selection whenadult survival (up to the next reproductive bout) is predictably very low relative tojuvenile survival (Cole 1954; Charnov & Schaffer 1973; Young 1981; Ranta et al. 2002;Zeineddine & Jansen 2009). However, which ecological factors determine predictablylow survival in adults of erythraeid predators remains to be elucidated. High soil sur-face temperatures may cause dramatic mortality, but at least some Balaustium spp.seem to resist temperatures up to 48-52 ºC and have a high dehydration tolerance(Hedges et al. 2012). Critical factors determining survival of B. leanderi clearly needto be investigated.

Effect of diet on life table parametersIn the case of big-bang reproduction, life-history theory predicts the existence of aswitch point in the relative importance of net reproduction (R0) and the generationtime (T) (Caswell 1982). Here, we investigate the existence of such a switch point inB. leanderi. The relative contribution of R0 and T to rm can be assessed from the ratioof r-values obtained after a fixed proportional increase (a) in R0 [referred to as r(R0)]and in 1/T [referred to as r(T)]:

r(R0)/r(T) = [ln(aR0)/(T+1)]/[aln(R0)/(T+1)] = ln(aR0)/aln(R0). (2)

In order to study the relative importance of R0 and T for the population growth rateof B. leanderi, we assessed the change of rm as a result of changing from feedingand developing on one diet to another diet. Because the change from diet j to diet iand the change from diet i to j would result in the same absolute value of the changein rm, we took the subset of changes from an inferior diet to a superior diet. Hence,each combination of diets was represented once.

Expression (2) can be used to calculate the relative contribution of R0 and T to rmas a function of R0 for equal proportional changes (a) in both parameters. However,the changes in R0 and T were not equal in our data set. The R0 on the best diet

(whiteflies plus thrips) was 21.29 and on the worst diet (thrips) 9.35. Hence, the max-imum proportional change in R0 was a = 2.277 (= 21.29/9.35). Likewise, T on the bestdiet (whiteflies plus spider mites) was 43.75, and on the worst diet (thrips) it was 63.0,resulting in a proportional change of b = 1.44 (= 63.0/43.75). We used these valuesto calculate the relative contribution of R0 and T to rm as a function of R0 using:

r(R0)/r(T) = ln(aR0)/bln(R0). (3)

The curve resulting from this (FIGURE 1.6) shows that R0 contributes more toincreases of rm for values of R0<7, whereas T is more important for higher values.Hence, there is indeed a switch point from R0 being more important to T being moreimportant for realistic values of these parameters.

We subsequently considered all possible diet changes of B. leanderi that resultedin an increase of the intrinsic growth rate (hence, changing to a better diet) and cal-culated the contribution of R0 and T to the increase in rm using expression (3), usingthe relative changes in R0 and T for this particular diet shift as estimates of a and b.The results are shown as points in FIGURE 1.6. Our data indeed show that the contri-bution of T to rm is higher than that of R0 for higher values of R0, whereas the oppo-site is true for low values of R0 (FIGURE 1.6). However, the switch point seems to occurfor somewhat higher values of the R0 on the inferior diet than predicted (cf. pointsand curve in FIGURE 1.6). The existence of such a switch point has interesting conse-quences for the evolution of the life-history of B. leanderi, which will experience

40


0 5 10 15 20 25R0 on inferior diet

0.0

0.5

1.0

1.5

rela

tive

cont

riut

ion

R 0T



0.0

0.5

1.0

1.5

rela

tive

cont

riut

ion

R 0T

0.0

0.5

1.0

1.5

rela

tive

cont

riut

ion

R 0T

FIGURE 1.6. The relative contribution of R0 and T to changes in intrinsic growth rate as a result of changing from

an inferior diet to a superior diet. Above the interrupted line, the contribution of R0 to an increase in rm as a result

from the switch from an inferior diet to a superior diet is more important, below this line, the contribution of T is

larger. The curve is calculated using expression (3) (see text), points were calculated from the data of each pos-

sible shift from an inferior to a superior diet.

stronger selection for higher net reproduction than for a shorter developmental peri-od on some diets, and the reverse on other diets. This could be verified by setting upan experimental evolution approach with samples from the same population of thispredator on different diets.

Thrips and spider mites as inferior preyTo explain the low performance on a diet of thrips larvae only, the above analysis indi-cates that we should focus on the factors that decrease net reproduction of the ery-thraeid predator (R0). Thrips are known to counter-attack predatory mites, especial-ly their eggs (Faraji et al. 2002a,b; Janssen et al. 2002; Magalhães et al. 2005), andthis may offer explanations for the low numbers of eggs found for predators on a dietof thrips. Firstly, the female predators may oviposit normally on a diet of thrips lar-vae, but counterattacks by the thrips may have resulted in egg mortality, thus false-ly giving the impression that the oviposition rate of the predators was low. Second,it is known that predators may retain eggs when eggs run the risk of being preyedupon (Montserrat et al. 2007), and this may have occurred here. Third, it is possiblethat thrips also injure adult predators, which are therefore unable to produce eggs(for such effects on phytoseiid predators, see Bakker & Sabelis 1989). Furthermore,the low juvenile survival on a diet of thrips larvae (FIGURE 1.2) may directly result fromattacks on predator larvae by thrips larvae. In particular, the larvae of the erythraeidpredator are vulnerable because they are similar in size to thrips larvae. The deu-tonymphs, however, are larger and therefore suffer less from counterattack by thripslarvae. All these possible effects warrant a more detailed study of counterattackingbehaviour of the thrips and the impact on the predatory mites.

On the other hand, the decreased performance on a diet of spider mites which ledto a rm as low as the one obtained for the predators fed on thrips (FIGURE 1.4) wasnot due to low survival (FIGURE 1.2) but to the long generation and pre-ovipositiontime, the low oviposition, and the high number of individuals that did not reproduce.This could indicate that there is a deficiency in the nutrients found in the spider miteeggs: this possible lack of nutrients may have affected the development and repro-duction of B. leanderi.

The benefit of a mixed dietThe major outcome of our study is that the intrinsic rate of population increase wassignificantly higher when the predators fed on a mixture of whitefly eggs and spider-mite eggs compared to other diets. Such effects of mixed diets have been reportedbefore (Toft & Wise 1999; Oelbermann & Scheu 2002; Messelink et al. 2008; Harwoodet al. 2009). Because we used a replacement design for the number of each preyspecies in the mixtures and because spider-mite eggs are smaller than whitefly eggs

41


(thus containing less food), the higher reproductive capacity must be due to the nutri-tional composition of the mixed diet and not due to an increased availability of food.However, the quality of the diet is probably not only determined by nutritional con-tent, but also by interference among the prey species, i.e. intraguild predation bythrips larvae, and by prey defences, such as the web of spider mites which protectsthe eggs, but can also be used as a refuge by thrips larvae. These latter factors werenot taken into account in our experiments (since the web was removed) and there-fore need to be considered in future experiments.

Clearly, the role of mixed diets for predatory arthropods that are used as biologi-cal control agents needs more attention. Mixed diets may boost populations of pred-ators, thereby increasing their impact on pest populations (see also Messelink et al.2008), even on pests that are difficult to attack by the predators (Messelink et al.2010). We therefore advocate elaborate testing of dietary effects of all arthropodsthat are sufficiently abundant in the litter and on the plant and may serve as prey forbiological control agents. Moreover, it would be worthwhile to investigate which alter-native foods/prey can be manipulated to boost predator populations in the litter layer,thereby increasing their impact on pests in the crop.

Acknowledgements

We thank Paul van Rijn for comments and especially for some pertinent suggestionswith respect to FIGURE 1.5. KM was supported by Colciencias (Colombia) (Programa‘Francisco José de Caldas’ 2011).

ReferencesBakker FM, Sabelis MW (1989) How larvae of Thrips tabaci reduce the attack success of phytoseiid preda-

tors. Entomol Exp Appl 50:47-51Blommers LHM, van Arendonk RCM (1979) The profit of senescence in phytoseiid mites. Oecologia 44:87-

90Bruce WA, Wrensch DL (1990) Reproductive potential, sex ratio, and mating efficiency of the straw itch mite

(Acari: Pyemotidae). J Econ Entomol 83:384-391Cadogan B, Laing J (1977) A technique for rearing the predaceous mite Balaustium putmani (Acarina:

Erythraeidae), with notes on its biology and life history. Can Entomol 109:1535-1544Carey JR (1993) Applied demography for biologists, with special emphasis on insects. Oxford University

Press, New York, USACaswell H (1982) Life history theory and the equilibrium status of populations. Am Nat 120:317-339Chang GC, Kareiva P (1999) The case for indigenous generalists in biological control. In: BA Hawkins and HV

Cornell (eds) Theoretical approaches to biological control. Cambridge University Press, Cambridge,England, pp. 103-115

Charnov EL, Schaffer WM (1973) Life-history consequences of natural selection: Cole's result revisited. AmNat 107:791-793

Childers CC, Rock GC (1981) Observations on the occurrence and feeding habits of Balaustium pulmani(Acari: Erythraeidae) in North Carolina apple orchards. Int J Acarol 7:63-68

Chiverton PA (1986) Predator density manipulation and its effects on populations of Rhopalosiphum padi(Hom, Aphididae) in spring barley. Ann Appl Biol 109:49-60

42


Cole LC (1954) The population consequences of life history phenomena. Quarterly Rev Biol 29:103-137Crawley MJ (2007) The R Book. John Wiley & Sons, Chichester, EnglandDiamond JM (1982) Big-bang reproduction and ageing in male marsupial mice. Nature 298:115-116Faraji F, Janssen A, Sabelis MW (2002a) The benefits of clustering eggs: the role of egg predation and larval

cannibalism in a predatory mite. Oecologia 131:20-26Faraji F, Janssen A, Sabelis MW (2002b) Oviposition patterns in a predatory mite reduce the risk of egg pre-

dation caused by prey. Ecol Entomol 27:660-664Fuentes Quintero LS, Muñoz-Cárdenas K, Combita O, Jimeno E, Getiva de la Hoz JC, Cantor F, Rodriguez

D, Makol J (2014) A re-description of Balaustium leanderi Comb. Nov. (Actinotrichida, Erythraeidae) withfirst report on characteristics of all active instars and taxonomic notes on the genus. Fla Entomol 97:937-951

Getiva J, Acosta A (2004) Taxonomía de ácaros asociados a cultivos de flores. Asocolflores 66:59-68Halliday R (2005) Predatory mites from crops and pastures in South Africa: potential natural enemies of red-

legged earth mite Halotydeus destructor (Acari: Penthaleidae). Zootaxa 1079:11-64Harwood JD, Phillips SW, Lello J, Sunderland KD, Glen DM, Bruford MW, Harper GL, Symondson WOC

(2009) Invertebrate biodiversity affects predator fitness and hence potential to control pests in crops. BiolControl 51:499-506

Hassell MP, May RM (1986) Generalist and specialist natural enemies in insect predator-prey interactions. JAnim Ecol 55:923-940

Hautekeete NC, Piquot Y, Van Dijk H (2001) Investment in survival and reproduction along a semelparity-iteroparity gradient in the Beta species complex. J Evol Biol 14:795-804

Hayes JL (1985) The predator-prey interaction of the mite Balaustium sp. and the pierid butterfly Coliasalexandra. Ecology 66:300-303

Hedges BZ, Rosselot AE, Tomko PM, Yoder JA, Benoit JB (2012) High temperature and dehydration toler-ance of the red velvet mite, Balaustium sp. (Erythraeidae), permit the exploitation of extremely hot, drymicrohabitats. Int J Acarol 38:89-95

Hosmer DWJ, Lemeshow S (1999) Applied Survival Analysis. Regression Modeling of Time to Event Data.Wiley-Interscience Publication, New York, USA

Hothorn T, Bretz F, Westfall P (2008) Simultaneous inference in general parametric models. Biometrical J50:346-363

Janssen A, Faraji F, van der Hammen T, Magalhães S, Sabelis MW (2002) Interspecific infanticide deters pred-ators. Ecol Lett 5:490-494

Janssen A, Montserrat M, HilleRisLambers R, de Roos AM, Pallini A, Sabelis MW (2006) Intraguild predationusually does not disrupt biological control. In: Brodeur J, Boivin G (eds) Trophic and guild interactions inbiological control. Vol. 3. Springer, Dordrecht, The Netherlands, pp. 21-44

Kaliszewski M, Athias-Binche F, Lindquist EE (1995) Parasitism and parasitoidism in Tarsonemina (Acari:Heterostigmata) and evolutionary considerations. Advances in Parasitology 35:335-367

Makol J, Arijs Y, Wäckers FL (2012) A new species of Balaustium von Heyden, 1826 (Acari: Actinotrichida,Erythraeidae) from Spain. Zootaxa 3178:1-21

Magalhães S, Janssen A, Montserrat M, Sabelis MW (2005) Prey attack and predators defend: counterattack-ing prey trigger parental care in predators. Proc R Soc B 272:1929-1933

Magalhães S, van Rijn PC, Montserrat M, Pallini A, Sabelis MW (2007) Population dynamics of thrips preyand their mite predators in a refuge. Oecologia 150:557-568

Messelink G, van Maanen R, van Holstein-Saj R, Sabelis MW, Janssen A (2010) Pest species diversityenhances control of spider mites and whiteflies by a generalist phytoseiid predator. BioControl 55:387-398

Messelink GJ, Bloemhard CMJ, Cortes JA, Sabelis MW, Janssen A (2011) Hyperpredation by generalistpredatory mites disrupts biological control of aphids by the aphidophagous gall midge Aphidoletesaphidimyza. Biol Control 57:246-252

Messelink GJ, Sabelis MW, Janssen A (2012) Generalist predators, food web complexities and biological pestcontrol in greenhouse crops. In: M.L. Larramendy and S. Soloneski (eds) Integrated Pest Managementand Pest Control – Current and Future Tactics. InTech, Rijeka, Croatia, pp. 191-214

43


Messelink GJ, van Maanen R, van Steenpaal SEF, Janssen A (2008) Biological control of thrips and whitefliesby a shared predator: two pests are better than one. Biol Control 44:372-379

Montserrat M, Bas C, Magalhaes S, Sabelis MW, de Roos AM and Janssen A (2007) Predators induce eggretention in prey. Oecologia 150:699-705

Muñoz K, Fuentes L, Cantor F, Rodríguez D, Cure JR (2009) Feeding preferences of the mite Balaustium sp.under controlled conditions. Agron Colomb 27:95-103

Murdoch WW (1994) Population regulation in theory and practice. Ecology 75:271-287Murdoch WW, Chesson J, Chesson PL (1985) Biological control in theory and practice. Am Nat 125:344-366Murdoch WW, Scott MA, Ebsworth P (1984) Effects of the general predator, Notonecta (Hemiptera) upon a

fresh-water community. J Anim Ecol 53:791-808Nomikou M, Janssen A, Schraag R, Sabelis MW (2001) Phytoseiid predators as potential biological control

agents for Bemisia tabaci. Exp Appl Acarol 25:271-291Nomikou M, Janssen A, Schraag R, Sabelis MW (2002) Phytoseiid predators suppress populations of Bemisia

tabaci on cucumber plants with alternative food. Exp Appl Acarol 27:57-68Oelbermann K, Scheu S (2002) Effects of prey type and mixed dites on survival, growth and development of

a generalist predator, Pardosa lugubris (Araneae: Lycosidae). Basic Appl Ecol 3:285-291Pozzebon A, Duso C (2008) Grape downy mildew Plasmopara viticola, an alternative food for generalist

predatory mites occurring in vineyards. Biol Control 45:441-449Putman WL (1969) Life history and behavior of Balaustium putmani (Acarina: Erythraeidae). Ann Entomol Soc

Am 63:76-81Ranta E, Tesar D, Kaitala V (2002) Environmental variability and semelparity vs. iteroparity as life histories. J

Theor Biol 217:391-396Roff DA (1992) The evolution of life histories: theory and analysis. Chapman & Hall, New York, USARosenheim JA (1998) Higher-order predators and the regulation of insect herbivore populations. Annu Rev

Entomol 43:421-447Rosenheim JA, Kaya HK, Ehler LE, Marois JJ, Jaffee BA (1995) Intraguild predation among biological control

agents – Theory and evidence. Biol Control 5:303-335Rosenheim JA, Wilhoit LR, Armer CA (1993) Influence of intraguild predation among generalist insect preda-

tors on the suppression of an herbivore population. Oecologia 96:439-449Sabelis MW (1991) Life-history evolution of spider mites. In: R. Schuster and P.W. Murphy (eds) The Acari.

Reproduction, Development and Life-history Strategies. Chapman & Hall, London, UK, pp. 23-49Sabelis MW, Bruin J (1996) Evolutionary ecology: Life history patterns, food plant choice and dispersal. In:

E.E. Lindquist, M.W. Sabelis and J. Bruin (eds) World Crop Pests. Vol. 6. Elsevier, Amsterdam, TheNetherlands, pp. 329-366

Sabelis MW, Janssen A (1994) Evolution of life history patterns in the Phytoseiidae. In: Houck MA (ed) Mites:ecological and evolutionary analyses of life history patterns. Chapman & Hall, New York, USA, pp. 70-98

Scheu S (2001) Plants and generalist predators as links between the below- ground and above-ground sys-tem. Basic Appl Ecol 2:3-13

Settle WH, Ariawan H, Astuti ET, Cahyana W, Hakim AL, Hindayana D, Lestari AS (1996) Managing tropicalrice pests through conservation of generalist natural enemies and alternative prey. Ecology 77:1975-1988

Sonenshine DE (1991) Biology of ticks. Oxford University Press, New York, USAStearns SC (1976) Life-history tactics: a review of the ideas. Q Rev Biol 51:3-47Stearns SC (1992) The evolution of life histories. Oxford University Press, Oxford, EnglandSymondson WOC, Sutherland KD, Greenstone MH (2002) Can generalist predators be effective biocontrol

agents? Annu Rev Entomol 47:561-594Toft S, Wise DH (1999) Growth, development, and survival of a generalist predator fed single- and mixed-

species diets of different quality. Oecologia 119:191-197Torrado E, Pérez M, Cure J, García M, Echeverri C (2001) Evaluación de sistemas de control biológico utiliza-

dos comercialmente en Europa para el control de plagas de rosa bajo invernadero, en la Sabana deBogotá. Asocolflores 61:34-45

44



Welbourn WC (1983) Potential use of trombidoid and erythraeoid mites as biological control agents of insectpests. In: M.A. Hoy, G.L. Cunningham and L. Knutson (eds) Biological Control of Pests and Mites.University of California, Berkeley, USA, pp. 103-140

Welbourn WC, Jennings DT (1991) Two new species of Erythraeidae (Acari, Prostigmata) associated with thespruce budworm, Choristoneura fumiferana (Clemens) (Lepidoptera, Tortricidae), in Maine. Can Entomol123:567-580

Young TP (1981) A general model of comparative fecundity for semelparous and iteroparous life histories. AmNat 118:27-36

Zeineddine M, Jansen VA (2009) To age, to die: parity, evolutionary tracking and Cole's paradox. Evolution63:1498-1507

SUPPLEMENTARY MATERIAL 1.1 – Details of culture methods

45


Karen Muñoz1,3

Luz Fuentes2

Fernando Cantor3

Daniel Rodríguez3

Arne Janssen1

Maurice W. Sabelis1

Photography: Jan van Arkel1

1. IBED, University of Amsterdam. 2. Centro de Biosistemas, Universidad Jorge Tadeo Lozano

3. Facultad de Ciencias Básicas, Universidad Militar Nueva Granada

contact: [email protected]

Generalist red velvet mitepredator (Balaustium

leanderi) performs better on a mixed diet

46


Predator cultures

Leaves with prey

Predatory mites

Humid peat

Towel paper

Experimental units

Petri dish lid wrapped with paper towel

Opaque PVC tube fragment

Plastic wrap barrier

Piece of cork

Plastic tube with water

Piecee

pppp bbbbbbbbbbarrie

Cork provides oviposition sites

47


Experimental units

Petri dish lid wrapped with paper towel

Transparent PVC tube fragment

Plastic wrap barrier

Plastic tube with water

pppp bbbbbbbbbbbarrie

Arenas to assess juvenile development

Eggs must be kept on wet cotton

Experimental units

48


Balaustium sp.

Life cycle

Eggs

49


Prelarva

Larva

50


Protonymph

Deutonymph

51


Tritonymph

Emerging adult

52


Balaustium leanderi preying on agreenhouse whitefly adult

Balaustium leanderi preying onwestern flower thrips larva

53


Balaustium leanderi adult preying ontwo-spotted spider mite adult

0.166 mmx

0.20 mmEgg size

54


Balaustium leanderi exhibits big-bang reproduction

SUPPLEMENTARY MATERIAL 1.2 – Prey mortality

MethodsAs explained in the Methods section of the main text, we provided sufficient amountsof each prey species and stage to allow for maximum predation rates. Each day, thenumber of surviving prey individuals was assessed per arena. Mortality rates per preyspecies were first averaged per predator individual per stage and were comparedamong stages with a mixed effects model with the individual predator as random fac-tor (LME of the nlme package of R). For these comparisons, only the single diets (sonot the mixed diets) were used.

ResultsAll mobile stages of the predator consumed prey. Prey mortality rates increased withthe developmental stage of predator, irrespective of the prey offered (FIGURE S1.2.1;TABLE S1.2.1a).

Because we supplied sufficient amounts of prey, we expected consumption ofeach prey species in a prey mixture to be lower than with single prey diets due topredator satiation. This was indeed the case for the mixed diet of eggs of whitefliesand of spider mites (FIGURE S1.2.1), but more eggs of spider mites and more eggs ofwhiteflies were killed in the presence of thrips larvae than when these eggs wereoffered alone (FIGURE S1.2.1; TABLE S1.2.1b). This may have been caused by thripslarvae acting as intraguild predators on the eggs of whiteflies and spider mites, there-

55


FIGURE S1.2.1. Average (+ s.e.) prey mortality (prey/predator/day) in the presence of different developmental

stages (larva, deutonymph, adult) of Balaustium leanderi when offered single or mixed diets of whitefly eggs (W),

thrips larvae (T) and spider-mite eggs (S).

by increasing overall levels of mortality of these latter two prey. Therefore, we referto prey mortality rates instead of predation rates throughout the manuscript.

56


TABLE S1.2.1a. Comparisons among the number of prey consumed by different stages of Balaustium leanderi

(mixed effects model with repeated measures).

Prey Overall effect and comparisons df Likelihood ratio PWhitefly eggs Overall 2 109.0 <0.0001

Larvae-Nymphs 1 77.7 <0.0001Nymphs-Adults 1 24.4 <0.0001Adults-Larvae 1 105.5 <0.0001

Thrips larvae Overall 2 30.2 <0.0001Larvae-Nymphs 1 1.11 0.29Nymphs-Adults 1 20.7 <0.0001Adults-Larvae 1 27.3 <0.0001

Spider mite eggs Overall 2 187.8 <0.0001Larvae-Nymphs 1 135.7 <0.0001Nymphs-Adults 1 56.9 <0.0001Adults-Larvae 1 181.5 <0.0001

TABLE S1.2.1b. Comparison of the prey mortality in treatments with whitefly eggs (W) or spider mite eggs (S) alone

and in combination with western flower thrips (T) (linear mixed effects model with repeated measures).

Predator stage Prey Comparison df Likelihood ratio PLarva Whitefly eggs W vs. W+T 1 46.9 <0.0001

Spider mite eggs S vs. S+T 1 77.2 <0.0001Nymph Whitefly eggs W vs. W+T 1 15.9 <0.0001

Spider mite eggs S vs. S+T 1 44.9 <0.0001Adult Whitefly eggs W vs. W+T 1 14.4 <0.0001

Spider mite eggs S vs. S+T 1 63.1 <0.0001

Supplying high-quality alternative preyin the litter increases control of anabove-ground plant pest by a generalistpredator

Karen Muñoz-Cárdenas, Firdevs Ersin, Juliette Pijnakker, Yvonne van Houten, Hans

Hoogerbrugge, Ada Leman, Maria L. Pappas, Marcus V. A. Duarte, Gerben J.

Messelink, Maurice W. Sabelis & Arne Janssen

Supplying predators with alternative food can have short-term positive effects on

prey densities through predator satiation (functional response) and long-term neg-

ative effects through increases of predator populations (numerical response). In

biological control, alternative food sources for predators are normally supplied on

the crop plants; using the litter-inhabiting food web as a source of alternative food

for plant-inhabiting predators has received less attention. We investigated the

effect of supplying plant-inhabiting predatory mites with alternative prey (astigmat-

ic mites) in the litter on pest control. Predator (Amblyseius swirskii) movement and

population dynamics of the pest (western flower thrips) and predators were stud-

ied on rose plants in greenhouses. Predators commuted between the above-

ground plant parts where they controlled thrips, and the litter, where they fed on

alternative prey, although the latter were a superior diet. Predators controlled thrips

better in the presence of the astigmatic mites than in their absence. We show that

predatory mites can form a link between above-ground pests and the litter food

web, and propose that adding alternative prey to the litter of ornamental green-

houses can result in higher predator densities and increased biological control.

Biological Control (2017) 105: 19-26

59

2

IntroductionTraditionally, above-ground and below-ground interactions involving the food webassociated with plants were studied independently, but it has become clear thatthese two food webs are connected (A’Bear et al. 2014; Bezemer & van Dam 2005;Gange & Brown 1989; van der Putten et al. 2001). Generalist predators can linkabove-ground and below-ground food webs by attacking prey in both habitats, thusshaping the composition and structure of the communities (Scheu 2001; Wardle etal. 2004). Such links between spatially coupled food webs may affect the stability ofprey dynamics in ecosystems (de Roos et al. 1998; McCann et al. 2005). Here, weinvestigate whether such predator-mediated links can be used to improve biologicalcontrol.

Generalist predators are commonly used in biological control of crop pests(Symondson et al. 2002). They can feed on both the target pest and non-pest preyor other food sources (English-Loeb et al. 1993). Supplying alternative food to pred-ators can affect biological control positively by increasing predator survival andreproduction when target pests are scarce (Eubanks & Denno 2000). If populationsof predators and prey do not exhibit sustained oscillations, adding extra food willresult in an increase in the densities of predators in the longer term through thenumerical response. This, in turn, will result in a decrease of the densities of both thepest and the alternative prey (apparent competition; Holt 1977). However, the addi-tion of alternative prey or food may initially decrease predation on target pests whenpredators concentrate feeding on the alternative food, or because predators becometemporarily satiated (Abrams & Matsuda 1996; Holt 1977; van Baalen et al. 2001; vanMaanen et al. 2012). These positive effects of adding alternative food on a prey arereminiscent of apparent mutualism (Holt 1977; Abrams & Matsuda 1996).

Whereas apparent mutualism is undesired in biological control, apparent compe-tition is beneficial. For example, Liu et al. (2006) showed better control of spidermites on apple trees in the presence of alternative prey in the longer term (3.5months), whereas there was no evidence of apparent mutualism in the shorter term.Messelink et al. (2008) found better control of one pest in the presence of anotherpest, both of which were attacked by the same predator species. Here too, there wasno evidence of short-term apparent mutualism, but pest densities of the initial 4weeks were lacking. A follow-up of this study indeed did show short-term apparentmutualism (van Maanen et al. 2012). Messelink et al. (2008) furthermore showed thatthe predators performed better on a mixed diet of two pest species, an effect thathas not been included in the theory of apparent competition, but which results ineven further reductions of prey densities.

Another strategy to improve biological control is to boost the populations of pred-ators not with alternative prey, but with alternative food that does not damage the

60

CHAPTER 2 | ALTERNATIVE PREY IN THE LITTER HELPS PEST CONTROL IN THE CANOPY

plant, such as pollen (Calvert & Huffaker 1974; van Rijn et al. 2002; Janssen & Sabelis2015). However, the application of pollen to a crop can sometimes be risky becausepests can also benefit from the presence of pollen (Chitturi et al. 2006; Leman &Messelink 2015; Vangansbeke et al. 2016; but see van Rijn et al. 2002).

We studied a novel approach to improve biological control in ornamental green-houses, which is supplying alternative prey for predators in the soil/litter layer. Plant-inhabiting predators can feed on alternative prey belonging to the decomposer com-munity, which can feed and reproduce on the available organic material (Settle et al.1996). The advantage of using the litter food web to supply alternative food is that itleaves no residues on the plant parts to be commercialized, which is desirable in orna-mental crops. The risk is that the above-ground target pest and the alternative foodare spatially separated, and predators may not commute sufficiently between theabove-ground plant parts and the litter, thus either not benefitting fully from the alter-native food, or reducing their attacks on the pest. This will even be more risky whenthe alternative food is of better quality for the predator than the plant-inhabiting pest.

Links between the above-ground and below-ground food webs through general-ist predators have been observed in different systems (Moore et al. 2003), such ascrops in the field (Settle et al. 1996), forests (Miyashita et al. 2003), grasslands (Huntet al. 1987; Wardle et al. 2005) and organic farms (Birkhofer et al. 2008). Howeverthere are only a few examples of applications of such links between above-groundand below-ground food webs for biological control (Birkhofer et al. 2008; Halaj &Wise 2002; Settle et al. 1996). Settle et al. (1996) added organic matter in rice fields,thus boosting populations of decomposers and plankton feeders and significantlyincreasing the abundance of predators and enhancing pest control. In contrast,Birkhofer et al. (2008) showed that increasing decomposer prey densities in wheatfields resulted in increases of populations of herbivores above-ground, and conclud-ed that the predators switched from feeding on herbivores to decomposers (appar-ent mutualism; Holt 1977).

Hence, contrasting effects have been found of adding alternative prey on biolog-ical control. We therefore tested the effect of adding alternative prey to the litter ondensities of plant-inhabiting predators and an above-ground plant pest. The studysystem consisted of rose plants, the pest species Frankliniella occidentalis(Pergande), which cause economic damage in many different crops (Loomans &Murai 1997), the predatory mite species Amblyseius swirskii (Athias-Henriot), apredatory mite used to control thrips and whiteflies (Buitenhuis et al. 2015; Calvo etal. 2015; Messelink et al. 2006; Nomikou et al. 2001; Pijnakker & Ramakers 2008) andseveral species of soil-inhabiting predators. We assessed whether predators com-muted between the above-ground plant parts and the litter layer with alternative prey(astigmatic mites). Subsequently, we tested whether the alternative prey is of supe-

61

ALTERNATIVE PREY IN THE LITTER HELPS PEST CONTROL IN THE CANOPY | CHAPTER 2

rior or inferior quality compared to the pest. Finally, we carried out experiments in thelaboratory and greenhouse to test whether control of thrips was improved by supply-ing alternative prey for their predators in the litter.

Materials and methodsRearing methodsRose plants for cut flower production were purchased when they were 4 weeks oldand kept in acclimatized rooms or in a greenhouse. To avoid contamination witharthropods, we removed all leaves and washed the stems with running tap water.Subsequently, the stems and roots were allowed to regrow in an acclimatized room (25°C, RH 60%, photoperiod L16:D8, Sylvania GRO-LUX F58W/GRO 5FT T8 58W) or agreenhouse compartment (22 °C, RH 70%, photoperiod L14:D10) inside cages. Theywere either planted in pots (26 cm diameter, 25 cm high) with peat as substrate (50%coco peat, 15% white peat, 35% frozen black peat; Jongkind grond BV, Aalsmeer, TheNetherlands) or in a rockwool strip (Grodan® Delta) inside a plastic tray (42 × 8 × 30).Plants were watered twice a week, and macro- and micronutrients (0.5 g N-P-K and0.5 g micronutrients mix/200-500 ml of water/plant) were applied once a week.

Frankliniella occidentalis came from the stock colony of Koppert BiologicalSystems (Berkel en Rodenrijs, The Netherlands), where they had been reared on run-ner bean pods Phaseolus vulgaris L. plus Typha latifolia L. pollen for more than 10years. The thrips were reared in the laboratory (at 23 °C, RH 65%) in plastic contain-ers (8 × 8 × 12 cm), with openings covered with mite-proof mesh (30 μm) for ventila-tion (adapted from Loomans & Murai 1997). Around 200 adult thrips were introducedinto each container with four half bean pods, in which the females laid eggs for 1 day.Subsequently, the pods were replaced with fresh ones for further egg laying. Theegg-infested pods were transferred to clean rearing containers, giving rise to individ-uals of the same age. Young first-instar thrips were taken from these containers andused for experiments on food quality.

Amblyseius swirskii used for laboratory experiments were originally collected fromcotton in Redavim, Israel (Nomikou et al. 2001). They were reared on plastic arenas(8 × 15 cm), placed on a wet sponge in a plastic tray containing water, followingOvermeer (1985). Strips of wet tissue were placed along the periphery of the arenato provide water. A tent-shaped piece of plastic sheet (1–2 cm2) was placed on eacharena and functioned as a shelter for the mites. A few cotton wool threads were putunderneath the shelter as oviposition substrates. Cultures were fed with T. latifoliapollen. To produce cohorts of predators of the same age, the threads were trans-ferred to a new rearing arena every 1-2 days. Ten days later, adults from these are-nas were used for experiments. For greenhouse experiments, commercially availablepredators were used (Swirski-mite®, Koppert Biological Systems).

62


63


We used two species of astigmatic mites as alternative prey: Tyrophagus putres-centiae (Schrank), the most abundant astigmatic species in the litter of commercialrose crops in greenhouses in the Netherlands (K. Muñoz-Cárdenas, pers. obs.), andCarpoglyphus lactis L., also commonly occurring in rose litter (K. Muñoz-Cárdenas,pers. obs.) and used for mass rearings of A. swirskii (Bolckmans & van Houten 2006).Astigmatic mites were provided by Koppert Biological Systems and were reared incylindrical plastic containers (8 cm diameter, 12 cm high). For T. putrescentiae 10 gof bran (De Halm, Heeswijk, The Netherlands) was added as food; C. lactis culturesreceived 5 g of bran and 5 g of yeast (De Halm), all once per week. The containerswere kept in the laboratory inside styrofoam boxes at 21 °C with wet tissue paper atthe bottom to increase humidity.

Movement of Amblyseius swirskiiBecause A. swirskii is mainly known to occur on above-ground plant parts, we test-ed whether they moved to the litter to feed on alternative prey. In a greenhouse com-partment at the University of Amsterdam, 4-week-old rose plants (var. Avalanche; OlijRozen Int., de Kwakel, The Netherlands) with 3-4 leaves (c. 15 cm high) planted inrock wool were placed inside small insect-proof cages (30 × 30 × 30 cm). Small Petridishes (2.5 cm) with c. 100 adult females of T. putrescentiae mixed with bran wereplaced on the rock wool under each plant. There were four treatments: a control with-out thrips and three treatments with 5, 10 or 25 first-instar thrips larvae on plantleaves. Immediately after releasing the thrips (at 21:00 h), we released five A. swirskiifemales (10-12 days old) per plant on the leaves. Thrips and predators were trans-ferred to the plants using a fine brush. There were two blocks (October 3 and 24,2012), with three replicates of each treatment per block. Predatory mites in the Petridishes and on the rock wool were counted 10 h later using a stereo-microscope.

The numbers of predators found under the plant per cage were log(x+1) trans-formed to stabilize variance, and the transformed data were compared among treat-ments using a linear mixed effects model (LME, package nlme of R; Pinheiro et al.2014) with the number of thrips as fixed factor and block as random factor. Residualswere analysed to check the suitability of the models (Crawley 2013). All statisticalanalyses were done using R (R Development Core Team 2013).

Effect of alternative prey on thrips densitiesTwo 10-week-old rose plants (var. White Naomi, Olij Rozen Int, c. 70 cm high) in potswith peat as substrate were placed in mite-proof cages (47.5 × 47.5 × 92 cm) in anacclimatized room at the University of Amsterdam (25 °C, RH 60%, photoperiodL16:D8). Fallen, dried leaves were kept in the pots and flower buds were removed;both are common growers’ practices during the early stages of plants development

(Yong 2004). There were two treatments, cages to which A. swirskii was added andcages with A. swirskii plus alternative prey (T. putrescentiae) added. The experimentwas done in two blocks (May and August, 2012), with three replicates (cages) of eachtreatment per block. Thrips were collected from the rearing unit with a disposablepolypropylene pipette tip covered at the wide end with a piece of gauze (mesh 30μm) and connected to a flexible plastic tube, which was either connected to a pumpor used as mouth piece. Air carrying thrips or predators was sucked through the tip,and the tip was subsequently closed at both sides with Parafilm®. A small piece ofyarn was taped to the pipette tip to suspend it from a branch of a rose plant. Afterremoving the Parafilm, the thrips could move onto the plants. In the first week, 40adult thrips were released per cage. During the first, second and third week, 50 adultfemale T. putrescentiae, mixed with 2 g of bran serving as food, were dispersed onthe substrate under the plants of the respective treatment. In the third and fifth week,40 female A. swirskii (10-12 days old) were released onto the leaves with a fine brush.

In the second week, 10 leaves/cage were collected, five leaves from the upperpart and five from the lower part of the plants to confirm establishment of the thrips.This was repeated in the fourth week, 1 week after the first predator release. Thenumbers of thrips and the proportion of leaves with thrips damage were scored usinga stereo-microscope. The plants did not produce any flowers during the experiment.Because most leaves were already damaged since the first sampling, the proportionsof damaged leaves were not analysed. We checked for the presence and identity ofA. swirskii in all the cages. After 6 weeks, all leaves were collected from each cageand the average numbers of thrips per leaf were scored. These numbers werelog(x+0.1) transformed and analysed with a linear mixed effects model (LME) withtreatment as fixed factor and block as random factor. Residuals were checked asabove.

Food quality of pest and alternative preyThe quality of astigmatic mites and thrips larvae as food for the predators wasassessed by measuring juvenile survival and development and oviposition of A.swirskii in the laboratory at the University of Amsterdam. Cohorts of C. lactis wereprepared by allowing 50-60 females to oviposit on a plastic arena (similar to thoseused for A. swirskii, see ‘Rearing methods’) with one gram of solid yeast during 24 h(25 °C, RH 70%). Five days later, larvae were taken from these arenas. One day oldthrips larvae were taken from the thrips colony.

Newly hatched predator larva were transferred each to a separate leaf disc (var.Avalanche) with either eight C. lactis immatures, four young first-instar thrips or fourC. lactis immatures plus two thrips. A surplus of prey was offered, based on report-ed consumption rates of A. swirskii (Bolckmans & van Houten 2006; Messelink et al.

64


65


2008). Every day, the juvenile predatory mites were transferred to a new leaf disc withthe same food. There were two blocks in time, with 14 or 15 predators (replicates)per treatment in the first block and six in the second block. Survival and the stage ofthe predator were recorded daily until all mites had reached adulthood or died.Juvenile development and survival were analyzed using time-to-event analysis (Coxproportional hazards, R package coxme; Therneau 2015), with diet as fixed factorand block as random factor. Contrasts among treatments were assessed throughgeneral linear hypothesis testing (glht of the lsmeans package with the ‘tukey’ adjust-ment of P values; Lenth 2016).

The oviposition of adult A. swirskii was measured during 10 days on the threediets mentioned above. To obtain adult females, rose leaves on pieces of wet cottonwool in Petri dishes were supplemented with around 50 predator eggs. When the lar-vae hatched, we daily added around 100 C. lactis and/or thrips larvae according tothe treatment until copulations of adult predators were observed (after c. 7 days).Twelve females per diet were transferred each to a separate leaf disc with a cottonthread (0.5 cm) as oviposition substrate. They were supplied with either 20 C. lactisimmatures, eight first instar thrips larvae or 10 C. lactis plus four thrips larvae per day.The cotton thread was replaced daily during 10 days and predator eggs were count-ed. The log(x+1)-transformed numbers of eggs produced per female per day wereanalyzed with an LME with individual as random factor and the age of the predatorand the treatment as fixed factors.

Thrips control by Amblyseius swirskii with alternative preyBecause soil-inhabiting predators were encountered above (‘Effect of alternativeprey on thrips densities’), another experiment was done to specifically study theeffect of A. swirskii with alternative food on thrips control. We therefore thoroughlywashed roots of rose plantlets with running tap water to remove soil mites. Plants(var. Avalanche, 4 weeks old) were planted in clean peat before placing them insidecages in a greenhouse compartment at the University of Amsterdam, where twoplants were allowed to grow for 4 weeks, having 10-12 leaves and an approximateheight of 30 cm. Litter collected from a commercial greenhouse was sterilized (108 °C,1 bar, 20 min) to kill mites, moistened (100 ml water / l litter) and placed at the baseof the plants.

There were four treatments, each replicated three times: thrips with bran; thrips,A. swirskii with bran; thrips with c. 4000 adult female C. lactis mixed with all otherstages; thrips, A. swirskii with bran and female C. lactis (same quantities).Carpoglyphus lactis was reared on bran, and was released on the litter with c. 100 gof bran for practical reasons. The same quantity of sterilized bran was added to thecontrol treatments. Forty predators (Swirski-mite®) were released on the litter of

each cage with a mix of bran with or without C. lactis. One week later, each cagereceived 60 adult thrips using pipette tips (‘Effect of alternative prey on thrips densi-ties’). From the third to the twelfth week, we counted thrips, thrips damage and pred-ators on six leaves (two from the top part, two from the middle and two from thelower leaves). All flowers were collected in plastic containers with alcohol (70%),washed with 70% alcohol on a mesh (100 μm) and the thrips and mites on the meshwere counted under a stereo-microscope. Once per week, we checked 50 ml of lit-ter and bran from each cage for the presence of astigmatic mites and predatorymites under a stereo-microscope. Each week, three of the adult female predatorsfound were identified (Chant & McMurtry 2007; Demite et al. 2014). Predatory mitesfrom the flowers and the litter were identified by Farid Faraji (MitoxConsultants/Eurofins).

The densities of thrips on leaves and in flowers and predators were log(x+0.1)transformed, the proportions of damaged leaves (judged by the presence of feedingscars) were not transformed, and all were analyzed with linear mixed effects models(LME) with the experimental unit (cage) as random factor and time (week) and treat-ment as fixed factors. The proportion of flowers with heavy thrips damage (>5 feed-ing scars/flower) were analyzed with a generalized linear model (GLM) with treatmentas factor and a quasi-binomial error distribution. Contrasts among treatments andchecking of the error distribution were done as above.

ResultsA pilot experiment showed that high densities of A. swirskii controlled thrips anddecreased the thrips damage of flowers after 4 weeks (SUPPLEMENTARY MATERIAL 2.1).We therefore investigated whether adding alternative food to the litter could boostdensities of this predator to sufficiently high levels for thrips control.

Movement of Amblyseius swirskiiOf the five mites released per plant on the above-ground parts, we recaptured onaverage 37.5% (s.e. 7.2%). The majority of these (1.4 mites/replicate) were found inthe Petri dishes with alternative prey; the others were found on the plants. Mostprobably, the rest of the mites were inside the rock wool, which was not sampledbecause of its complex structure. If they were indeed there, this would mean thatthey also moved down. The densities of thrips on the plants had no effect on themovement of predatory mites (FIGURE 2.1; LME: d.f. = 3, χ2 = 6.34, P = 0.10). Theseresults confirm observations of A. swirskii individuals in the litter in commercial rosegreenhouses in which they had been released on the plants (K. Muñoz-Cárdenas,pers. obs.).

66


67


Effect of alternative prey on thrips densitiesThere were no significant differences in the average numbers of thrips per leaf beforereleasing the predators in the second week (FIGURE 2.2; LME: d.f. = 1, χ2 = 0.17, P =0.67) or after predator release in the fourth week of the experiment (FIGURE 2.2; LME:d.f. = 1, χ2 = 0.05, P = 0.83). Destructive sampling in the sixth week showed signifi-cantly fewer thrips per leaf in treatments with A. swirskii plus alternative prey thanwith A. swirskii alone (FIGURE 2.2; LME: d.f. = 1, χ2 = 6.75, P = 0.009). Besides A.swirskii, naturally litter-inhabiting thrips similar densities of predators from the familyLaelapidae (genera Cosmolaelaps and Stratiolaelaps) were found in the litter in bothtreatments. Similar results were obtained in a larger scale experiment in the green-house (SUPPLEMENTARY MATERIAL 2.1).

FIGURE 2.2. Effect of addition of astigmatic mites (Tyrophagus putrescentiae) as alternative prey in the litter layer

under laboratory conditions on average numbers of thrips/leaf (± s.e.). Data of the second and fourth weeks cor-

respond to samplings of 10 leaves per cage; data of the sixth week correspond to destructive sampling in which

the number of thrips were scored on all leaves of each cage. The asterisks indicate a significant difference

between treatments (LME, P<0.01)..

A. swirskii

A. swirskii T. putrescentiae

FIGURE 2.1. Effect of the density of first instar larvae of thrips (horizontal axis) on leaves of rose plants on the aver-

age number (± s.e.) of Amblyseius swirskii moving down from the leaves to Petri dishes with astigmatic mites on

the substrate under the plant. There was no significant effect of the number of thrips larvae on predator move-

ment (LME, P>0.05).

Food quality of pest and alternative preyThe rate of juvenile development was affected by diet (FIGURE 2.3a; mixed-effectsCox model: d.f. = 2, χ2 = 25.1, P<0.001); it was similar on diets of C. lactis alone orcombined with thrips and lower on a diet of only thrips (FIGURE 2.3a). Survival washigher in the treatments with C. lactis alone (90%) or combined with thrips (80%)than in the treatment with thrips alone (77%), but these differences were notsignificant (FIGURE 2.3a; d.f. = 2, χ2 = 1.18, P = 0.55).

The average oviposition through time of A. swirskii on a diet of C. lactis immatures,first instar thrips larvae or on a mixed diet did also not differ significantly (FIGURE 2.3b;LME: χ2 = 3.03, P = 0.22).

68


FIGURE 2.3. (a) Development of juveniles of the predatory mite Amblyseius swirskii on a diet of either Carpoglyphus

lactis, thrips, or C. lactis plus thrips. Shown are the cumulative proportions of juveniles that developed into adults

(± s.e.). Different letters in the legend represent significant differences in developmental rate among treatments

(contrasts with general linear hypothesis testing after Cox proportional hazards, P<0.05). Differences in survival

until adulthood were not significant (P>0.05). ((b) Average numbers of eggs per female per day (± s.e.) of young

female A. swirskii during 10 days on a diet of C. lactis, thrips or C. lactis plus thrips. There was no significant effect

of diet on oviposition (LME, P>0.05).

C. lactis a

b

C. lactis

69


Thrips control by Amblyseius swirskii with alternative preyThere was a significant effect of the interaction between treatment and time on thedensities of thrips on leaves (FIGURE 2.4a; LME: d.f. = 3, χ2 = 222.7, P<0.0001). Thiswas caused by the densities of thrips remaining low in the treatment with A. swirskiiplus alternative food and increasing in the other treatments (FIGURE 2.4a). Thrips den-sities on leaves were significantly higher in the treatments with A. swirskii than in thecontrol (FIGURE 2.4a). The numbers of thrips in the flowers differed significantlyamong treatments (FIGURE 2.4b; d.f. = 3, χ2 = 9.26, P = 0.026) and with time (d.f. = 1,χ2 = 31.4, P<0.0001). The numbers of thrips in flowers were significantly lower withA. swirskii plus alternative food than in the control (FIGURE 2.4b). The interactionbetween treatment and time had no significant effect (d.f. = 3, χ2 = 7.66, P = 0.054).Alternative prey significantly increased densities of predators (FIGURE 2.4c; LME: d.f.= 1, χ2 = 11.8, P<0.001). We found A. swirskii in the litter throughout the experimentwhen it was released together with alternative prey.

The proportion of damaged leaves was significantly affected by the interactionbetween treatment and time (FIGURE 2.4d; d.f. = 3, χ2 = 12.6, P = 0.0057). Initially theproportion of damaged leaves was similar in all treatments, but overall, damage lev-

FIGURE 2.4. Effect of addition of Carpoglyphus lactis to a sterilized litter layer as alternative prey for Amblyseius

swirskii on thrips control under greenhouse conditions. ((a) The average number of thrips per leaf on rose plants

(± s.e.) inside cages. Thrips were released in the second week and larvae and adult thrips were sampled from the

third week on. ((b) The average number of thrips in flowers inside the same cages as ((a). ((c) The average number

of A. swirskii (± s.e., all stages). ((d) The average proportion of damaged leaves (± s.e.) per week. Different letters

in the legend represent significant differences among treatments through time (contrasts with general linear

hypothesis testing after LME, P<0.05).

ControlC. lactis A. swirskii A. swirskii C. lactis

A. s

wirs

kii

A. swirskiiA. swirskii + C. lactis



els were significantly lower when A. swirskii was released than in the other two treat-ments (FIGURE 2.4d).

The proportion of flowers with heavy thrips damage varied significantly with treat-ment (GLM, F3,8 = 5.11, P = 0.029). The treatment with A. swirskii with alternativefood (12 ± 8.5% flowers damaged) differed significantly from the control (73 ±11.6%). The treatments with A. swirskii alone (32 ± 13.4%) and with alternative foodalone (61 ± 14.6%) did not differ significantly from the other treatments. No litter-inhabiting predators were found in this experiment, confirming that sterilizing the lit-ter and washing the roots of the plants before the experiment had resulted in theirexclusion.

DiscussionThe use of generalist predators for biological pest control has become a common-place (Janssen & Sabelis 2015; Symondson et al. 2002). Generalists have the advan-tage that their populations can be maintained in a crop when no pests are present.To date, this was mainly achieved by supplying alternative food on the above-groundplant parts (Adar et al. 2014; Delisle et al. 2015; Duarte et al. 2015; Kumar et al. 2015;Leman & Messelink 2015; Liu et al. 2006; Messelink et al. 2008; Nomikou et al. 2010,2002; Pijnakker et al. 2016; van Rijn et al. 2002). The current study capitalized on theuse of soil or litter-inhabiting arthropods as food for predators. For this strategy to beeffective, generalist predators need to forage in both habitats, which was confirmedhere for A. swirskii. This resulted in better biological control of thrips in the currentstudy, an above-ground pest which passes part of its life cycle in the litter.

A risk of supplying generalist predators with alternative prey is the occurrence ofpositive effects on pest densities (apparent mutualism; Holt 1977), for which we didnot find evidence. Supplementary food or prey can hamper biological control in theshort term (Koss & Snyder 2005; Prasad & Snyder 2006), but the presence or addi-tion of alternative food or prey can also lead to high densities of predators (Nomikouet al. 2010). Supplying alternative food in the litter has two advantages. First, noalternative prey or food needs to be added to the above-ground crop parts, which isthe marketable part in roses. Second, no pollen needs to be dusted on the plants,which can be risky because thrips can also feed on pollen (Chitturi et al. 2006; vanRijn et al. 2002; Vangansbeke et al. 2016).

Some studies show that the presence of alternative food on plants did not resultin yield loss or decrease of plant damage, even if there were decreases in pest num-bers and increases in predator numbers (Delisle et al. 2015; Jaworski et al. 2015). Incontrast, we found that adding alternative prey for A. swirskii to the litter resulted inlower damage of leaves and flowers. When A. swirskii and alternative prey werereleased before thrips infestations, a significant reduction of pest damage was

70


observed (FIGURE 2.4c, FIGURE S2.1.1). We therefore suggest that predators and alter-native food should be added to the crop before pest invasion, which will decreaseplant damage by thrips. Furthermore, instead of the common practice of putting thelitter under the benches with plants, growers could add some humid litter (humidityis essential for survival and development of the alternative prey) at the base of theplants and provide alternative prey in this litter to increase pest control.

A possible disadvantage of adding alternative prey to the litter is that predatorswill have to commute between the litter and the above-ground plant parts wherepests reside. In the case of western flower thrips, this is not a disadvantage becausethrips prepupae and pupae live mainly in the litter and A. swirskii can attack thesestages (K. Muñoz-Cárdenas & M.V.A. Duarte, pers. obs.). Another disadvantage ofadding alternative prey is that predators may refrain from attacking the pest when thealternative prey in the litter is of superior quality, as was the case here (FIGURE 2.3).Nevertheless, we found better thrips control with alternative prey in the litter (FIGURE

2.4a). Moreover, there were four times more predators on the above-ground plantparts in the treatment with alternative prey in the litter (FIGURE 2.4b). This suggeststhat the predators commuted from the litter to the plants and fed on both prey ratherthan concentrating their attacks on the superior prey in the litter. However, the alter-native prey was added at the beginning of the experiments, and their densities mayhave been reduced towards the end of the experiment, resulting in hungry predatorsattacking thrips. Clearly, the dynamics and the timing of release of the alternativeprey deserve further study, because adding high-quality alternative prey may resultin a temporal release of thrips from predation (short-term apparent mutualism; Holt1977; Abrams & Matsuda 1996).

Another mechanism by which predatory mite populations can increase whenfeeding on two different prey is through diet supplementation (Marques et al. 2015;Messelink et al. 2008). We did not find evidence for better performance of A. swirskiion a mixed diet of thrips and the alternative prey C. lactis than on C. lactis only(FIGURE 2.3). Instead, the juvenile survival and developmental rate was highest ondiets containing C. lactis, irrespective of the addition of thrips (FIGURE 2.3a). In agree-ment with this, we found the highest densities of predators when this alternative preywas present (FIGURE 2.4b). This confirms that C. lactis is a better food source for A.swirskii than are thrips larvae.

It is not obvious that plant-inhabiting predators such as A. swirskii move down tothe litter to feed on alternative prey. Buitenhuis et al. (2010) reported that A. swirskiimoved down to the soil to disperse; we found that they move to this habitat providedthere are astigmatic mites in the litter on which they can feed (FIGURE 2.1). Possibly,the presence of alternative prey in the litter caused A. swirskii to forage more there,also attacking thrips pre-pupae and pupae (K. Muñoz-Cárdenas & M.V.A. Duarte,

71


pers. obs.). Other litter-inhabiting predatory mites can also feed on these thrips stagesand on the alternative prey in the litter. Therefore, further experiments should assessthe effect of these litter-inhabiting predators on pest populations and their interactionswith plant-inhabiting predatory mites, especially because it has been shown thatintraguild predation between plant-inhabiting and litter-inhabiting predators can occur(Messelink & van Holstein-Saj 2011) and might disrupt biological control (Vance-Chalcraft et al. 2007; Rosenheim et al. 1995; but see Janssen et al. 2006).

In conclusion, our results demonstrate that links can be established betweenabove-ground plant pests and the litter food web and that such links can benefitplant-inhabiting predators, resulting in an increase of predator densities andenhanced biological control. This confirms the importance of considering connec-tions between above-ground and below-ground food webs associated with plants(A’Bear et al. 2014; van der Putten et al. 2001), also for applied purposes.

Acknowledgements

We thank Peter de Ruiter for constructive comments. K.M.C was supported byColciencias (Colombia) (Programa ‘Francisco José de Caldas’ 2011). Thanks toÖmer Üçerler for support. The authors declare no conflict of interests.

ReferencesA’Bear AD, Johnson SN, Jones TH (2014) Putting the `upstairs-downstairs’ into ecosystem service: What can

aboveground-belowground ecology tell us? Biol Control 75:97-107Abrams PA, Matsuda H (1996) Positive indirect effects between prey species that share predators. Ecology

77:610-616Adar E, Inbar M, Gal S, Gan-Mor S, Palevsky E (2014) Pollen on-twine for food provisioning and oviposition

of predatory mites in protected crops. BioControl 59:307-317Bezemer TM, van Dam NM, 2005. Linking aboveground and belowground interactions via induced plant

defenses. Trends Ecol Evol 20:617-624Birkhofer K, Wise DH, Scheu S (2008) Subsidy from the detrital food web, but not microhabitat complexity,

affects the role of generalist predators in an aboveground herbivore food web. Oikos 117:494-500Bolckmans KJF, van Houten YM (2006) Mite composition, use thereof, method for rearing the phytoseiid

predatory mite Amblyseius swirskii, rearing system for rearing said phytoseiid mite and methods for bio-logical pest control on a crop. WO Patent WO/2006/057552

Buitenhuis R, Murphy G, Shipp L, Scott-Dupree C (2015) Amblyseius swirskii in greenhouse production sys-tems: a floricultural perspective. Exp Appl Acarol 65:451-464

Buitenhuis R, Shipp L, Scott-Dupree C (2010) Dispersal of Amblyseius swirskii Athias-Henriot (Acari:Phytoseiidae) on potted greenhouse chrysanthemum. Biol Control 52:110-114

Calvert D, Huffaker C (1974) Predator [Metaseiulus occidentalis]—Prey [Pronematus spp.] interactions undersulfur and cattail pollen applications in a noncommercial vineyard. Entomophaga 19:361-369

Calvo FJ, Knapp M, van Houten Y, Hoogerbrugge H, Belda J (2015) Amblyseius swirskii: What made thispredatory mite such a successful biocontrol agent? Exp Appl Acarol 65:419-433

Chant DA, McMurtry JA (2007) Illustrated keys and diagnoses for the genera and subgenera of thePhytoseiidae of the world (Acari: Mesostigmata). Indira Publishing House, Michigan, USA

Chitturi A, Riley DG, Joost PH (2006) Effect of pine pollen on settling behavior of Frankliniella occidentalis andFrankliniella fusca (Thysanoptera: Thripidae) on tomato and peanut. Environ Entomol 35:1396-1403

72


Crawley MJ (2013) The R book. John Wiley & Sons, Chichester, Englandde Roos AM, McCauley E, Wilson WG (1998) Pattern formation and the spatial scale of interaction between

predators and their prey. Theor Popul Biol 53:108-130Delisle JF, Shipp L, Brodeur J (2015) Apple pollen as a supplemental food source for the control of western

flower thrips by two predatory mites, Amblyseius swirskii and Neoseiulus cucumeris (Acari: Phytoseiidae),on potted chrysanthemum. Exp Appl Acarol 65:495-509

Demite PR, de Moraes GJ, McMurtry JA, Denmark HA, Castilho RC (2014) Phytoseiidae database.www.lea.esalq.usp.br/phytoseiidae.

Duarte MV, Venzon M, Bittencourt MCS, Rodríguez-Cruz FA, Pallini A, Janssen A (2015) Alternative food pro-motes broad mite control on chilli pepper plants. BioControl 60:817-825

English-Loeb GM, Karban R, Hougen-Eitzman D (1993) Direct and indirect competition between spider mitesfeeding on grapes. Ecol Appl 3:699-707

Eubanks MD, Denno RF (2000) Health food versus fast food: the effects of prey quality and mobility on preyselection by a generalist predator and indirect interactions among prey species. Ecol Entomol 25:140-146

Gange A, Brown V (1989) Effects of root herbivory by an insect on a foliar-feeding species, mediated throughchanges in the host plant. Oecologia 81:38-42

Halaj J, Wise DH (2002) Impact of a detrital subsidy on trophic cascades in a terrestrial grazing food web.Ecology 83:3141-3151

Holt RD (1977) Predation, apparent competition, and structure of prey communities. Theor Popul Biol 12:197-229Hunt H, Coleman D, Ingham E, Ingham RE, Elliott E, Moore J, Rose S, Reid C, Morley C (1987) The detrital

food web in a shortgrass prairie. Biol Fertil Soils 3:57-68Janssen A, Montserrat M, HilleRisLambers R, de Roos AM, Pallini A, Sabelis MW (2006) Intraguild predation

usually does not disrupt biological control, in: Brodeur, J., Boivin, G. (Eds.), Trophic and Guild Interactionsin Biological Control, Progress in Biological Control. Springer, Dordrecht, The Netherlands, pp. 21-44

Janssen A, Sabelis M (2015) Alternative food and biological control by generalist predatory mites: the caseof Amblyseius swirskii. Exp Appl Acarol 65:413-418

Jaworski CC, Chailleux A, Bearez P, Desneux N (2015) Apparent competition between major pests reducespest population densities on tomato crop, but not yield loss. J Pest Sci 88:793-803

Koss AM, Snyder WE (2005) Alternative prey disrupt biocontrol by a guild of generalist predators. BioControl32:243-251

Kumar V, Xiao Y, McKenzie C, Osborne L (2015) Early establishment of the phytoseiid mite Amblyseiusswirskii (Acari: Phytoseiidae) on pepper seedlings in a Predator-in-First approach. Exp Appl Acarol65:465-481

Leman A, Messelink G (2015) Supplemental food that supports both predator and pest: A risk for biologicalcontrol? Exp Appl Acarol 65:511-524

Lenth R (2016) Lsmeans: Least-Squares Means. J Stat Softw 69:1-33Liu CZ, Yan L, Li HR, Wang G (2006) Effects of predator-mediated apparent competition on the population

dynamics of Tetranychus urticae on apples. BioControl 51:453-463Loomans A, Murai T (1997) Culturing thrips and parasitoids, in: Lewis T (Ed.), Thrips as Crop Pests. CAB

International, Wallingford, England, pp. 477-503Marques RV, Sarmento RA, Lemos F, Pedro-Neto M, Sabelis MW, Venzon M, Pallini A, Janssen A (2015)

Active prey mixing as an explanation for polyphagy in predatory arthropods: synergistic dietary effects onegg production despite a behavioural cost. Funct Ecol 29:1317-1324

McCann KS, Rasmussen JB, Umbanhowar J (2005) The dynamics of spatially coupled food webs. Ecol Lett8:513-523

Messelink G, van Holstein-Saj R (2011) Generalist predator Stratiolaelaps scimitus hampers establishment ofthe bulb scale mite. Proc Neth Entomol Soc Meet 22:67-73

Messelink GJ, van Maanen R., van Steenpaal SEF, Janssen A (2008) Biological control of thrips and white-flies by a shared predator: two pests are better than one. Biol Control 44:372-379

Messelink GJ, van Steenpaal SEF, Ramakers PJM (2006) Evaluation of phytoseiid predators for control ofwestern flower thrips on greenhouse cucumber. BioControl 51:753-768

73


Miyashita T, Takada M, Shimazaki A (2003) Experimental evidence that aboveground predators are sustainedby underground detritivores. Oikos 103:31-36

Moore JC, McCann K, Setala H, De Ruiter PC (2003) Top-down is bottom-up: Does predation in the rhizos-phere regulate aboveground dynamics? Ecology 84:846-857




Overmeer WPJ (1985) Rearing and handling, in: Helle W, Sabelis MW (Eds.), Spider Mites, Their Biology,Natural Enemies and Control. Elsevier, Amsterdam, The Netherlands, pp. 162-170

Pijnakker J, Arijs Y, de Sousa A, Wackers F (2016) The use of Typha angustifolia (cattail) pollen to establishthe predatory mites Amblyseius swirskii, Iphiseius degenerans, Euseius ovalis and Euseius gallicus inglasshouse crops. IOBC/WPRS Bull 120:47-52


Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team (2014) nlme: Linear and Nonlinear Mixed EffectsModels. http://CRAN.R-project.org/package=nlme.

Prasad RP, Snyder WE (2006) Diverse trait-mediated interactions in a multi-predator, multi-prey community.Ecology 87:1131-1137

R Development Core Team (2013) R: A language and environment for statistical computing. R Foundation forStatistical Computing, Vienna, Austria. http://www.R-project.org.

Rosenheim JA, Kaya HK, Ehler LE, Marois JJ, Jaffee BA (1995) Intraguild predation among biological controlagents – Theory and evidence. Biol Control 5:303-335

Scheu S (2001) Plants and generalist predators as links between the below- ground and above-ground sys-tem. Basic Appl Ecol 2:3-13

Settle WH, Ariawan H, Astuti ET, Cahyana W, Hakim AL, Hindayana D, Lestari AS (1996) Managing tropicalrice pests through conservation of generalist natural enemies and alternative prey. Ecology 77:1975-1988

SymondsonWOC, Sutherland KD, Greenstone MH (2002) Can generalist predators be effective biocontrolagents? Annu Rev Entomol 47:561-594

Therneau TM (2015) Coxme: Mixed Effects Cox Models. http://CRAN.R project.org/package=coxmevan Baalen M, Krivan V, van Rijn PCJ, Sabelis MW (2001) Alternative food, switching predators, and the per-

sistence of predator-prey systems. Am Nat 157:512-524van der Putten WH, Vet LEM, Harvey JA, Wackers FL (2001) Linking above-ground and belowground multi-

trophic interactions of plants, herbivores, pathogens, and their antagonists. Trends Ecol Evol 16:547-554van Maanen R, Messelink GJ, Van Holstein-Saj, Sabelis MW, Janssen A (2012) Prey temporarily escape from

predation in the presence of a second prey species. Ecol Entomol 37:529-535van Rijn PCJ, van Houten YM, Sabelis MW (2002) How plants benefit from providing food to predators even

when it is also edible to herbivores. Ecology 83:2664-2679Vance-Chalcraft HD, Rosenheim JA, Vonesh JR, Osenberg GW, Sih A (2007) The influence of intraguild pre-

dation on prey suppression and prey release: a meta-analysis. Ecology 88:2689-2696Vangansbeke D, Nguyen DT, Audenaert J, Verhoeven R, Gobin B, Tirry L, De Clercq P (2016) Supplemental

food for Amblyseius swirskii in the control of thrips: feeding friend or foe? Pest Manag Sci 72:466-473Wardle DA, Bardgett RD, Klironomos JN, Setälä H, van der Putten WH, Wall DH (2004) Ecological linkages

between aboveground and belowground biota. Science 304:1629-1633Wardle DA, Williamson WM, Yeates GW, Bonner KI (2005) Trickle-down effects of aboveground trophic cas-

cades on the soil food web. Oikos 111:348-358Yong A (2004) Técnicas de formación e manejo del rosal. Cultiv Trop 25:53-60

74


SUPPLEMENTARY MATERIAL 2.1

Thrips control with high densities of Amblyseius swirksiiWe tested whether releasing high densities of A. swirskii (1000 mites/m2 of green-house, commercial recommendation is 100 mites/m2), could satisfactorily controlthrips in roses. If this were the case, it would be worthwhile to investigate which typeof alternative food can be used to boost the densities of this predatory mite speciesto sufficiently high levels for thrips control.

Methods

Rose plants (var. Avalanche) were placed in insect-proof cages with a wooden frame(140 × 160 × 80 cm) in a greenhouse compartment at Koppert Biological Systems inBerkel en Rodenrijs, The Netherlands (22.2 ± 2.5 °C, RH 61.8 ± 10.6%). Each cagewas considered an experimental unit and contained six rose plants (Avalanche, c. 18weeks old) in pots (26 cm diameter, 25 cm high) with peat as potting substrate (50%coco peat, 15% white peat, 35% frozen black peat; Jongkind grond BV, Aalsmeer,The Netherlands). Plants were watered with a drip irrigation system. According tocommercial practice, rose stems with shoots and buds were bent to allow moreshoots to develop. Because there was some risk of previous contamination of theplants, around 10 Phytoseiulus persimilis (Spidex®, Koppert Biological Systems) and50 Encarsia formosa (En-strip®, Koppert Biological Systems) were released per plantto control spider mites and whiteflies, respectively.

Thrips were collected from the stock colony of Koppert Biological Systems. Torelease 1000 A. swirskii/m2, we took 10 ml of the commercial product Swirski-mite®(containing 50.000 predators / 500 ml of bran). In each cage, 130 adult thrips werereleased in the first and second week. There were two treatments: one with onlythrips and one with thrips and 1000 adult female A. swirskii, released in the thirdweek. There were two trials with two replicates (cages), each lasting for 30 days(March to May 2011). Plants were sampled destructively at the end of the experi-ment, All leaves in the lower and upper stratum were counted, as well as the budsand flowers of all the plants. We counted the thrips and mites on 15 leaves of theupper stratum and 15 leaves of the lower stratum and in all flowers and buds undera stereo-microscope.

The numbers of thrips in the flowers and on the leaves (√x transformed) were com-pared between treatments using LME, with treatment as fixed factor and trial as ran-dom factor. The distribution of the residuals was checked for normality and treat-ments were compared by combining factor levels.

75


Results

Significantly lower densities of thrips were found on the leaves in the cages withpredatory mites than in the cages with only thrips (average ± s.e. = 0.058 ± 0.039 vs.0.24 ± 0.10 thrips/leaf; LME: d.f. = 1, χ2 = 4.04, P = 0.044). The difference in the aver-age densities of thrips in the flowers and buds was more pronounced than that onleaves (with predators: 2.23 ± 0.59 thrips/flower, without predators: 13.2 ± 5.0; LME:d.f. = 1, χ2 = 6.12, P = 0.013). This also resulted in less damage to the petals of theflowers (FIGURE S2.1.1). We found 157.2 (± 55.4) A. swirskii on the plants per cage.

Thrips control by predatory mites with alternative preyTo further test the effect of adding alternative food on thrips densities (cf. Effect ofalternative prey on thrips densities in the main text), we performed a greenhouseexperiment. This experiment was perfomed on a larger scale than the one in the maintext. The plants used here were older and had been sprayed with chemicals. So, wewanted to test whether under these conditions (similar to commercial rose produc-tion) we could observe a similar effect of alternative prey on thrips populations as inthe experiment perfomed in laboratory conditions and presented in the main text.

Methods

Insect-proof cages (120 × 90 × 80 cm) were placed on tables in a greenhouse com-partment at Wageningen UR Greenhouse Horticulture (Bleiswijk, The Netherlands)from September to December 2012 (20 °C, RH 85%, photoperiod L18:D6 artificial

76


FIGURE S2.1.1. Effect of high densities of predators on thrips damage to rose petals. Left: petals from flowers of

plants treated with high densities of Amblyseius swirskii. Right: control treatment with only thrips.

light: 10.000 lux). Each cage was considered an experimental unit and containedthree 1-year-old rose plants (var. Red Naomi; Schreurs, De Kwakel, The Netherlands).Each plant was placed in a plastic dish (40 cm diameter), which was placed in a plas-tic tray. Plants were sprayed before the experiment, once with a 0.25% solution ofDodemorf (Metaltox, BASF) against powdery mildew (3 weeks before the experiment)and twice (3 and 2 weeks before) with a 0.028% solution of flonicamid (Teppeki,Belchim) to kill whiteflies present on the plants. The plants were rooted in rockwoolstrips. Stems with shoots and buds were bent to allow development of more shoots.Flowers that developed from the bent shoots near the soil were not removed butwere kept in order to facilitate the establishment of thrips.

As in the previous experiment, P. persimilis and E. formosa were released preven-tively. Thrips were collected from rose flowers in a greenhouse of the experimentalstation using a pipette tip (cf. Effect of alternative prey on thrips densities). The tipswere inserted into the substrate under the plants and the wide end was opened torelease the thrips. Amblyseius swirskii (Swirski-mite®, Koppert Biological Systems)were collected on sweet pepper leaf discs (2.5 cm diameter, 15 females/leaf disc),which were put on the litter layer. Litter was collected from a commercial greenhousein Stompwijk (The Netherlands). In an attempt to remove arthropods, litter was incu-bated in Berlese funnels for 7 days before the experiment. Subsequently, 90 g of lit-ter was moistened with 60 g of water and was added at the base of each plant.Tyrophagus putrescentiae were obtained from Koppert Biological Systems andadded to the litter (TABLE S2.1.1). There were four treatments: only thrips, thrips +alternative prey, thrips + A. swirskii, and thrips + A. swirskii + alternative prey. Eachtreatment was replicated three times. The schedule of thrips and mite releases isshown in TABLE S2.1.1.

Every week from the 6th until the 9th week, 10 leaves from the upper stratum, 10from the middle and 10 from the lower stratum were taken randomly per cage. Fromthese 30 leaves collected, we counted the leaves damaged by thrips for each cagebut we did not count the thrips or predators on the leaves. We checked that A.swirskii was present in all treatments in which they were released but they were notcounted. In the 10th week, all leaves were collected and the numbers of thrips andpredators were scored under the stereo-microscope. Additionally, the number ofleaves with thrips damage was scored. Flowers and buds were sampled as explainedin the main text (cf. Thrips control by Amblyseius swirskii with alternative prey).

The proportions of damaged leaves from the 6th to the 9th week were arcsin √xtransformed, and compared using LME as explained in the main text, with experi-mental unit (cage) as random factor and treatment and time as fixed factors.Proportions of damaged leaves from the destructive sampling in the 10th week wereanalysed using a GLM with a quasibinomial error distribution with treatment as fixed

77


factor. The numbers of thrips on leaves were √x transformed and analysed with aGLM with a Gaussian error distribution. The numbers of thrips in flowers and budswere analysed with a GLM with a Poisson error distribution. The numbers of preda-tors per cage were compared with a GLM with treatment as fixed factor and a quasi-Poisson error distribution.

Results

There was a significant effect of the treatments on the proportion of leaves with thripsdamage from the 6th until the 9th week (FIGURE S2.1.2a; LME: d.f. = 3, χ2 = 11.2, P =0.01), with the lowest proportion of damaged leaves in the treatment with A. swirskiiand alternative prey (FIGURE S2.1.2a). In the 10th week, the proportion of damagedleaves was lowest in the treatment of A. swirskii with alternative food but there wasno significant effect of treatments (FIGURE S2.1.2a; right-hand data points, GLM: F3,8

= 1.75, P = 0.23). Densities of thrips on leaves were low (FIGURE S2.1.2b) and did notdiffer significantly among treatments (GLM: F3,8 = 2.61, P = 0.12). In contrast, thenumbers of thrips in flowers and buds were high (FIGURE S2.1.2b) and differed amongtreatments (GLM: d.f. = 3, χ2 = 40.1, P<0.0001), with the lowest densities in the treat-ment with A. swirskii with alternative food (FIGURE S2.1.2b). There was a trend of high-er densities of A. swirskii per cage in the treatment in which the alternative prey wasadded than in the treatment in which A. swirskii was released alone (71.3 ± 27.4 vs.16.7 ± 4.8; GLM: F1,4 = 6.15, P = 0.068).

In all treatments, we found on average two litter-inhabiting predators from thefamily Laelapidae per open flower close to the litter. Presumably, these mites camefrom the litter collected in the commercial greenhouse, indicating that removing thelitter-inhabiting arthropods in the Berlese funnels was not successful. These preda-tors may have affected the control of thrips.

78


Table S2.1.1. Schedule of releases and samplings of the western flower thrips Frankliniella occidentalis, predato-

ry mites (Amblyseius swirskii) and alternative prey (Tyrophagus putrescentiae) when assessing the effect of adding

alternative prey.

Time (weeks) 1 2 3 4 5 6 7 8 9 10A. swirskiia 105 105 30F. occidentalisb 5+10 5+15 10+15 20+20 10+0T. putrescentiaec 100 300 3000 3000 3000Sampling √ √ √ √ √a Number of adult females released per week.b Numbers of a mixture of adults + juveniles released.c Number of mites released with bran, adult females were counted, immatures were also present but notcounted.

79


Figure S2.1.2. Effect of addition of the astigmatic mite Tyrophagus putrescentiae as alternative prey in the litter

layer under greenhouse conditions on the proportion of damaged leaves ((a). Each point is the average proportion

of damaged leaves per treatment in a given week. Standard errors are not shown to facilitate the visualization of

the trends of each treatment. The data from the 6th to the 9th weeks are based on samples of 30 leaves per

cage. Different letters in the legend represent significant differences among treatments for these weeks (contrasts

with general linear hypothesis testing after LME, P<0.05). The data of the 10th week (right of the dashed line) are

based on destructive sampling. Treatments: Control: thrips alone; T. putrescentiae: thrips + alternative prey; A.

swirskii: thrips + predatory mites; A. swirskii + T. putrescentiae: thrips + predators + alternative prey. ((b) The aver-

age (± s.e.) number of thrips on leaves (light gray) and in flowers and buds (dark gray) in the 10th week. Different

letters inside the bars indicate significant differences among treatments (contrasts with general linear hypothesis

testing after GLM, P<0.05).

Control

A. swirskii A. swirskii T. putrescentiae

T. putrescentiae

A. swirskii T. putrescentiae A. swirskiiT. putrescentiae

Alternative food for litter-inhabitingpredators decreases pest densities andabove-ground plant damage

Karen Muñoz-Cárdenas, Diana Rueda-Ramirez, Firdevs Ersin, Farid Faraji & Arne

Janssen

Biological control using generalist predators is successful in many crops. Such

predators feed on pests and food or alternative prey that do not cause economic

damage to the plants. Adding food for predators to a crop has a positive effect on

their numerical response. Thus, more predators are available to potentially attack

the pest, reducing their densities (apparent competition). However, in the short

term, the addition of alternative food may decrease the attack of pests because

the predators are satiated or because they preferentially feed on the alternative

food (apparent mutualism). Such positive effects of alternative food on pest densi-

ties may occur repeatedly when populations of pests or of alternative food fluctu-

ate. We investigated whether providing generalist litter-inhabiting predators with

alternative prey in the litter can increase control of thrips, Frankliniella occidentalis.

The larvae of this pest feed on above-ground plant parts, causing significant dam-

age, but prepupae and pupae live in the litter layer. Densities of thrips and damage

on above-ground plant parts were significantly reduced when supplying litter-

inhabiting predators, predominantly Cosmolaelaps n. sp., with alternative food.

Subsequently, Cosmolaelaps n. sp. was tested separately; this species alone also

reduced thrips densities and damage significantly when it was supplied with alter-

native food. Concluding, litter-inhabiting predators can significantly reduce above-

ground plant damage when supplied with alternative food in the litter layer.

83

3

IntroductionGeneralist predators can efficiently control pest populations in crops (Fleschner1959; Ramakers 1990; Symondson et al. 2002; van Rijn et al. 2002; Messelink et al.2008, 2010). They can act as a link between above-ground and below-ground foodwebs (Scheu 2001; Muñoz-Cárdenas et al. 2014, 2017) when they feed on above-ground plant pests but also on below-ground prey such as herbivores and detriti-vores (Settle et al. 1996; Muñoz-Cárdenas et al. 2017). Alternative foods allow gen-eralist predators to survive and reproduce when densities of the target pest are low(Settle et al. 1996; Wäckers et al. 2005; Liu et al. 2006), and they can also result inlower pest densities in the long term due to a negative indirect interaction betweenthe pest and the alternative food. In short, this interaction works as follows: the addi-tion of alternative food results in higher densities of the predators, resulting in lowerdensities of the pest (so-called apparent competition; Holt 1977). However, addingfood for predators to the crop can also result in higher pest densities (so-calledapparent mutualism; Holt 1977; Abrams & Matsuda 1996), and this is obviously detri-mental for biological control (van Maanen et al. 2012). The mechanism behind this isthat in the short term, predators might concentrate their attacks on the alternativeprey or become satiated by feeding on this prey, freeing the target pest from attacks.Concluding, adding alternative food for predators may not always result in betterpest control and this needs to be verified for each combination of predator and preyspecies and alternative food type.

Several studies in vegetable crops have demonstrated the benefits of additionalherbivorous prey or supplementary food for plant-inhabiting predators on biologicalcontrol of above-ground pests (Collyer 1964; Karban et al. 1994; van Rijn et al. 1999,2002; Liu et al. 2006; Messelink et al. 2008). In contrast, biological control in orna-mentals has been less successful because of the much lower damage thresholds.This could in theory be solved by adding more alternative food, resulting in higherpredator populations. The obvious danger of adding more alternative food are that itmay result in decreased pest suppression (apparent mutualism). Another problemcould be that consumers might not be prepared to accept a product with alternativefood or alternative prey on it, even when these do not damage the product. Thus, asolution would be supplying alternative food to increase populations of potential bio-logical control agents away from the marketable part of the crop, for example in thelitter or in the parts of the plant that are not sold.

A few studies have investigated the effect of adding alternative food below-ground to control pests on the above-ground plant parts. Some of these studiesshowed a positive effect on biological control, other studies showed negative effects,reminiscent of apparent mutualism (Settle et al. 1996; Birkhofer et al. 2008; Muñoz-Cárdenas et al. 2017). In a previous study, we show that a generalist predatory mite

84

CHAPTER 3 | ALTERNATIVE PREY FOR PREDATORS IN LITTER REDUCES CANOPY DAMAGE

[Amblyseius swirskii (Athias-Henriot)] can act as a link between above-ground andbelow-ground food webs by feeding on an above-ground plant pest [Frankliniellaoccidentalis (Pergande)] but making excursions to feed on alternative prey (astigmat-ic mites) below-ground (Muñoz-Cárdenas et al. 2017). Adding this alternative preyresulted in improved biological control of thrips and higher densities of predators onthe above-ground plant parts. In some experiments, we observed that predators thatnaturally inhabit the litter also fed on the alternative prey and on pupae of F. occiden-talis (western flower thrips), which occur in the litter (Muñoz-Cárdenas et al. 2017). Inthis paper, we investigate the suitability of these predators for biological control ofthrips in roses.

Some predatory mites from the soil and litter are known to play an important rolein controlling insect pests with edaphic stages such as thrips (Wu et al. 2014), espe-cially when they are released in combination with other biological control agents(Gillespie & Quiring 1990; Premachandra et al. 2003; Wu et al. 2016). In some cases,releases of commercially available soil-inhabiting predators have resulted in satisfac-tory reduction of thrips populations, but only when released in high numbers(Wiethoff et al. 2004). Such high densities of soil-inhabiting predators could also beobtained by adding alternative prey to the soil or litter, thus increasing predator pop-ulations. However, the effect of adding alternative prey on the population dynamicsof litter-inhabiting predators and above-ground pests has not been studied. The goalof this study is to assess whether biological control of thrips can be increased bysupplying litter-inhabiting predators with alternative prey. We collected litter withpredatory mites from a commercial greenhouse and tested the two most commonpredatory mites for their capacity to feed on the alternative prey and on thrips.Subsequently, two experiments were done to assess the capacity of the predators tocontrol thrips damage on rose plants; one experiment with multiple predatory mitespecies and one with the most promising species.

Materials and methodsThe experimental systemWe studied rose plants, which are often damaged by the western flower thrips (Parket al. 2002; Manners et al. 2013). Above-ground plant parts of many crops are affect-ed by this thrips species, which spends part of its life cycle in the soil or litter (Tom -ma sini & Maini 1995; Lewis 1997). The alternative prey used here was Carpoglyphuslactis (L.), an astigmatic mite commonly found in the litter of commercial rose cropsin The Netherlands (Muñoz-Cárdenas et al. 2017). These mites are not considered apest in roses and are used for mass-rearing of predatory mites (Ramakers & vanLieburg 1982; Bolckmans & van Houten 2006).

85

ALTERNATIVE PREY FOR PREDATORS IN LITTER REDUCES CANOPY DAMAGE | CHAPTER 3

CulturesRose plantlets were bought when they were 4 weeks old and had 4-5 leaves (Rosasp. var. Avalanche) (Olij Rozen Int. BV, De Kwakel, The Netherlands). Because theseplants had been kept in a commercial greenhouse, there was a risk that they werecontaminated with pests. In order to eliminate these, we removed the leaves andwashed the stems and roots under running tap water. Subsequently, the remainingplant parts were dried and planted in clean peat. They were allowed to regrow for4 weeks in a greenhouse compartment (22 °C, RH 70%, photoperiod L14:D10) insideinsect-proof cages to avoid new contaminations. The regrown plants were inserted inrock wool strips (Grodan®Delta: 40 × 6 × 28), were watered twice a week, and macro-and micronutrients were applied dissolved in water once a week (0.5 g N-P-K and 0.5 gmicronutrients mix/200-500 ml of water/plant). Before the first experiment, P. per -similis (Spidex®, Koppert Biological Systems, Berkel en Rodenrijs, The Nether lands)was preventively released (20 per cage) to avoid contamination with spider mites.

Frankliniella occidentalis was taken from the stock colony of Koppert BiologicalSystems. They were reared (at 23 °C, RH 65%) using a method adapted fromLoomans & Murai (1997). In short, thrips were reared on fresh bean pods inside plas-tic containers (8 × 8 × 12 cm) with openings covered with mite-proof mesh (30 μm)for ventilation. Around 200 adult thrips were introduced in a rearing unit and allowedto oviposit in two bean pods for 1 or 2 days to obtain cohorts to be used in the exper-iments. Before use, the pods were washed with soap, rinsed and dried to removeresidues of pesticides. Typha latifolia pollen was provided as food for the thrips. Thispollen was collected during spring from plants growing on the campus of theUniversity of Amsterdam. Carpoglyphus lactis were provided by Koppert BiologicalSystems and were reared as explained in Muñoz-Cárdenas et al. (2017), inside cylin-drical plastic containers (8 cm diameter, 12 cm high) to which 5 g of bran plus 5 g ofyeast were added as food once per week. To facilitate the release of C. lactis inexperiments, we counted C. lactis per mg of the mixture of bran and yeast from therearing unit and added the appropriate weight of the mixture to the experimentalunits to end up with the desired numbers of C. lactis females per experimental unit.

Litter-inhabiting predators that benefit from alternative preyIn previous experiments, we observed that predators inhabiting the moss and litter incommercial greenhouses might play a role in controlling thrips when alternative food(C. lactis) was added to the litter (Muñoz-Cárdenas et al. 2017). For this reason, wedid not collect the mite fauna directly from the samples taken in the commercialgreenhouses, but we selected mites that could reproduce on C. lactis. To this end,we collected five samples of 1.5 l of various materials from different rose greenhous-es in The Netherlands: 1) litter covering the soil in Amstelveen in October, 2011 and

86


87


2) in December, 2013; 3) litter from Stompwijk, July 2012; 4) moss growing on thesurface of rock wool strips where plants were grown from Aalsmeer, January 2013;and 5) old, open flowers from Amstelveen, July 2013. These latter flowers are non-marketable, and their stems are bent as a common practice. When those flowersdevelop they can be found touching the litter layer, thus serving as bridges betweenthe above-ground plant parts and the litter. Samples were placed on top of humidrock wool (Grodan®Delta: 28 × 6 × 28) in a plastic tray and spread c. 100 g of food(c. 4000 C. lactis females plus bran) on top of the samples. Each tray was placedinside a mite-proof cage (47.5 × 47.5 × 92 cm) in a greenhouse compartment. After10 days, we collected each sample and placed it in a Berlese funnel for 5 days.

The predatory mites were collected in vials with alcohol 70% for taxonomic iden-tification. Most predators were found in the litter samples; only few were found in themoss and open flowers, we therefore refer to them as litter-inhabiting predators.Predatory mites were mounted on glass slides with Hoyer´s medium. Subsequently,they were observed under a phase contrast microscope and taxonomic keys wereused for identification (Krantz & Ainscough 1990; Lindquist et al. 2009). Adult femaleswere identified to species, based on the original and complementary descriptions. Anextra sample (1.5 l of litter) was taken from the greenhouse in Amstelveen inDecember 2013 in order to extract live mites to set-up cultures of the most commonpredatory mites (Cosmolaelaps n. sp. and Proctolaelaps pygmaeus) in the laborato-ry (22 °C). Rearing units consisted of plastic containers (125 ml) with humid plasterof Paris mixed with coal at the bottom (Lesna et al. 2009). Peat mixed with auto-claved rose litter was added on top of the plaster. We introduced 50 adult females ofCosmolaelaps n. sp. or P. pygmaeus to each rearing unit. The plastic containers withthe colonies were placed inside a tray with water to increase humidity, the tray wascovered with another plastic tray to shield from direct light. A mix of bran with freshyeast and C. lactis (30 g) were added to each container as food for the predatorsonce per week. We collected around 10 individuals from the colonies, three timesduring this study for identification.

Predation of Frankliniella occidentalis and Carpoglyphus lactis byCosmolaelaps n. sp. and Proctolaelaps pygmaeusThis experiment was performed to assess whether the litter-inhabiting predatory mitesCosmolaelaps n. sp. and P. pygmaeus feed on thrips and the alternative prey C. lactis.Twenty P. pygmaeus females and 30 Cosmolaelaps n. sp. females were taken from therearing units and starved individually by placing each in a vial (3 cm diameter, 4 cmhigh) with a moistened bottom of plaster of Paris mixed with charcoal (Lesna et al.2009). The next day, we added 10 protonymphs of C. lactis to the vials of one group offemales of each species (10 females of P. pygmaeus and 15 females of Cosmolaelaps

n. sp.), and we added three thrips pre-pupae to the vials of second groups of preda-tors (same numbers as above). We added pieces of dried leaves to simulate the struc-ture of the litter. One day later, we scored how many individuals of each prey wereeaten. This experiment was carried out at 25 °C, RH 60%, and photoperiod L16:D8.

Thrips control by litter-inhabiting predatorsLitter was collected in plastic bags from a commercial greenhouse (Amstelveen, April,2014) and was kept in a cold room (4 °C) until the beginning of this experiment in May,2014. This litter harboured different predator species inhabiting the commercialgreenhouse. In a greenhouse compartment at the University of Amsterdam, experi-mental units consisted of two rose plants in rock wool in a mite-proof cage (47.5 ×47.5 × 92 cm). During the first week of the experiment, 1 l of litter moistened with 100ml of water was placed at the foot of the rose plants. Around 2 h later, c. 100 g of branwith alternative prey (c. 4000 C. lactis females) was dispersed on the litter of three ofthe experimental units; six other cages received bran without alternative prey. Thethree treatments (predators with or without alternative prey and the control with onlythrips) were replicated three times (three cages). Subsequently, for the treatment withalternative prey, 30 adult females of Cosmolaelaps n. sp. were transferred from thecolonies to plastic Petri dishes with bran and C. lactis with a fine brush. For the treat-ment without alternative prey, three Petri dishes were prepared with only bran and 30Cosmolaelaps. For the control treatment, we removed all the visible predatory mitesfrom the litter with an aspirator and we prepared Petri dishes with only bran. The con-tent of the Petri dishes was subsequently dispersed on the litter inside the threecages with alternative prey and in six cages without alternative prey. In the secondweek, the rose plants were infested with 60 adult thrips/cage, which were collectedfrom the colonies using an aspirator made of a disposable polypropylene pipette tipcovered at the wide end with a piece of gauze (mesh 30 μm) (Muñoz-Cárdenas et al.2017). This wide end was connected with a long, flexible piece of tubing to a pump.Thrips were sucked up through the tip, which was subsequently closed withParafilm® (Sigma-Aldrich, The Netherlands). To release the thrips, we taped a pieceof yarn to the pipette tip and suspended it from a branch of the plant. Subsequently,the Parafilm was removed and thrips were free to infest the plant.

From the third to the tenth week, five leaves were collected from each of three plantstrata (higher, medium and lower) and thrips larvae and adults were counted using astereo-microscope. The total number of leaves and the number of leaves with thripsdamage were also recorded. As soon as the plants produced flowers (after 4 weeks),these were collected once every 2 weeks, until the tenth week, placed in plastic con-tainers with alcohol (70%), washed with 70% alcohol on a mesh (100 μm) and the thripsremaining on the mesh were counted under a stereo-microscope. Flowers were also

88


89


checked for damage under the stereo-microscope. We recorded a flower as damagedwhen it had more than five scars on petals and sepals (each scar of around 2 mm diam-eter). We selected this number and size of scars because previous studies showed thatthis area of damage is not visible to the naked eye (Muñoz-Cárdenas et al. 2017). Thepresence of astigmatic mites and predators in the litter was confirmed by checking50 ml of humid litter with bran from each cage under a stereo-microscope every2 weeks (from the fourth to the tenth week), but mites were not counted. On the tenthweek we identified up to 10 predatory mite adults from the litter samples per cage.

The average numbers of thrips per leaf were log(x+0.1) transformed and analysedwith a linear mixed effects model (package LME of R; Pinheiro et al. 2014; RDevelopment Core Team 2015) with replicate (cage) as random factor and time andtreatment as fixed factors. The significance of factors and their interaction was deter-mined by comparing models with and without them with the anova function of R.Contrasts between treatments were assessed with a general linear hypothesis test (glhtfunction of the package lsmeans of R; Lenth 2016) with the Tukey method for correctionof multiple comparisons. The proportions of damaged flowers and the average numbersof thrips per flower were √x transformed. The proportions of damaged leaves wereanalysed using a similar model as above on untransformed proportions.

Thrips control by Cosmolaelaps n. sp.For this experiment, we excluded predators from the litter (from a commercial green-house in Amstelveen, September 2014) by sterilizing it (108 °C, ± 1 bar pressure, 20min), thus avoiding possible negative effects of competition among predator specieson biological control due to competition or intraguild predation among predatorspecies (Rosenheim et al. 1995; Messelink & van Holstein-Saj 2011). We usedCosmolaelaps n. sp. as litter predator because it inhabits the litter in greenhousesand it feeds on prepupae of thrips. Moreover other species of this genus have alsoshown potential for thrips control (Furtado et al. 2015). This experiment was per-formed from September to November 2014 (22 °C, RH 85%, photoperiod L16:D8) atthe University of Amsterdam. There were three treatments: Cosmolaelaps n. sp. withalternative prey (C. lactis + bran in sterilized litter), Cosmolaelaps n. sp. without alter-native prey (only bran in sterilized litter), and a control with only thrips (60/cage) withbran in sterilized litter. Each treatment was replicated three times (cages).

Besides the sterilization of the litter, the experimental set-up and the samplingswere similar to those in the experiment above. The alternative prey and the predatorswere released as above, for the treatment with alternative food (c. 100 g of bran +4000 C. lactis females + 30 Cosmolaelaps n. sp.) and for the treatment without alter-native food we used bran without C. lactis; thrips were released as above (secondweek, 60 thrips/cage). Sampling was similar as above, but we sampled nine leaves

per week per cage, three leaves per stratum (higher, medium and lower plant parts).As soon as the plants produced flowers (week 4), these were sampled every week.Here, we identified and counted predators per 50 ml of litter sample every week. Wechecked for the presence of C. lactis but we did not count the mites.

The average numbers of thrips and damage on leaves and flowers were analysedas above. The numbers of predators per sample were log(x+1) transformed andanalysed as above for thrips.

ResultsLitter-inhabiting predators that benefit from alternative preyWe found two species of thrips predators benefiting from the alternative prey added tothe greenhouse litter. Cosmolaelaps n. sp. naturally occurs in the litter layer in rosegreenhouses and was found in high percentages in all samples (TABLE 3.1).Stratiolaelaps scimitus was found in low numbers in two samples (TABLE 3.1), this pred-ator is commercially available in The Netherlands, possibly, it was present because ofprevious releases by the grower (K. Muñoz-Cárdenas, pers. comm.). The other speciesencountered are not known to feed on thrips but on fungi, nematodes, collembolans,mites, insects other than thrips, or different combinations of these groups (TABLE 3.1).

Predation of Frankliniella occidentalis and Carpoglyphus lactis byCosmolaelaps n. sp. and Proctolaelaps pygmaeusCosmolaelaps n. sp. fed on C. lactis immatures (average ± s.e.: 8.2 ± 0.4 per day) andF. occidentalis pre-pupae (1.3 ± 0.2 per day). Proctolaelaps pygmaeus did not feed onF. occidentalis pre-pupae but it fed on C. lactis immatures (3.5. ± 0.5 per day).

Thrips control by litter-inhabiting predatorsThrips densities on leaves differed significantly through time among treatments(FIGURE 3.1a; LME: interaction of treatment with time: χ2 = 25.5, d.f. = 2, P<0.0001).This interaction was caused by the numbers of thrips increasing through time with-out alternative food and in the control and remaining low when predators were sup-plied with alternative food. Overall, the two treatments without alternative prey dif-fered significantly from the treatment with alternative prey (FIGURE 3.1a; contrastsafter LME). Thrips densities in the flowers did not differ significantly amongtreatments (FIGURE 3.1b; LME: χ2 = 4.8, d.f. = 2, P = 0.09). There was no significanteffect of the interaction of treatment with time (LME: χ2 = 1.97, d.f. = 2, P = 0.37).

The proportion of damaged leaves varied significantly through time (LME: χ2 =5.70, d.f. = 2, P = 0.17), but there was no significant effect of the interaction oftreatment and time (LME: χ2 = 2.22, d.f. = 2, P = 0.33; FIGURE 3.1c). The proportionsof damaged leaves differed significantly among treatments (LME: χ2 = 26.7, d.f. = 2,

90


91


P<0.0001). The proportions of damaged leaves were lower in the treatment withalternative prey than in the two other treatments (FIGURE 3.1c; contrasts after LME).

The proportion of damaged flowers differed significantly among treatments (LME:χ2 = 13.7, d.f. = 2, P = 0.0011). The proportion of flowers with thrips damage was

TABLE 3.1. Relative abundance and number of predatory mite species found in litter, old flowers and moss of a

rose greenhouse, after offering astigmatic mites as food for 10 days.

Habitat/Samplea Species %b Feeding habits (reference)c

Litter/A Benoinyssus sp. 9.3 Fungi / Yeast (1)Cosmolaelaps n. sp. 14.8 Collembolans / Insects: Thysanoptera / Mites

Nematodes (2)Gaeolaelaps queenslandica 8.6 Nematodes (3)Gamasellodes sp. 1.9 Collembolans / Mites / Nematodes (4)Lasioseius subterraneus 29.0 Collembolans / Insects: Hemiptera and Diptera /

Mites / Nematodes (5)Proctolaelaps pygmaeus 32.7 Fungi / Mites / Nematodes (5)Robustocheles mucronata 3.7 No info foundTotal (N) 162

Litter/B Ameroseius pseudoplumosus 6.7 Fungi / Mites (6)Cosmolaelaps n. sp. 40.0Lasioseius floridensis 6.7 Fungi / Mites (5,7)Proctolaelaps pygmaeus 6.7Stratiolaelaps scimitus 40.0 Collembolans / Insects: Thysanoptera, Diptera,

Lepidoptera / Mites / Nematodes (2,8)Total (N) 15

Litter/C Benoinyssus sp. 1.2Cosmolaelaps n. sp. 50.9Lasioseius subterraneus 9.0Proctolaelaps pygmaeus 38.3Robustocheles mucronata 0.6Total (N) 167

Moss/D Cosmolaelaps n. sp. 100.0Total (N) 7

Flowers/E Cosmolaelaps n. sp. 55.6Stratiolaelaps scimitus 33.3Proctolaelaps pygmaeus 5.6Ameroseius pseudoplumosus 5.6Total (N) 18

a Plant material collected in commercial greenhouses in The Netherlands (A: Litter, Amstelveen, October2011; B: Litter, Stompwijk, July 2012; C: Litter, Amstelveen, December 2013; D: Moss, Aalsmeer, January2013. E: Old flowers, Amstelveen, July 2013). b Relative abundance (%) calculated as the proportion of each species in each sample, based on the total(N) given for each sample. c References for feeding habits for each genus: (1) Abou-Awad et al. 2008, (2) Furtado & de Moraes 2015,(3) Salehi et al. 2014, (4) Evans 1987, (5) de Moraes et al. 2015, (6) Flechtmann 1985, (7) Britto et al. 2012,(8) Wu et al. 2014.

lower in the treatment with alternative prey than in the control treatments (FIGURE 3.1d;contrasts after LME). The proportion did not vary significantly through time (LME: χ2 =1.54, d.f. = 2, P = 0.21; FIGURE 3.1d) and there was no significant effect of the interac-tion of treatment and time on flower damage (LME: χ2 = 2.38, d.f. = 2, P = 0.30).

In week 2, we observed low numbers of the astigmatic mite Tyrophagus putres-centiae in all treatments, this mite can also serve as alternative prey for predators.After 8 weeks, we found unidentified predatory mites in the control (without preda-tors), but these were not encountered in the tenth week. From the samples of preda-tory mites taken in the last (tenth) week (maximum 10 individuals), we found mainlyCosmolaelaps n. sp. and P. pygmaeus in the treatment with alternative food (average± s.e.: 6 ± 1.15 and 3 ± 0.58, respectively) and in the treatment without alternativefood (1.33 ± 0.33 and 0.66 ± 0.66). The numbers of other predatory mite specieswere lower; these were not identified.

Thrips control by Cosmolaelaps n. sp.The results with Cosmolaelaps n. sp. as the only predator were comparable to thosedescribed in the previous experiment. The effect of treatment on the average num-

92


FIGURE 3.1. Effect of addition of Carpoglyphus lactis as alternative prey for litter-inhabiting predators on ((a) the

average numbers of thrips on leaves (± s.e.); ((b) the average numbers of thrips in the flowers (± s.e.); ((c) the aver-

age proportion of damaged leaves (± s.e.) and ((d) the average proportion of damaged flowers per week (± s.e.).

Predators were released in the first week, thrips were released in the second week, leaf sampling started in the

third week and flower sampling in the fourth week. Different letters indicate significant differences among treat-

ments (contrasts with general linear hypothesis testing after LME, P<0.05).

C. lactis

93


bers of thrips per leaf varied significantly through time (FIGURE 3.2a; LME: interactionof treatment with time: χ2 = 13.3, d.f. = 2, P<0.0013). This interaction occurredbecause the densities of thrips remained low in the treatment with alternative foodbut increased in the other treatments, which differed significantly from the former(FIGURE 3.2a). The effect of treatments on the average numbers of thrips in the flow-

FIGURE 3.2. Effect of Cosmolaelaps n. sp. with or without alternative prey (Carpoglyphus lactis) on ((a) the average

(± s.e.) number of thrips on leaves; (± s.e.); ((b) the average numbers of thrips in the flowers (± s.e.); ((c) the aver-

age proportion of damaged leaves (± s.e.); ((d) the average proportion of damaged flowers per week (± s.e.) and

(e) the average number of Cosmolaelaps n. sp. adults in litter samples (± s.e.). Predators were released in the first

week, thrips were released in the second week, leaf sampling started in the third week and flower sampling in the

fourth week. Different letters indicate significant differences among treatments (contrasts with general linear

hypothesis testing after LME, P<0.05).

CosmolaelapsCosmolaelaps +C. lactis

Cos

mol

aela

ps

ers varied significantly through time (FIGURE 3.2b; LME: interaction of treatment withtime: χ2 = 7.4, d.f. = 2, P<0.02). This interaction occurred because the densities ofthrips remained low in the treatments with Cosmolaelaps n. sp. but increased signif-icantly in the treatment without predators (FIGURE 3.2b).

The proportions of damaged leaves varied differently through time among treat-ments (FIGURE 3.2c; LME: interaction of treatment with time: χ2 = 19.9, d.f. = 2,P<0.0001). It was significantly lower in the treatment with alternative food than in thetwo other treatments (FIGURE 3.2c). There was also a significant effect of treatmentson flower damage (FIGURE 3.2d; LME: χ2 = 14.0, P<0.001). Releasing Cosmolaelapsn. sp. independently of the addition of alternative prey resulted in flowers with lessthrips damage. The proportion of damaged flowers varied significantly through time(LME: χ2 = 4.26, d.f.= 2, P = 0.039; FIGURE 3.2d), and there was no significant effectof the interaction between treatment and time (χ2 = 0.49, d.f. = 2, P = 0.78).

The addition of alternative prey significantly increased the densities of predators(FIGURE 3.2e; LME: χ2 = 20.5, d.f. = 1, P<0.0001), which did not vary significantlythrough time (χ2 = 1.57, d.f. = 1, P = 0.21) and there was no significant effect of theinteraction of treatment with time (χ2 = 0.74, d.f. = 1, P = 0.39). We did not find otherspecies of predators in these samples. We found C. lactis in all replicates in which itwas released.

DiscussionRose is a prime example of an ornamental that must be damage-free to be mar-ketable (Parrella et al. 1999). Introducing generalist predators has been a successfulmethod of biological control (Symondson et al. 2002), mainly in vegetable crops. Oneway of improving biological control to meet the standards for ornamentals is to boostthe densities of predators in the crop before pest invasion. Nowadays, growers addfood to above-ground plant parts to increase the densities of various generalist pred-ators (Vangansbeke et al. 2014). Studies have shown that adding food or prey canenhance biological control (van Rijn et al. 2002; Nomikou et al. 2002, 2010; Janssen& Sabelis 2015; Kumar et al. 2015). These studies, however, concerned above-ground food webs. Only few studies show that biological control can be improved bylinking food webs inhabiting the litter to above-ground food webs (Muñoz-Cárdenaset al. 2017). Here, we show that adding food for predators in the litter reduces pestdamage on above-ground plant parts.

We found different species of predators in the litter samples collected in commer-cial greenhouses after adding alternative prey (TABLE 3.1). The dominating specieswas Cosmolaelaps n. sp. (TABLE 3.1), from a genus that is known to occur in litter ofnatural forests (Furtado & de Moraes 2015). Another species that was abundant infour of the five samples was P. pygmaeus (TABLE 3.1), which has been reported from

94


soil in general and in association with plants, mammals, bark beetles and their gal-leries, birds, insects, rotting fruits, stored products and bird and mammal nests (deMoraes et al. 2015). It is possible that the different predatory mites in the litter of thegreenhouse came from the soil or were carried by insects invading the crop. In thelitter, these predatory mites might feed on other arthropods and fungi that are com-monly associated with the decomposition of litter (Walter & Proctor 2013). We foundthat both species fed on the alternative prey C. lactis but only Cosmolaelaps n. sp.fed on western flower thrips. This coincides with findings of other authors (de Moraeset al. 2015).

We investigated the effect of Cosmolaelaps n. sp plus alternative food on thripspopulations because we previously observed that these predators benefit from thealternative prey C. lactis and we suspected that they might play a role in thrips con-trol (Muñoz-Cárdenas et al. 2017). Another reason for investigating Cosmolaelaps n.sp. is that plants that were left in the greenhouse outside cages and that harboredCosmolaelaps n. sp. predators were less likely to become contaminated by thrips (K.Muñoz-Cárdenas, pers. obs.).

The various predator species present in the litter in the experiment ‘Thrips controlby litter-inhabiting predators’ could have engaged in negative predator-predatorinteractions such as competition or intraguild predation, which could have affectedbiological control (Rosenheim et al. 1995; Messelink & van Holstein-Saj 2011; but seeJanssen et al. 2006). However, thrips densities on leaves were lower when multiplepredators plus alternative prey were present than in the controls without predators orwithout alternative food (FIGURE 3.1). As was found here, several studies show thatthe effect of the presence of multiple predators on pest control can be neutral or pos-itive (Gillespie & Quiring 1990; Losey & Denno 1998; Sokol-Hessner & Schmitz 2002;Wiethoff et al. 2004; Thoeming & Poehling 2006; Snyder et al. 2008; Wu et al. 2016);however, none of these publications used alternative food. Yet, other studies shownegative effects of the presence of multiple predators on pest control (Rosenheim etal. 1993, 2004; Rosenheim 2001; Schausberger & Walzer 2001). In another study, wealso studied the effect of multiple predators on pest control, showing that releasingthe predatory mites A. swirskii on the plants and M. robustulus in the litter, each withtheir own alternative food, also resulted in a positive or neutral effect on thrips con-trol (CHAPTER 4).

Because the experiments of population dynamics of thrips with several predatorspecies and with Cosmolaelaps n. sp alone were carried out at different times forlogistical reasons, assessing the effect of single or multiple predators on thrips con-trol is not straightforward. To nevertheless compare thrips control with multiple andsingle predator species, we calculated the proportional decrease of thrips densitiesby predators with alternative food relative to the treatment without predators for both

95


experiments. Thus, we compared two time series with relative changes of the thripspopulations. The two time series did not differ significantly for thrips found in theflowers (glm with Gaussian error distribution: F1,9 = 0.007, P = 0.93) of for the thripson leaves (F1,14 = 0.79, P = 0.38), suggesting that Cosmolaelaps n. sp. alone was asefficient in reducing thrips densities as it was together with the other predatorspecies. Further research is needed to assess the effects of interactions among soilpredators on biological control.

In both experiments reported here, thrips densities decreased when adding foodto a single predator or to a community of predators in the litter. Leaf damage wasalso reduced compared with the control and a treatment without alternative prey(FIGURE 3.2). We did not observe short-term increases of thrips densities (as in appar-ent mutualism), perhaps because predators were released before the thrips. Inanother study, where thrips were released before the predators, we found a short-term negative effect when adding alternative food for thrips, as in apparent mutual-ism (CHAPTER 4).

In the current experiment, there were high densities of thrips in the flowers caus-ing damage in all treatments with Cosmolaelaps n. sp., and this is not acceptable forcommercial growers. Hence, further studies are needed to investigate the possibili-ties of optimizing control methods with this predator species. Moreover, it is worth-while to investigate the potential of adding alternative food for Cosmolaelaps n. sp.to control thrips in crops with higher damage tolerance, but also for the control ofother pests with stages living in the soil. It is estimated that around 75% of all insectpests inhabit the litter or soil at some stage of their lives (Wahab 2010) and havinghigh densities of Cosmolaelaps n. sp. as a result of adding alternative prey couldincrease pest control.

Another avenue to increase control of thrips in roses would be to combineCosmolaelaps n. sp. with canopy-dwelling predators plus alternative food. In thecase of the plant-dwelling predator A. swirskii, there can be negative interactionswith Cosmolaelaps n. sp., such as competition or intraguild predation that couldaffect thrips biological control. This topic deserves further study; however, we expectthat the simultaneous use of these two species would not affect thrips control basedon a study we performed in which we released A. swirskii together with anotherspecies of litter-inhabiting predator (CHAPTER 4).

Summarizing, we showed that adding alternative food in the litter increased thecontrol of an important pest species by litter-inhabiting predators. Most studies offood webs associated with plants focus on above-ground and below-ground com-munities that are linked via the plant (Scheu 2001), for example, how below-groundherbivores or mycorrhizae modify plant quality, thus affecting the performance ofabove-ground herbivores (Masters et al. 1993; Gange & West 1994; Brussard 1998;

96


Mortimer et al. 1999; A'Bear et al. 2014). Less attention has been given to herbivoreswith stages that feed on above-ground plant parts and other stages that occurbelow-ground (Johnson et al. 2016). Here, we show that such herbivores can form adirect link between above-ground and below-ground food webs. We suggest thatadding alternative prey for litter-inhabiting predators could also increase biologicalcontrol of other pests that inhabit the litter or soil at some stage of their lives, also inother crops.

Acknowledgements

K.M.C and D.R.R. were supported by Colciencias (Colombia) (Programa ‘FranciscoJosé de Caldas’ 2011). The authors declare no conflict of interests.

ReferencesA'Bear AD, Johnson SN, Jones TH (2014) Putting the 'upstairs-downstairs' into ecosystem service: What can

aboveground-belowground ecology tell us? Biol Control 75:97-107Abou-Awad BA, El-Sawaf BM, Hassan MM, Abdel-Khalek AA (2008) Comparative morphological and biolog-

ical studies of two eupodid mites: Benoinyssus momeni and Eupodes bakeri. Acarologia 48:47-54Abrams PA, Matsuda H (1996) Positive indirect effects between prey species that share predators. Ecology

77:610-616Birkhofer K, Wise DH, Scheu S (2008) Subsidy from the detrital food web, but not microhabitat complexity,

affects the role of generalist predators in an aboveground herbivore food web. Oikos 117:494-500Bolckmans KJF, van Houten YM (2006) Mite composition, use thereof, method for rearing the phytoseiid

predatory mite Amblyseius swirskii, rearing system for rearing said phytoseiid mite and methods for bio-logical pest control on a crop. WO patent WO/2006/057552

Britto EP, Gago E, de Moraes GJ (2012) How promising is Lasioseius floridensis as a control agent ofPolyphagotarsonemus latus? Exp Appl Acarol 56:221-231

Brussaard L (1998) Soil fauna, guilds, functional groups and ecosystem processes. Appl Soil Ecol 9:123-135Collyer E (1964) The effect of an alternative food supply on the relationship between two Typhlodromus

species and Panonychus ulmi (Koch) (Acarina). Entomol Exp Appl 7:120-124de Moraes GJ, Venancio R, dos Santos VLV, Paschoal AD (2015) Potential of Ascidae, Blattisociidae and

Melicharidae (Acari: Mesostigmata) as biological control agents of pest organisms. In: Carrillo D, deMoraes GJ, Peña JE (Eds.) Prospects for biological control of plant feeding mites and other harmfulorganisms. Springer International Publishing, Cham, Switzerland, pp. 77-102

Evans D (1987) Life history, trophic behavior, and description of Gamasellodes vermivorax n. sp.(Mesostigmata: Ascidae), a predator of nematodes and arthropods in semiarid grassland soils. Can J Zool65:1689-1695

Flechtmann CAH (1985) On the biology of Ameroseius dendrovagans (Acari, Mesostigmata, Ameroseiidae).Rev Bras Zool 2:397-399

Fleschner CA (1959) Biological control of insect pests. Science 129:537-544Furtado G, de Moraes GJ (2015) The potential of free-living Laelapid mites (Mesostigmata: Laelapidae) as

biological control agents. In: Carrillo D, de Moraes GJ, Peña JE (Eds.) Prospects for biological control ofplant feeding mites and other harmful organisms. Springer International Publishing, Cham, Switzerland,pp. 33-75

Gange AC, West HM (1994) Interactions between arbuscular mycorrhizal fungi and foliar feeding insects inPlantago lanceolata L. New Phytol 128:79-87

Gerson U, Weintraub PG (2007) Mites for the control of pests in protected cultivation. Pest Manag Sci 63:658-676

97


Gillespie DR, Quiring DMJ (1990) Biological control of fungus gnats, Bradysia spp. (Diptera: Sciaridae), and west-ern flower thrips Frankliniella occidentalis (Pergande) (Thysanoptera: Thripidae), in greenhouses using a soil-dwelling predatory mite, Geolaelaps sp. nr. aculeifer (Canestrini) (Acari: Laelapidae). Can Ent 122:975-983


Janssen A, Montserrat M, HilleRisLambers R, de Roos AM, Pallini A, Sabelis MW (2006) Intraguild predationusually does not disrupt biological control, in: Brodeur, J., Boivin, G. (Eds.), Trophic and Guild Interactionsin Biological Control, Progress in Biological Control. Springer, Dordrecht, The Netherlands, pp. 21-44

Janssen A, Sabelis MW (2015) Alternative food and biological control by generalist predatory mites: the caseof Amblyseius swirskii. Exp Appl Acarol 65:413-418

Johnson SN, Erb M, Hartley SE (2016) Roots under attack: contrasting plant responses to below- and above-ground insect herbivory. New Phytol 210:413-418

Karban R, Hougen-Eitzman D, English-Loeb G (1994) Predator-mediated apparent competition between her-bivores that feed on grapevines. Oecologia 97:508-511

Krantz GW, Ainscough B (1990) Mesostigmata. In: Dindal DL (ed.), Soil Biology Guide. John Wiley & Sons,Brisbane, Australia, pp. 583-665

Kumar V, Xiao Y, McKenzie CL, Osborne LS (2015) Early establishment of the phytoseiid mite Amblyseiusswirskii (Acari: Phytoseiidae) on pepper seedlings in a Predator-in-First approach. Exp Appl Acarol65:465-481

Lenth R (2016) Lsmeans: Least-Squares Means. J Stat Softw 69:1-33Lesna I, Wolfs P, Faraji F, Roy L, Komdeur J, Sabelis MW (2009) Candidate predators for biological control of

the poultry red mite Dermanyssus gallinae. Exp Appl Acarol 48:63-80Lewis T (1997) Pest thrips in perspective, in: Lewis, T. (Ed.), Thrips as Crop Pests. CAB International,

Wallingford, UK, pp. 1-13Lindquist EE, Krantz GW, Walter DE (2009) Order Mesostigmata. In: Krantz, GW, Walter DE (Eds.) A Manual

of Acarology. 3rd Edition. Texas Tech University Press, Texas, USA, pp. 124-232Liu CZ, Yan L, Li HR, Wang G (2006) Effects of predator-mediated apparent competition on the population

dynamics of Tetranychus urticae on apples. BioControl 5:453-463Loomans AJM, Murai T (1997) Culturing thrips and parasitoids. In: Lewis T (ed.) Thrips as Crop Pests. CAB

International, Wallingford, UK, pp. 477-503Losey JE, Denno RF (1998) Positive predator–predator interactions: enhanced predation rates and synergis-

tic suppression of aphid populations. Ecology 79:2143-2152Manners AG, Dembowski BR, Healey MA (2013) Biological control of western flower thrips, Frankliniella occi-

dentalis (Pergande) (Thysanoptera: Thripidae), in gerberas, chrysanthemums and roses. Aust J Entomol52:246-258

Masters GJ, Brown VK, Gange AC (1993) Plant mediated interactions between above-and below-groundinsect herbivores. Oikos 66:148-151

Messelink G, van Maanen R, van Holstein-Saj R, Sabelis MW, Janssen A (2010) Pest species diversityenhances control of spider mites and whiteflies by a generalist phytoseiid predator. BioControl 55:387-398

Messelink GJ, van Holstein-Saj, R (2011) Generalist predator Stratiolaelaps scimitus hampers establishmentof the bulb scale mite. Proc Neth Entomol Soc Meet 22:67-73


Mortimer SR, Van der Putten WH, Brown VK (1999). Insect and nematode herbivory below ground: interac-tions and role in vegetation succession. In: Olff H, Brown VK, Drent RH (Eds.) Herbivores: between plantsand predators. British Ecological Society Symposium, Blackwell Science, Oxford, UK, pp. 205-238

Muñoz-Cárdenas K, Ersin F, Pijnakker J, van Houten Y, Hoogerbrugge H, Leman A, Pappas M, Duarte M,Messelink G, Sabelis MW, Janssen A (2017) Supplying high-quality alternative prey in the litter increasescontrol of an above-ground plant pest by a generalist predator. Biol Control 105:19-26

Muñoz-Cárdenas K, Fuentes L, Cantor F, Rodríguez D, Janssen A, Sabelis MW (2014) Generalist red velvetmite predator (Balaustium sp.) performs better on a mixed diet. Exp Appl Acarol 62:19-32

98




Park J, Kim S, Kim D, Cho K (2002) Population dynamics of Frankliniella occidentalis on different rose culti-vars and flowering stages. J Asia Pac Entomol 5:97-102

Parrella MP, Stengard L, van Lenteren J (1999) Glasshouse environments. In: Bellows TS, Fisher TW Jr. (eds.)Handbook of Biological Control: Principles and Applications. Academic Press, San Diego, CA, USA, pp.819-839


Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team (2014) nlme: Linear and Nonlinear Mixed EffectsModels. http://CRAN.R-project.org/package=nlme

Premachandra W, Borgemeister C, Berndt O, Ehlers RU, Poehling HM (2003) Combined releases of ento-mopathogenic nematodes and the predatory mite Hypoaspis aculeifer to control soil-dwelling stages ofwestern flower thrips Frankliniella occidentalis. BioControl 48:529-541

Ramakers PMJ (1990) Manipulation of phytoseiid thrips predators in the absence of thrips. SROP/WPRS Bull13:169-172

Ramakers PMJ, van Lieburg MJ (1982) Start of commercial production and introduction of Amblyseiusmckenziei Sch. & Pr. (Acarina: Phytoseiidae) for the control of Thrips tabaci Lind. (Thysanoptera:Thripidae) in glasshouses. Mededelingen van de Faculteit der Landbouwwetenschappen van deRijksuniversiteit te Gent 47:541-545

Rosenheim JA (2001) Source-sink dynamics for a generalist insect predator in habitats with strong higher-order predation. Ecol Monogr 71:93-116

Rosenheim JA, Glik TE, Goeriz RE, Rämert B (2004) Linking a predator's foraging behavior with its effects onherbivore population suppression. Ecology 85:3362-3372

Rosenheim JA, Kaya HK, Ehler LE, Marois JJ, Jaffee BA (1995) Intraguild predation among biological controlagents: theory and evidence. Biol Control 5:303-335

Rosenheim JA, Wilhoit LR, Armer CA (1993) Influence of intraguild predation among generalist insect preda-tors on the suppression of an herbivore population. Oecologia 96:439-449

Salehi A, Ostovan H, Modarresi M (2014) Evaluation of the efficiency of Gaeolaelaps aculeifer in control ofplant parasitic nematode Tylenchulus semipenetrans under greenhouse conditions. J Entomol Nematol6:150-153

Schausberger P, Walzer A (2001) Combined versus single species release of predaceous mites: Predator-predator interactions and pest suppression. Biol Control 20:269-278

Scheu S (2001) Plants and generalist predators as links between the below-ground and above-ground sys-tem. Basic Appl Ecol 2:3-13

Settle WH, Ariawan H, Astuti EI, Cahyana W, Hakim A, Hindayana D, Lestari A (1996) Managing tropical ricerests through conservation of generalist natural enemies and alternative prey. Ecology 77:1975-1988

Snyder GB, Finke DL, Snyder WE (2008) Predator biodiversity strengthens aphid suppression across single-and multiple-species prey communities. Biol Control 44:52-60

Sokol-Hessner L, Schmitz OJ (2002) Aggregate effects of multiple predator species on a shared prey. Ecology83:2367-2372



Tommasini MG, Maini S (1995) Frankliniella occidentalis and other thrips harmful to vegetable and ornamen-tal crops in Europe. In: Loomans AJL, van Lenteren JC, Tommasini MG, Maini S, Riudavets J, Eds.Biological control of thrips pests, Wageningen Agricultural University Papers 95:1-42

99


van Maanen R, Messelink GJ, van Holstein-Saj R, Sabelis MW, Janssen A. (2012) Prey temporarily escapefrom predation in the presence of a second prey species. Ecol Entomol 37:529-535

van Rijn PCJ, van Houten YM, Sabelis MW (1999) Pollen improves thrips control with predatory mites.IOBC/WPRS Bull 22:209-212


Vangansbeke D, Nguyen DT, Audenaert J, Verhoeven R, Gobin B, Tirry L, De Clercq P (2014) Performance ofthe predatory mite Amblydromalus limonicus on factitious foods. BioControl 59:67-77

Wäckers FL, van Rijn PCJ, Bruin J, Eds. (2005) Plant-Provided Food for Carnivorous Insects: a ProtectiveMutualism and Its Applications. Cambridge University Press, Cambridge, UK

Wahab S (2010) Integrated pest management strategies for sustainability of agriculture. In: Sharma AK,Wahab S, Srivastava R, Eds. Agriculture diversification: Problems and perspectives. I.K. InternationalPublihing House Pvt. Ltd. New Delhi, India, pp. 29-43

Walter DE, Proctor HC (2013) Mites: ecology, evolution, and behaviour (2nd edn). Springer, Dordrecht, TheNetherlands, pp. 161-228

Wiethoff J, Poehling HM, Meyhöfer R (2004) Combining plant- and soil-dwelling predatory mites to optimizebiological control of thrips. Exp Appl Acarol 34:239-261

Wu S, Gao Y, Xu X, Wang E, Wang Y, Lei Z (2014) Evaluation of Stratiolaelaos scimitus and Neoseiulus bark-eri for biological control of thrips on greenhouse cucumbers. Biocontrol Sci Techn 24:1110-1121

Wu S, Zhang Z, Gao Y, Xu X, Lei Z (2016) Interactions between foliage- and soil-dwelling predatory mites andconsequences for biological control of Frankliniella occidentalis. BioControl 61:717-727

100


Single and combined predator releaseswith alternative food increases thripscontrol in an ornamental crop

Karen Muñoz-Cárdenas, Ada Leman, Marcus V. A. Duarte, Gerben J. Messelink &

Arne Janssen

Many pest species spend part of their life cycle in the soil or litter and another part

on the above-ground plant parts. Therefore, a viable strategy to control such pests

may be to release a combination of canopy-dwelling predators and litter-inhabiting

predators. However, when predators attack the same pest population, there will be

competition at some point, even if they attack different stages and different parts

of the population. This competition may be reduced when predators have alterna-

tive food sources. The effect of adding alternative food for predators can have pos-

itive or negative effects on pest control, depending on the time frame and fluctua-

tions of prey and predator populations. We tested the effect of single or combined

predator releases with alternative food on thrips control on cut roses in an experi-

mental greenhouse. In a first experiment, we released thrips before the predators

Amblyseius swirskii and the soil-dwelling predatory mites Macrocheles robustulus

and Stratiolaelaps scimitus, using Acarus siro or pollen as alternative food. In a

second experiment, we released A. swirskii and M. robustulus before the thrips

and supplied Typha spp. pollen on the plants for A. swirskii and A. siro in the litter

for M. robustulus. We found best thrips control when A. swirskii plus pollen were

released, independently of the release of soil-dwelling predatory mites.

Furthermore, better thrips control was achieved when predators were released

before the thrips. Our results suggest that soil-dwelling predatory mites, which are

present in the litter layer in many commercial greenhouses, do not affect thrips

control by A. swirskii. Hence, growers can use this latter species independent of

the presence of these soil-dwelling predators. Growers should preferably release

A. swirskii plus pollen before thrips invade the crop.

103

4

IntroductionPrey are often attacked by several species of predators (Polis 1991; Polis & Strong1996; Sih et al. 1998). Different species of natural enemies can have additive effectson prey populations, meaning that their combined effect equals the sum of theeffects of the different natural enemy species (Sih et al. 1998; Casula et al. 2006).Multiple natural enemies can also cause further increases in prey densities, for exam-ple because of the occurrence of interactions such as intraguild predation amongpredators (Ferguson & Stiling 1996; Rosenheim et al. 1993, 2004; Rosenheim 2001;Snyder & Ives 2001), or because the response of the prey to a species of naturalenemy decreases the risk of being attacked by the other species (Magalhães et al.2002; Sih et al. 1998). Furthermore, the combined effect of several species of natu-ral enemies can cause further decreases in prey densities than their additive effect,for example because they feed on different stages of the prey (Takagi & Hirose 1994),or because the response of prey to one natural enemy species makes them more vul-nerable to the other species (Losey & Denno 1998, 1999).

In biological control, the effect of multiple natural enemies versus the effect of onenatural enemy varies according to the crop and the species involved (Chang 1996;Ferguson & Stiling 1996; Sih et al. 1998). There are examples of disruption of biolog-ical control by multiple natural enemies (Finke & Denno 2004), of neutral effects,where the addition of other species of natural enemies does not result in furtherreduction of pest populations (Cardinale et al. 2003), and positive effects, which canbe additive or synergistic, increasing pest control (Losey & Denno 1998). Studiesshowing interactions among predators that have a positive effect on pest controlinvolve systems in which predators attack the pest at different locations or on differ-ent plant parts, different seasons, or at different life stages (Murdoch et al. 1984;Onzo et al. 2004). However, when predators attack the same pest population, theywill compete in the end, even when they attack different stages or different parts ofthe population. Such competition may be reduced when the predators have alterna-tive food (Onzo et al. 2005). Here, we investigated the effect of adding alternativefood for a combination of predator species that attack different stages of a pest andat different locations (i.e. the canopy and the litter).

The effects of adding alternative food, such as pollen, to increase biological con-trol by generalist predators has been widely investigated (McMurtry & Scriven 1966;van Rijn et al. 2002; Nomikou et al. 2002, 2010). There can be positive effects on bio-logical control of adding alternative food, because the addition of food results in anincrease in predator densities in the long term, after one or a few predator genera-tions. These higher predator densities result in more attacks on pest individuals(apparent competition; Holt 1977), enhancing biological control. However, during thefirst few generations of the predator, biological control might also be negatively

104

CHAPTER 4 | THRIPS CONTROL IN CUT ROSES

affected by adding alternative food, because predators may feed on the alternativeprey instead of feeding on the target pest, or because adding alternative food resultsin predator satiation (apparent mutualism; Holt 1977; Abrams & Matsuda 1996).

Several studies have demonstrated that adding pollen as food for predators to theplant canopy can indeed increase pest control (McMurtry & Scriven 1966; Kennett etal. 1979; van Rijn et al. 2002; Maoz et al. 2009; Nomikou et al. 2010; Adar et al. 2014;Leman & Messelink 2015; Duarte et al. 2015). Alternative prey for predators can alsobe added to the litter. In previous studies, we showed that litter-inhabiting predatorsthat feed on litter-inhabiting stages of an important plant pest (western flower thrips)can benefit from alternative prey supplied in the litter, resulting in increased thripscontrol (Muñoz-Cárdenas et al. 2017; CHAPTER 3). So far, studies focusing on releas-es of combinations of predators on plants and in the litter or soil have not addressedthe effect of alternative food sources for both predators on biological control. Thus,our question is whether a combination of predators on the plants and in the litter,each with their own alternative food, would increase biological control.

We studied the control of Western flower thrips (Frankliniella occidentalis) in green-house cut roses. In the particular case of this pest, adding pollen as alternative foodcan be risky, because pollen is also a high-quality food source for thrips (van Rijn etal. 2002; Chitturi et al. 2006; Leman & Messelink 2015; Vangansbeke et al. 2016).There are only few studies on the effect of adding pollen for predatory mites on thripscontrol in ornamentals (Delisle et al. 2015; Leman & Messelink 2015). In one of thesestudies, potted chrysanthemum plants received more damage when the predatorymite Amblyseius swirskii (Athias-Henriot) (Acari: Phytoseiidae) was released togetherwith pollen than without pollen (Delisle et al. 2015). However, this experiment lastedfor 4 weeks (Delisle et al. 2015), so the evaluation period may have been too short toobserve the long term positive effect of adding pollen for predators on thrips control.The study of Leman & Messelink (2015) on chrysanthemum lasted for 8 weeks andshows enhanced control of thrips when predatory mites are provided with pollen, butplant damage was not assessed. Another study, in rose plants, shows increases indensities of A. swirskii and other phytoseiid predators when adding pollen, but thripsdensities and plant damage were not assessed (Pijnakker et al. 2016). Because thripsspend part of their life cycle in the litter or soil, attempts have been made to increasecontrol of thrips by using canopy-dwelling predators that attack thrips larvae togeth-er with soil-inhabiting predators that attack prepupae and pupae (Wiethoff et al.2004; Thoeming & Poehling 2006; Manners et al. 2013; Pozzebon et al. 2015). Thesestudies did not find a significant increase in thrips control using such combinationsof predators. However, they did not supply alternative food for the predators. Wetherefore conducted two experiments. In the first experiment, we supplied one typeof alternative prey in the litter for a combination of predators. Previous studies

105

THRIPS CONTROL IN CUT ROSES | CHAPTER 4

(Muñoz-Cárdenas et al. 2017; CHAPTER 3) show that both the soil-inhabiting and thecanopy-dwelling predators do use this alternative food when supplied to the litter.Here, we assessed the combined effects of these predators with alternative food onthrips control. In the second experiment, we supplied each predator with its ownalternative food; alternative prey was again added to the litter, but we also addedpollen to the above-ground plant parts.

Materials and methodsThe experimental systemPlants were grown in rockwool (Grodan®) and nutrients were applied in the irrigationsystem. Rose plants in rockwool slabs were placed on top of gutters(JBHydroponics®) at around 60 cm from the ground. In the Netherlands, cut roseplants are grown in this way for 5-10 years, resulting in the accumulation of litter onthe greenhouse floor. Some plant branches touch this litter layer, forming a connec-tion between the above-ground and below-ground habitats. In this study we simulat-ed these conditions in an experimental greenhouse under conditions resemblingthose of commercial crop production systems.

The western flower thrips Frankliniella occidentalis (Pergande) causes importanteconomic damage in many different crops, including roses (Park et al. 2002; Mannerset al. 2013). The predators used in this study were A. swirskii, which is commonlyreleased to control thrips and whiteflies in various crops (Nomikou et al. 2001, 2002;Messelink et al. 2008), and the soil-dwelling predatory mites Macrocheles robustulus(Berlese) and Stratiolaelaps scimitus (Womersley), which can also control thrips(Berndt et al. 2004; Messelink & van Holstein-Saj 2011) and sciarids (Grosman et al.2011). We used Typha latifolia L. or Typha angustifolia L. pollen as alternative food forA. swirskii. Both predators feed on the alternative prey Acarus siro (L.), which wasreleased in the litter.

Arthropod cultures and pollenWestern flower thrips (F. occidentalis) were reared on flowering chrysanthemumplants (Dendranthema grandiflora Tzelev, var. Miramar) in a greenhouse compart-ment. The flour mite A. siro was reared on wheat bran (Ramakers & van Lieburg1982). Amblyseius swirskii (Swirski-mite®), S. scimitus (Entomite M®) and M. robus-tulus (Macro-mite®) were obtained from Koppert Biological Systems (Berkel enRodenrijs, The Netherlands). For the first experiment, we used T. latifolia pollen, col-lected from plants in the surroundings of the experimental station (Bleiswijk, TheNetherlands). It was dried and stored at -20 °C. For the second experiment, we usedT. angustifolia pollen (Nutrimite®) provided by Biobest (Westerlo, Belgium).

106


107


Study area and plant materialThe experiments were performed in a greenhouse compartment at Wageningen URGreenhouse Horticulture in Bleiswijk, The Netherlands. The compartment was 96 m2,the average temperature was 22.2 °C, RH 80%, photoperiod L13:D11. Artificial lightwas 10.000 lux, and was switched on when natural light in the greenhouse was lessthan 150 Watt/m2. The windows of the greenhouse were provided with insect gauze(0.40 × 0.45 mm) to prevent arthropod invasions. Twenty cages, consisting of a plasticframe (1 × 1.5 × 2 m) covered with insect-proof mesh (mesh size 0.22 × 0.31 mm), wereplaced inside this compartment. Cages were located about 50 cm from each other.Each cage had two zippers at the front for easy access. In March 2014, 10 rose plants(1 year old, var. Avalanche ++; Leo Ammerlaan, Bleiswijk) rooted in rock wool wereplaced in each cage on top of two gutters (JBHydroponics®) (5 plants/gutter). The plantdensity (6 plants/m2) is characteristic of Dutch commercial roses. Litter from a commer-cial greenhouse (Royal Roses, Stompwijk, The Netherlands) was sterilized (120 °C, 20min) and placed below the gutters of the cages. To start the experiment with plants thatwere free of pests and diseases, we treated all plants prior to the experiment with a mixof pyridylethylamide (Luna privilage, Bayer) against powdery mildew and lufenuron(Match, Syngenta), abamectin (Vertimec gold, Syngenta) and pyridalyl (Nocturn,Nufarm) against western flower thrips. To prevent powdery mildew infection, we used aSulphur evaporator until the week before the start of the experiment. Because thecages had no bottom, we built a barrier of sand (10 cm wide, 10 cm high) on the flooragainst the cage walls on the inside and outside of all the cages to prevent escapes oflitter-inhabiting mites. During both experiments, a standard nutrient solution was pro-vided to the plants with a drip irrigation system. Roses were harvested once per week,stems were bent and buds from these stems were removed according to commercialpractice. As a preventive measurement, each cage was weekly supplied with Encarsiaformosa Gahan (150 individuals/cage, En-strip®, Koppert) against whiteflies(Trialeurodes vaporariorum Westwood) and Phytoseiulus persimilis Athias-Henriot (200individuals/cage, Spidex®, Koppert) against spider mites (Tetranychus urticae Koch).

ExperimentsIn previous studies (Muñoz-Cárdenas et al. 2017; CHAPTER 3) we found that the pos-itive effect of adding alternative prey on biological control of thrips in roses isachieved only when the predator populations are established before pest invasion(also known as the predator-in-first approach; Ramakers 1990; Kutuk & Yigit 2011;Kumar et al. 2015). However, in the case of ornamentals, there are no studies com-paring the effect of predator pre-establishment with releases of predators after pestinvasions on biological control. For this reason, we carried out the first experiment byreleasing predators with alternative food after thrips. For the second experiment,

predator populations were allowed to establish by releasing them with their alterna-tive food before releasing the pest.

Thrips control with two predators and one alternative food sourceThis experiment was carried out for 19 weeks from June to October of 2014.Frankliniella occidentalis adults were released during the first 3 weeks of the experi-ment to simulate a thrips invasion (see TABLE 4.1 for numbers). Thrips adults were col-lected from the colony in 10 ml pipette points with a vacuum pump. The tips wereclosed with cotton wool and transferred to the greenhouse for release of thrips insidethe cages. In the fifth week, the cages received one of five treatments, each with fourreplicates (cages): 1) no predators, 2) A. swirskii, 3) A. swirskii + pollen, 4) M. robus-tulus + S. scimitus + A. siro, or 5) M. robustulus + S. scimitus + A. siro + A. swirskii.Females of A. swirskii were taken from the bottles of the commercial product usinga fine brush and placed on sweet pepper leaf discs (Capsicum annuum L.).Subsequently the discs were placed on the rose plants in the cages. These preda-tors were released in week 5 and week 12 (TABLE 4.1). As alternative food source forA. swirskii, T. latifolia pollen (1 g/cage) was evenly distributed with a fine brush on theplants in the cages according to the treatment (TABLE 4.1).

Food and substrate for the alternative prey A. siro consisted of sterilized wheatbran, coarse vermiculite, peat, water and yeast. Subsequently A. siro were added tothis mix. Individuals of the soil-dwelling predatory mites S. scimitus and M. robustu-lus were added to this mix of food and alternative prey. The release of A. siro as alter-native prey was repeated in the twelfth week (TABLE 4.1).

Cages were sampled every 2 weeks (from weeks 7 to 19), by haphazardly collect-ing twenty leaves from the top and lower parts of the plants per cage. The thrips (lar-vae and adults) and predatory mites (larvae, nymphs and adults) were counted oneach leaf under a stereo-microscope in the laboratory and the predatory mites weremounted on slides for identification. The number of damaged leaves per cage wasalso recorded.

For sampling litter-inhabiting predators, we took 250 ml of litter from each cagetwo times during the experiment (weeks 10 and 17) and incubated it in Tullgren fun-nels during 1 week. Subsequently, the samples were sieved, predatory mites andalternative prey were counted and the predatory mites found were mounted on slidesfor identification. Mature flowers were also sampled; five open flowers per cage werecollected in plastic bags every week (from weeks 9 to 19), and the presence of scarson the sepals and on the petals of the flowers was recorded. Flowers were consid-ered as damaged if they had any deformations or scars on the petals. Flowers weresubsequently washed with 70% alcohol on a mesh (100 μm) and the thrips and pred-ators remaining on the mesh were counted under a stereo-microscope.

108


109


Average numbers of thrips per flower and per leaf were log(x+1) transformed, theaverage numbers of thrips per leaf and the average number of A. swirskii per leaf andper flower were log(x+0.01) transformed. Numbers of M. robustulus and S. scimituswere log(x+1) transformed before analysis. Because almost no A. swirskii wereencountered in the flowers during the first 14 weeks of the experiment, only datafrom week 15 onwards were evaluated for this species. All data were analyzed withlinear mixed effects models (LME of the nlme package of R, Pinheiro et al. 2014) withexperimental unit (cage) as random factor and time and treatment and their interac-tion as fixed factors. Contrasts among treatments were assessed with the glht func-tion with Tukey HSD (package lsmeans of R; Lenth 2016). We found A. siro individu-als in the litter only in week 17, not in week 10. The numbers per sample of thesemites were therefore compared using generalized lineal models (GLM instead ofLME) with treatment as factor and a quasi-Poisson error distribution. All statisticalanalyses were done using R (R Development Core Team 2015). Leaf sampling start-ed in week 7, flower sampling in week 9 when the first flowers could be harvested.

Thrips control with two predators and two alternative food sourcesThis experiment was performed using the same cages with the same plants as in theprevious experiment. To remove arthropods and mildew, we sprayed with a solutioncontaining the insecticides flonicamid (Teppeki®, Belchim) and abamectin (Vertimecgold®, Syngenta), the fungicide dodemorf (Meltatox®, BASF) against powderymildew, and heptamethyltrisiloxane (Silwet gold®, Certis Europe) as wetting agent.Furthermore, the plants were pruned, the stems that had flowered were bent, all col-ored buds were removed, and the litter from the bottom of the cages was removedfor sterilization and placed back under the gutters in the cages. After allowing theplants to regrow, an experiment was done from April to July of 2015. There were fivetreatments with four repetitions (cages): 1) untreated (no predators), 2) M. robustulus+ A. siro, 3) A. swirskii + pollen weekly, 4) M. robustulus + A. siro + A. swirskii + pollenweekly, or 5) A. swirskii + pollen (interrupted). We released only M. robustulus as lit-ter predators in this experiment because they were the dominant species in the pre-vious experiment. To test whether the frequency of adding pollen had an effect onthrips densities, pollen was added weekly in two of the treatments, and in one othertreatment, the pollen supply was interrupted (TABLE 4.2). Thrips, alternative prey andpredators were released as in the previous experiment, timing of their release can befound in TABLE 4.2. Leaves, flowers and litter were sampled as in the previous exper-iment. For data analysis, average numbers of thrips per flower were log(x+1) trans-formed, thrips per leaf were log(x+0.01) transformed. The densities of A. swirskii inflowers were very low during the first 9 weeks of the experiment; these data weretherefore excluded from the analysis. The remaining densities were log(x+0.1) trans-

formed, as were the densities of this predator per leaf. Densities of M. robustulus andA. siro were log(x+1) transformed. Proportions of damaged leaves and flowers werearcsin √x transformed. All transformed data were compared using LME as in the pre-vious experiment.

ResultsThrips control with combined predators and one alternative food sourceThe average numbers of thrips per flower differed significantly among treatmentsthrough time (FIGURE 4.1a; LME, interaction of treatment with time: d.f. = 4, χ2 = 74.5,P<0.0001). This was because the average thrips densities increased in most treat-ments, but decreased in the treatment with A. swirskii plus pollen. Thrips densitieswere on average highest in the control and with A. swirskii without pollen, and low-est with A. swirskii plus pollen (FIGURE 4.1a). Overall, thrips densities in the flowerswere high, with unsatisfactory control, except in the treatment with A. swirskii pluspollen during the last 3 weeks of the experiment (FIGURE 4.1a; contrasts with glhtfunction of package lsmeans).

The average numbers of thrips per leaf also differed significantly among treat-ments through time (FIGURE 4.1b; LME, interaction of treatment with time: d.f. = 4, χ2

= 17.1, P = 0.0018). This may have been because the densities of thrips dropped inthe treatment with A. swirskii plus pollen during the last few weeks, as it did in theflowers. However, none of the contrasts among treatments was significant (FIGURE

4.1b; contrasts).Amblyseius swirskii was practically absent in the flowers during the first 14 weeks

of the experiment (FIGURE 4.1c). Analysis of the data from week 15 onwards showeda significant difference among treatments through time (LME: d.f. = 2, χ2 = 6.91, P =0.032), caused by the higher numbers of A. swirskii in the presence of pollen duringthe last 4 weeks than in the other treatments (FIGURE 4.1c). The average density of A.swirskii per leaf also differed significantly among treatments through time (FIGURE

4.1d; LME, interaction: d.f. = 2, χ2 = 10.8, P = 0.0044), likely caused by the high num-bers of predators towards the end of the experiment in the treatment with A. swirskiiplus pollen, but not in the other treatments (FIGURE 4.1d; contrasts).

By releasing the thrips before the predators and the alternative food, all the flowershad thrips damage on the petals, independently of the treatment (K. Muñoz-Cárdenas,pers. obs.). Moreover, the proportion of leaves damaged by thrips was high in all treat-ments (FIGURE 4.1e), but differed significantly among treatments through time (LME,interaction: d.f. = 4, χ2 = 13.3, P = 0.01). This was because the proportion of damagedleaves in the treatment with A. swirskii plus pollen was the highest at the start of theexperiment and lowest at the end of the experiment. Nevertheless, the differencesamong treatments through time were not significant (FIGURE 4.1e; contrasts).

110


111


A. swirskii A. swirskii M. robustulus S. scimitus + A. siro M. robustulus S. scimitus A. siro A. swirskii

A. swirskiiA. swirskii M. robustulus S. scimitus A. siroM. robustulus S. scimitus A. siro +A. swirskii

A. s

wirs

kii

A. s

wirs

kii

A. swirskii A. swirskiiM. robustulus S. scimitus A. siro A. swirskii

A. swirskiiA. swirskiiM. robustulus S. scimitus + A. siroM. robustulus + S. scimitus + A. siro A. swirskii

M. robustulus S. scimitus +

A. swirskii A. swirskii A. siro A. swirskii

FIGURE 4.1. Effect of addition two predators and one alternative food source on thrips densities, predator densi-

ties and proportion of damaged leaves on rose plants. In all the treatments thrips were released before predators

plus alternative food (TABLE 4.1): 1) no predators; 2) Amblyseius swirskii, without alternative food; 3) A. swirskii +

pollen on leaves; 4) Macrocheles robustulus + Stratiolaelaps scimitus + Acarus siro in the litter; or 5) M. robustu-

lus + S. scimitus + A. swirskii + A. siro. Shown are average numbers (± s.e.) of ((a) thrips/flower, ((b) thrips/leaf,

(c) A. swirskii/flower, ((d) A. swirskii/leaf, and ((e) proportion of damaged leaves. Different letters in the legends rep-

resent significant differences among treatments through time.

The numbers of M. robustulus in litter samples differed significantly among treat-ments through time (LME, interaction: d.f. = 1, χ2 = 9.7, P = 0.0018). This was becausethe numbers of this predator did not differ between treatments in the tenth week of theexperiment, whereas the numbers were significantly higher in the treatment without A.swirskii at 17 weeks (FIGURE 4.2a). The densities of S. scimitus in the litter samples didnot differ significantly among treatments through time (LME, interaction: d.f. = 1, χ2 =3.5, P = 0.061), and there was no significant effect of treatment (FIGURE 4.2b; LME: d.f.= 1, χ2 = 0.0036, P = 0.95). There were significantly more S. scimitus present in the lit-ter after 17 weeks than after 10 weeks (FIGURE 4.2b; LME: d.f. = 1, χ2 = 10.8, P = 0.001).Besides the species of litter predators released, we found high densities of Lasioseiusfimetorum (Acari: Podocinidae) (average ± s.e. = 68.1 ± 37.3) in the treatments with soildwelling predators in week 10. Probably they came from the soil in the greenhouse andtheir populations increased as a result of the alternative prey added (A. siro). After 17weeks, we found high densities of Ameroseius sp. (Acari: Ameroseiidae) (112.3 ± 54.9)in the treatments without releases of soil dwelling predators. At the same time, wefound low numbers of Neoseiulus barkeri (Acari: Phytoseiidae) (6 ± 3.5) and Geolaelapsaculeifer (Acari: Laelapidae) (5.3 ± 4) in the litter samples of all treatments.

Acarus siro, the alternative prey added to the litter, was not encountered after 10weeks and was therefore added another time in week 12 (TABLE 4.1). This resulted intheir presence in the litter after 17 weeks, and their numbers did not differ significant-ly between treatments [M. robustulus + A. siro: (average ± s.e.) 847.5 ± 465.1, M.robustulus + A. siro + A. swirskii: 535 ± 185.7; GLM: d.f. = 1, F1,6 = 0.45, P = 0.53].

112


FIGURE 4.2. Average numbers (±s.e.) of mites found in 250 ml of litter samplings after 10 and 17 weeks in the treat-

ments in which Macrocheles robustulus + Stratiolaelaps scimitus were released with the alternative prey Acarus

siro. (a) M. robustulus/litter sample, ((b) S. scimitus/litter sample. Different letters above the bars in week 17 indi-

cate a significant difference between treatments.

M. robustulus + S. scimitus + A. siroM. robustulus+ S. scimitus + A. siro + A. swirskii

S. s

cim

itus

/ sam

ple

[#]

0

20

40

60

80

100

120

10 17

a

b

M. r

obus

tulu

s / s

ampl

e [#

]

(a)

0

5

10

15

20

25

30

35

40

10 17

(b)

Time [weeks]

113


Thrips control with two predators and two alternative food sourcesThe average numbers of thrips per flower differed significantly among treatmentsthrough time (FIGURE 4.3a; LME, interaction: d.f. = 4, χ2 = 16.1, P = 0.0029). Whereasthrips densities did not differ during the first few weeks, they subsequently increasedin the treatments without A. swirskii plus pollen, whereas they remained low in alltreatments with A. swirskii plus pollen (FIGURE 4.3a).

Densities of thrips per leaf did not vary significantly through time (FIGURE 4.3b;LME: d.f. = 4, χ2 = 0.35, P = 0.56), nor was there a significant interaction of treatmentwith time (d.f. = 4, χ2 = 8.61, P = 0.072). There was a significant effect of treatments(d.f. = 4, χ2 = 11.6, P = 0.021), but contrasts among treatments did not reveal any sig-nificant differences (FIGURE 4.3b).

The densities of A. swirskii per flower were very low during the first 9 weeks of theexperiment (FIGURE 4.3c), and did not differ significantly among treatments from week10 onwards (LME: d.f. = 2, χ2 = 1.32, P = 0.52). There was no significant effect of time(d.f. = 2, χ2 = 0.87, P = 0.35) or of the interaction of treatment with time (d.f. = 2, χ2

= 1.97, P = 0.37). Densities of A. swirskii on the leaves varied significantly amongtreatments (FIGURE 4.3d; LME: d.f. = 2, χ2 = 25.5, P<0.0001), and there was no signif-icant effect of time (d.f. = 2, χ2 = 0.045, P = 0.83) or of the interaction of treatmentwith time (d.f. = 2, χ2 = 3.65, P = 0.16). Densities of A. swirskii on leaves were muchlower when pollen supply was interrupted (FIGURE 4.3d).

The proportion of damaged flowers varied significantly through time among treat-ments (FIGURE 4.3e; LME, interaction of treatment with time: d.f. = 4, χ2 = 40.1,P<0.0001). Initially, there were few damaged flowers in all treatments, but this sub-sequently increased in the treatments without A. swirskii plus pollen (FIGURE 4.3e).The average proportion of damaged leaves varied significantly among treatments

TABLE 4.1. Thrips control with combined predator species and one alternative food source. The timing of releas-

es of the western flower thrips Frankliniella occidentalis, predatory mites (Amblyseius swirskii and Macrocheles

robustulus) and alternative prey (Acarus siro) or food (pollen) on rose plants or in the litter.

Week 1 2 3 5 12 14 16F. occidentalisa 15 15 15A. swirskiia 100 200Pollenb 1 1 1 0.5M. robustulusa 200S. scimitusa 200A. siroc 90000 15000Food in the litterd 1.75 0.35a Number of adults released per week per cage. b Grams added per cage. c Number of immatures and adults. d Liters of food and substrate for alternative prey containing 1 l of sterilized wheat bran + 200 ml of coarsevermiculite + 300 ml of peat + 250 ml of water + 25 g of dried baker’s yeast.

114


No predators (a)M. robustulus + A. siro (a)A. swirskii + pollen (weekly) (b)M. robustulus + A. siro + A. swirskii +

A. swirskii + pollen (interrupted) (b)pollen (weekly) (b)

No predatorsM. robustulus + A. siroA. swirskii + pollen (weekly)M. robustulus + A. siro + A. swirskii +

A. swirskii + pollen (interrupted)pollen (weekly)

0

1

2

3

4

5

5 6 7 8 9 10 11 12 13 14A. s

wirs

kii/

flow

er [#

]

(c) A. swirskii + pollen (interrupted)A. swirskii + pollen (weekly)

A. swirskii + pollen (weekly)M. robustulus + A. siro +

Time [weeks]A. swirskii + pollen (weekly) (a)M. robustulus + A. siro+ A. swirskii +

A. swirskii + pollen (interrupted) (b)pollen (weekly) (a)

Time [weeks]

No predators (a) M. robustulus + A. siro (a)A. swirskii + pollen (weekly) (b)M. robustulus + A. siro + A. swirskii +

A. swirskii + pollen (interrupted) (b)pollen (weekly) (b)

Time [weeks]

No predators (a)M. robustulus + A. siro (ab)A. swirskii + pollen (weekly) (bc)M. robustulus + A. siro + A. swirskii +

A. swirskii + pollen (interrupted) (c)pollen (weekly) (c)

0

10

20

30

40

50

60

70

80

90

5 6 7 8 9 10 11 12 13 14

Thrip

s / f

low

er [#

]

(a)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

5 6 7 8 9 10 11 12 13 14

Thrip

s / l

eaf [

#]

(b)

0

1

2

3

4

5

5 6 7 8 9 10 11 12 13 14

A. s

wirs

kii/

leaf

[#]

(d)

00.10.20.3

0.40.50.60.70.8

0.9

1

5 6 7 8 9 10 11 12 13 14

Dam

aged

flow

ers

[pro

porti

on]

(e)

00.10.20.30.40.50.60.70.80.9

1

5 6 7 8 9 10 11 12 13 14

Dam

aged

leav

es [p

ropo

rtion

]

(f)

FIGURE 4.3. Effect of addition two predators and two alternative food sources on thrips densities, predator densities

and proportions of damaged rose leaves and flowers. Predators plus alternative food were allowed to establish

before thrips releases (TABLE 4.2). Treatments were: 1) No predators; 2) Macrocheles robustulus + Acarus siro in the

litter; 3) Amblyseius swirskii + pollen weekly applied on leaves; 4) M. robustulus + A. siro + A. swirskii + pollen week-

ly; or 5) A. swirskii + pollen (interrupted): pollen was added four times during the experiment (TABLE 4.2). Predators

and alternative food were released before thrips (TABLE 4.2). Shown are average numbers (± s.e.) of ((a) thrips/flower,

(b) thrips/leaf, ((c) A. swirskii/flower, ((d) A. swirskii/leaf, ((e) proportion of damaged flowers, and ((f) proportion of dam-

aged leaves. Different letters in the legends represent significant differences among treatments through time.

through time (FIGURE 4.3f; LME, interaction: d.f. = 4, χ2 = 30.8, P<0.0001): the propor-tion of damaged leaves was initially low in all treatments, but increased considerablyin the treatments without A. swirskii plus pollen (FIGURE 4.3f). It is noteworthy that theproportion of damaged leaves increased considerably in the last week when theaddition of pollen to A. swirskii was interrupted (FIGURE 4.3f).

There was no statistically significant effect of the treatments on the numbers of M.robustulus in the litter (FIGURE 4.4; LME: d.f. = 1, χ2 = 0.10, P = 0.75), the numbersvaried significantly with time (d.f. = 1, χ2 = 4.20, P = 0.041), but the interaction

115


FIGURE 4.4. Average numbers (± s.e.) of mites found in 250 ml of litter samplings in the treatments in which

Macrocheles robustulus was released with the alternative prey Acarus siro.

M. r

obus

tulu

s / s

ampl

e [#

]

0

10

20

30

40

50

60

70

6 10

M. robustulus + A. siroM. robustulus + A. siro + A. swirskii + pollen (weekly)

Time [weeks]

TABLE 4.2. Thrips control with two predators and two alternative food sources. The timing of releases of the west-

ern flower thrips Frankliniella occidentalis, predatory mites (Amblyseius swirskii and Macrocheles robustulus) and

alternative prey (Acarus siro) or food (pollen) on rose plants. Pollen was either added weekly or interrupted.

Week 1 2 3 5 7 9 10 11 12 13F. occidentalisa 15 15 15A. swirskiia 100 100 100Pollen (interrupted)b 0.2 0.2 0.2Pollen (weekly) 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2M. robustulusa 400A. siroc 90000Food in the litterd 1.75a Number of adults released per week per cage. b Grams added per cage c Number of immatures and adults. d Liters of food and substrate for alternative prey containing 1 L of sterilized wheat bran + 200 ml of coarsevermiculite + 300 ml of peat + 250 ml of water + 25 g of baker’s yeast.

between treatment and time was not significant (d.f. = 1, χ2 = 0.92, P = 0.34). Thedensities of the alternative prey A. siro did not differ between treatments (LME: d.f. = 1,χ2 = 0.25, P = 0.61), but varied significantly with time (d.f. = 1, χ2 = 35.8, P<0.0001;interaction treatment and time: d.f. = 1, χ2 = 0.84, P = 0.36). The averages of A. siroin week 6 were higher than in week 10 (week 6: M. robustulus + A. siro: (average ±s.e.) 1787 ± 477; M. robustulus + A. siro + A. swirskii + pollen: 2225 ± 607; week 10:M. robustulus + A. siro: 10 ± 6; M. robustulus + A. siro + A. swirskii + pollen: 5 ± 5).Again, we found predatory mites of other families in all treatments (probablyPhytoseiidae, Ameroseiidae and Laelapidae) (39.7 ± 21).

Summarizing both experiments, independent of the presence of M. robustulusand A. siro, releasing A. swirskii and adding pollen as alternative food resulted in thelargest decrease in thrips densities, the largest increase in A. swirskii densities andthe lowest plant damage, compared to the rest of the treatments.

DiscussionCombining predators does not increase or hampers thrips controlIn the experiment with combined predators and one type of alternative food, wefound lower thrips densities in the flowers of plants with A. siro and M. robustulus inthe litter than in the treatment without predators. However, these densities did notdiffer significantly from those in other treatments with predators (FIGURE 4.1a). At theend of this experiment, we found significantly fewer M. robustulus when it wasreleased in combination with A. swirskii and S. scimitus than when it was releasedwith S. scimitus only (FIGURE 4.2a). This may have been caused by negative preda-tor-predator interactions, such as competition for food or intraguild predation. Thistopic deserves further study, as well as the role of the other litter-inhabiting mitesencountered in biological pest control.

We observed low numbers of S. scimitus when they were released together withM. robustulus. Elsewhere (Berndt et al. 2004), this species (synonym Hypoaspismiles) was shown to be an effective predator of thrips in the litter. In their experi-ments, the plants were confined to small tubes and the thrips mainly pupated in thesoil where S. scimitus and M. robustulus were released. In our experiments, thripshad more options for pupation, for example at the base of the plants or in the flow-ers, thus they may have escaped from predation by these litter-inhabiting predators.This may have caused the low effect of the litter-inhabiting predators on thrips con-trol.

The combined release of the predatory mites A. swirskii with pollen as alternativefood on the plant canopy and M. robustulus with A. siro in the litter resulted inincreased biological control compared to the release of M. robustulus with A. siro as

116


alternative prey. However, the combined release did not result in better control thanreleases of only A. swirskii with pollen. Similar effects of combined releases of otherspecies of canopy-dwelling and soil-inhabiting predators on thrips control werefound by Wiethoff et al. (2004) and Thoeming & Poehling (2006), but these authorsdid not supply the predators with alternative food. Our results show that singlereleases of A. swirskii combined with pollen resulted in increased thrips control(FIGURES 4.1 and 4.3). There was no synergistic effect of the two predators (facilita-tion), as has been found in another system with canopy-dwelling and soil-dwellingpredators (Losey & Denno 1998, 1999). However, we also did not find evidence thatthe soil predator hampered pest control by the canopy-dwelling predator, in contrastto a study in another system (Messelink & van Holstein-Saj 2011).

Many arthropod plant pests spend part of their life cycles in the litter or soil(Wahab 2010). This means that biological control could be directed to the pest inboth habitats by releasing predators in the plant canopy and in the litter or soil. In theexperiment with two predators and two alternative food sources, we did not find asignificant effect of M. robustulus on thrips populations in the flowers, whereasMesselink & van Holstein-Saj (2008) show that this predator can substantially reducethrips populations. In a previous study (CHAPTER 3), we found that another litter-inhabiting predatory mite, Cosmolaelaps n.sp., can control thrips in rose plants whenreleased with alternative food in the litter at the foot of the plant. In the present study,we released the predators in the litter layer under the gutters on which rose plantswere located, thus the predators were not in direct contact with the plants. Again, itis possible that some thrips pupated at the base of the plants, thus escaping frompredation by M. robustulus. Further studies could therefore include treatments inwhich litter-inhabiting predators with alternative food are added at the foot of theplant, just below the canopy.

Supplying pollen to Amblyseius swirskii increases thrips controlOur results confirm that adding alternative food for predators can increase biologicalcontrol even if the pest can also benefit from the alternative food (van Rijn et al.2002). At the start of the experiment with one type of alternative food, however, wefound the largest numbers of thrips in the flowers and a larger proportion of damagedleaves in the presence of A. swirskii and pollen (FIGURE 4.1a,e). This could have beencaused by a direct short-term positive effect of pollen on thrips densities, by appar-ent mutualism (i.e. an indirect short-term positive effect mediated by the predators,Holt 1977), or both.

In the experiment with two types of alternative food, we found that adding pollenon plants weekly or interrupted (4 times in 14 weeks) caused decreases of thripsdensities and plant damage (FIGURE 4.3). There were no significant differences in

117


thrips densities or plant damage between the treatments with weekly or interruptedpollen supply (FIGURE 4.3e,f), but there were significantly lower numbers of A. swirskiiwith interrupted pollen supply (FIGURE 4.3d) and an increase in damaged leaves in thistreatment (FIGURE 4.3f). This suggests that it is not a good idea to interrupt pollensupply when thrips invade the crop. To our knowledge, there are no studies thatinvestigate the most adequate frequency of food supply to predators; further studiesshould be devoted to this topic. The commercial recommendation for growers is toadd pollen to the crop every 2 weeks (Pijnakker et al. 2016), according to our results,growers could add pollen weekly, which results in more predators on the leaves com-pared with an interrupted pollen supply (FIGURE 3d).

We found no negative effect of M. robustulus on thrips control by A. swirskii. Thissuggests that growers can preventively release A. swirskii and add pollen on theplants for control of F. occidentalis in roses even when these predators are in the lit-ter.

Pre-establishment of predators is necessary for thrips controlBecause the two experiments reported here were carried out at different times forlogistical reasons, assessing the effect of releasing predators before or after releas-ing thrips is not straightforward. To nevertheless compare thrips control in the twoexperiments, we calculated the proportional decrease of thrips densities inside flow-ers, the marketable plant part, and on the leaves. Specifically, we calculated thisdecrease in the treatments with A. swirskii with pollen relative to the control (withoutpredators and alternative food) for both experiments. This yielded two time serieswith relative changes of the thrips populations in the flowers, and two time series forthe leaves. The average proportional thrips density in the flowers was 0.74 (s.e. 0.18)when thrips were released first and 0.34 (s.e. 0.078) when the predators werereleased first. Overall, this difference was significant (GLM with Gaussian error distri-bution: F1,18 = 6.41, P = 0.021). For the leaves, we did not find a significant differ-ence in thrips densities (GLM with gamma error distribution: F1,10 = 2.56, P = 0.14).These comparisons suggest that pre-establishment of A. swirskii with pollen resultsin better control of thrips in rose flowers, confirming the findings of Muñoz-Cárdenaset al. (2017) and the results of other authors in vegetable crops regarding the impor-tance of pre-establishment of predators for thrips control (Ramakers 1990; Kutuk &Yigit 2011; Kumar et al. 2015).

Acknowledgements

K.M.C was supported by Colciencias (Colombia) (Programa ‘Francisco José deCaldas’ 2011). We thank Peter de Ruiter for constructive comments.

118


ReferencesAbrams PA, Matsuda H (1996) Positive indirect effects between prey species that share predators. Ecology


of predatory mites in protected crops. BioControl 59:307-317Berndt O, Meyhöfer R, Poehling HM (2004) The edaphic phase in the ontogenesis of Frankliniella occidental-

is and comparison of Hypoaspis miles and Hypoaspis aculeifer as predators of soil-dwelling thrips stages.Biol Control 30:17-24

Cardinale BJ, Harvey CT, Gross K, Ives AR (2003) Biodiversity and biocontrol: emergent impacts of a multi-enemy assemblage on pest suppression and crop yield in an agroecosystem. Ecol Lett 6:857-865

Casula P, Wilby A, Thomas MB (2006) Understanding biodiversity effects on prey in multi–enemy systems.Ecol Lett 9:995-1004

Chang GC (1996) Comparison of single versus multiple species of generalist predators for biological control.Environm Entomol 25:207-212

Chitturi A, Riley DG, Houston JP (2006) Effect of pine pollen on settling behavior of Frankliniella occidental-is and Frankliniella fusca (Thysanoptera: Thripidae) on tomato and peanut. Environ Entomol 35:1396-1406

Delisle JF, Shipp L, Brodeur J (2015) Apple pollen as a supplemental food source for the control of westernflower thrips by two predatory mites, Amblyseius swirskii and Neoseiulus cucumeris (Acari: Phytoseiidae),on potted chrysanthemum. Exp App Acarol 65:495-509

Duarte MV, Venzon M, Bittencourt MCDS, Rodríguez-Cruz FA, Pallini A, Janssen A (2015) Alternative foodpromotes broad mite control on chilli pepper plants. BioControl 60:817-825

Ferguson KI, Stiling P (1996) Non additive effects of multiple natural enemies on aphid populations.Oecologia 108:375-379

Finke DL, Denno RF (2004) Predator diversity dampens trophic cascades. Nature 429:407-410Grosman A, Messelink G, Groot E (2011) Combined use of a mulch layer and the soil-dwelling predatory mite

Macrocheles robustulus (Berlese) enhance the biological control of sciarids in potted plants. IOBC/WPRSBull 68:51-54


Kennett CE, Flaherty DL, Hoffmann RW (1979) Effect of wind-borne pollens on the population dynamics ofAmblyseius hibisci (Acarin: Phytoseiidae). Entomophaga 24:83-98


Kutuk H, Yigit A (2011) Pre-establishment of Amblyseius swirskii (Athias–Henriot) (Acari: Phytoseiidae) usingPinus brutia (Ten.) (Pinales: Pinaceae) pollen for thrips (Thysanoptera: Thripidae) control in greenhousepeppers. Intern J Acarol 37:95-101

Leman A, Messelink G (2015) Supplemental food that supports both predator and pest: A risk for biologicalcontrol? Exp Appl Acarol 65:511-524

Lenth R (2016) Lsmeans: Least-Squares Means. J Stat Softw 69:1-33Losey JE, Denno RF (1998) Positive predator–predator interactions: enhanced predation rates and synergis-

tic suppression of aphid populations. Ecology 79:2143-2152Losey JE, Denno RF (1999) Factors facilitating synergistic predation: the central role of synchrony. Ecol Appl

9:378-386Magalhães S, Janssen A, Hanna R, Sabelis MW (2002) Flexible antipredator behaviour in herbivorous mites

through vertical migration in a plant. Oecologia 132:143-149Manners AG, Dembowski BR, Healey MA (2013) Biological control of western flower thrips, Frankliniella occi-

dentalis (Pergande) (Thysanoptera: Thripidae), in gerberas, chrysanthemums and roses. Aust J Entomol52:246-258

119


Maoz Y, Gal S, Abrahams J, Gan-Mor S, Coll M, Palevsky E, Vincent C (2009) Pollen provisioning enhancesEuseius scutalis (Phytoseiidae) populations and improves control of Oligonychus perseae (Tetranychidae).In Proceedings of the 3rd International Symposium on Biological Control of Arthropods, Christchurch,New Zealand 8-13:339-346

McMurtry JA, Scriven GT (1966) Studies on predator–prey interactions between Amblyseius hibisci andOligonychus punicae (Acarina: Phytoseiidae, Tetranychidae) under greenhouse conditions. Ann EntomolSoc Am 59:793-800

Messelink GJ, van Holstein-Saj R (2008) Improving thrips control by the soil-dwelling predatory miteMacrocheles robustulus (Berlese). IOBC/WPRS Bull 32:135-138

Messelink GJ, van Holstein-Saj R (2011) Generalist predator Stratiolaelaps scimitus hampers establishment ofthe bulb scale mite predator Neoseiulus barkeri in Hippeastrum. Proc Neth Entomol Soc Meet 22:67-73


Muñoz-Cárdenas K, Ersin F, Pijnakker J, van Houten Y, Hoogerbrugge H, Leman A, Pappas ML, Duarte MVA,Messelink GJ, Sabelis MW, Janssen A (2017) Supplying high-quality alternative prey in the litter increas-es control of an above-ground plant pest by a generalist predator. Biol Control 105:19-26

Murdoch WW, Reeve JD, Huffaker B, Kennett CE (1984) Biological control of olive scale and its relevance toecological theory. Am Nat 123:371-392




Onzo A, Hanna R, Janssen A, Sabelis MW (2004) Interactions between two neotropical phytoseiid predatorson cassava plants and consequences for biological control of a shared spider mite prey – a screenhouseevaluation. Biocont Sci Tech 14:63-76

Onzo A, Hanna R, Negloh K, Toko M, Sabelis MW (2005) Biological control of cassava green mite with exot-ic and indigenous phytoseiid predators – effects of intraguild predation and supplementary food. BiolControl 33:143-152

Park J, Kim S, Kim D, Cho K (2002) Population dynamics of Frankliniella occidentalis on different rose culti-vars and flowering stages. J Asia Pac Entomol 5:97-102

Pijnakker J, Arijs Y, de Sousa A, Wackers F (2016) The use of Typha angustifolia (cattail) pollen to establishthe predatory mites Amblyseius swirskii, Iphiseius degenerans, Euseius ovalis and Euseius gallicus inglasshouse crops. IOBC/WPRS Bull 120:47-52

Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team (2014) nlme: Linear and Nonlinear Mixed EffectsModels. http://CRAN.R-project.org/package=nlme

Polis GA (1991) Complex trophic interactions in deserts: an empirical critique of food-web theory. Am Nat138:123-155

Polis GA, Strong DR (1996) Food web complexity and community dynamics. Am Nat 147:813-846Pozzebon A, Boaria A, Duso C (2015) Single and combined releases of biological control agents against

canopy- and soil-dwelling stages of Frankliniella occidentalis in cyclamen. BioControl 60:341-350R Development Core Team (2013) R: A language and environment for statistical computing. R Foundation for

Statistical Computing, Vienna, AustriaRamakers PMJ (1990). Manipulation of phytoseiid thrips predators in the absence of thrips. IOBC/WPRS Bull

13:169-172Ramakers PMJ, van Lieburg MJ (1982) Start of commercial production and introduction of Amblyseius

mckenziei Sch. & Pr. (Acarina: Phytoseiidae) for the control of Thrips tabaci Lind. (Thysanoptera: Thripi -dae) in glasshouses. Mededelingen van de Faculteit der Landbouwwetenschappen van de Rijksuni ver -siteit te Gent 47:541-545

120


Rosenheim JA (2001) Source-sink dynamics for a generalist insect predator in habitats with strong higher-order predation. Ecol Monogr 71:93-116

Rosenheim JA, Glik TE, Goeriz RE, Rämert B (2004) Linking a predator’s foraging behavior with its effects onherbivore population suppression. Ecology 85:3362-3372

Rosenheim JA, Wilhoit LR, Armer CA (1993) Influence of intraguild predation among generalist insect preda-tors on the suppression of an herbivore population. Oecologia 96:439-449

Sih A, Englund G, Wooster D (1998) Emergent impacts of multiple predators on prey. TREE 13: 350-355Snyder WE, Ives AR (2001) Generalist predators disrupt biological control by a specialist parasitoid. Ecology

82:705-716Takagi M, Hirose Y (1994) Building parasitoid communities: the complementary role of two introduced para-

sitoid species in a case of successful biological control. In Hawkins BA, Sheehan W (eds.). Parasitoidcommunity ecology. Oxford University Press, New York, USA, pp. 437-448



Vangansbeke D, Nguyen DT, Audenaert J, Verhoeven R, Gobin B, Tirry L, De Clercq P (2016) Supplementalfood for Amblyseius swirskii in the control of thrips: feeding friend or foe? Pest Manag Sci 72:466-473

Wahab S (2010) Integrated pest management strategies for sustainability of agriculture. In: Sharma AK,Wahab S, Srivastava R, Eds. Agriculture Diversification: Problems and Perspectives. I.K. InternationalPublishing House Pvt. Ltd. New Delhi, India, pp. 29-43

Wiethoff J, Poehling HM and Meyhöfer R (2004) Combining plant- and soil-dwelling predatory mites to opti-mize biological control of thrips. Exp Appl Acarol 34:239-261

121


General discussion

Biological control of pests is an important component in crop management (Bale etal. 2008). Usually, crops are attacked by various pest species, therefore, generalistpredators have been considered as a solution for pest control (Symondson et al.2002). Contrary to specialist predators, generalist predators can feed on differentprey species and on plant-provided foods (Symondson et al. 2002). For a long time,there was scepticism regarding the use of generalists for pest control: their ability tofeed on different sources was sometimes considered a disadvantage for biologicalcontrol because it was interpreted as generalist predators not being well adapted tothe pest (Huffaker et al. 1969). However, since the 1950s, numerous studies haveshown that the presence of alternative food for generalist predators on the plant didnot have negative effects on biological control, but rather improved it (Huffaker &Kennett 1956; Collyer 1964; van Rijn et al. 2002). In theory, two prey populations canaffect each other’s densities indirectly when they are attacked by the same popula-tion of natural enemies (Holt 1977). The fact that generalist predators feed on differ-ent prey or other foods could negatively affect biological control because they mightfeed on the alternative food or prey, thus releasing the target pest from attacks, thisshort-term effect is called apparent mutualism (Murdoch 1969; Abrams & Matsuda1996). This positive effect of adding alternative prey or food on pest densities is obvi-ously detrimental for biological control. In the long-term, however, the densities ofprey may decrease in response to increases of predator populations, increasing bio-logical control (Collyer 1964; Karban et al. 1994; Müller & Godfray 1997). Hence, theeffect of the alternative food or prey on each pest densities can be positive (appar-ent mutualism) or negative (apparent competition). When population densities do notreach equilibria, but show cycles, apparent mutualism can occur also in the long-term, resulting in repeated satiation of the shared predators and repeated reducedpredation on the other prey, thus affecting biological control negatively (Abrams et al.1998). Moreover, the use of generalist predators might be also detrimental for biolog-ical control because they can compete for prey or food with other natural enemiesand they can also engage in negative interactions such as intraguild predation(Rosenheim et al. 1995). However, the use of combinations of species of generalistpredators or combinations with other natural enemies do not always result in nega-tive effects on biological control (Cardinale et al. 2003; Losey & Denno 1998).Summarizing, generalist predators can mediate interactions between prey (apparentmutualism, apparent competition; Holt et al. 1977) and can also engage in positive

123

or negative interactions with other natural enemies. Thus, the outcome of using gen-eralist predators for biological control can best be predicted when studying the dif-ferent interactions among predators and prey.

Because generalist predators interact with several species, they increase connec-tivity within and among food webs. In the past, above-ground and below-groundfood webs were studied independently. However, it has become clear that these twofood webs are connected, which affects the composition and structure of the com-munities in both habitats (Gange & Brown 1989; van der Putten et al. 2001; Bezemer& van Dam 2005; A’Bear et al. 2014). Generalist predators can form direct linksbetween above-ground and below-ground food webs when they feed on pests pres-ent on above-ground plant parts and on pest or alternative prey from the decompos-er community below-ground (Settle et al. 1996; Scheu 2001). As a result, there arecomplex food web interactions in crops due to the presence of generalist predators,pests, other natural enemies and alternative prey not only on above-ground plantparts but also in the below-ground plant system. In CHAPTER 1, I show an example ofa generalist predator that lives in the soil or litter layer and also preys on pests on theplant. This predator’s life history inspired me to investigate the possibilities of con-necting above-ground and litter food webs. Therefore, the question addressed in therest of the dissertation was how connecting above-ground and below-ground plantsystems can be used to improve biological control by generalist predators.Furthermore, I study how supplying natural enemies with alternative prey or food canimprove biological control. In CHAPTERS 2 and 3, I show how alternative prey addedto the litter increases the biological control of thrips, a pest that causes damage onabove-ground plant parts and spends part of its life cycle in the litter or soil. InCHAPTER 4, I show how a combination of generalist predator species and alternativefoods in the litter and on above-ground plant parts affects the biological control ofthrips.

In CHAPTER 1 I investigated how the generalist predator Balaustium leanderi, whichinhabits the below-ground plant system and forages on above-ground plant parts,was affected by feeding on a mixed diet of different species of important crop peststhat affect above-ground plant parts. Several generalist predators have been shownto reproduce better on mixtures of prey species than on each prey species alone(Bilde & Toft 1994; Toft & Wise 1999; Messelink et al. 2008; Lefcheck et al. 2013,Marques et al. 2015). A special characteristic of B. leanderi is that it exhibits big-bangreproduction (semelparous reproduction), which means that females lay all their eggsat once and die soon after ovipositing. I show that B. leanderi females reproduce ear-lier and lay more eggs on a mixed diet of whitefly eggs and spider mite eggs.Predator females that reproduce early consequently have higher population growthrates than late reproducers (CHAPTER 1). The effect of a mixed diet on B. leanderi is

124

GENERAL DISCUSSION

important for biological control because the mixed diet of whiteflies with spider mitesboosts its population densities, possibly increasing the biological control of the threeprey species (see Messelink et al. 2008 for an example). Further experiments shouldbe carried out to confirm whether apparent competition mediated by B. leanderioccurs between the pests under cropping conditions.

In CHAPTER 2, I further investigated how connecting above-ground and below-ground food webs can serve to improve biocontrol with the predatory miteAmblyseius swirskii, which was known as an above-ground predator. I studied howthe indirect interaction between the thrips and an alternative prey affected thrips con-trol. Astigmatic mites (Carpoglyphus lactis or Tyrophagus putrescetiae) were used asalternative prey and were released in the litter, because during the litter decomposi-tion process, astigmatic mites can feed on fungi and other organisms, ensuring ade-quate conditions for the alternative prey to establish. I found that this predator devel-oped faster when feeding on C. lactis either alone or in combination with the canopy-inhabiting thrips larvae. Hence, C. lactis was of higher nutritional value for A. swirskiithan the pest. Adding a high-quality alternative prey may result in the predators con-centrating their attacks on the alternative prey, resulting in a temporal release ofthrips from predation (short-term apparent mutualism; Holt 1977; Abrams & Matsuda1996). However, I found no evidence for this effect when performing populationdynamics experiments in the greenhouse: when astigmatic mites were present asalternative prey for A. swirskii, the population densities of thrips decreased and pred-ator densities increased (CHAPTER 2). Hence, there was a predator-mediated negativeinteraction between prey (apparent competition; Holt 1977). Amblyseius swirskii wereinitially released in the litter together with the alternative prey and were subsequent-ly found on the above-ground plant parts, showing that it commutes between theabove-ground and below-ground food web. To conclude, the predatory mite A.swirskii acts as a link between the above-ground and litter food webs and mediatesa negative indirect interaction between prey, enhancing thrips control.

During the experiments performed in CHAPTER 2, I observed that litter-inhabitingpredatory mites found in commercial greenhouses (i.e. other species than A. swirskii)can also feed on thrips pupae and pre-pupae that inhabit the litter and on the alter-native prey in the litter. Therefore, the next question was whether the negative indi-rect interaction among thrips and astigmatic mites could also be mediated by theselitter-inhabiting predators. In CHAPTER 3, I therefore selected a predatory mite species,Cosmolaelaps n. sp. from the community of predators found in the litter in commer-cial rose crops. This predator benefits from an alternative prey added to the litter, theastigmatic mite C. lactis, and it also feeds on thrips. I studied the population dynam-ics of this predator and the pest, thrips. Similar to results in CHAPTER 2, I showed bet-ter thrips control by either the community of litter-inhabiting predator species or by

125

GENERAL DISCUSSION

Cosmolaelaps n. sp. alone when astigmatic mites were added as alternative prey tothe litter. I found that another species of predatory mite, Proctolaelaps pygmaeus,was abundant when the alternative prey was added to the community of predators,and that this predator feeds on the alternative prey C. lactis but not on thrips. It is notknown to which extent the presence of predators like P. pygmeaus can affect thripscontrol by species such as Cosmolaelaps n. sp. Predator-predator interactions suchas competition for food or intraguild predation among these litter-inhabiting preda-tors should be addressed in future studies. Thus, in CHAPTER 3, I showed that litter-inhabiting predators can also mediate a negative interaction between prey, resultingin increased biological control of thrips.

Adding alternative food for predators does not always result in better pest control:some studies have shown that offering two prey to a generalist predator did result inincreases in numbers of predators and decreases in the pest number, but not indecreases of plant damage or yield loss (Delisle et al. 2015; Jaworski et al. 2015). InCHAPTERS 2 and 3, I show that releasing A. swirskii or litter-inhabiting predatorstogether with alternative prey in the litter results in decreases in thrips damage com-pared to treatments in which alternative prey was not added. However, damage toflowers, the marketable part of the crop, is more effective when the canopy-dwellingpredator A. swirskii was released with alternative food (CHAPTER 2) than litter-inhabit-ing predators or Cosmolaelaps n. sp. with alternative prey (CHAPTER 3). Thus, thecombination canopy-dwelling predators and litter-inhabiting predators with alterna-tive prey might lead to lower damage of flowers.

Therefore, the question that I addressed in CHAPTER 4 was whether it would bepossible to further decrease thrips damage when using different kinds of alternativefood and a combination of canopy-dwelling and litter-inhabiting predators. I alsotested thrips control when releasing predators preventively or only after thrips hadestablished, which has shown to affect thrips control in other crops (Ramakers 1990;Kutuk & Yigit 2011; Kumar et al. 2015).

Previous experiments (CHAPTERS 2 and 3) were done with litter added to the baseof the plants, but the litter in commercial Dutch greenhouses is not found at the baseof the plants, but further down, under the tables on which the plants are grown. InCHAPTER 4, I therefore tested the effects of addition of food under more realistic con-ditions, including the presence of a common soil predator that might improve orinterfere with control. Thus, the distance between the litter and the plants was largerthan in the previous chapters.

In the first experiment in CHAPTER 4, the question was whether adding alternativefood either in the litter (Acarus siro) or on the plant (pollen) for a canopy-dwellingpredator (A. swirskii) and litter-inhabiting predators would increase thrips control. Inthis experiment, thrips were released before the predators with alternative prey or

126

GENERAL DISCUSSION

127

GENERAL DISCUSSION

pollen, and as a result, damage was too high to find differences among treatments.Decreases in thrips numbers were observed only during the last 3 weeks of theexperiment, and only in the treatment with A. swirskii plus pollen (CHAPTER 4). All theleaves and flowers were damaged by thrips and I observed higher densities of thripsand higher plant damage during the first weeks on plants to which pollen was addedas alternative food. This could have been an effect of apparent mutualism, of thripsfeeding on the pollen, or of both. These results are similar to those obtained byDelisle et al. (2015), whom found more thrips damage on plants on which pollen wasadded than on plants without pollen. As explained above, apparent mutualism canoccur when the shared predator becomes satiated or switches to feed on the alter-native prey or food decreasing the attacks to the alternative prey (Murdoch 1969;Abrams & Matsuda 1996). This effect occurs in the short-term but can also occur inthe long term when population densities fluctuate. Curiously, I did not observe appar-ent mutualism-like phenomena in CHAPTERS 2 and 3. One obvious difference betweenthe experiment described in CHAPTER 4 and those of the previous chapters was thedistance between the litter and the above-ground plant parts. Possibly, A. swirskii didnot commute as much between the litter and the canopy because of this increaseddistance, and instead concentrated their attacks initially on the alternative prey in thelitter (apparent mutualism; Holt 1977; Abrams & Matsuda 1996). Holt (1984) studiedmodels of prey that occurred in different habitats, but were attacked by the samepredators. He predicted that negative effects of prey species on each other’s densi-ties decrease when migration by the shared predator is slow, because it decouplesthe dynamics of the predators in the two habitats. This may have occurred in myexperiments as well, which may explain the less effective control of thrips.Additionally, thrips perhaps did not only pupate in the litter under the tables withplants, but also on the tables at the base of the plant, where they would experienceless predation by the litter predators. Summarizing, this experiment shows that thedistance between the sites where alternative food or prey is supplied and the part ofthe crop where the pest occurs may be critical for efficient pest control. It also showsthat by releasing thrips before predators, plant damage is too high in all treatmentswith or without predators and alternative food.

In a second experiment in CHAPTER 4, I tested the effect of the combination ofpredators (A. swirskii and litter-inhabiting predators) and two alternative foods (A. siroand pollen) on thrips control. The differences with the previous experiment were thata combination of predators and their own alternative food or prey were added simul-taneously on the plant and in the litter, and that predators with alternative food werereleased before the thrips. Moreover, we included two frequencies of addition ofpollen. The results of this experiment show that adding pollen to the plants for theplant-inhabiting predator A. swirskii leads to high predator densities and decreases

in thrips densities and their damage (CHAPTER 4). Adding litter-inhabiting predatorsplus A. siro did not result in better thrips control than releases of only A. swirskii pluspollen. In fact, the presence of litter-inhabiting predators plus alternative prey did nothave any effect on thrips densities. This is in contrast with the findings of CHAPTER 3when releasing another litter-inhabiting predator, Cosmolaelaps n. sp. plus alterna-tive prey in the litter resulted in a decrease in thrips numbers and thrips plant dam-age (CHAPTER 3). A possible reason for litter-inhabiting predators not contributingsubstantially to the control thrips is that the litter-inhabiting predators were farremoved from the above-ground plant parts in CHAPTER 4. However, the presence oflitter-inhabiting predators plus the alternative prey did not have negative effects onthrips control by A. swirskii (CHAPTER 4). Future experiments should investigatewhether placing some litter with litter-inhabiting predators plus alternative prey at thebase of the rose plants would result in better control. Many pest species spend partof their life cycles in the litter or soil (Wahab 2010), others drop to the soil whenthreatened (Losey & Denno 1998). Thus, having high quantities of litter or soil-inhab-iting predators could contribute to the control of various pest species in many crops.

Another aspect that should receive more attention in future studies is the timingat which alternative food should be added to the crop. Many studies have shown thatadding alternative food can improve biological control (McMurtry & Scriven 1966;Kennett et al. 1979; van Rijn et al. 2002; Maoz et al. 2009; Nomikou et al. 2010; Adaret al. 2014; Duarte et al. 2015), but none have investigated the amounts and supplyfrequencies on biological control. In the second experiment presented in CHAPTER 4,two frequencies of pollen addition were tested. In one treatment, pollen was addedduring 4 weeks out of the 14 weeks of the experiment. In other treatments pollen forA. swirskii was added weekly. There were no differences among treatments concern-ing thrips numbers or thrips plant damage. However, there were differences in thenumber of A. swirskii. Leaves of plants to which pollen was added every week har-boured significantly more predators than plants with the interrupted feeding treat-ment. Future studies should further explore ‘optimal’ frequencies of addition of alter-native food on crop protection.

Summarizing, in this thesis, I present experiments that contribute to the test offood web theory. I studied aspects such as apparent competition, apparent mutual-ism, effects of combined releases of predators on prey on above-ground and below-ground plant systems. The results in this thesis contribute to the understanding ofthe dynamics of above-ground and below-ground food webs and they are first stepsin the development of a new biological control strategy in ornamentals, which is theaddition of alternative food for predators in the litter. Comparing experimental resultsand existing theory of population dynamics can be challenging because of differ-ences in time scale. Experiments are usually short, but most theory is based on long-

128

GENERAL DISCUSSION

129

GENERAL DISCUSSION

term equilibrium dynamics (Briggs & Borer 2005). The experiments performed in thisthesis were long enough to record positive and negative effects of the addition ofalternative food or prey on pest populations and on plant damage, hence they suc-ceed in assessing whether such effects as apparent competition and apparent mutu-alism do occur. Moreover, they reveal connections between above-ground andbelow-ground food webs through the foraging of generalist predators. A questionthat arises at this point is how these findings can help growers to enhance biologicalcontrol. In CHAPTER 2, I proposed a novel way to enhance pest control in ornamentalcrops by supplying alternative food in the litter. Adding alternative prey in the litter atthe base of the plant, far from the commercial plant parts may be an important con-tribution to sustainable production of ornamentals. It has the advantage that noresidues of alternative food will be left on the marketable plant parts (CHAPTERS 2 and3). As a conclusion, links can be established between above-ground plant pests andthe litter food web and such links can benefit above-ground and below-ground pred-ators, resulting in an increase of predator densities and enhanced biological control.Thus, connections between above-ground and below-ground food webs associatedwith plants may play a key role not only in natural communities (van der Putten et al.2001; A’Bear et al. 2014), but also in agricultural systems. They should thereforereceive more attention when designing biological control programs.

ReferencesA’Bear AD, Johnson SN, Jones TH (2014) Putting the ‘upstairs-downstairs’ into ecosystem service: What can

aboveground-belowground ecology tell us? Biol Control 75:97-107Abrams PA, Holt RD, Roth JD (1998) Apparent competition or apparent mutualism? Shared predation when

populations cycle. Ecology 79:201-212Abrams PA, Matsuda H (1996) Positive indirect effects between prey species that share predators. Ecology


of predatory mites in protected crops. BioControl 59:307-317Bale JS, van Lenteren JC, Bigler F (2008) Biological control and sustainable food production. Phil Trans R

Soc B 363:761-776Bezemer TM, van Dam NM (2005) Linking aboveground and belowground interactions via induced plant

defenses. Trends Ecol Evol 20:617-624Bilde T, Toft S (1994) Prey preference and egg production of the carabid beetle Agonum dorsale. Entomol Exp

Appl 73:151-156Briggs, CJ, Borer ET (2005) Why short-term experiments may not allow long-term predictions about intraguild

predation. Ecol Appl 15:1111-1117Cardinale BJ, Harvey CT, Gross K, Ives AR (2003) Biodiversity and biocontrol: emergent impacts of a multi-

enemy assemblage on pest suppression and crop yield in an agroecosystem. Ecol Lett 6:857-865Collyer E (1964) The effect of an alternative food supply on the relationship between two Typhlodromus

species and Panonychus ulmi (Koch) (Acarina). Entomol Exp Appl 7:120-124Delisle JF, Shipp L, Brodeur J (2015) Apple pollen as a supplemental food source for the control of western

flower thrips by two predatory mites, Amblyseius swirskii and Neoseiulus cucumeris (Acari: Phytoseiidae),on potted chrysanthemum. Exp App Acarol 65:495-509

Duarte MV, Venzon M, Bittencourt MCDS, Rodríguez-Cruz FA, Pallini A, Janssen A (2015) Alternative foodpromotes broad mite control on chilli pepper plants. BioControl 60:817-825

Gange A, Brown V (1989) Effects of root herbivory by an insect on a foliar-feeding species, mediated throughchanges in the host plant. Oecologia 81:38-42


Holt RD (1984) Spatial heterogeneity, indirect interactions, and the coexistence of prey species. Am Nat124:377-406

Huffaker CB, Kennett C (1956) Experimental studies on predation: predation and cyclamen-mite populationson strawberries in California. Hilgardia 26:191-222

Huffaker C, van de Vrie M, McMurtry J (1969) The ecology of tetranychid mites and their natural control. AnnuRev Entomol 14:125-174

Jaworski CC, Chailleux A, Bearez P, Desneux N (2015) Apparent competition between major pests reducespest population densities on tomato crop, but not yield loss. J Pest Sci 88:793-803

Karban R, Hougen-Eitzman D, English-Loeb G (1994) Predator mediated apparent competition between her-bivores that feed on grapevines. Oecologia 97:508-511

Kennett CE, Flaherty DL, Hoffmann RW (1979) Effect of wind-borne pollens on the population dynamics ofAmblyseius hibisci (Acarin: Phytoseiidae). Entomophaga 24:83-98


Kutuk H, Yigit A (2011) Pre-establishment of Amblyseius swirskii (Athias–Henriot) (Acari: Phytoseiidae) usingPinus brutia (Ten.) (Pinales: Pinaceae) pollen for thrips (Thysanoptera: Thripidae) control in greenhousepeppers. Intern J Acarol 37:95-101

Lefcheck JS, Whalen MA, Davenport TM, Stone JP, Duffy JE (2013) Physiological effects of diet mixing onconsumer fitness: a meta-analysis. Ecology 94:565-572

Losey JE, Denno RF (1998) Positive predator–predator interactions: enhanced predation rates and synergis-tic suppression of aphid populations. Ecology 79:2143-2152

Maoz Y, Gal S, Abrahams J, Gan-Mor S, Coll M, Palevsky E, Vincent C (2009) Pollen provisioning enhancesEuseius scutalis (Phytoseiidae) populations and improves control of Oligonychus perseae (Tetranychidae).In Proceedings of the 3rd International Symposium on Biological Control of Arthropods, Christchurch,New Zealand 8:339-346

Marques RVR, Sarmento A, Lemos F, Pedro-Neto M, Sabelis MW, Venzon M, Pallini A, Janssen A (2015)Active prey mixing as an explanation for polyphagy in predatory arthropods: synergistic dietary effects onegg production despite a behavioural cost. Func Ecol 29:1317-1324

McMurtry JA, Scriven GT (1966) Studies on predator–prey interactions between Amblyseius hibisci andOligonychus punicae (Acarina: Phytoseiidae, Tetranychidae) under greenhouse conditions. Ann EntomolSoc Am 59:793-800


Müller CB, Godfray HCJ (1997) Apparent competition between two aphid species. J Anim Ecol 66:57-64Murdoch WW (1969) Switching in general predators: experiments on predator specificity and stability of prey

populations. Ecol Monogr 39:335-354Nomikou M, Sabelis MW, Janssen A (2010) Pollen subsidies promote whitefly control through the numerical

response of predatory mites. BioControl 55:253-260Ramakers PMJ (1990) Manipulation of phytoseiid thrips predators in the absence of thrips. IOBC/WPRS Bull

13:169-172Rosenheim JA, Kaya HK, Ehler LE, Marois JJ, Jaffee BA (1995) Intraguild predation among biological control

agents: theory and evidence. Biol Control 5:303-335Scheu S (2001) Plants and generalist predators as links between the below-ground and above-ground sys-

tem. Basic Appl Ecol 2:3-13

130

GENERAL DISCUSSION

131

GENERAL DISCUSSION

Settle WH, Ariawan H, Astuti EI, Cahyana W, Hakim A, Hindayana D, Lestari A (1996) Managing tropical ricerests through conservation of generalist natural enemies and alternative prey. Ecology 77:1975-1988


Toft S, Wise DH (1999) Growth, development and survival of a generalist predator fed single- and mixed-species diets of different quality. Oecologia 119:191-197

van der Putten WH, Vet LEM, Harvey JA, Waeckers FL (2001) Linking above-ground and belowground mul-titrophic interactions of plants, herbivores, pathogens, and their antagonists. Trends Ecol Evol 16:547-554


Wahab S (2010) Integrated pest management strategies for sustainability of agriculture. In: Sharma AK,Wahab S, Srivastava R, Eds. Agriculture Diversification: Problems and Perspectives. I.K. InternationalPublishing House Pvt. Ltd. New Delhi, India, pp. 29-43

Author contributions and project funding

1 – Generalist red velvet mite predator (Balaustium leanderi) performs better on amixed dietK Muñoz-Cárdenas, LS Fuentes, RF Cantor, CD Rodríguez, A Janssen & MW Sabelis

KMC, LSF, RFC & CDR – planned the experiments; KMC & LSF – conductedexperiments; KMC, AJ & MWS – analysed the data; KMC, AJ & MWS – wrote themanuscript

2 – Supplying high-quality alternative prey in the litter increases control of an above-ground plant pest by a generalist predatorK Muñoz-Cárdenas, F Ersin, J Pijnakker, Y van Houten, H Hoogerbrugge, A Leman, ML Pappas, MVA

Duarte, GJ Messelink, MW Sabelis & A Janssen

KMC, MWS, YvH, HH & AJ – planned the experiments; KMC, FE, JP, AL MLP &MVAD – conducted experiments; KMC, GJM & AJ – analysed the data; KMC, MWS& AJ – wrote the manuscript

3 – Alternative food for litter-inhabiting predators decreases pest densities andabove-ground plant damageK Muñoz-Cárdenas, D Rueda-Ramirez, F Ersin, F Faraji & A Janssen

KMC & AJ – planned the experiments; KMC DRR, FE & FF – conducted experi-ments; KMC & AJ – analysed the data; KMC & AJ – wrote the manuscript

4 – Single and combined predator releases with alternative food increases thripscontrol in an ornamental cropK Muñoz-Cárdenas, A Leman, MVA Duarte, GJ Messelink & A Janssen

KMC, AL, GJM & AJ – planned the experiments; KMC, AL & MVAD – conductedexperiments; KMC, GJM & AJ – analysed the data; KMC, AL, GJM & AJ – wrote themanuscript

Financial funding of the projectKaren Andrea Muñoz Cárdenas was supported by Colciencias (Colombia) (Programa‘Francisco José de Caldas’ 2011). Koppert Biological Systems (Berkel en Rodenrijs,The Netherlands) sumministered materials and locations for experiments. Biobest(Westerlo, Belgium) provided materials for experiments.

132

Summary

Arthropod food webs associated with plants are commonly composed of severalspecies of herbivores, the detritivore community, and specialist and generalist pred-ators and parasitoids that feed on the first two groups and on each other. In the past,specialist natural enemies were preferably used for biological control because theyare adapted to their prey. However, they cannot persist in a crop when these prey arescarce or absent. Thus, specialist natural enemies can only be successfully releasedafter the pest has invaded the crop, or must be released repeatedly, which can becostly. In contrast, populations of generalist predators can persist in crops by feed-ing on different species of pests, plant-provided food sources such as honey andpollen, and on organisms of the detritivore community. Thus, generalist natural ene-mies mediate a number of direct and indirect interactions within and between arthro-pod communities associated with plants. These interactions can have positive ornegative effects on biological control. For example, a prey species is often attackedby several species of natural enemies, which thus compete for this prey. As a con-sequence, different species of natural enemies can have additive effects on preypopulations: their combined effects on a prey population equal the sum of the effectsof the different natural enemy species, thereby increasing biological control. Multiplenatural enemies can also disrupt biological control, for example because of interac-tions such as intraguild predation among natural enemies, or because the responseof the prey to one species of natural enemy decreases the risk of being attacked bythe other species. Furthermore, the combined effect of several species of naturalenemies can cause further decreases in prey densities than their additive effect, forexample because they feed on different stages of the prey, or because the responseof prey to one natural enemy species makes them more vulnerable to the otherspecies. Thus, in biological control the effect of multiple natural enemies versus theeffect of a single natural enemy varies and must be investigated before implement-ing biological control strategies in crops. Studies that show interactions among pred-ators that have a positive effect on pest control involve systems in which predatorsattack the pest at different locations or on different plant parts, in different seasons,or at different life stages. An aspect that has not received enough attention and Iinvestigated in this thesis is the effect of the addition of alternative food for differentgeneralist predators (on the canopy and in the litter) on biological control.

The effects of adding alternative food such as pollen to increase biological controlby a generalist predator has been widely investigated. There can be positive effects

133

of adding alternative food or prey on biological control because it results in anincrease in predator densities in the long term, after one or a few predator genera-tions. Alternative prey and food thus affect the density of a pest indirectly by chang-ing the population densities of the generalist natural enemies. Adding alternativefood or prey to a system of one predator and one pest species increases the densi-ty of the predator, and this numerical response leads to a decrease in pest densities.This is called ‘apparent competition’, because the effect of the alternative food on tothe pest is reminiscent of competition for resources. These negative effects of alter-native food or prey on the pest are a consequence of the numerical response of thepredators, hence, become manifest after a few generations of the predator. In theshort term, when the populations of predators have not reached dynamic equilibria,the natural enemy-mediated indirect interaction may cause the opposite effect: theaddition of alternative prey or food can result in satiation of the predator population,which will decrease predation of the pest population. In this case, the pest benefitsfrom the addition of alternative food (so-called ‘apparent mutualism’). Such effectsmay also occur repeatedly in predator-prey systems that show long-term persistentfluctuations. Hence, positive or negative consequences can be expected when usinggeneralist predators for biological control. Studying the effects of the interactions ofgeneralist predators with other members of the food webs is therefore important topredict the effects of releasing these predators and supplying them with alternativefood on pest control.

Generalist predators can also connect above-ground and below-ground foodwebs when they feed on prey and other food sources in both food webs. The effectof the connection of these food webs by generalist predators on pest control has notreceived much attention, especially in the case of ornamentals. Thus, the mainresearch question of this thesis was how interactions between above-ground and lit-ter food webs affect biological control. In CHAPTER 1, I investigated the effect of amixed diet of above-ground plant pests (spider mites, thrips, whiteflies) on life histo-ry traits of a generalist below-ground predator, Balaustium leanderi, which also for-ages on the canopy of crop plants. I recorded its life history traits (reproductive per-formance, survivorship and development) when fed on mixed diets of three pestspecies that inhabit above-ground plant parts. I showed that B. leanderi benefitsfrom a mixed diet of two of the species that inhabit above-ground plant parts. Thepredators reproduce more and faster when feeding on this mixed diet than on singlediets. In CHAPTER 2, I investigated Amblyseius swirskii, another generalist predatorthat mainly inhabits the canopy. I show that it makes excursions to the litter layer tofeed or disperse. From this chapter on, I focused on the control of the western flowerthrips Frankliniella occidetalis because it is one of the most damaging pests in orna-mentals. Like many other pests, this thrips spends its life cycle partly on above-

134

SUMMARY

ground plant parts and partly in the liter or soil. In the specific case of thrips, theymove to the soil or litter to pupate. I show that by adding astigmatic mites (Acari:Astigmata) as alternative prey for the predators to the litter, the densities of A. swirskiiwere boosted, resulting in decreases of thrips densities (apparent competition) anddecreases in plant damage. In CHAPTER 3, I investigated the effect of predators thatinhabit the litter of commercial rose production greenhouses on thrips control. I showthat supplying astigmatic mites in the litter as alternative food for either a communi-ty of predators or for a single predator species, Cosmoalelaps n. sp., results indecreases in thrips densities (apparent competition) and plant damage.

In commercial crops, canopy-dwelling predators (such as A. swirskii) are com-monly released to control pests, and these may interact with the litter-inhabitingpredators that are already present (such as Macrocheles robustulus, Stratiolaelapsscimitus and Cosmolaelaps n. sp.), and this might affect biological control. For thisreason, I investigated control of thrips by the combination of the canopy-dwellingpredator A. swirskii and litter-inhabiting predatory mite species such as M. robustu-lus and S. scimitus in CHAPTER 4. Because pests such as thrips inhabit above-groundplant parts and the litter or soil, attempts have been made to increase thrips controlby combining canopy-dwelling predators and soil-inhabiting predators. The noveltyhere is that I investigated whether biological control can be enhanced by supplyingdifferent alternative foods for canopy-dwelling and litter-inhabiting predators (pollenand astigmatic mites, respectively). Moreover I assessed whether the presence ofthrips before predator releases would affect biological control and whether the fre-quency of addition of pollen for canopy-dwelling predators would affect thrips con-trol. I conclude that adding pollen on the plants for canopy-dwelling predators result-ed in the best thrips control (as in apparent competition), either with or without litter-dwelling predators, and these latter predators did not interfere with thrips control. Inthis experiment, litter was placed on the soil, under the tables on which the plantsare grown, as is customary in commercial rose production. Because the litter preda-tors were released in that litter, the distance from the canopy was rather large.Further studies could therefore include treatments in which litter-dwelling predatorswith alternative food are added at the base of the plant, just below the canopy. Thiswould create a tighter link between the food webs above-ground and in the litter. Thestudy of these litter-dwelling predators to control other pests deserves further studybecause they are well adapted to the crop and their populations can be boosted withalternative food (CHAPTER 3). The release of predators after thrips was introducedproved detrimental for biological control, with high plant damage in all the treat-ments. In contrast, plants were protected from thrips damage when the predatorswere released with pollen before introducing the thrips. Furthermore, I found no dif-ferences in thrips densities or thrips damage either by adding pollen weekly or when

135

SUMMARY

interrupting the addition of pollen. However, there were significantly more predatorson plants that received pollen weekly. Concluding, in CHAPTER 4 I show that in orderto decrease flower damage by thrips in ornamentals such as roses, canopy-dwellingpredators should be released before thrips invasions. Also, pollen should be addedon a weekly basis. I suggest that more studies on the frequency and quantity of sup-plemented alternative food should be conducted. Moreover, the role of litter-dwellingpredators in control of thrips and other pests in crops with higher damage thresholdsas in ornamentals deserves further study.

Thus, in this thesis I present different aspects that can be used by biological con-trol practitioners. I also demonstrate that greenhouse experiments can help to testecological theories: I show that adding alternative food for generalist predators main-ly results in apparent competition, apparent mutualism was found only in few occa-sions for short periods of time, when the pest was present before the predators werereleased. Based on this thesis I recommend two factors that deserve further theoret-ical exploration. The first is that the effect of mixed diets on apparent competitionshould be further investigated. The second is the effect of the frequency of supply-ing alternative food on the dynamics of the pest and natural enemy. I also show thatthe interactions between above-ground and below-ground food webs affect pestdensities. In this regard, I recommend that theoretical approaches could be directedto study the impact of multiple generalist predators and food sources in above-ground and below-ground plant parts on pest control.

Summarizing, I conclude that litter and canopy food webs can be linked and thiscan result in increased biological control in an ornamental crop and that greenhouseexperiments evaluating the population dynamics of multiple predators with multiplefood sources are crucial for the development of new biological control strategies. Atthe same time, such experiments are excellent test cases for ecological theories.

136

SUMMARY

Samenvatting

Voedselwebben van geleedpotigen op planten bestaan meestal uit diverse soortenplanteneters, detritivoren en specialistische en generalistische natuurlijke vijanden:predators en parasieten, die zich voeden met leden van de eerste twee groepen enmet elkaar. Gewasbeschermers hadden in het verleden een voorkeur voor hetgebruik van specialistische natuurlijke vijanden voor biologische plaagbestrijdingomdat die zijn aangepast aan hun prooi, maar zulke specialisten kunnen zich niethandhaven in het gewas als de plaagdichtheid laag is. Ze kunnen daarom alleenmaar worden losgelaten nadat de plaag het gewas al heeft geïnvadeerd, of ze moe-ten herhaaldelijk worden losgelaten, hetgeen kosten met zich meebrengt.Generalistische predators daarentegen kunnen zich in het gewas handhaven door-dat zij zich kunnen voeden met diverse plaagsoorten, met plantaardig voedsel zoalshoning en stuifmeel, en met detritivoren. Generalisten veroorzaken hierdoor allerleidirecte en indirecte interacties binnen het voedselweb geassocieerd met planten.

Deze interacties kunnen zowel positieve als negatieve gevolgen hebben voor bio-logische plaagbestrijding. Bijvoorbeeld, een prooisoort kan vaak door meerderesoorten predators worden aangevallen, waardoor die predators concurreren om dezeprooi. Ten gevolge hiervan kunnen verschillende soorten natuurlijke vijanden eenadditief effect hebben op prooidichtheden: prooidichtheden in aanwezigheid van ver-schillende soorten predators zijn dan lager dan in de aanwezigheid van iedere vijandapart. Er kunnen echter ook andere interacties dan concurrentie om voedsel optre-den tussen de verschillende predators. Zo kunnen ze bijvoorbeeld elkaar aanvallen.Ook kan het effect van een soort predator op het gedrag van de prooi (bijvoorbeeldvluchtgedrag) ertoe leiden dat andere soorten vijanden meer moeite hebben met hetvangen van de prooi. Door deze indirecte interacties kan biologische bestrijding metmeerdere soorten natuurlijke vijanden minder efficiënt zijn dan met iedere vijandapart. Als laatste kan het zo zijn, dat het gebruik van verschillende soorten vijandeneen synergistisch effect heeft op plaagbestrijding, doordat de vijanden bijvoorbeeldverschillende ontwikkelingsstadia van de prooi aanvallen, of doordat ze gedragsver-anderingen in de prooi veroorzaken waardoor het makkelijker voor andere vijandenwordt om de prooi te vangen.

Voor efficiënte biologische bestrijding is het nodig dat de effecten van verschillen-de natuurlijke vijanden, alleen of gezamenlijk, op plaagdichtheden worden onder-zocht. Betere plaagbestrijding door het gebruik van meerdere natuurlijke vijandenwordt vooral gevonden in systemen waarin de diverse vijanden de plaag op verschil-

137

lende plekken aanvallen, bijvoorbeeld op verschillende delen van de plant, waar zeverschillende stadia van de plaag aanvallen, of in systemen waarin de vijanden in ver-schillende seizoenen actief zijn. Een aspect dat tot nu toe nog niet veel onderzochtis, en wat ik in dit proefschift onderzoek, is het effect op biologische bestrijding vanhet toevoegen van alternatief voedsel voor verschillende soorten natuurlijke vijanden.Dit voedsel wordt zowel toegediend op de bovengrondse plantendelen als in destrooisellaag aan de voet van de planten.

De effecten van het toedienen van alternatief voedsel, zoals pollen, op biologischebestrijding zijn al vaak onderzocht. Er kan een positief effect van het toevoegen vanpollen op biologische bestrijding zijn omdat het na een aantal generaties van natuur-lijke vijanden leidt tot hogere dichtheden van de vijanden. Het alternatieve voedselzorgt dan indirect voor lagere plaagdichtheden door hogere dichtheden van de vijan-den via de numerieke respons. Dit fenomeen wordt wel ‘apparent competition’genoemd, omdat de dynamiek van de plaag lijkt op die van een populatie die con-curreert met een andere populatie, in dit geval lijkt het of de plaag concurreert methet alternatieve voedsel. Op kortere termijn echter kan de indirecte interactie tussenhet alternatieve voedsel en de plaag leiden tot het tegenovergestelde effect: het toe-voegen van alternatief voedsel leidt tot verzadiging van de aanwezige predators endat leidt vervolgens tot lagere predatie van de plaagpopulatie. In dat geval profiteertde plaag dus van het toevoegen van het alternatieve voedsel (zogenaamd ‘apparentmutualism’). Zulke verzadigingseffecten kunnen ook herhaaldelijk optreden wanneerde dynamica van de natuurlijke vijand en de plaag persisterende oscillaties vertoont.Concluderend, het toevoegen van alternatief voedsel voor predators kan tot positie-ve, maar ook tot negatieve effecten op biologische bestrijding leiden. Het is daarombelangrijk voor efficiënte biologische bestrijding om de effecten van de interactiesvan generalistische predators met andere soorten in het voedselweb en met alterna-tief voedsel te bestuderen.

Generalistische predators kunnen ook een indirecte interactie tussen verschillen-de voedselwebben bewerkstelligen, bijvoorbeeld het voedselweb op de boven-grondse plantendelen en dat onder de grond. Hoe biologische bestrijding wordtbeïnvloed doordat generalistische predators deze voedselwebben verbinden is nogniet uitgebreid bestudeerd, al helemaal niet in siergewassen. De voornaamste onder-zoeksvraag van dit proefschrift is daarom hoe biologische bestrijding van een plan-tenplaag wordt beïnvloed door de interacties tussen het voedselweb op de boven-grondse plantendelen en het voedselweb in de strooisellaag aan de voet van deplant. In HOOFDSTUK 1 onderzoek ik het effect van een gemengd diet van verschillen-de bovengrondse plantenplagen (spintmijten, wittevlieg en trips) op de levensge-schiedenis van de generalistische predator Balaustium leanderi. Deze roofmijt leeftvoornamelijk ondergronds, maar zoekt ook naar voedsel op bovengrondse planten-

138

SAMENVATTING

delen. De levensgeschiedeniskenmerken die ik bestudeerde waren reproductie,overleving en ontwikkeling. De predators blijken meer en sneller te reproduceren alsze zich voeden met een gemengd dieet dan op een dieet van ieder van de drie soor-ten afzonderlijk. In HOOFDSTUK 2 bestudeer ik Amblyseius swirskii, een andere gene-ralistische roofmijt, die voornamelijk op bovengrondse plantendelen leeft. Ik laat ziendat deze predator excursies naar de strooisellaag maakt om daar te voeden of te dis-pergeren. Vanaf dit hoofdstuk concentreer ik me op de Californische tripsFrankliniella occidentalis omdat het een van de belangrijkste plagen in de sierteelt is.Zoals vele andere plagen, brengt deze trips een deel van zijn levenscyclus in destrooisellaag door, in dit geval om te verpoppen. De trips valt echter vooral boven-grondse plantendelen aan. Ik laat zien dat het toevoegen van alternatief voedsel inde vorm van astigmate mijten (Acari: Astigmata) aan de strooisellaag resulteerde inhoge dichtheden van A. swirskii, met als gevolg dat trips-dichtheden en de schadeaan planten afnemen (‘apparent competition’). In HOOFDSTUK 3 onderzoek ik het effectvan predatoren uit de strooisellaag van commerciële rozenkassen op de bestrijdingvan trips. Ik toon aan dat de roofmijt Cosmolaelaps n. sp., alleen of in combinatie metandere bodembewonende roofmijten, tripspopulaties beter bestrijdt en plantenscha-de meer beperkt als astigmate mijten aan de strooisellaag werden toegevoegd alsvoedsel voor deze predators (‘apparent competition’).

Bladbewonende rovers zoals A. swirskii worden vaak losgelaten in commerciëlegewassen om plagen te bestrijden, en deze rovers kunnen interacties aangaan metde predators die zich in de strooisellaag bevinden, zoals Macrocheles robustulus,Stratiolaelaps scimitus en Cosmolaelaps n. sp., en dit kan biologische bestrijdingbeïnvloeden. Omdat plaaginsecten zoals thrips zowel bovengrondse plantendelenals de strooisellaag bewonen, zijn er verschillende pogingen geweest om trips tecontroleren met een combinatie van plantbewonende en bodembewonende natuur-lijke vijanden. In HOOFSTUK 4 onderzoek ik de bestrijding van trips met een combina-tie van A. swirskii en bodembewonende roofmijten zoals M. robustulus and S. scimi-tus. Nieuw in deze experimenten is dat ik verscheidene typen alternatief voedsel toe-dien voor de twee typen rovers (pollen bovengronds en astigmate mijten in de strooi-sellaag). Ik onderzoek bovendien het effect van het loslaten van de rovers voor danwel na een tripsinfectie, en van de frequentie waarmee pollen wordt toegevoegd optripsbestrijding. De voornaamste conclusie is dat het toevoegen van pollen voor debladbewonende rovers resulteert in de beste tripsbestrijding (zoals bij ‘apparentcompetition’), onafhankelijk van de aanwezigheid van rovers in de strooisellaag.Bovendien blijken de rovers in de strooisellaag de bestrijding van trips niet te beïn-vloeden. Zoals gebruikelijk in commerciële kweken, bevond de strooisellaag in ditexperiment zich op de bodem, onder de tafels waarop de planten groeien. Hierdoorwas de afstand van de planten tot die strooisellaag vrij groot. Toekomstige experi-

139

SAMENVATTING

menten zouden daarom kunnen achterhalen wat het effect is van het toevoegen vanstrooisel met rovers aan de voet van de planten, hetgeen zou kunnen resulteren ineen sterkere verbinding tussen de beide voedselwebben. Bestrijding van andere pla-gen door de predators die in de strooisellaag leven verdient meer studie omdat zegoed zijn aangepast aan het gewas en omdat hun dichtheden kunnen worden ver-hoogd door toediening van alternatief voedsel (HOOFDSTUK 3). Het loslaten van depredators nadat tripsaantasting was opgetreden resulteerde in onvoldoende bestrij-ding, met veel schade aan de planten. Wanneer predators werden losgelaten en bij-gevoerd met pollen voor de tripsaantasting, waren planten wél beschermd. Ik vondgeen verschil in tripsdichtheden en schade wanneer pollen wekelijks werd toege-diend en wanneer die toediening werd onderbroken. Er waren echter wel significantmeer rovers aanwezig op de planten die continu pollen ontvingen. Concluderend,laat ik in HOOFDSTUK 4 zien dat bladbewonende rovers moeten worden losgelatenvoordat trips invadeert om schade aan bloemen van siergewassen zoals roos te ver-minderen. Bovendien zou pollen wekelijks moeten worden toegevoegd. Ik geef aandat meer studie nodig is naar de frequentie van toediening en de kwaliteit van alter-natief voedsel. Ook de rol van predators in de strooisellaag bij de bestrijding van tripsen andere plagen verdient meer onderzoek.

In dit proefschrift presenteer ik diverse methoden die kunnen worden gebruikt bijbiologische bestrijding. Ik laat ook zien dat kasexperimenten kunnen helpen bij hettesten van ecologische theorie: ik toon aan dat het toevoegen van alternatief voed-sel voor generalistische predators voornamelijk resulteert in ‘apparent competition’;‘apparent mutualism’ werd af en toe waargenomen, en gedurende korte perioden,vooral wanneer de plaag aanwezig was voordat de predators werden losgelaten. Opbasis van dit proefschrift, suggereer ik twee factoren die meer theoretische aandachtverdienen: (1) het effect van gemengde diëten op ‘apparent competition’ en (2) heteffect van de frequentie van toediening van alternatief voedsel op de dynamica vanpredators en prooien. Ik laat ook zien dat de interacties tussen bovengrondse enondergrondse voedselwebben effect hebben op plaagdichtheden. Met betrekking totdit laatste onderwerp, beveel ik theoretische studies aan over het effect van meerde-re soorten van generalistische predators en het toedienen van voedsel bovengrondsen ondergronds op plaagbestrijding.

Samenvattend, concludeer ik dat voedselwebben in de strooisellaag en op debovengrondse plantendelen kunnen worden verbonden en dat dit kan resulteren inbetere biologische bestrijding in een siergewas. Kasexperimenten naar de dynamicavan meerdere predatorsoorten met meerdere voedselbronnen zijn cruciaal voor hetontwikkelen van nieuwe bestrijdingsmethoden. Zulke experimenten vormen tegelij-kertijd goede tests voor ecologische theorie.

140

SAMENVATTING

Curriculum vitae

Karen was born in Bogotá, Colombia, surrounded by her beloved – and luckily a bitcrazy and adventurous – relatives. She went to a catholic primary school in Soacha,a town near Bogotá. Karen and her family moved to Bogotá when she was 12 yearsold. There she attended a public technical high school. After graduating from highschool, Karen travelled to the USA to work and to learn English. In 2002, Karenreturned to Colombia to study a bachelor’s degree in Applied Biology at the NuevaGranada Military University (UMNG). Karen was, is and will be a pacifist, so going toa military university may sound paradoxical. However, the UMNG was affordable, andtheir biology program focused on the important role biodiversity plays in agriculture.

Karen became especially interested in the subject of biological control: visiting andconducting experiments in ornamental and aromatic crops, Karen observed that incrops with low pesticide use, there was a higher diversity of beneficial insects and mites,and a better quality of life for the growers. This had a strong impact on Karen’s life. That’swhy she decided to pursue a master’s degree focused on plant protection and biologi-cal control at the UMNG. During her postgraduate studies, Karen continued working asa research assistant for the Colombian Association of Flower Growers (ASOCOLFLO-RES). Her master’s thesis dealt with the integrated pest management in rose crops.

In 2008, Karen attended a workshop on agricultural acarology at Ohio State Univer -sity. There she met Prof. Maurice Sabelis. He accepted Karen in his team. In 2010, sheearned a scholarship from the Colombian Department of Science and Technology(COLCIENCIAS) to pursue a PhD program. In the year 2011, Karen did an internshipin a biological control company, met her soul mate, and began her PhD research.

Karen now works as an Environmental Risk Evaluator at the Dutch Board for theAuthorization of Plant Protection Products and Biocides (CTGB). In her new role, Karenwishes to continue highlighting the importance of biodiversity for sustainable agriculture.

Publications

Muñoz K, Fuentes L, Cantor F, Rodriguez D, Cure J (2009) Preferencia alimenticia del ácaro depredadorBalaustium sp. en condiciones controladas. Agronomía Colombiana 27: 95-103

Buitrago I, Muñoz K, Bustos A, Cantor F (2010) Evaluación de plantas hospederas para la producción deltrips Frankliniella occidentalis (Thysanoptera: thripidae) como suministro de presas para sus contro-ladores. Revista de la Facultad de Ciencias de la Universidad Militar Nueva Granada 6:12-23

Muñoz-Cárdenas K, Fuentes L, Cantor F, Rodríguez D, Janssen A, Sabelis MW (2014) Generalist red velvet mitepredator (Balaustium sp.) performs better on a mixed diet. Experimental and Applied acarology 62:19-32

141

Fuentes Quintero LS, Muñoz-Cárdenas K, Combita O, Jimeno E, Getiva de la Hoz JC, Cantor F, RodriguezD, Makol J (2014) A re-description of Balaustium leanderi Comb. Nov. (Actinotrichida, Erythraeidae) withfirst report on characteristics of all active instars and taxonomic notes on the genus. The FloridaEntomologist 97:937-951

Muñoz-Cárdenas K, Fuentes L, Rueda-Ramirez D, Rodríguez D, Cantor F (2015) The Erythraeoidea(Trombidiformes: Prostigmata) as biological control agents, with special reference to the genusBalaustium. In: Prospects for biological control of plant feeding mites and other harmful organisms, edit-ed by D Carrillo, GJ de Moraes, JE Peña, Chapter 8, Springer

Muñoz-Cárdenas K, Ersin F, Pijnakker J, Van Houten Y, Hoogerbrugge H, Leman A, Pappas M, Duarte M,Messelink G, Sabelis MW, Janssen A (2017) Supplying high-quality alternative prey in the litter increasescontrol of an above-ground plant pest by a generalist predator. Biological Control 105:19-26

142

CURRICULUM VITAE AND PUBLICATIONS

Acknowledgements

People say that a community is needed to raise a child and I think that a communi-ty is also needed to make a happy doctor in philosophy. I am deeply grateful with myfamily, friends and colleagues from all over the world that supported me during thechallenging years of my PhD.

I have no words to thank my supervisor Maurice Sabelis (Maus), sadly he is nothere to celebrate the end of this stage. He gave me the opportunity to join his teamin The Netherlands. I feel very fortunate to have listened to his talks and to have hadvery exciting discussions about biological control and ecology. To my co-supervisorArne, I am also deeply grateful, I learnt a lot from him and I really appreciate that hesupported me always in all my academic adventures. Special thanks also go to Peterde Ruiter for accepting the role of supervisor and for contributing during the writingperiod of my thesis.

Juliette, Ada, Firdevs, Maria, Gerben and Marcus, it was so cool to work with youand get to learn from your experience in laboratory and field experiments. Thanks forengaging conversations, meetings and presentations to my colleagues from IBED orthat passed by IBED, Iza, Martijn, Farid, Isabel, Michiel, Anieke, Paulien, Andre, Tom,Nienke, Zepeng, Dan Li, Yukie, George, Nicky, Sabina, Masoumeh, Seye, Merijn andhis team, Juan Ma, Bart, Alessandra, Carlos, Rachid, Jie and Josefine. I am gratefulfor all help and support from Mary, Maria, Tanya, Ludek and Harold. Paul, manythanks, I highly admire your work and I feel fortunate to have received your advicesregarding experiments which were always super useful. Louise, Bram and Floor, itwas fun to work together in the Ph(D) council. Thanks to Yvonne, Markus, Hans andKarel for your support during the internship before starting my PhD. To Tom and hisresearch team in Koppert, thanks for your support with materials and discussionsalso during my PhD. To Jan vA, your work with images and photos is excellent andit made me very happy. Jan B, thanks for editing the thesis before printing. Hecticor,gracias por ayudarme tanto con la escritura de mi tesis. To all amazing friends that Imet in The Netherlands, Laura, Patricia, Saioa, Nina, Heike, Kirsten, Monika, Hector,Dim, Rocío, Marian, Sofi, Daniela, Fernando, Joris, Livia, Felipe, Catalina, Thomas,Alexandra, Josinaldo, Oziel, Hidde, Leena, Ana, Paola, Aija and Olivier and otherfriends that I might forget to name, thanks for caring and for all the moments wespent together.

A mis amigos y colegas acarólogos, Luz Stella, Orlando, Elisa y Diana, siempre esmuy divertido trabajar juntos, espero que lo sigamos haciendo. Abuelita, tía Abelina,tía Ana, tío Jairo, Claudia, Danny, Betty, Isabella, Nana y Manu, Andrés, Will, Yheimy,

143

Jessica, Catalina, Alejita, Angie, Alex, Dieguito, Elizabeth y Maria Claudia, tía Rosa,tía Elizabeth y los Muñoz, doña María y los Gutiérrez, gracias por su amor y apoyoincondicionales. Tineke, Erkal en Emre dank jullie wel dat jullie mijn familie inNederland zijn. Mamá, Padre y Marquito, gracias por la fortaleza emocional quesiempre me transmiten, por ser fuertes y aventureros, y por enseñarme que todo enla vida se puede lograr. And finally, last but by no means least, mi amor, gracias porhacerme tan feliz.

144

ACKNOWLEDGEMENTS

UvA-DARE (Digital Academic Repository) What lies …K.A. Muñoz Cárdenas, 2017. What lies beneath?...

Documents

Transcript of UvA-DARE (Digital Academic Repository) What lies …K.A. Muñoz Cárdenas, 2017. What lies beneath?...