Challenges and Prospects in the Telemetry of Insects

Biol. Rev. (2014), 89, pp. 511–530. 511 doi: 10.1111/brv.12065 Challenges and prospects in the telemetry of insects

W. Daniel Kissling1,2,∗, David E. Pattemore3 and Melanie Hagen4 1Ecoinformatics & Biodiversity, Department of Bioscience, Aarhus University, Ny Munkegade 114, DK-08000 Aarhus C, Denmark 2Institute for Biodiversity and Ecosystem Dynamics (IBED), University of Amsterdam, PO Box 94248, 1090 GE Amsterdam, The Netherlands 3The New Zealand Institute for Plant & Food Research Limited, Private Bag 3230, Waikato Mail Centre, Hamilton 3240, New Zealand 4Genetics & Ecology, Department of Bioscience, Aarhus University, Ny Munkegade 114, DK-08000 Aarhus C, Denmark

ABSTRACT

Radio telemetry has been widely used to study the space use and movement behaviour of vertebrates, but transmitter sizes have only recently become small enough to allow tracking of insects under natural field conditions. Here, we review the available literature on insect telemetry using active (battery-powered) radio transmitters and compare this technology to harmonic radar and radio frequency identification (RFID) which use passive tags (i.e. without a battery). The first radio telemetry studies with insects were published in the late 1980s, and subsequent studies have addressed aspects of insect ecology, behaviour and evolution. Most insect telemetry studies have focused on habitat use and movement, including quantification of movement paths, home range sizes, habitat selection, and movement distances. Fewer studies have addressed foraging behaviour, activity patterns, migratory strategies, or evolutionary aspects. The majority of radio telemetry studies have been conducted outside the tropics, usually with beetles (Coleoptera) and crickets (Orthoptera), but bees (Hymenoptera), dobsonflies (Megaloptera), and dragonflies (Odonata) have also been radio-tracked. In contrast to the active transmitters used in radio telemetry, the much lower weight of harmonic radar and RFID tags allows them to be used with a broader range of insect taxa. However, the fixed detection zone of a stationary radar unit (< 1 km diameter) and the restricted detection distance of RFID tags (usually < 1–5 m) constitute major constraints of these technologies compared to radio telemetry. Most of the active transmitters in radio telemetry have been applied to insects with a body mass exceeding 1 g, but smaller species in the range 0.2–0.5 g (e.g. bumblebees and orchid bees) have now also been tracked. Current challenges of radio-tracking insects in the field are related to the constraints of a small transmitter, including short battery life (7–21 days), limited tracking range on the ground (100–500 m), and a transmitter weight that sometimes approaches the weight of a given insect (the ratio of tag mass to body mass varies from 2 to 100%). The attachment of radio transmitters may constrain insect behaviour and incur significant energetic costs, but few studies have addressed this in detail. Future radio telemetry studies should address (i) a larger number of species from different insect families and functional groups, (ii) a better coverage of tropical regions, (iii) intraspecific variability between sexes, ages, castes, and individuals, and (iv) a larger tracking range via aerial surveys with helicopters and aeroplanes equipped with external antennae. Furthermore, field and laboratory studies, including observational and experimental approaches as well as theoretical modelling, could help to clarify the behavioural and energetic consequences of transmitter attachment. Finally, the development of commercially available systems for automated tracking and potential future options of insect telemetry from space will provide exciting new avenues for quantifying movement and space use of insects from local to global spatial scales.

Key words: automated telemetry, body size, dispersal, invertebrates, landscape ecology, radio tag, radio tracking, receiver, satellite.

CONTENTS I. Introduction ...... 512 II. History of insect telemetry ...... 513 (1) Historical development ...... 514

* Author for correspondence (Tel: +31 20 525 6635; E-mail: [email protected]).

(a) The first pioneering studies ...... 514 (b) The last 10 years ...... 514 (c) Summary across taxa and studies ...... 515 (2) Main study aims ...... 516 (a) Movement ...... 516 (b) Habitat use ...... 517 (c)Behaviour ...... 517 (d)Migration ...... 518 (e) Evolution ...... 518 III. Comparison of radio telemetry with other techniques ...... 518 (1) Harmonic radar ...... 518 (2) Radio frequency identification ...... 520 (3) Comparison of passive tags to active transmitters ...... 521 (4) Alternatives to radio telemetry, harmonic radar and RFID ...... 521 IV. Challenges of tracking insects with active transmitters ...... 521 (1) Constraints of small size ...... 521 (a) Battery size ...... 521 (b) Limited tracking range ...... 522 (2) Behavioural effects of transmitter attachment ...... 522 (a) Observational and indirect evidence ...... 522 (b) Quantitative tests ...... 522 (3) Transmitter weights and energetic costs ...... 522 (a) Ratio of tag mass to body mass ...... 522 (b) Potential energetic costs of transmitter attachment ...... 523 V. Future prospects in radio telemetry with insects ...... 523 (1) Field observational studies ...... 523 (a) Expanding species coverage ...... 523 (b) Future priorities for field studies ...... 523 (2) Laboratory and field experiments and theoretical studies ...... 524 (a) Laboratory experiments ...... 524 (b) Theoretical models ...... 525 (3) Automated tracking systems ...... 525 (a) Automated data logging versus automated tracking ...... 525 (b) Presence/absence design ...... 525 (c) Triangulation design ...... 525 (4) Tracking from space ...... 526 (a) Space-based tracking systems ...... 526 (b) Prospects and limitations of tracking insects from space ...... 526 VI. Conclusions ...... 527 VII. Acknowledgements ...... 527 VIII. References ...... 527

I. INTRODUCTION Radio telemetry involves three primary components: an active (battery-powered) transmitter affixed to the animal An understanding of the movement and space use of animals that emits a radio signal [usually in the very high frequency is fundamental to basic and applied ecology (Kareiva & (VHF; 30–300 MHz) range], an antenna system, and a radio Shigesada, 1983; Turchin, 1998; Schick et al., 2008; Bowlin receiver (with or without a datalogger). The latter two detect et al., 2010; Manly et al., 2010). Techniques such as radio and process the radio signals that are emitted from the telemetry have been used for more than 50 years for studying radio transmitters (Fig. 1). The use of radio transmitters local and landscape-scale movements, habitat use, and has provided novel and unexpected insights into space dispersal of animals (Cochran & Lord, 1963; Craighead use and movement of free-ranging wild animals and has & Craighead, 1963; Millspaugh & Marzluff, 2001). They become a standard technique in wildlife biology, especially can also be used to measure physiological and energetic when applied to large-bodied vertebrates (White & Garrott, variables remotely (commonly referred to as biotelemetry; 1990; Kenward, 2001; Millspaugh & Marzluff, 2001). For Cooke et al., 2004), and this has been an important aspect of large-bodied birds, mammals, fishes and reptiles, basic radio wildlife telemetry since the first use of this technology (Essler transmitters have been extended to global positioning system & Folk, 1961; Mackay, 1964). (GPS) dataloggers, biotelemetry tools and satellite-based

(A) (B) (C)

Fig. 1. Principal elements of radio telemetry illustrated with a radio-tagged bumblebee. (A) An active radio transmitter is glued onto the body of the insect and emits a radio signal with regular pulses at a fixed radio frequency and a predetermined pulse rate. (B) Strength of the radio signal. It is strongest when an antenna is pointing directly at the transmitter. The pink line indicates signal gain in decibels arriving at different angles to the antenna’s main axis. (C) A yagi antenna picks up the radio signal broadcast by the transmitter. The tracking person can sweep the antenna from side-to-side to determine the direction of the strongest signal. tracking systems (Cooke et al., 2004; Cagnacci et al., 2010; well as some other approaches (e.g. homing experiments, Bridge et al., 2011). theoretical models, genetic microsatellite approaches, etc.) The key limiting factor in radio telemetry studies is the are also briefly mentioned. trade-off between transmitter weight versus power (i.e. range) Here, we provide an overview of radio telemetry studies and battery life (Wikelski et al., 2007). Because of this, the with insects. First, we review published studies of insect majority of studies to date have been conducted on large- telemetry and summarize the historical development and bodied vertebrates. However, because of continued advances the main study aims in this research field. Second, we in technology, weights of the most basic radio transmitters for provide an overview of key findings from harmonic radar vertebrates have been reduced to less than 5 g. Transmitters and RFID studies, compare advantages and disadvantages of weighing below 1 g are now available (Table 1), which allow both techniques to radio telemetry with active transmitters, the tracking of small-bodied animals such as insects and and highlight a few other alternatives for studying insect invertebrates in the field (Fig. 2). These small transmitters movement. Third, we identify current challenges for tracking are glue-on models that are simply attached to the body insects with active transmitters, including the constraints of animals with an adhesive, sticky paste or glue (Fig. 2). of a small battery, the behavioural effects of transmitter Additional electronics for GPS, archival loggers or satellite attachment, and the potential effects of transmitter weights retrieval are usually not included as they are too heavy for on the insect’s energetic costs. Finally, we highlight future small-bodied species in most cases (Bridge et al., 2011). prospects of this research field, including the extension of Active transmitters (as used in radio telemetry) are powered fieldwork, experimental and theoretical studies, and future by batteries that are the main contributor to the weight possibilities such as the development of automated tracking of the tags. As an alternative, there are also passive tags systems or the tracking of insects from space. Throughout which do not incorporate their own power source and our review, we consider both local spatial extents (e.g. movements within habitat patches or at grain sizes below hence can be reduced to a fraction of the weight of active 2 transmitters. These passive tags can also be used to study 10 km ) and continental to global spatial scales (e.g. when movement behaviour of animals. For instance, harmonic studying migration), and highlight the utility of different radio telemetry approaches relative to the scale of the study radar tags transpose an incoming signal (sent out by a radar objective. station) to a different frequency. The reflected signal then returns to the radar station and can be distinguished from other radar-reflective objects. Radio frequency identification (RFID) tags involve a circuit that only transmits when a II. HISTORY OF INSECT TELEMETRY powerful radio signal is sent to it. These circuits are very light (< 10 mg) and allow individual identification of animals, The following historical overview of case studies on insect but the range between RFID tag and transceiver usually telemetry is based on a literature search in the Web of Science needs to be below 1 m. Both harmonic radar and RFID (WoS)andScopus database (September and October 2012). have been used widely to study insect movements. We focus Using the key words ‘insect telemetry’, we searched WoS our review on radio telemetry using active transmitters, but and Scopus across all journals. In addition, we used the key key findings from harmonic radar and RFID studies as words ‘telemetry’, ‘radio tracking’ or ‘radiotracking’ to search

Table 1. Characteristics of the smallest and lightest currently commercially available (October 2012) active radio transmitters. The transmitters listed here are provided by companies that have been mentioned in published case studies on insect telemetry (see Table 2)

Battery Pulse Pulse lifespanb rateb widthb Tag name Mass (g) Sizea (mm) (days) (ppm) (ms) Companyc Homepage A2412 0.20 12 × 5 × 1.5 9–22 15–40 15 Advanced Telemetry Systems www.atstrack.com LB-2X 0.22 8 × 5.3 × 2.8 4–8 40 20 Holohil Systems www.holohil.com PicoPip 0.19–0.24 12 × 5 × 2 3–5 30–50 20 Biotrack www.biotrack.co.uk LT6-337 0.47 NA 12–22 40–57 10–20 Titley Scientific www.titley-scientific.com ZV1G 102 0.95 14 × 8 × 5 26–32 30–40 18 Sirtrack www.sirtrack.com G3-1 V > 113× 8 × 5 7–10 50 15–20 AVM Instrument Company www.avminstrument.com aThe size of a transmitter is given as length × width × height. bThe battery lifespan depends on the pulse rate and pulse width of the transmitter. For a fast pulse rate, the battery life span is usually shorter than for a slow pulse rate. A slower pulse rate can thus lead to a longer lifespan of the battery. ppm, pulses per minute. cThe companies listed here include all those mentioned in published case studies on insect telemetry (except Sparrow Systems, which had no accessible information on its homepage: www.sparrowtracking.com). NA, information not available. specific insect journals, including Journal of Orthoptera Research, invertebrates such as freshwater crabs (Gherardi & Vannini, Journal of Insect Conservation, Ecological Entomology, European 1989). Journal of Entomology, Apidologie, Environmental Entomology, For terrestrial insects, the first telemetry study was Journal of Insect Science, Insectes Sociaux,andInternational Journal published at the beginning of the 1990s (Fig. 3A). It of Tropical Insect Science. Finally, we used our own literature focused on a carabid beetle (Carabus coriaceus) with the databases and other studies we were aware of. aim of testing the potential of small radio transmitters for We identified 27 case studies of insect radio telemetry tracing the movements of non-flying terrestrial insects at (summarized in Table 2), i.e. studies where active (battery- the landscape scale (Riecken & Ries, 1992). The study was powered) radio transmitters had been used to study insect later extended to quantify the habitat use and dispersal movement, habitat use, behaviour, migration or evolution. of this large (30–40 mm long) Central European carabid We focused on insects and included species from both the beetle (Riecken & Raths, 1996). A first preliminary radio aquatic and terrestrial realms. Unless otherwise stated, we telemetry study on the threatened New Zealand giant weta did not include other invertebrates such as spiders (Janowski- (Deinacrida heteracantha) was also conducted at the end of Bell & Horner, 1999) or freshwater invertebrates (Gherardi & the 1990s, but the few individuals available only provided Vannini, 1989; Robinson, Thom & Lucas, 2000). We also did limited insights into habitat use and movement (Gibbs & not specifically search for unpublished theses or unpublished McIntyre, 1997). After the turn of the millennium, the reports because they are hard to find and gaining access to first radio telemetry studies on flying terrestrial insects were them is difficult. Studies using harmonic radar, RFID tags published (Fig. 3A). They focused on large-bodied beetles or other alternative approaches are briefly summarized in (Lucanus cervus, Osmoderma eremita, Scapanes australis)andwere Section III. mainly interested in local movements and habitat use, either from a habitat fragmentation and conservation perspective (Sprecher-Uebersax & Durrer, 2001; Hedin & Ranius, 2002) (1) Historical development or from an interest in pest control (Beaudoin-Ollivier et al., (a) The first pioneering studies 2003). At the same time, studies on a non-flying bush-cricket (Anabrus simplex) revealed differences in movement rates and More than two decades ago, the first radio telemetry studies directionality in outbreak and non-outbreak populations were conducted with free-ranging insects (Fig. 3A). These (Lorch & Gwynne, 2000; Lorch et al., 2005). pioneering studies (Hayashi & Nakane, 1988, 1989) were focused on the non-flying larvae of an aquatic insect rather than on terrestrial species. They aimed at understanding (b) The last 10 years the foraging movements and activity patterns of dobsonfly Over the last 5–10 years, there has been a marked increase larvae (Protohermes grandis) in their natural stream habitats. in insect telemetry studies compared to previous years (Fig. The transmitter was already very small (7 × 18 × 2 mm) 3A). Besides new telemetry studies on beetles (Rink & Sinsch, and light (0.25 g in water), and their mass approximated c. 2007; Dubois & Vignon, 2008; Hedin et al., 2008; Negro 10–40% of the body mass of the larvae. The tags did not et al., 2008; Svensson et al., 2011) and crickets (Kelly, Bussiere have an antenna, so the signal range of the transmitter was & Gwynne, 2008; Stringer & Chappell, 2008; Sword, Lorch very limited (2 m). At the same time (in the late 1980s), the & Gwynne, 2008; Watts & Thornburrow, 2011; Fornoff, first telemetry studies were published on other (flightless) Dechmann & Wikelski, 2012; Watts et al., 2012), the range of

(A) (B)

(E) (F)

Fig. 2. Examples of active radio transmitters attached to different insect species. (A) Male orchid bee (Exaerete frontalis)witha 300 mg transmitter attached to the thorax. (B) Bumblebee (Bombus terrestris) with a 200 mg transmitter attached to the dorsal upper abdomen. (C) Neotropical orchid bee (Eulaema sp.) with a 220 mg transmitter attached to its upper dorsal abdomen. (D) Neotropical katydid (Philophyllia ingens; Orthoptera: Tettigoniidae) with a 300 mg transmitter attached to the upper thorax. (E) Monarch butterfly (Danaus plexippus) with a 200 mg transmitter attached to the lower frontal abdomen. (F) Neotropical moth (Sphingidae) with a 220 mg transmitter attached to the lower abdomen. Photograph credits: Christian Ziegler (A, C), W. Daniel Kissling (B), and Martin Wikelski (D–F). radio-tagged insect taxa has been expanded to other groups bumblebees (Hagen et al., 2011) and dragonflies (Levett & such as dragonflies (Wikelski et al., 2006; Levett & Walls, Walls, 2011). In addition, a number of telemetry studies 2011) and bees (Pasquet et al., 2008; Wikelski et al., 2010; have focused on several species of giant weta (Orthoptera: Hagen, Wikelski & Kissling, 2011). The first study on drag- Anostostomatidae) endemic to New Zealand, mostly with onflies revealed individual migration tactics of green darners the aim of understanding the movements, habitat use and (Anax junius), their distinct stopover and migration periods, behaviour of these insects after translocation to islands or and the dependence of individual migratory decisions on mammal-free mainland sanctuaries (Stringer & Chappell, specific weather patterns (Wikelski et al., 2006). The first 2008; Watts & Thornburrow, 2011; Watts et al., 2012). telemetry study on bees was published 2 years later (Fig. 3A), with a focus on carpenter bees (Xylocopa flavorufa)andtheir role in pollinating cowpeas (Vigna unguiculata) (Pasquet et al., (c) Summary across taxa and studies 2008). Subsequently, other studies have demonstrated the Out of the 27 case studies reviewed here (Table 2), 20 (74%) great potential of radio-tracking for understanding space use have been conducted with either beetles or crickets (Fig. 3B). and movement behaviour of bees and dragonflies, including Radio-tagged beetles include ground beetles (Coleoptera: tropical orchid bees (Wikelski et al., 2010) as well as temperate Carabidae), scarab beetles (Coleoptera: Scarabaeidae),

Table 2. Summary of case studies (N = 27) using radio telemetry of insects. The case studies are grouped alphabetically within insect orders (beetles, crickets, bees, dobsonflies, and dragonflies). Organism characteristics describe species names, insect family, and whether flying (yes/no) or of tropical affinity (yes/no). The aims of each case study (yes/no) are summarized within five categories: HAB, habitat use and selection; MOV, movement distance; BEH, behaviour; MIG, migration; EVO, evolution

Organism characteristics Study aims Case study Species Family Flying Tropical HAB MOV BEH MIG EVO Beetles (Coleoptera) Beaudoin-Ollivier et al. Scapanes australis Scarabaeidae Yes Yes Yes Yes Yes No No (2003) Dubois & Vignon (2008) Osmoderma eremita Scarabaeidae Yes No Yes Yes No No No Hedin & Ranius (2002) Osmoderma eremita Scarabaeidae Yes No No Yes No No No Hedin et al. (2008) Osmoderma eremita Scarabaeidae Yes No No Yes No No No Negro et al. (2008) Carabus olympiae Carabidae No No Yes Yes Yes No No Riecken & Ries (1992) Carabus coriaceus Carabidae No No No Yes No No No Riecken & Raths (1996) Carabus coriaceus Carabidae No No Yes Yes Yes No No Rink & Sinsch (2007) Lucanus cervus Lucanidae Yes No No Yes Yes No No Sprecher-Uebersax & Lucanus cervus Lucanidae Yes No Yes Yes No No No Durrer (2001) Svensson et al. (2011) Osmoderma eremita Scarabaeidae Yes No No Yes No No No Crickets (Orthoptera) Fornoff et al. (2012) Philophyllia ingens Tettigoniidae Yes Yes Yes Yes Yes No No Gibbs & McIntyre (1997) Deinacrida heteracantha Anostostomatidae No No No Yes No No No Kelly et al. (2008) Deinacrida rugosa Anostostomatidae No No No Yes No No Yes Lorch & Gwynne (2000) Anabrus simplex Tettigoniidae No No No Yes No Yes No Lorch et al. (2005) Anabrus simplex Tettigoniidae No No No Yes No Yes No Srygley et al. (2009) Anabrus simplex Tettigoniidae No No No No No Yes No Stringer & Chappell Motuweta isolata Anostostomatidae No No No Yes No No No (2008) Sword et al. (2008) Anabrus simplex Tettigoniidae No No No Yes No Yes No Watts & Thornburrow Deinacrida heteracantha Anostostomatidae No No Yes Yes Yes No No (2011) Watts et al. (2012) Deinacrida rugosa Anostostomatidae No No Yes Yes Yes No No Bees (Hymenoptera) Hagen et al. (2011) Bombus terrestris, B. hortorum, Apidae Yes No Yes Yes Yes No No B. ruderatus Pasquet et al. (2008) Xylocopa flavorufa Apidae Yes Yes No Yes No No No Wikelski et al. (2010) Exaerete frontalis Apidae Yes Yes Yes Yes No No No Dobsonflies (Megaloptera) Hayashi & Nakane (1988) Protohermes grandis Corydalidae No No No Yes Yes No No Hayashi & Nakane (1989) Protohermes grandis Corydalidae No No No Yes Yes No No Dragonflies (Odonata) Levett & Walls (2011) Anax imperator Aeshnidae Yes No Yes Yes No No No Wikelski et al. (2006) Anax junius Aeshnidae Yes No No Yes No Yes No

and stag beetles (Coleoptera: Lucanidae), while crickets evolution (Table 2). The most common study objective is include bush-crickets (Orthoptera: Tettigoniidae) and wetas to obtain quantitative estimates of insect movement (Fig. (Orthoptera: Anostostomatidae) (Table 2). A smaller number 3C), especially average and maximum movement distances of studies (N = 7) have focused on other insect orders such under natural conditions. For instance, recent landscape- as bees (Hymenoptera: Apidae), dobsonflies (Megaloptera: level telemetry studies of bees have revealed maximum flight Corydalidae), and dragonflies (Odonata: Aeshnidae) (Fig. distances of 2.5, 5 and 6 km within a few days for bumblebees 3B). An almost equal number of studies has addressed flying (Hagen et al., 2011), orchid bees (Wikelski et al., 2010), and and non-flying insects (Fig. 3B), but most studies have been carpenter bees (Pasquet et al., 2008), respectively. Other stud- conducted in non-tropical regions (Fig. 3B). ies on ground-living carabid beetles (Riecken & Raths, 1996; Negro et al., 2008) or flightless crickets (Kelly et al., 2008; (2) Main study aims Watts & Thornburrow, 2011; Watts et al., 2012) suggest movement distances to be less than 100 m per day (usually at (a) Movement night, with males covering greater distances than females). The main study aims in insect telemetry can be grouped Total lengths of movement paths can be up to several hun- into movement, habitat use, behaviour, migration, and dred metres within a 2–3 week time period (Riecken & Raths,

Bees and 1996; Negro et al., 2008; Watts & Thornburrow, 2011) or dragon- even over 1 km within 1–2 months (Watts et al., 2012). In con- Aquatic Flying flies trast to flightless insects, beetles with flight abilities have been non-flying terrestrial (A) insect observed to cover larger distances, with single long-distance 12 beetles larvae and flights of 50–1700 m (Sprecher-Uebersax & Durrer, 2001; crickets Beaudoin-Ollivier et al., 2003; Rink & Sinsch, 2007; Dubois & 10 Non- flying Vignon, 2008) and maximum recorded movement distances 8 terrestrial of over 2 km (Rink & Sinsch, 2007). However, movement 6 insects distances of flying insects might also differ between sexes (as for non-flying insects), with males covering larger distances 4 than females or vice versa, depending on the species considered

Number of studies Number of 2 (Sprecher-Uebersax & Durrer, 2001; Beaudoin-Ollivier et al., 2003; Rink & Sinsch, 2007). Besides horizontal movements, 0 radio telemetry has also been used to study vertical movements. For instance, radio-tracking of a bush-cricket (Philo- phyllia ingens) in a tropical rainforest revealed that these insects 1990-1994 2000-2004 2010-2012 1985-1989 1995-1999 2005-2009 are more or less stationary at 10 m above ground throughout most of the day, but that they perform upward movements (B) on trees starting at sunrise, followed by distinct downward 25 movements beginning at noon (Fornoff et al., 2012). 20 (b) Habitat use 15 Another widely addressed study aim in insect telemetry 10 is habitat use (Fig. 3C), including studies on proportional habitat preferences, home range requirements, and resource 5 Number of studies Number of selection of insects (Table 2). For instance, radio-tracking 0 has been used to study the habitat selection of a ground beetle (Carabus olympiae) at a landscape scale, suggesting that Bee Flying

Beetle beech forests and subalpine shrubs are the preferred habitat Cricket Tropical Dragonfly

Dobsonfly for this endemic and endangered species (Negro et al., 2008). Non-flying

Non-tropical At finer spatial scales, the microhabitat selection of insects (C) can also be studied with radio-tracking. For instance, the proportional use of specific food plants has been studied in 25 giant wetas (Deinacrida heteracantha) (Watts & Thornburrow, 20 2011), whereas significant differences between sexes in the use of woodpiles, stumps, open areas, and trees were revealed for 15 astagbeetle(Lucanus cervus) (Sprecher-Uebersax & Durrer, 10 2001). Other studies have used telemetry to quantify the occupancy rate of hollow trees by a threatened beetle

Number of studies Number of 5 (Dubois & Vignon, 2008), the precise location of feeding and mating sites of pest beetles (Beaudoin-Ollivier et al., 2003), 0 or the preference and avoidance of certain habitat types by bumblebees in a Central European landscape mosaic (Hagen et al., 2011). Radio telemetry studies have also obtained Evolution Migration Behaviour Movement Habitat use first quantitative estimates of home range sizes in beetles Fig. 3. Number of case studies on insect telemetry versus (A) time (Sprecher-Uebersax & Durrer, 2001), bees (Wikelski et al., (year of publication), (B) characteristics of radio tagged insects, 2010; Hagen et al., 2011), crickets (Fornoff et al., 2012), and and (C) main study aims. The black arrows in A denote the first dragonflies (Levett & Walls, 2011). These results on home time at which a specific ecological group of insects was radio- range sizes are likely to be underestimates because of the tracked in the field. In B, the characteristics of studied species short battery lifespan, the limited tracking range of the are insect order (beetle, bee, cricket, dragonfly, dobsonfly), flying transmitters, and the effects of tag weight on behaviour and versus non-flying, and tropical versus non-tropical species. In C, energetic costs (see Section IV). the main study aims are divided into movement, habitat use, behaviour, migration, and evolution. In A and B, the number of studies adds up to the total (N = 27) within those bars that (c) Behaviour have the same colour. In C, a single study can count multiple Besides movement distances and habitat use, radio telemetry times (i.e. have multiple study aims). Compare with Table 2. has also been used to understand the basic behaviour of

Biological Reviews 89 (2014) 511–530 © 2013 The Authors. Biological Reviews © 2013 Cambridge Philosophical Society 518 W. Daniel Kissling and others insects (Fig. 3C). The first studies of radio telemetry in insects (e) Evolution were aimed at understanding the foraging behaviour of the Finally, radio telemetry can also be used to study questions aquatic, stream-dwelling dobsonfly larvae (Protohermes grandis) in evolution, although this has been done rarely to date (Hayashi & Nakane, 1988, 1989). The results showed that (Fig. 3C). An interesting study on the Cook Strait giant these predatory insect larvae do not actively search for prey weta (Deinacrida rugosa) in New Zealand tested hypotheses of but instead use ambushing as the mode of foraging. For a sexual (female-biased) size dimorphism (Kelly et al., 2008). terrestrial beetle (Lucanus cervus), radio telemetry showed that Using radio telemetry, the authors found that adult male males (but not females) climb up to elevated structures (trees, giant weta with longer legs and smaller bodies travelled shrubs) before they take off for flight (Rink & Sinsch, 2007). significantly further per night and accrued significantly Radio tags further allowed quantification of the relative greater insemination success. This suggests a strong mobility- amount of time spent for different behaviours, including driven selection pressure on males, with greater mobility resting, foraging and moving of bumblebees (Hagen et al., favouring smaller males, whereas fecundity selection might 2011) and giant wetas (Watts et al., 2012). However, the favour larger females (Kelly et al., 2008). Hence, radio large mass of transmitters relative to the body mass of telemetry has also the potential to contribute to our insects might influence such behaviour (Hagen et al., 2011). understanding of evolution and natural selection if this is Field studies on a scarab beetle (Scapanes australis) revealed related to space use and movement of insects. differences in flight behaviour between sexes, with males typically flying within 5 m of the ground and females flying to above canopy level (> 20 m) and then further away (Beaudoin-Ollivier et al., 2003). Radio telemetry has also III. COMPARISON OF RADIO TELEMETRY been used to quantify activity periods of beetles (Riecken & WITH OTHER TECHNIQUES Raths, 1996; Sprecher-Uebersax & Durrer, 2001; Beaudoin- Ollivier et al., 2003; Negro et al., 2008), crickets (Fornoff et al., Besides radio telemetry, there are two established tagging 2012) and dragonflies (Levett & Walls, 2011) in day and alternatives which have also been used extensively to study night time. While studies on carabid and scarab beetles the movement and activity patterns of insects: harmonic mostly confirmed their predominantly nocturnal activity, radar and radio frequency identification (RFID) (Fig. 4). radio telemetry of a tropical bush-cricket (Philophyllia ingens) Both use passive tags which do not have their own power revealed peaks of activity during the day (probably feeding source (i.e. no battery as in an active transmitter). Harmonic and/or stridulating activity), despite the insect being assumed radar and RFID tags can therefore be reduced to a fraction to be predominantly nocturnal (Fornoff et al., 2012). Radio of the weight of an active transmitter (Table 3). We here give telemetry studies can therefore reveal unexpected insights a short overview of these two technologies, provide some into insect behaviour. examples of their application in insect movement studies, and compare their advantages and disadvantages to radio (d) Migration telemetry with active transmitters. We also briefly highlight alternatives to radio telemetry, harmonic radar and RFID. Another important study aim addressed in five case studies is insect migration (Table 2). Countless numbers of insects (1) Harmonic radar migrate every year between and within continents, but little is known about individual long-distance trajectories and the Radar has been used by entomologists for over 40 years ultimate reasons for mass movements (Holland, Wikelski & (Chapman, Drake & Reynolds, 2011; Drake & Reynolds, Wilcove, 2006). A ground-breaking telemetry study in insect 2012). The first introduction of a radar system to entomology migration followed the migratory pathways of a dragonfly allowed observing high-flying insects at distances of up to using Cessna airplanes equipped with receivers and ground 2 km (Schaefer, 1969; Riley, 1975). This technique (often teams (Wikelski et al., 2006). The study showed that the green referred to as ‘vertical-looking entomological radar’, VLR) darner (Anax junius) migrated every 3 days with an average has been used successfully to study movements (e.g. flight net advance of 58 km in 6 days, having distinct stopover orientation or migratory activity) of large (> 100 mg) insects and migration days similar to those proposed for songbirds. (e.g. moths or locusts) in the air without the need to attach Besides dragonfly migration, four case studies have focused a tag (Chapman, Reynolds & Smith, 2003a;Chapman on the migratory activity of mormon crickets (Anabrus simplex) et al., 2010; Chilson et al., 2012; Drake & Reynolds, 2012). (Table 2). Average migration distances, directionality, and For low-flying insects, such classical radar observations are intrinsic versus extrinsic (environmental) cues were studied only possible over very flat and featureless terrain because in outbreak and non-outbreak populations during mass terrain features and vegetation strongly echo the radar signals migrations (Lorch & Gwynne, 2000; Lorch et al., 2005; (Riley & Smith, 2002). In the mid-1980s, the technique of Sword et al., 2008), and the immediate effects of protein and harmonic radar was therefore introduced to study movement carbohydrate diets on migratory activity could be quantified behaviour and space use of low-flying insects (Mascanzoni (Srygley et al., 2009). Radio telemetry thus has great potential & Wallin, 1986; Riley et al., 1996). This approach uses a to gain insights into animal migration, although few studies small passive transponder (attached to the insect) that re- have so far used this methodology with insects (Fig. 3C). radiates transmissions at exactly half the wavelength of the

(A) (B) (D)

(C)

(E) (F) (G)

(H)

Fig. 4. Equipment for studying insect movement and foraging activity with passive tags, including scanning harmonic radar (A–D) and radio frequency identification (RFID, E–H). The black arrows in photographs of insects indicate the position of tags. (A) Scanning harmonic radar from Rothamsted, Hertfordshire, UK. Signals returned from tagged insects are captured with the smaller (0.76 m diameter) parabolic reflector mounted above the larger transmitting antenna (for details see Riley & Smith, 2002; Ovaskainen et al., 2008). (B) A honeybee carrying a harmonic radar transponder designed for the use with the Rothamsted scanning harmonic radar. (C) A large white butterfly (Pieris brassicae) with a harmonic radar transponder. (D) A bumblebee (Bombus terrestris) with a harmonic radar transponder. (E) Two microsensys RFID readers linked by colourless Perspex tubing and connected to the entrance of a bumblebee (Bombus terrestris) nest (for details see Gill et al., 2012). (F) A honeybee carrying a RFID tag (see Henry et al., 2012). (G) Bumblebee (Bombus terrestris) workers with RFID tags. (H) An eusocial paper wasp (Polistes canadensis)withaRFIDtag attached (for details see Sumner et al., 2007). Photo credits: Alan D. Smith (A), Andrew Martin (B), Juliet Osborne (C), Stephan Wolf (D), Oscar Ramos-Rodriguez (E), ACTA & Axel Decourtye (F), Nigel Raine (G), Aidan Weatherill (H). original transmitted wave, and thus allows tagged animals method are up to almost 1 km (Table 3) when assuming a to be distinguished from other radar-reflecting objects such flat terrain and a transmitted wavelength of 3.2 cm at 25 kW as terrain features or vegetation (Mascanzoni & Wallin, (Riley & Smith, 2002). The technique requires a powerful 1986; Chapman, Reynolds & Smith, 2004; Chapman et al., (and costly) radar device, but the costs for individual tags 2011). The very simple tags (including a wire and diode) are minimal which allows many replicates once the initial weigh only a few milligrams (6–20 mg) and can therefore system is purchased and set up. A number of studies have be carried by many insects without any obvious effects on used this ground-based scanning harmonic radar to obtain behaviour. However, the length of the wire/antenna needed insights into the movement and flight behaviour of insects to achieve sufficient range can be problematic for some insect (for good overviews see Chapman et al., 2011; Drake & species. Reynolds, 2012). For instance, studies with honeybees have Two forms of harmonic radar can be distinguished. The provided insights into their waggle-dance communication more sophisticated version consists of a ground-based (i.e. (Riley et al., 2005), their ability to compute optimum flight stationary) scanning harmonic radar station in which the vectors between resources (Menzel et al., 2011), their re- movement of a tagged insect is tracked on a circular radar orientation following displacement (Riley et al., 2003, 2005; display (Riley & Smith, 2002; Drake & Reynolds, 2012). Reynolds et al., 2007), and their map-like spatial memory This radar system allows recording both the range and for navigation (Menzel et al., 2005). Ground-based scanning the direction of the insect. The detection distances for this harmonic radar has also been used successfully with other

Table 3. Characteristics of three key techniques that employ active transmitters (radio telemetry) or passive tags (scanning harmonic radar, radio frequency identiﬁcation) to study the movement and activity patterns of insects. Harmonic radar refers here to ground-based scanning but does not include handheld receivers (see text for details)

Scanning harmonic Radio frequency Characteristic Radio telemetry (RT) radar (SHR) identification (RFID) Type of tag Active (with battery) Passive (without power source) Passive (without power source) Tag mass (mg)a 200–1000 1–20 0.9–100 Signal range (m) on groundb 100–500 250–900 Usually < 1 (often < 0.1, sometimes < 0.001) Spatial extent of study focus Landscape to global Landscape Local Differentiates between Yes No Yes individuals (unique signal) Mobility of system Mostly mobile (rarely fixed) Fixed Mobile or fixed Early applications (year of 1988 1996 2003 publication)c Main limitationsd Heavy transmitters, long Long antenna, terrain and Very short signal range antenna vegetation as barriers, signal without identification information, technically demanding aRT (see references in Table 4); SHR (O’Neal et al., 2004; Riley et al., 2005); RFID (Streit et al., 2003; Robinson et al., 2009a; Silcox et al., 2011; Schneider et al., 2012). bRT (see text for details); SHR: (Riley et al., 1996; Osborne et al., 1999; Riley & Smith, 2002; Menzel et al., 2005); RFID (Streit et al., 2003; Silcox et al., 2011; Schneider et al., 2012). cRT (Hayashi & Nakane, 1988); SHR (Riley et al., 1996); RFID (Streit et al., 2003). dRT (Hagen et al., 2011); SHR (O’Neal et al., 2004; Chapman et al., 2011); RFID (Streit et al., 2003; Silcox et al., 2011; Gill et al., 2012). insects such as bumblebees (Osborne et al., 1999; Riley et al., identification. The main use of RFID has been for the long- 1999; Lihoreau et al., 2012), moths (Riley et al., 1998), and term identification of a large number of individuals. The butterflies (Cant et al., 2005; Ovaskainen et al., 2008). effective signal range between the reader and tag varies with A more cost-effective version of harmonic radar is a light- the power output of the antenna, the radio frequency, and weight handheld receiver (O’Neal et al., 2004; Psychoudakis the size of tag (Table 3). It can reach up to 30 m or more, et al., 2008), often referred to as a ‘harmonic direction but usually detection is below 1–5 m (Domdouzis, Kumar finder’ (Chapman et al., 2011). This form of harmonic & Anumba, 2007). For insects, RFID tags have to be very radar transmits at a longer wavelength compared to the small and their detection is usually in the range of a few ground-based scanning radar. The detection distances for centimetres or even millimetres (Reynolds & Riley, 2002; tags held above the ground are therefore only in the range Silcox et al., 2011). They can weigh as little as 0.9–4 mg of 30–60 m and even shorter (< 10 m) when the tag is on (Streit et al., 2003; Robinson et al., 2009a; Schneider et al., the ground (Mascanzoni & Wallin, 1986; Lovei¨ et al., 1997; 2012) and their small volume and low profile means they can O’Neal et al., 2004; Psychoudakis et al., 2008). Moreover, the be glued onto a diverse range of insect species. handheld harmonic devices used with insects often only allow RFID tags have been used in a number of insect measurement of the direction of the signal, but not the range studies. For instance, impacts of pesticides on individual (Drake & Reynolds, 2012). Ground-based scanning radar, survival rates, foraging behaviour, homing ability and activity by contrast, measures range as well as direction. A series of patterns have been studied in honeybee and bumblebee studies have used handheld harmonic radar to investigate colonies using RFID (Decourtye et al., 2011; Gill, Ramos- the short-term and short-distance movement behaviour of Rodriguez & Raine, 2012; Henry et al., 2012; Schneider beetles including both ground-dwelling (e.g. Mascanzoni & et al., 2012). In such studies, readers were placed at the Wallin, 1986; Lovei¨ et al., 1997; O’Neal et al., 2004) as well as entrance of colonies to assess in- and out-going movements flying species (e.g. Boiteau & Colpitts, 2001, 2004). However, from colonies or specially designed feeding stations (e.g. Fig. due to the limited detection range this system is applied less 4E). Similar approaches have been applied to study forager often to insect taxa than is the ground-based scanning radar. recruitment via pheromones in bumblebee colonies (Molet et al., 2008) or nest-drifting behaviour in natural populations (2) Radio frequency identification of eusocial paper wasps (Sumner et al., 2007). RFID tags RFID is a wireless sensor technology based on detecting have also been successfully applied in studies with ants electromagnetic signals. The antenna and receiver unit that (Robinson et al., 2009a, b), beetles (Vinatier et al., 2010), and reads the RFID tags is called a ‘reader’. The reader powers other surface-dwelling or subterranean insects (Silcox et al., a passive RFID tag wirelessly to emit a radio signal for 2011).

(3) Comparison of passive tags to active minimum (rather than maximum) movement distances (as transmitters do harmonic radar and RFID). Homing experiments have been used to approximate The principal advantage of passive tags (harmonic radar, maximum flight distances of central place foragers (mostly RFID) over active transmitters (radio telemetry) is their hymenopterans), i.e. species that bring resources back to much smaller weight (Table 3). This allows the application their nests. This approach allows investigation of an animal’s of passive tags to a much broader range of insects including ability to return to its home after displacement to unfamiliar the economically important honey bee; weight constrains regions. It has been used, for instance, with honeybees the use of active transmitters to the largest insect species (Capaldi & Dyer, 1999), solitary Osmia spp. bees (Guedot,´ (see Section IV). The lack of a battery also greatly extends Bosch & Kemp, 2009), social bumblebees (Goulson & Stout, the potential operational lifespan of the passive tags and 2001), and ants (Collett, Graham & Harris, 2007). Similarly, considerably reduces their costs. This means that a much theoretical models can be used to predict maximum flight larger number of individuals can be tracked over a longer ranges of insects, e.g. by correlating flight range with body time period once the harmonic radar station or RFID reader size (Greenleaf et al., 2007) or with energy budgets based on has been purchased. optimal foraging theory (e.g. Cresswell, Osborne & Goulson, The fixed range of a stationary radar unit limits the 2000). flexibility of the system compared to a hand-held antenna Several techniques are available to measure minimum and receiver unit as used in radio telemetry. Ground-based (rather than maximum) movement distances. For instance, scanning harmonic radar can have a signal range of almost genetic microsatellite approaches are used to determine 1 km (Table 3), but detection of larger scale movements sibling relationships of foraging insects and therefore allow is problematic due to the stationary design of the system. quantifying foraging range and nest density (Chapman, Moreover, harmonic radar does not return a uniquely Wang & Bourke, 2003b; Darvill, Knight & Goulson, 2004; identifiable signal and is thus unable to identify individuals Kraus, Wolf & Moritz, 2009). Insects can also be marked in most cases. Individuals can only be identified if tags are individually with simple labels or tags to follow their local used sequentially rather than simultaneously, or if the studied movements (Hagler & Jackson, 2001). Combined with mark- insect behaviour occurs over a small spatial and temporal reobservation methods, this allows tracking of the movement scale so that each individual can be tracked continuously. of marked individuals within or across study sites (e.g. Ground-based scanning harmonic radar is best suited for Dramstad, 1996; Walther-Hellwig & Frankl, 2000). This the analysis of movement and foraging behaviour in flat, tagging technique is inexpensive, but usually tedious and open habitat of < 10 ha. Harmonic radar using a handheld time consuming, and restricted to the local scale. LIDAR receiver is even more limited in space. It is ideally suited for uses light (laser) to measure reflectance of objects, in contrast studying movements of ground-dwelling insects at the scale to radar which uses radio waves. It has been applied to track of a few metres. insect movements, e.g. those of honeybees at a patch scale Like active transmitters, RFID tags allow the identification (Shaw et al., 2005). LIDAR does not permit identification of of individuals. An advantage is that tags are available at individuals if more than one individual is tracked at the same low cost and with almost unlimited lifespan. The main time. A number of additional computer-based technologies limitation is the very small range at which the tags can are available, but they have been less widely applied to be detected (Table 3). Careful placements of readers can observe insect movements (for a review see Reynolds & permit investigations of spatial movements at very fine scales Riley, 2002). (e.g. when entering and leaving colonies). However, the high costs of readers make applications across a large spatial extent beyond the reach of most research budgets. New IV. CHALLENGES OF TRACKING INSECTS experimental designs with mobile readers can increase the WITH ACTIVE TRANSMITTERS spatial coverage and allow monitoring of large populations of individuals over long periods of time (Moreau et al., 2011). (1) Constraints of small size (a) Battery size (4) Alternatives to radio telemetry, harmonic radar and RFID One of the key challenges in insect telemetry is the need for a very small and light transmitter so that the animal is able Besides radio telemetry, harmonic radar, and RFID, a to carry it (Fig. 2). This currently precludes the use of GPS number of alternative techniques can be used to quantify technology (Cagnacci et al., 2010), satellite tracking (Bridge insect movements and space use. These include homing et al., 2011) or biotelemetry (Cooke et al., 2004). The weights experiments, theoretical models, genetic microsatellite of such transmitters are simply too heavy for insects. For the approaches, mark-reobservation studies, and Light Detection simplest transmitters, a lighter weight can best be achieved and Ranging (LIDAR). While the first two approaches tend by reducing the battery size, as this is the heaviest part of to overestimate the routine movement behaviour of insects an active radio transmitter. However, a smaller battery also under natural conditions, the latter three tend to measure generally means a shorter lifespan (Fig. 5). As a consequence,

(2) Behavioural effects of transmitter attachment 50 r = 0.76 (a) Observational and indirect evidence A key challenge in the radio telemetry of insects is to 40 understand whether and how the attachment of transmitters affects their behaviour and other aspects of their life history. 30 In most cases, authors only report anecdotally about how transmitters might or might not affect insect behaviour. For instance, radio-tagged carabid beetles (Carabus coriaceus)were 20 observed in the laboratory to experience problems when digging themselves into the ground (Riecken & Ries, 1992), Battery lifespan (days) 10 and scarab beetles (Osmoderma eremita) can be blocked by thin twigs or leaves when walking on the ground or climbing 0 trunks (Dubois & Vignon, 2008). Some studies use indirect evidence to support the assumption that transmitters have no 0.0 0.2 0.4 0.6 0.8 1.0 or only negligible effects on insect behaviour. For the scarab Transmitter mass (g) beetle Osmoderma eremita, movement rates and ranges as well Fig. 5. Relationship between transmitter mass and battery as dispersal kernels obtained from telemetry data were similar lifespan for active radio transmitters used in insect telemetry to those estimated from mark-recapture experiments (Hedin studies. The heavier a transmitter, the longer the battery lifespan & Ranius, 2002; Dubois & Vignon, 2008; Svensson et al., and the better the possibilities for data collection. Values were 2011), and hence it was suggested that dispersal behaviour is obtained from 17 insect telemetry case studies for which data not strongly affected by transmitter attachment. on tag mass and battery lifespan were available in the original publications (Table 2). The Pearson correlation (r) between log-transformed variables is provided. (b) Quantitative tests Potential behavioural effects of transmitter attachment insect telemetry studies have been conducted over short time have been tested quantitatively in only a few cases. For periods, mostly over 7–21 days (Fig. 5). Reducing the pulse instance, relationships between handling time and relative rate can increase battery life, but if the pulses are too far prey size for dobsonflies have been reported to be similar apart, it is difficult for a tracking person to determine the for individuals with and without transmitters (Hayashi & direction of the signal. Nakane, 1988, 1989). For a bush-cricket (Anabrus simplex), the effects of transmitter attachment on six movement (b) Limited tracking range behaviours were experimentally tested in a laboratory arena, including three turning-related behaviours (turn Another consequence of the requirement to use a small angle, turning rate, directional change relative to distance) transmitter is that the tracking range is reduced compared and three walking-related (walk time, speed, and distance) with that of larger transmitters that can send out a stronger behaviours (Lorch et al., 2005). No significant differences in signal. For instance, the maximum tracking range at ground these movement parameters were found between groups level over flat terrain that has been reported in insect of males or females either with or without transmitters. telemetry case studies is 500 m (Beaudoin-Ollivier et al., Bumblebee (Bombus terrestris) workers fitted with transmitters 2003; Kelly et al., 2008; Levett & Walls, 2011). Usually exhibited significantly lower flower visitation rates and spent maximum tracking ranges of 100–400 m are mentioned. significantly more time foraging on individual flower heads The tracking range depends not only on battery power but than individuals without transmitters (Hagen et al., 2011). also on other factors such as topography and vegetation Additionally, bumblebees with transmitters rested for long structure, and routine tracking distances in the field are time periods (> 45 min), suggesting that tag attachment therefore often much shorter than the maximum tracking increased their energy usage in the field (Hagen et al., 2011). distances reported by radio tag suppliers. One way to increase Overall, the attachment of transmitters can affect insect the area of tracking is to use helicopters and aeroplanes behaviour in the short or long term, but it is not easy to equipped with external antennae, as there are usually fewer generalize about these effects. More quantitative tests on obstacles in the line-of-sight between transmitter and antenna how transmitters affect various aspects of insect behaviour (depending on terrain and vegetation). Aerial tracking has been used in four insect telemetry studies (Wikelski and life history are therefore needed. et al., 2006, 2010; Pasquet et al., 2008; Hagen et al., 2011) and enabled researchers to obtain better information on (3) Transmitter weights and energetic costs maximum flight distances at the landscape scale. Moreover, (a) Ratio of tag mass to body mass it is currently the only possible option to track long-distance migration events of insects with radio telemetry (Wikelski The weight of a transmitter might not only affect behaviour, et al., 2006). but also the energy budget and metabolism of insects. For

Biological Reviews 89 (2014) 511–530 © 2013 The Authors. Biological Reviews © 2013 Cambridge Philosophical Society Insect telemetry 523 instance, the energy costs of flight, movement, and burrowing V. FUTURE PROSPECTS IN RADIO TELEMETRY might be increased if transmitters are attached to the body WITH INSECTS of insects. For terrestrial vertebrates such as birds and mammals, a ‘< 5% body size’ rule is often used, suggesting To date, there have been few published case studies on that the weight of the transmitter should not exceed 5% of a insect telemetry. However, given the increasing availability of species’ body weight (Brander & Cochran, 1969; Murray & miniaturized active radio transmitters and the famously high Fuller, 2000). This rule is arbitrary and has been questioned taxonomic richness of insects, there is considerable potential because it ignores aerodynamic relationships and fails to to develop this research field further. Below we highlight provide a direct estimate of energetic costs of transporting a future prospects related to field studies, experimental and transmitter (Caccamise & Hedin, 1985). Nevertheless, it has theoretical approaches, the development of automated been used as a rule of thumb for attaching transmitters tracking systems, and potential future options for tracking to vertebrates. For insects, much less is known about insects from space. the relationship between tag mass and insect body mass and its effect on insect metabolism. The smallest active radio transmitters (despite being very small) are still heavy (1) Field observational studies compared to the body mass of insects, and the ratio of tag (a) Expanding species coverage mass to body mass in insect telemetry studies has ranged from 2 to 100% (Table 4). In most cases, the weight of a The limited number of insect telemetry studies currently transmitter has been less than one third of the body weight precludes broad generalisations across species, functional of a species (Table 4), but for some species (e.g. bees), the groups, biomes and habitat types. Species coverage therefore weight of the transmitter has exceeded one third or even needs to be expanded to obtain more comprehensive half of the insect’s body weight (Table 4). Some species such quantitative data on movement behaviour and space use at as bumblebees are known to be able to carry such heavy the landscape and patch scales. This should cover aspects of loads (e.g. nectar and pollen loads up to 100% of their space use and movement such as home range sizes (Wikelski body weight; Heinrich, 1979), but this additional weight is et al., 2010; Hagen et al., 2011; Levett & Walls, 2011; Fornoff likely significantly to affect their normal energy expenditure et al., 2012), local movement paths (Negro et al., 2008; Hagen if carried beyond the time span of a single foraging bout. et al., 2011), habitat selection (Negro et al., 2008; Hagen Nevertheless, no study has yet tested how the additional et al., 2011; Watts & Thornburrow, 2011), and routine flight weight of transmitters affects energy costs of insects, both in distances (Pasquet et al., 2008; Wikelski et al., 2010; Hagen the short and the long term. et al., 2011) of insects from different taxonomic groups. Such studies might be particularly important for conservation (Negro et al., 2008), pest control (Beaudoin-Ollivier et al., (b) Potential energetic costs of transmitter attachment 2003), and landscape ecological management (Osborne et al., Some insect telemetry studies have made a first attempt 1999), and will provide the basis for understanding the towards understanding the potential energetic costs of survival of insects (e.g. pollinators) in human-dominated and carrying a transmitter. For instance, changes in body weight fragmented landscapes (Hagen et al., 2012). Furthermore, before and after radio-tagging have been used to quantify radio telemetry needs to be extended to a larger number net effects on metabolism. For wetas (Gibbs & McIntyre, of migratory insects, to gain an understanding of migration 1997) and carabid beetles (Riecken & Ries, 1992), an strategies in insects at regional or even continental and increase in body mass of radio-tagged insects from the global scales (Holland et al., 2006; Wikelski et al., 2006). The start to the end of the tracking period has been considered as use of helicopters and airplanes equipped with external evidence that transmitter weights do not strongly reduce food antennae has great potential for covering larger areas, uptake and foraging success (Riecken & Ries, 1992; Gibbs e.g. for tracking regional migration or landscape-level & McIntyre, 1997). By contrast, a radio-tagged individual movements. of the scarabid beetle Osmoderma eremita lost 13% of its body mass from the beginning of the tracking until its death (b) Future priorities for field studies (Dubois & Vignon, 2008). This could be attributed to the increased energy costs of carrying the additional transmitter From our review above, some specific recommendations weight, but other explanations are also possible. For both for future field studies emerge that could help to obtain wetas (Watts & Thornburrow, 2011) and dobsonfly larvae more balanced insights into insect ecology and behaviour. (Hayashi & Nakane, 1988, 1989), the measured movement First, a broader coverage of species from different insect distances of radio-tagged individuals were not related to the families and functional groups seems to be possible with ratio of the transmitter mass to body mass, suggesting no the smallest currently available radio tags. To date, most bias in distance estimates due to additional energetic costs. radio-tagged insect species have a body mass in excess of By contrast, long resting periods during flight movements 1 g, but smaller species with a body mass of 0.2–0.5 g (e.g. of radio-tagged bumblebee individuals could indicate that bumblebees and orchid bees) have now also been tracked transmitter weights may incur significant energy costs for under natural (field) conditions (Table 4). This opens up new these insects (Hagen et al., 2011). opportunities to track species with a wider body size range

Table 4. Body mass, tag mass, and their ratio (tag mass/body mass) for insect species which have been tracked with active radio transmitters in the ﬁeld. Values represent the mean or range (if available) across individuals and sexes as reported in the original publications (compare with Table 2). Note that not all case studies from Table 2 reported both body mass and tag mass

Species Sex-specific tags Body mass (g) Tag mass (g) Ratio (%) Reference Beetles (Coleoptera) Carabus coriaceus No 1.37–1.79 0.60–0.70 34–51 Riecken & Raths (1996) Carabus olympiae No 0.75 0.30 40 Negro et al. (2008) Lucanus cervus No 1.30–4.90 0.35 7–27 Rink & Sinsch (2007) Lucanus cervus No 1.90–3.30 0.40 12–21 Sprecher-Uebersax & Durrer (2001) Osmoderma eremita No 1.70–2.84 0.41 15–24 Dubois & Vignon (2008) Osmoderma eremita No 2.20–2.40 0.48–0.52 20–24 Hedin & Ranius (2002) and Hedin et al. (2008) Osmoderma eremita No 1.41–2.93 0.41 14–29 Svensson et al. (2011) Scapanes australis No 6.43–8.95 0.47–0.49 5–8 Beaudoin-Ollivier et al. (2003) Crickets (Orthoptera) Anabrus simplex No 2–3 0.85 28–43 Lorch & Gwynne (2000) Anabrus simplex No 2–3 0.45 15–23 Lorch et al. (2005) and Sword et al. (2008) Deinacrida heteracantha Yes 9; 35 0.87; 1.08 10; 3 Watts & Thornburrow (2011) Deinacrida rugosa No 10–20 0.40 2–4 Kelly et al. (2008) Deinacrida rugosa Yes 9.6; 19.2 0.87; 1.08 9; 6 Watts et al. (2012) Motuweta isolata No 8.6–34 1.70 5–20 Stringer & Chappell (2008) Philophyllia ingens No 3.50 0.30–0.35 9–10 Fornoff et al. (2012) Bees (Hymenoptera) Bombus hortorum No 0.30–0.45 0.20 44–66 Hagen et al. (2011) Bombus ruderatus No 0.27–0.40 0.20 50–74 Hagen et al. (2011) Bombus terrestris No 0.20–0.30 0.20 66–100 Hagen et al. (2011) Exaerete frontalis No 0.49–0.69 0.30 43–61 Wikelski et al. (2010) Xylocopa flavorufa No 1.01 0.35 35 Pasquet et al. (2008) Dobsonflies (Megaloptera) Protohermes grandis No 0.80–2.26 0.25–0.33 11–41 Hayashi & Nakane (1988) Protohermes grandis No 1.00–1.89 0.19 10–19 Hayashi & Nakane (1989) Dragonflies (Odonata) Anax junius No 1.20 0.3 25 Wikelski et al. (2010)

‘Sex-specific tags’ (yes/no) indicates whether a case study used different tag sizes for different sexes because of size dimorphism between males and females. Male values are then reported before the semicolon, and female values after. Species are ordered alphabetically within orders (beetles, crickets, bees, dobsonflies and dragonflies). than possible previously. Second, relatively few studies have observations for individuals with and without transmitters, to applied insect telemetry to tropical species (Fig. 3B, Table assess the potential negative effects of transmitter attachment 2), which is surprising given that most insect species occur under field conditions. at tropical latitudes (Godfray, Lewis & Memmott, 1999). We suggest that considerable progress in insect ecology and (2) Laboratory and field experiments and movement could be made if telemetry studies were not theoretical studies predominantly conducted in extra-tropical regions. Third, the use of radio telemetry has great potential to allow In addition to field observational studies, there is also great further insights into insect behaviour, but the intraspecific potential for conducting experimental and theoretical studies variability in ecological and behavioural characteristics, such to quantify how transmitters affect or might change insect as those between sexes, ages, castes, and individuals, has only energy budgets and behaviour. A key factor here is to rarely been investigated. For instance, various studies using understand the energy costs of carrying a transmitter, which different approaches have obtained data on landscape-level probably varies with insect body size (Chown & Gaston, movements and foraging ranges of social bee workers (e.g. 2010) and across taxonomic groups (e.g. beetles, bees, Osborne et al., 1999; Goulson, 2003; Greenleaf et al., 2007), dragonflies). For flying insects, the aerodynamics of flight but we have surprisingly little knowledge about the detailed need to be considered when assessing transmitter effects, space use of queens, males or solitary bees beyond the scale including flight behaviour and manoeuvrability (Caccamise of local foraging patches (Goulson, 2003; Hagen et al., 2011, & Hedin, 1985). 2012). Finally, few studies have quantified the effects of (a) Laboratory experiments transmitter attachment on insect behaviour in the field (e.g. Hagen et al., 2011). We recommend that future studies obtain To date, only a few studies have used detailed laboratory basic measures of condition and body mass before and after experiments to assess how transmitter attachment affects insect telemetry, and conduct quantitative behavioural field movement behaviour (e.g. Lorch et al., 2005). Laboratory or

flight cage experiments could help to test how transmitters individuals. Automated data logging has been used since affect insect energy expenditure. For instance, increased the start of wildlife radio telemetry (Cochran et al., 1965; flower handling times in bumblebees fitted with transmitters Voegeli et al., 2001) and has also been used with insects, e.g. have been demonstrated in the field (Hagen et al., 2011), to measure daily activity patterns of bush-crickets (Fornoff but calibrated feeder experiments in a flight cage could et al., 2012) or dragonflies (Levett & Walls, 2011). By contrast, additionally assess whether this change in behaviour reduces automated tracking systems (both radio and acoustic) have overall nectar consumption or whether it demonstrates so far only been applied to track large-bodied animals, behavioural plasticity to maintain overall nectar intake including sharks (Heupel & Hueter, 2001; Klimley et al., (Pflumm, 1977). Experiments in flight cages (e.g. with 2001), ungulates (Johnson et al., 2000; Ager et al., 2003), increasing transmitter weights) could also be used to identify primates (Crofoot et al., 2010), bats (Dechmann et al., 2009; maximum weights of tags that have no or minimal impact on Holland et al., 2011), and other vertebrates (Kays et al., 2011). insect behaviour (Boiteau & Colpitts, 2001) and to assess the We include acoustic tracking systems in aquatic animals physical constraints that the tags place on animals’ movement because the principles are similar and informative for the and flight. Alternatively, experimental laboratory and field design of future automated radio telemetry systems. studies can be conducted by adding tiny extra weights to the body of insects to test the effect of increasing load weight (b) Presence/absence design on foraging behaviour and lifespans (Schmid-Hempel, 1986; Wolf & Schmid-Hempel, 1989). Experiments with radio The simplest automated tracking system to obtain spatially transmitters could also be combined with other techniques explicit data involves a series of receivers that are spread that allow studying insect movements in a laboratory setting across a study site to record the presence or absence of a (for an overview see section 8 in Reynolds & Riley, 2002). signal within range (Fig. 6A). For instance, the space use and movement of a marine cephalopod in and around a harbour was recorded using 64 automated acoustic receivers (b) Theoretical models installed on a network of moored buoys (Stark, Jackson Theoretical models have been used to analyse economics & Lyle, 2005). Such an approach could, in principle, be and distances of foraging and flight range (Viswanathan applied to terrestrial insects (using radio rather than acoustic et al., 1999; Cresswell et al., 2000). Such models can be tested signals), specifically for understanding the use of specific and compared with experimental and observational data habitats or the movements to and from clearly defined and allow theoretical predictions about how energy intake locations such as nest sites or foraging areas. The method and foraging strategies affect movement. A combination could be refined by carefully considering how overlaps in of short-term experimental studies with theoretical models detection could indicate location, and by potentially using of energy budgets and metabolism in relation to foraging signal strength as a measure of distance (given flat terrain and (Viswanathan et al., 1999; Cresswell et al., 2000) and flight no major obstacles that would affect detectability, Fig. 6A). behaviour (Caccamise & Hedin, 1985; Fischer & Ebert, 1999; However, this method is only useful for research questions Fischer & Kutsch, 2000; Ando, Shimoyama & Kanzaki, that are focused on movements of the order of hundreds of 2002) could potentially permit predictions about the short- metres, because this is the detectable distance of the radio and long-term sub-lethal effects of transmitters on insect transmitters. Consequently, smaller scale movements cannot individuals and populations. be distinguished using a system that detects the presence or absence of a transmitter within range. For studies where accuracy in the order of metres is required, passive RFID (3) Automated tracking systems tag technologies are more appropriate (see above). The radio telemetry studies discussed herein have relied on manual tracking with hand-held or aircraft-mounted (c) Triangulation design antennae (Fig. 1). This limits sample sizes and temporal resolution while requiring considerable labour. Automated To obtain more precise spatial locations of moving recording and triangulation of radio signals from small radio individuals, triangulation with three or more antenna and transmitters could overcome these limitations. Moreover, receiver units is required (Fig. 6B). Such automated tracking by removing the need for subjective assessments of signal systems that simultaneously record signal bearings from volume, an automated system would allow the pulse rate at least three receiver stations and then triangulate the of transmitters to be reduced significantly, thus lengthening location of the transmitter have been used successfully transmitter life. to track larger animals (e.g. Kays et al., 2011; Fig. 6B). They allow multiple individuals to be tracked simultaneously and provide data at high spatial and temporal resolution, (a) Automated data logging versus automated tracking 24 h a day, for the lifetime of the transmitter (Kays We distinguish ‘automated data logging’ from ‘automated et al., 2011). A new receiver (SRX 600, Lotek Wireless, tracking systems’ because the former only detects the www.Lotek.com) has now become commercially available presence and absence of movements and activity patterns that could be used to compute bearings of signals from while the latter determines the location and movement of coded transmitters, as it automatically assesses relative signal

(A) (B)

Field A 12 WWW

Fallow margin

3 4 Field B 900 MHz

Fig. 6. Examples of designs for automated tracking systems. (A) Presence/absence design where an automated tracking of movement and space use could be approximated using the presence/absence of a radio signal from a transmitter within the range (red circle outlines) of a receiver with a simple dipole antenna (red dot with black outline, numbered 1–4). In this hypothetical example, the careful placement of receiver and data-logger units could allow broad discrimination between animal use of Field A (detected by 1 or 2), Field B (detected by 3 or 4), or fallow margin (detected by either 1 and 3 or 2 and 4). The limitations of this system design are that the radius around each receiver is determined by the detectible distance of the radio transmitter (likely to be in the order of hundreds of metres). Detectability at the limit of the signal range can be variable, and it requires flat ground with no obstacles to radio transmission. (B) Triangulation design which allows triangulating the position of an object to achieve high spatial precision and to track movement within habitats. The example illustrates the automated radio telemetry system (ARTS) established on Barro Colorado Island in Panama (Kays et al., 2011). Each antenna tower and receiver computes the signal bearing based on relative signal strength recorded at six directional yagi antennae set at fixed angles. These data are fed via a wireless network back to the base station and then made accessible online to researchers. Once the raw signal-bearing data are cleaned, locations of individual animals can be computed using the principles of triangulation. This system permits simultaneous tracking of 20 or more individuals 24 h per day for the lifetime of the transmitters. This type of system could be developed for the smaller signal power and range of < 1 g radio transmitters to answer questions about habitat use and movement patterns of insects at fine spatial scales. Transportable antenna towers could permit the system to be used at multiple sites, but site-specific signal distortions would need to be mapped at each site. (Image in B is reproduced from Kays et al. (2011), used with permission from R. Kays and Oxford University Press.). strength from multiple inputs. Coded transmitters allow www.microwavetelemetry.com). As an alternative to satellite multiple identifiable transmitters to use the same frequency tracking systems such as ARGOS, the ICARUS initiative and the receiver is thus able simultaneously to record (International Cooperation of Animal Research Using Space; signal strength for many coded transmitters using the same http://icarusinitiative.org) has been started with the aim frequency. This improves the temporal resolution of fixes of establishing a new space-based system that allows the beyond that obtained by previous systems (Kays et al., 2011). global tracking of small-bodied vertebrates and insects. The drawback of coded transmitters is that the extra circuitry The principal idea is to establish sensitive receivers with needed adds additional transmitter weight. However, very sophisticated antennae in near-earth orbit that are able to small coded transmitters have now also become available detect signals from small radio tags attached to animals (0.25 g, www.Lotek.com). We are not aware of any published (Pennisi, 2011). The shorter distance from the Earth to the study that has used such automated systems to track insects. space station (∼320 km) compared with that to the ARGOS satellites (∼850 km) would demand less energy to send data, (4) Tracking from space potentially reducing battery requirements. (a) Space-based tracking systems (b) Prospects and limitations of tracking insects from space The emergence of satellite tracking systems such as ARGOS (www.argos-system.org) in the 1980s has been a The current challenge for tracking insects with space-based breakthrough for the monitoring and analysis of global tracking systems such as ARGOS and ICARUS is to develop movements of large-bodied vertebrates. The transmitters transmitters that are small (but still powerful) enough to be needed for such systems are relatively large and their detected from space. Transmitters with a mass below 5 g application has long been limited to species that weigh are envisioned (Wikelski et al., 2007) or currently developed more than approximately 300 g (Wikelski et al., 2007). (Microwave Telemetry, www.microwavetelemetry.com) and However, new developments have further reduced the mass a further reduction to below 1 g might be expected in the of solar-powered transmitters operating via the ARGOS near future if the current rate of technological advancement satellite systems to 5 g and this allows e.g. birds down continues. Such developments could potentially create to almost 100 g to be tracked (Microwave Telemetry, a breakthrough for the global tracking of small-bodied

Biological Reviews 89 (2014) 511–530 © 2013 The Authors. Biological Reviews © 2013 Cambridge Philosophical Society Insect telemetry 527 vertebrates and, if technology develops further, might also be analysis of regional and continental movements of large- applied to insects. Nevertheless, even transmitters in the 1 g bodied insects. range will only be applicable to the largest insects, and global space-based tracking of insects will probably only be possible for specific species, e.g. those that live in open habitats such VII. ACKNOWLEDGEMENTS as deserts (Wikelski et al., 2007). The spatial resolution (i.e. accuracy) of such systems is likely to be in the range of a few kilometres, which would be suitable for detecting large- We thank two anonymous reviewers for constructive and scale movements of insects such as migrations. For more stimulating comments. This review was initiated by a local and landscape-level applications, field-based manual workshop on ‘Frontiers in Telemetry’ (organized by Toke T. tracking, aerial surveys, and automated tracking systems (as Høye, M.H. and W.D.K., and held at the Department of discussed above) will remain the key for providing data at Bioscience, Aarhus University, Denmark on 3 November the appropriate resolution. 2011) and benefited from an invitation to give a talk at the Biodiversity Institute Symposium on ‘Biodiversity Technologies’ (organized by Kathy Willis and Gillian Petrokofsky, held at the Department of Zoology, University VI. CONCLUSIONS of Oxford, UK, on 27–28 September 2012). We thank Martin Wikelski for continuous inspiration and support, (1) Radio telemetry is applicable for a wide range Tony Corbett for creating Fig. 1, and Roland Kays, of large-bodied insects and allows researchers to gain Barbara Corbett Kermeen, Sarah Levett, Ulrike Raths, Don insights into insect ecology, behaviour, and evolution, Reynolds, Nigel Raine, Juliet Osborne, and Seirian Sumner especially when addressing questions about space use, habitat for providing information. We acknowledge support from selection, mobility, movement distances, activity patterns and the Danish Council for Independent Research — Natural migration strategies at local, landscape and even regional Sciences (M.H., and via a starting independent researcher scales. Such data are fundamental for basic and applied grant # 11-106163 to W.D.K.). biodiversity science, including species conservation, pest control, landscape ecology, and agricultural and horticultural management. VIII. REFERENCES (2) The limited number of insect telemetry studies currently precludes broader generalisations of movement behaviour Ager,A.A.,Johnson,B.K.,Kern,J.W.&Kie, J. G. (2003). Daily and seasonal and space use across species, functional groups, biomes movements and habitat use by female Rocky Mountain elk and mule deer. Journal and habitats. We recommend (i) extending coverage to of Mammalogy 84, 1076–1088. a wider range of species from different insect families and Ando,N.,Shimoyama,I.&Kanzaki, R. (2002). A dual-channel FM transmitter for acquisition of flight muscle activities from the freely flying hawkmoth, Agrius functional groups, (ii) putting more emphasis on (sub-)tropical convolvuli. Journal of Neuroscience Methods 115, 181–187. regions, and (iii) quantifying intraspecific variability between Beaudoin-Ollivier,L.,Bonaccorso, F., Aloysius,M.&Kasiki, M. (2003). sexes, ages, castes, and individuals. Aerial surveys from Flight movement of Scapanes australis australis (Boisduval) (Coleoptera: Scarabaeidae: Dynastinae) in Papua New Guineaa radiotelemetry study. Australian Journal of helicopters and aeroplanes equipped with external antennae Entomology 42, 367–372. can help to increase the limited tracking range of active Boiteau,G.&Colpitts, B. (2001). Electronic tags for the tracking of insects in flight: radio transmitters, especially when quantifying maximum effect of weight on flight performance of adult Colorado potato beetles. Entomologia Experimentalis et Applicata 100, 187–193. flight distances and home range sizes at the landscape scale, Boiteau,G.&Colpitts, B. (2004). The potential of portable harmonic radar or when tracking long-distance migration events at regional technology for the tracking of beneficial insects. International Journal of Pest Management and continental scales. 50, 233–242. Bowlin,M.S.,Bisson,I.A.,Shamoun-Baranes,J.,Reichard,J.D.,Sapir,N., (3) Quantifying the effects of transmitter attachment on Marra,P.P.,Kunz,T.H.,Wilcove,D.S.,Hedenstrom,A.,Guglielmo,C. insect behaviour and energy costs is crucial, but has been G., Akesson,S.,Ramenofsky,M.&Wikelski, M. (2010). Grand challenges in widely neglected to date. We suggest that more quantitative migration biology. Integrative and Comparative Biology 50, 261–279. Brander,R.B.&Cochran, W. W. (1969). Radio-location telemetry. In Wildlife tests are needed to assess the consequences of transmitter Management Techniques (ed. R. H. Giles), pp. 95–103. The Wildlife Society, attachment, including basic measurements of condition Washington. Bridge,E.S.,Thorup,K.,Bowlin,M.S.,Chilson,P.B.,Diehl,R.H.,Fleéron´ , and body mass before and after telemetry, behavioural R. W., Hartl,P.,Roland,K.,Kelly, J. F., Robinson,W.D.&Wikelski,M. field studies for individuals with and without transmitters, (2011). Technology on the move: recent and forthcoming innovations for tracking laboratory or flight cage experiments, or theoretical studies migratory birds. Bioscience 61, 689–698. Caccamise,D.F.&Hedin, R. S. (1985). An aerodynamic basis for selecting on foraging and flight behaviour in relation to energy budgets transmitter loads in birds. Wilson Bulletin 97, 306–318. and metabolism. Cagnacci, F., Boitani,L.,Powell,R.A.&Boyce, M. S. (2010). Animal (4) Finally, we see great potential for moving the frontiers ecology meets GPS-based radiotelemetrya perfect storm of opportunities and challenges. Philosophical Transactions of the Royal Society B: Biological Sciences 365, of insect telemetry considerably further if automated systems 2157–2162. for tracking insect movement, activity and behaviour are Cant,E.T.,Smith,A.D.,Reynolds,D.R.&Osborne, J. L. (2005). Tracking implemented and applied. The further development of space- butterfly flight paths across the landscape with harmonic radar. Proceedings of the Royal Society B: Biological Sciences 272, 785–790. based tracking systems such as ARGOS or ICARUS could Capaldi,E.A.&Dyer, F. C. (1999). The role of orientation flights on homing ultimately provide a breakthrough for the monitoring and performance in honeybees. Journal of Experimental Biology 202, 1655–1666.

Chapman,J.W.,Drake,V.A.&Reynolds, D. R. (2011). Recent insights from Greenleaf,S.S.,Williams,N.M.,Winfree,R.&Kremen, C. (2007). Bee radar studies of insect flight. Annual Review of Entomology 56, 337–356. foraging ranges and their relationship to body size. Oecologia 153, 589–596. Chapman,J.W.,Nesbit,R.L.,Burgin,L.E.,Reynolds,D.R.,Smith,A.D., Guedot´ ,C.,Bosch,J.&Kemp, W. P. (2009). Relationship between body size and Middleton,D.R.&Hill, J. K. (2010). Flight orientation behaviors promote homing ability in the genus Osmia (Hymenoptera; Megachilidae). Ecological Entomology optimal migration trajectories in high-flying insects. Science 327, 682–685. 34, 158–161. Chapman,J.W.,Reynolds,D.R.&Smith, A. D. (2003a). Vertical-looking radara Hagen,M.,Kissling,W.D.,Rasmussen,C.,De Aguiar,M.A.M.,Brown,L. new tool for monitoring high-altitude insect migration. BioScience 53, 503–511. E., Carstensen,D.W.,Alves-Dos-Santos,I.,Dupont,Y.L.,Edwards,F.K., Chapman,R.E.,Wang,J.&Bourke,A.F.G.(2003b). Genetic analysis of spatial Genini,J.,Guimaraes˜ ,P.R.Jr.,Jenkins,G.B.,Jordano,P.,Kaiser-Bunbury, foraging patterns and resource sharing in bumble bee pollinators. Molecular Ecology C. N., Ledger,M.E.,Maia,K.P.,Marquitti,F.M.D.,McLaughlin, O.,´ 12, 2801–2808. Morellato,L.P.C.,O’Gorman,E.J.,Trøjelsgaard,K.,Tylianakis,J. Chapman,J.,Reynolds,D.&Smith, A. (2004). Migratory and foraging movements M., Vidal,M.M.,Woodward,G.&Olesen, J. M. (2012). Biodiversity, species in beneficial insects: a review of radar monitoring and tracking methods. International interactions and ecological networks in a fragmented world. Advances in Ecological Journal of Pest Management 50, 225–232. Research 46, 89–210. Chilson,P.B.,Bridge,E.,Frick, W. F., Chapman,J.W.&Kelly, J. F. (2012). Hagen,M.,Wikelski,M.&Kissling, W. D. (2011). Space use of bumblebees Radar aeroecology: exploring the movements of aerial fauna through radio-wave (Bombus spp.) revealed by radio-tracking. PLoS One 6, e19997. remote sensing. Biology Letters 8, 698–701. Hagler,J.R.&Jackson, C. G. (2001). Methods for marking insects: current Chown,S.L.&Gaston, K. J. (2010). Body size variation in insectsa macroecological techniques and future prospects. Annual Review of Entomology 46, 511–543. perspective. Biological Reviews 85, 139–169. Hayashi,F.&Nakane, M. (1988). A radio-tracking study on the foraging movements Cochran,W.W.&Lord, R. D. Jr. (1963). A radio-tracking system for wild animals. of the dobsonfly larva, Protohermes grandis (Megaloptera, Corydalidae). Japanese Journal The Journal of Wildlife Management 27, 9–24. of Entomology 56, 417–429. Cochran,W.,Warner,D.,Tester,J.&Kuechle, V. (1965). Automatic radio- Hayashi,F.&Nakane, M. (1989). Radio tracking and activity monitoring of tracking system for monitoring animal movements. Bioscience 15, 98–100. the dobsonfly larva, Protohermes grandis (Megaloptera: Corydalidae). Oecologia 78, Collett,T.S.,Graham,P.&Harris, R. A. (2007). Novel landmark-guided routes 468–472. in ants. Journal of Experimental Biology 210, 2025–2032. Hedin,J.&Ranius, T. (2002). Using radio telemetry to study dispersal of the beetle Cooke,S.J.,Hinch,S.G.,Wikelski,M.,Andrews,R.D.,Kuchel,L.J., Osmoderma eremita, an inhabitant of tree hollows. Computers and Electronics in Agriculture Wolcott,T.G.&Butler, P. J. (2004). Biotelemetrya mechanistic approach to 35, 171–180. ecology. Trends in Ecology & Evolution 19, 334–343. Hedin,J.,Ranius,T.,Nilsson,S.G.&Smith, H. G. (2008). Restricted Craighead,E.C.Jr.&Craighead, J. J. (1963). Radiotracking of Grizzly Bears: dispersal in a flying beetle assessed by telemetry. Biodiversity and Conservation 17, Grizzly Bear Ecological Findings Obtained by Biotelemetry. Montana Cooperative Wildlife 675–684. Research Unit, University of Montana, Missoula. Heinrich, B. (1979). Bumblebee Economics. Harvard University Press, Cambridge. Cresswell,J.E.,Osborne,J.L.&Goulson, D. (2000). An economic model of Henry,M.,Beguin´ ,M.,Requier, F., Rollin, O., Odoux, J.-F., Aupinel,P., the limits to foraging range in central place foragers with numerical solutions for Aptel,J.,Tchamitchian,S.&Decourtye, A. (2012). A common pesticide bumblebees. Ecological Entomology 25, 249–255. decreases foraging success and survival in honey bees. Science 336, 348–350. Crofoot,M.C.,Lambert,T.D.,Kays,R.&Wikelski, M. C. (2010). Does Heupel,M.R.&Hueter, R. E. (2001). Use of an automated acoustic telemetry watching a monkey change its behaviour? Quantifying observer effects in habituated system to passively track juvenile blacktip shark movements. In Electronic Tagging wild primates using automated radiotelemetry. Animal Behaviour 80, 475–480. and Tracking in Marine Fisheries (eds J. R. Sibert and J. L. Nielsen), pp. 217–236. Darvill,B.,Knight,M.E.&Goulson, D. (2004). Use of genetic markers to Kluwer Academic Publishers, Dordrecht. quantify bumblebee foraging range and nest density. Oikos 107, 471–478. Holland,R.A.,Meyer,C.F.J.,Kalko,E.K.V.,Kays,R.&Wikelski,M. Dechmann,D.K.N.,Heucke,S.L.,Giuggioli,L.,Safi,K.,Voigt,C.C.& (2011). Emergence time and foraging activity in Pallas’ mastiff bat, Molossus molossus Wikelski, M. (2009). Experimental evidence for group hunting via eavesdropping (Chiroptera: Molossidae) in relation to sunset/sunrise and phase of the moon. Acta in echolocating bats. Proceedings of the Royal Society B: Biological Sciences 276, 2721–2728. Chiropterologica 13, 399–404. Decourtye,A.,Devillers,J.,Aupinel,P.,Brun, F., Bagnis,C.,Fourrier,J.& Holland,R.A.,Wikelski,M.&Wilcove, D. S. (2006). How and why do insects Gauthier, M. (2011). Honeybee tracking with microchips a new methodology to migrate? Science 313, 794–796. measure the effects of pesticides. Ecotoxicology 20, 429–437. Janowski-Bell,M.E.&Horner, N. V. (1999). Movement of the male brown Domdouzis,K.,Kumar,B.&Anumba, C. (2007). Radio-Frequency Identification tarantula, Aphonopelma hentzi (Araneae, Theraphosidae), using radio telemetry. Journal (RFID) applications a brief introduction. Advanced Engineering Informatics 21, 350–355. of Arachnology 27, 503–512. Drake,V.A.&Reynolds, D. R. (2012). Radar Entomology: Observing Insect Flight and Johnson,B.K.,Kern,J.W.,Wisdom,M.J.,Findholt,S.L.&Kie, J. G. (2000). Migration. CABI, Wallingford. Resource selection and spatial separation of mule deer and elk during spring. Journal Dramstad, W. E. (1996). Do bumblebees (Hymenoptera: Apidae) really forage close of Wildlife Management 64, 685–697. to their nests? Journal of Insect Behavior 9, 163–182. Kareiva,P.M.&Shigesada, N. (1983). Analyzing insect movement as a correlated Dubois,G.&Vignon, V. (2008). First results of radio-tracking of Osmoderma eremita random walk. Oecologia 56, 234–238. (Coleoptera: Cetoniidae) in French chestnut orchards. Revue D Ecologie-La Terre Et La Kays,R.,Tilak,S.,Crofoot,M.,Fountain,T.,Obando,D.,Ortega, Vie 63, 131–138. A., Kuemmeth, F., Mandel,J.,Swenson,G.,Lambert,T.,Hirsch,B.& Essler,W.O.&Folk, G. E. (1961). Determination of physiological rhythms of Wikelski, M. (2011). Tracking animal location and activity with an automated unrestrained animals by radio telemetry. Nature 190, 90–91. radio telemetry system in a tropical rainforest. The Computer Journal 54, 1931–1948. Fischer,H.&Ebert, E. (1999). Tegula function during free locust flight in relation Kelly,C.D.,Bussiere,L.F.&Gwynne, D. T. (2008). Sexual selection for male to motor pattern, flight speed and aerodynamic output. Journal of Experimental Biology mobility in a giant insect with female-biased size dimorphism. American Naturalist 202,711–721. 172, 417–423. Fischer,H.&Kutsch, W. (2000). Relationships between body mass, motor output Kenward, R. E. (2001). Wildlife Radio Tagging. Academic Press, San Diego. and flight variables during free flight of juvenile and mature adult locusts, Schistocerca Klimley,A.P.,Le Boeuf,B.J.,Cantara,K.M.,Richert,J.E.,Davis,S.F. gregaria. Journal of Experimental Biology 203, 2723–2735. & Van Sommeran, S. (2001). Radio-acoustic positioning as a tool for studying Fornoff, F., Dechmann,D.K.N.&Wikelski, M. (2012). Observation of movement site-specific behavior of the white shark and other large marine species. Marine and activity via radio-telemetry reveals diurnal behaviour of the neotropical katydid Biology 138, 429–446. Philophyllia ingens (Orthoptera: Tettigoniidae). Ecotropica 18,27–34. Kraus,F.B.,Wolf,S.&Moritz, R. F. A. (2009). Male flight distance and population Gherardi,F.&Vannini, M. (1989). Spatial behavior of the freshwater crab, Potamon substructure in the bumblebee Bombus terrestris. Journal of Animal Ecology 78, 247–252. fluviatile—a radio-telemetric study. Biology of Behaviour 14,28–45. Levett,S.&Walls, S. (2011). Tracking the elusive life of the Emperor Dragonfly Gibbs,G.W.&McIntyre, M. E. (1997). Abundance and Future Options for Wetapunga on Anax imperator Leach. Journal of the British Dragonfly Society 27, 59–68. Little Barrier Island. Department of Conservation, Wellington. Lihoreau,M.,Raine,N.E.,Reynolds,A.M.,Stelzer,R.J.,Lim,K.S.,Smith, Gill,R.J.,Ramos-Rodriguez,O.&Raine, N. E. (2012). Combined pesticide A. D., Osborne,J.L.&Chittka, L. (2012). Radar tracking and motion-sensitive exposure severely affects individual- and colony-level traits in bees. Nature 491, cameras on flowers reveal the development of pollinator multi-destination routes 105–108. over large spatial scales. PLoS Biology 10, e1001392. Godfray,H.C.J.,Lewis,O.T.&Memmott, J. (1999). Studying insect diversity Lorch,P.D.&Gwynne, D. T. (2000). Radio-telemetric evidence of migration in in the tropics. Philosophical Transactions of the Royal Society of London. Series B, Biological the gregarious but not the solitary morph of the Mormon cricket (Anabrus simplex: Sciences 354, 1811–1824. Orthoptera: Tettigoniidae). Naturwissenschaften 87, 370–372. Goulson, D. (2003). Bumblebees: Behaviour, Ecology and Conservation. Second Edition. Lorch,P.D.,Sword,G.A.,Gwynne,D.T.&Anderson, G. L. (2005). Oxford University Press, Oxford. Radiotelemetry reveals differences in individual movement patterns between Goulson,D.&Stout, J. C. (2001). Homing ability of the bumblebee Bombus terrestris outbreak and non-outbreak Mormon cricket populations. Ecological Entomology 30, (Hymenoptera: Apidae). Apidologie 32, 105–111. 548–555.

Lovei¨ ,G.L.,Stringer,I.A.N.,Devine,C.D.&Cartellieri, M. (1997). Riley,J.R.&Smith, A. D. (2002). Design considerations for an harmonic radar to Harmonic radar—a method using inexpensive tags to study invertebrate movement investigate the flight of insects at low altitude. Computers and Electronics in Agriculture on land. New Zealand Journal of Ecology 21, 187–193. 35, 151–169. Mackay, R. S. (1964). Galapagos tortoise and marine iguana deep body temperatures Riley,J.R.,Smith,A.D.,Reynolds,D.R.,Edwards,A.S.,Osborne,J.L., measured by radio telemetry. Nature 204, 355–358. Williams,I.H.,Carreck,N.L.&Poppy, G. M. (1996). Tracking bees with Manly,B.F.G.,McDonald,L.L.,Thomas,D.L.,McDonald,T.L.&Erickson, harmonic radar. Nature 379, 29–30. W. P. (2010). Resource Selection by Animals: Statistical Design and Analysis of Field Studies. Riley,J.R.,Valeur,P.,Smith,A.D.,Reynolds,D.R.,Poppy,G.M.& Second Edition. Kluwer Academic Publishers, Dordrecht. Lofstedt¨ , C. (1998). Harmonic radar as a means of tracking the pheromone- Mascanzoni,D.&Wallin, H. (1986). The harmonic radar a new method of tracing finding and pheromone-following flight of male moths. Journal of Insect Behavior 11, insects in the field. Ecological Entomology 11, 387–390. 287–296. Menzel,R.,Greggers,U.,Smith,A.,Berger,S.,Brandt,R.,Brunke,S., Rink,M.&Sinsch, U. (2007). Radio-telemetric monitoring of dispersing stag beetles: Bundrock,G.,Hulse¨ ,S.,Plumpe¨ ,T.,Schaupp, F., Schuttler¨ ,E.,Stach, implications for conservation. Journal of Zoology 272, 235–243. S., Stindt,J.,Stollhoff,N.&Watzl, S. (2005). Honey bees navigate according Robinson,E.H.,Richardson,T.,Sendova-Franks,A.,Feinerman,O.& toamap-likespatialmemory.Proceedings of the National Academy of Sciences of the United Franks, N. (2009a). Radio tagging reveals the roles of corpulence, experience States of America 102, 3040–3045. and social information in ant decision making. Behavioral Ecology and Sociobiology 63, Menzel,R.,Kirbach,A.,Haass,W.-D.,Fischer,B.,Fuchs,J.,Koblofsky,M., 627–636. Lehmann,K.,Reiter,L.,Meyer,H.,Nguyen,H.,Jones,S.,Norton,P.& Robinson,E.J.H.,Smith,F.D.,Sullivan,K.M.E.&Franks, N. R. (2009b). Greggers, U. (2011). A common frame of reference for learned and communicated Do ants make direct comparisons? Proceedings of the Royal Society B: Biological Sciences vectors in honeybee navigation. Current Biology 21, 645–650. 276, 2635–2641. Millspaugh,J.J.&Marzluff,J.M.(2001).Radio Tracking and Animal Populations. Robinson,C.A.,Thom,T.J.&Lucas, M. C. (2000). Ranging behaviour of a large Academic Press, San Diego. freshwater invertebrate, the white-clawed crayfish Austropotamobius pallipes. Freshwater Molet,M.,Chittka,L.,Stelzer,R.J.,Streit,S.&Raine, N. E. (2008). Colony Biology 44, 509–521. nutritional status modulates worker responses to foraging recruitment pheromone Schaefer, G. (1969). Radar studies of locust, moth and butterfly migration in the in the bumblebee Bombus terrestris. Behavioral Ecology and Sociobiology 62, 1919–1926. Sahara. Proceedings of the Royal Entomological Society of London 34, 39–40. Moreau,M.,Arrufat,P.,Latil,G.&Jeanson, R. (2011). Use of radio-tagging to Schick,R.S.,Loarie,S.R.,Colchero, F., Best,B.D.,Boustany,A.,Conde, map spatial organization and social interactions in insects. The Journal of Experimental D. A., Halpin,P.N.,Joppa,L.N.,McClellan,C.M.&Clark, J. S. (2008). Biology 214, 17–21. Understanding movement data and movement processes: current and emerging Murray,D.L.&Fuller, M. R. (2000). A critical review of the effects of marking on directions. Ecology Letters 11, 1338–1350. the biology of vertebrates. In Research Techniques in Animal Ecology (ed. F. T. Boitani), Schmid-Hempel, P. (1986). Do honeybees get tired? The effect of load weight on pp. 15–64. Columbia University Press, New York. patch departure. Animal Behaviour 34, 1243–1250. ,C.W., ,J., ,B.& , S. (2012). RFID tracking Negro,M.,Casale,A.,Migliore,L.,Palestrini,C.&Rolando, A. (2008). Schneider Tautz Grunewald¨ Fuchs Apis Habitat use and movement patterns in the endangered ground beetle species, Carabus of sublethal effects of two neonicotinoid insecticides on the foraging behavior of mellifera. PLoS One 7, e30023. olympiae (Coleoptera: Carabidae). European Journal of Entomology 105, 105–112. Shaw,J.A.,Seldomridge,N.L.,Dunkle,D.L.,Nugent,P.W.,Spangler,L. O’Neal,M.E.,Landis,D.A.,Rothwell,E.,Kempel,L.&Reinhard,D. H., Bromenshenk,J.J.,Henderson,C.B.,Churnside,J.H.&Wilson,J. (2004). Tracking insects with harmonic radar a case study. American Entomologist 50, J. (2005). Polarization lidar measurements of honey bees in flight for locating land 212–218. mines. Optics Express 13, 5853–5863. Osborne,J.L.,Clark,S.J.,Morris,R.J.,Williams,I.H.,Riley,J.R.,Smith, Silcox,D.E.,Doskocil,J.P.,Sorenson,C.E.&Brandenburg, R. L. (2011). A. D., Reynolds,D.R.&Edwards, A. S. (1999). A landscape-scale study of Radio frequency identification tagging a novel approach to monitoring surface and bumble bee foraging range and constancy, using harmonic radar. Journal of Applied subterranean insects. American Entomologist 57, 86–93. Ecology 36, 519–533. Sprecher-Uebersax,E.&Durrer, H. (2001). Verhaltensstudienuber ¨ den Ovaskainen, O., Smith,A.D.,Osborne,J.L.,Reynolds,D.R.,Carreck,N.L., hirschkafer¨ Lucanus cervus L. mit hilfe der telemetrie und videobeobachtung. Martin,A.P.,Niitepold˜ ,K.&Hanski, I. (2008). Tracking butterfly movements Mitteilungen der Naturforschenden Gesellschaften beider Basel 5, 161–182. with harmonic radar reveals an effect of population age on movement distance. Srygley,R.B.,Lorch,P.D.,Simpson,S.J.&Sword,G.A.(2009).Immediate Proceedings of the National Academy of Sciences 105, 19090–19095. protein dietary effects on movement and the generalised immunocompetence of ,R.M.S., ,A., ,M.B., ,E., ,J., , Pasquet Peltier Hufford Oudin Saulnier Paul migrating Mormon crickets Anabrus simplex (Orthoptera: Tettigoniidae). Ecological L., Knudsen,J.T.,Herren,H.R.&Gepts, P. (2008). Long-distance pollen Entomology 34, 663–668. flow assessment through evaluation of pollinator foraging range suggests transgene Stark,K.E.,Jackson,G.D.&Lyle, J. M. (2005). Tracking arrow squid movements escape distances. Proceedings of the National Academy of Sciences 105, 13456–13461. with an automated acoustic telemetry system. Marine Ecology Progress Series 299, Pennisi, E. (2011). Global tracking of small animals gains momentum. Science 334, 167–177. 1042. Streit,S.,Bock, F., Pirk,C.W.W.&Tautz, J. (2003). Automatic life-long Pflumm, W. (1977). Factors determining volume of sugar solution bees and wasps monitoring of individual insect behaviour now possible. Zoology 106, 169–171. collect per trip at food source. Apidologie 8, 401–411. Stringer,I.A.N.&Chappell, R. (2008). Possible rescue from extinction: transfer Psychoudakis,D.,Moulder,W.,Chi-Chih,C.,Heping,Z.&Volakis,J.L. of a rare New Zealand tusked weta to islands in the Mercury group. Journal of Insect (2008). A portable low-power harmonic radar system and conformal tag for insect Conservation 12, 371–382. tracking. IEEE Antennas and Wireless Propagation Letters 7, 444–447. Sumner,S.,Lucas,E.,Barker,J.&Isaac, N. (2007). Radio-tagging technology Reynolds,D.R.&Riley, J. R. (2002). Remote-sensing, telemetric and computer- reveals extreme nest-drifting behavior in a eusocial insect. Current Biology 17, 140–145. based technologies for investigating insect movement a survey of existing and Svensson,G.P.,Sahlin,U.,Brage,B.&Larsson, M. C. (2011). Should I stay or potential techniques. Computers and Electronics in Agriculture 35, 271–307. should I go? Modelling dispersal strategies in saproxylic insects based on pheromone Reynolds,A.M.,Smith,A.D.,Menzel,R.,Greggers,U.,Reynolds,D.R.& capture and radio telemetry a case study on the threatened hermit beetle Osmoderma Riley, J. R. (2007). Displaced honey bees perform optimal scale-free search flights. eremita. Biodiversity and Conservation 20, 2883–2902. Ecology 88, 1955–1961. Sword,G.A.,Lorch,P.D.&Gwynne, D. T. (2008). Radiotelemetric analysis of the Riecken,U.&Raths, U. (1996). Use of radio telemetry for studying dispersal and effects of prevailing wind direction on mormon cricket migratory band movement. habitat use of Carabus coriaceus L. Annales Zoologici Fennici 33, 109–116. Environmental Entomology 37, 889–896. Riecken,U.&Ries, U. (1992). Untersuchung zur raumnutzung von laufkafern¨ Turchin, P. (1998). Quantitative Analysis of Movement: Measuring and Modeling Population (Col.: Carabidae) mittels radio-telemetrie. Methodenentwicklung und erste Redistribution in Animals and Plants. Sinauer Associates, Sunderland. freilandversuche. Zeitschrift für Okologie¨ und Naturschutz 1, 147–149. Vinatier, F., Chailleux,A.,Duyck, P. F., Salmon, F., Lescourret,F.& Riley, J. R. (1975). Collective orientation in night-flying insects. Nature 253, Tixier, P. (2010). Radiotelemetry unravels movements of a walking insect species 113–114. in heterogeneous environments. Animal Behaviour 80, 221–229. Riley,J.R.,Greggers,U.,Smith,A.D.,Reynolds,D.R.&Menzel, R. (2005). Viswanathan,G.M.,Buldyrev,S.V.,Havlin,S.,da Luz,M.G.E.,Raposo, The flight paths of honeybees recruited by the waggle dance. Nature 435, 205–207. E. P. & Stanley, H. E. (1999). Optimizing the success of random searches. Nature Riley,J.R.,Greggers,U.,Smith,A.D.,Stach,S.,Reynolds,D.R.,Stollhoff, 401, 911–914. N., Brandt,R.,Schaupp,F.&Menzel, R. (2003). The automatic pilot of Voegeli,F.A.,Smale,M.J.,Webber,D.M.,Andrade,Y.&O’Dor, R. K. (2001). honeybees. Proceedings of the Royal Society of London. Series B: Biological Sciences 270, Ultrasonic telemetry, tracking and automated monitoring technology for sharks. 2421–2424. Environmental Biology of Fishes 60, 267–281. Riley,J.R.,Reynolds,D.R.,Smith,A.D.,Edwards,A.S.,Osborne,J.L., Walther-Hellwig,K.&Frankl, R. (2000). Foraging distances of Bombus muscorum, Williams,I.H.&McCartney, H. A. (1999). Compensation for wind drift by Bombus lapidarius,andBombus terrestris (Hymenoptera, Apidae). Journal of Insect Behavior bumble-bees. Nature 400, 126. 13, 239–246.

Watts,C.,Empson,R.,Thornburrow,D.&Rohan, M. (2012). Movements, Wikelski,M.,Moskowitz,D.,Adelman,J.S.,Cochran,J.,Wilcove,D.S. behaviour and survival of adult Cook Strait giant weta (Deinacrida rugosa; & May, M. L. (2006). Simple rules guide dragonﬂy migration. Biology Letters 2, Anostostomatidae: Orthoptera) immediately after translocation as revealed by 325–329. radiotracking. Journal of Insect Conservation 16, 763–776. Wikelski,M.,Moxley,J.,Eaton-Mordas,A.,Lopez-Uribe,M.M.,Holland, Watts,C.&Thornburrow, D. (2011). Habitat use, behavior and movement R., Moskowitz,D.,Roubik,D.W.&Kays, R. (2010). Large-range movements patterns of a threatened New Zealand giant weta, Deinacrida heteracantha of neotropical orchid bees observed via radio telemetry. PLoS One 5, e10738. (Anostostomatidae: Orthoptera). Journal of Orthoptera Research 20, 127–135. Wolf,T.J.&Schmid-Hempel, P. (1989). Extra loads and foraging life span in White,G.C.&Garrott, R. A. (1990). Analysis of Wildlife Radio-tracking Data. honeybee workers. Journal of Animal Ecology 58, 943–954. Academic Press, San Diego. Wikelski,M.,Kays,R.W.,Kasdin,N.J.,Thorup,K.,Smith,J.A.&Swenson, G. W. (2007). Going wild: what a global small-animal tracking system could do for experimental biologists. Journal of Experimental Biology 210, 181–186.

(Received 21 December 2012; revised 2 September 2013; accepted 5 September 2013; published online 8 October 2013)