RESEARCH ARTICLE

Bee and floral traits affect the characteristics of the vibrationsexperienced by flowers during buzz pollinationBlanca Arroyo-Correa12 Ceit Beattie1 and Mario Vallejo-Marın1

ABSTRACTDuring buzz pollination bees use their indirect flight muscles toproduce vibrations that are transmitted to the flowers and result inpollen release Although buzz pollination has been known forgt100 years we are still in the early stages of understanding howbee and floral characteristics affect the production and transmissionof floral vibrations Here we analysed floral vibrations produced byfour closely related bumblebee taxa (Bombus spp) on two buzz-pollinated plants species (Solanum spp) We measured floralvibrations transmitted to the flower to establish the extent to whichthemechanical properties of floral vibrations depend on bee and plantcharacteristics By comparing four bee taxa visiting the same plantspecies we found that peak acceleration root mean-squaredacceleration (RMS) and frequency vary between bee taxa but thatneither bee size (intertegular distance) nor flower biomass (dry mass)affects peak acceleration RMS or frequency A comparison of floralvibrations of two bee taxa visiting flowers of two plant species showedthat while bee species affects peak acceleration RMS andfrequency plant species only affects acceleration (peakacceleration and RMS) not frequency When accounting fordifferences in the transmission of vibrations across the two types offlower using a species-specific lsquocoupling factorrsquo we found that RMSacceleration and peak displacement do not differ between plantspecies This suggests that bees produce the same initialacceleration in different plants but that transmission of thesevibrations through the flower is affected by floral characteristics

KEY WORDS Floral biomechanics Bombus Substrate-bornevibrations Solanum Sonication

INTRODUCTIONBuzz pollination is a fascinating interaction between flowers withspecialised morphologies and bees that use vibrations to removepollen (Buchmann 1983 Nunes-Silva et al 2013 Vallejo-Mariacuten2019) Buzz pollination is by no means rare The ability to producevibrations to remove pollen from flowers has evolved at least 45times in the evolutionary history of bees and it is estimated thatapproximately 6 of flowering plants are buzz pollinated (Cardinalet al 2018) In most buzz-pollinated plants pollen is kept lockedinside tubular non-dehiscent anthers (ie poricidal anthersBuchmann 1983 Harris 1905) or inside closed corolla tubes

with small apical openings (Corbet and Huang 2014 Kawai andKudo 2008 Macior 1968) that restrict direct access to pollenPollen grains in buzz-pollinated flowers are most efficientlyremoved by bees that are capable of using floral vibrations ndash alsocalled lsquosonicationsrsquo or lsquobuzzesrsquo ndash to release pollen from poricidalanthers and flowers (De Luca and Vallejo-Mariacuten 2013) Despite thewidespread occurrence of the ability to produce floral vibrationsamong bees (Hymenoptera Apoidea Anthophila) little is knownabout why some bee species buzz pollinate (eg carpenter beesbumblebees several sweat bees) while others seem incapable ofdoing so (eg honeybees and most leaf-cutter bees) (Cardinal et al2018 De Luca and Vallejo-Mariacuten 2013) Moreover we are still inthe early stages of understanding the extent to which floralvibrations produced by bees vary between bee species andbetween different types of plants

The production of floral vibrations by bees is a relativelystereotyped behaviour (De Luca and Vallejo-Mariacuten 2013) Duringa typical buzz-pollinating visit a bee embraces one ormore poricidalanthers holding on to the anthers using its mandibles curls its bodyaround the anthers and produces one or more bursts of vibrationsthat can be heard by a human observer as lsquobuzzesrsquo of a higher pitchthan those heard during flight (Macior 1964 Russell et al 2016)(Fig 1) A bee generates floral vibrations using its thoracicmusculature which causes rapid deformation of the thorax whilethe wings remain undeployed (King and Buchmann 2003) Thesevibrations are transmitted as substrate-borne vibrations to the anthersthrough the mandibleshead thorax and abdomen of the bee (King1993 Vallejo-Mariacuten 2019) The vibrations result in pollen ejectionfrom the anther tips (Fig 1) probably as a consequence of energytransfer from the bee to the pollen grains (Buchmann and Hurley1978) which should be a function of themechanical properties of thesubstrate-borne vibrations as well as the characteristics of both thebee and flower (Vallejo-Mariacuten 2019) Determining exactly whatcharacteristics of floral vibrations are most important for pollenrelease is an area that requires further theoretical and empirical work

Floral vibrations produced by bees are a type of substrate-bornevibration and can be described by biophysical properties such asduration frequency and amplitude (Cocroft and Rodriguez 2005Mortimer 2017) During a single visit a bee may produce one toseveral lsquobuzzesrsquo lasting from tens of milliseconds to a couple ofseconds (Buchmann andCane 1989Vallejo-Mariacuten 2019) (Fig 2A)Floral vibrations containmultiple harmonics (Buchmann et al 1977)although usually the fundamental frequency (first harmonic)contributes most to the overall magnitude of substrate-borne floralvibrations (De Luca and Vallejo-Mariacuten 2013 King 1993) In suchcases the fundamental frequency is also the dominant or peakfrequency (Fig 1) The fundamental frequency of floral vibrationstypically ranges between 100 and 500 Hz (De Luca and Vallejo-Mariacuten 2013) Floral vibrations can also be characterised by theiramplitude (Fig 2BC) Amplitude describes the magnitude ofoscillatory movement and can itself be defined in terms ofReceived 12 December 2018 Accepted 17 January 2019

1Biological and Environmental Sciences School of Natural Sciences University ofStirling Stirling FK9 4LA UK 2Integrative Ecology Group Estacion Biologica deDon ana (EBD-CSIC) 21760 Sevilla Spain

Author for correspondence (mariovallejostiracuk)

BA-C 0000-0002-9402-3013 CB 0000-0003-1309-5873 MV-M 0000-0002-5663-8025

1

copy 2019 Published by The Company of Biologists Ltd | Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

acceleration velocity and displacement (Sueur 2018) In many typesof oscillatory movement frequency acceleration velocity andamplitude are interrelated and a full description of an oscillationthus requires knowledge of the absolute value of more than one ofthese variables However assuming simple harmonic oscillations it ispossible to use two of these variables (eg frequency and acceleration)to calculate another one (eg displacement) (Vallejo-Mariacuten 2019)An approach often taken to characterise floral vibrations during buzzpollination is to use acoustic recordings Audio recordings provideaccurate estimates of some aspects of substrate-borne vibrations suchas frequency and duration but are less reliable in estimating theamplitude component (De Luca et al 2018) As floral vibrations arein essence a substrate-borne phenomenon measurements of thevibrations experienced by flowers thus require the vibrationstransmitted to the flowers to be assessed directly using for exampleaccelerometers or laser vibrometers (Vallejo-Mariacuten 2019) To date

very few studies have attempted to compare floral vibrations producedby different species of bee visiting different species of plant usingdirect measurements of substrate-borne vibrations

Our study focused on bumblebees (Bombus spp Latreille) whichare well known for their ability to buzz pollinate and are importantbuzz pollinators in both temperate and tropical regions (De Lucaet al 2014 Mesquita-Neto et al 2018 Rosi-Denadai et al 2018)There are approximately 250 species of bumblebees (BombusBombini) worldwide comprising bees of medium to very large size(9ndash22 mm long) (Michener 2000) Besides their important role aspollinators in natural systems some species of bumblebees arecommercially bred and used widely around the world for thepollination of some crops including soft fruits such as raspberries(Lye et al 2011) and buzz-pollinated plants such as tomatoes(Morandin et al 2001) Previous work has shown that bumblebeespecies differ in the acoustic frequency of floral vibrations producedwhile visiting buzz-pollinated flowers (Corbet and Huang 2014 DeLuca et al 2014 Switzer and Combes 2017 P A De Luca S LBuchman C Galen A Mason and MV-M unpublished) but itremains unclear how other characteristics of substrate-borne floralvibrations (eg acceleration and displacement) vary betweenbumblebee species De Luca et al (2013) found a positivecorrelation between body size and amplitude (peak velocityamplitude) of substrate-borne floral vibrations within a singlebumblebee species The frequency of floral vibrations (estimatedfrom acoustic recordings) usually shows no correlation with bodysizewithin bumblebee species visiting a single plant (P A De LucaS L Buchman C Galen A Mason and MV-M unpublished)and shows conflicting patternswhen comparing the same bumblebeespecies visiting different plants (Corbet and Huang 2014 Switzerand Combes 2017) Although wingbeat frequency during flight isnegatively associated with body size the relationship between bodysize and the acoustic component of floral vibrations seems to bemuch weaker (P A De Luca S L Buchman C Galen A Masonand MV-M unpublished) Comparisons of the properties of

05 10 15 20 25Frequency (kHz)

Rel

ativ

e am

plitu

de

0

02

04

06

08

10

Fig 1 Frequency spectrum of a single floral vibration of the buff-tailedbumblebee (Bombus terrestris ssp audax hereafter B audax) on aflower of the buzz-pollinated herb buffalo bur (Solanum rostratum)The peak with the highest relative amplitude is the dominant frequencywhich often corresponds to the fundamental frequency of the vibration (355 Hzin this example) This floral vibration corresponds to the one highlighted inFig 2 The inset shows how the floral vibrations produced by the bumblebeeresult in pollen ejection from pores located at the tip of the anthers

0 01 02 03 04 05

minus15

minus10

minus5

0

5

10

15

Acc

eler

atio

n (m

sndash2

)

Time (s)0 002 004 006 008

minus15

minus10

minus5

0

5

10

15

0 002 004 006 008

minus15

minus10

minus5

0

5

10

15

A B C

Fig 2 Floral vibration of B audax on a flower of S rostratum (A) Oscillogram of the recorded floral vibration The vertical dashed lines indicate theregion of the vibration selected for analysis (B) The selected vibration on an enlarged scale Root mean square acceleration (RMS) was calculated fromthese data and is shown by a horizontal dashed line (C) Amplitude envelope of the vibration shown in B calculated with a window size of 2 and a 75 overlapMaximum peak acceleration was calculated from this amplitude envelope and is shown by a dashed line

2

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

substrate-borne vibrations produced by different species of bee in thesame flower and among different flowers are needed to betterunderstand how bee species and individual characteristics influencethe mechanical vibrations transmitted during buzz pollinationOne of the largest gaps in our current knowledge on buzz

pollination is the effect of plant characteristics on the transmission ofbee vibrations Characteristics of the flower such as mass stiffnessgeometry and other material properties of anthers and associatedfloral structures are expected to affect the transmission of vibrations(Michelsen et al 1982) Studies of animal communication usingplant-borne substrates have shown that plant architecture (eg branchdiameter number of internodes between caller and receiver) affectsthe transmission of insect vibrations (Gibson and Cocroft 2018)However it is less clear whether plant characteristics influence thevibrations transmitted over short distances within a single flower Todate only one study has quantitatively compared the transmission offloral vibrations among different plant species King (1993) studiedthe differences in amplitude of floral vibrations measured as peakacceleration amplitude produced by bumblebees visiting flowers ofSymphytum sp (comfrey) andActinidia deliciosa (kiwi)King (1993)calculated the ratio between the peak acceleration of vibrations ofknown amplitude applied to the anthers using a shaker table with thepeak acceleration of vibrations detected in the stem adjacent to theflowermeasuredwith an accelerometer The ratiowhich he called thelsquocoupling factorrsquo provides an estimate of the attenuation observedbetween the source of the vibration and the accelerometer and isexpected to be a function of the filtering and attenuating properties ofthe flower and intervening plant tissue (Cocroft andRodriguez 2005Mortimer 2017) King (1993) showed that the coupling factor differsbetween specieswith radically distinct floralmorphologies (comfrey264 kiwi 182) The results from this pioneering study raise twosimple but key points First plant identity and floral characteristicsmediate the transmission of vibrations through the flower Secondsignal attenuation must be taken into account when trying to infer thecharacteristics of vibrations experienced by the anthers frommeasurements taken in other parts of the flower (eg stem pedicelcalyx petals)Here we investigated the characteristics of substrate-borne floral

vibrations produced by four closely related bumblebee taxa(Bombus ss) (Cameron et al 2007) on two buzz-pollinatedplants in the genus Solanum L Section Androceras (Solanaceae)We used an accelerometer to collect measurements of accelerationand frequency from three taxa in the Bombus terrestris speciesaggregate (B terrestris B terrestris ssp audax and B terrestris sspcanariensis) as well as in B ignitus while visiting flowers ofSolanum rostratum and Solanum citrullifolium We used these datato address four questions (1) Do the properties of substrate-bornefloral vibrations differ between bumblebee taxa when visitingflowers of the same plant species (2) Do floral vibrations producedby the same bee taxon depend on the species of flower visited (3)Do closely related plant species differ in their transmissionproperties of substrate-borne vibrations as estimated using Kingrsquoscoupling factor (4) Does bee size andor floral biomass influencethe characteristics of floral-borne vibrations transmitted to theflower during buzz pollination

MATERIALS AND METHODSBee materialWe studied four closely related bumblebee taxa in the Bombussubgenus Bombus Latreille (Hymenoptera Apidae) All beecolonies used were reared and supplied by Biobest (WesterloBelgium) Three taxa belong to the B terrestris L species

aggregate two of these are B terrestris ssp audax (Harris 1776)(hereafter B audax) and B terrestris ssp canariensis Peacuterez 1895(hereafter B canariensis) (Rasmont et al 2008) These two taxahave different distributions with B audax native to the British Islesand B canariensis native to the Canary Islands (Rasmont et al2008) The third taxon belongs to B terrestris but its exact originremains uncertain (hereafter B terrestris) it is probably B terrestrisssp dalmatinusDalla Torre 1882 (Lecocq et al 2016 Velthuis andVan Doorn 2006) which has a European continental distributionDifferences among subspecies within the B terrestris aggregateinclude coloration and behavioural traits (Rasmont et al 2008) Thefourth taxon studied was B ignitus (Smith 1869) which can befound in China Korea and Japan (Shao et al 2004)

We used five colonies from each the four taxa (six for B audax)When the colonies were received (3 July 2018) each colonyconsisted of a single queen and approximately 20ndash40 workers Anadditional colony of B audax was added to the experiment onAugust 2018 Upon receipt the colonies were placed underenclosed laboratory conditions at room temperature (20ndash23degC)and fed ad libitum with sugar solution from a plastic containerplaced under the colony (Biogluc Biobest) Additionally weprovided each colony with 2 g of ground honeybee-collected pollenevery week throughout the experiment Bee experiments wereconducted with approval from the Ethics Committee of theUniversity of Stirling

Plant materialWe studied two closely related species of buzz-pollinated plants inthe genus Solanum (Solanaceae) Solanum flowers are nectarlessand offer pollen as the main reward to attract pollinators Pollen iscontained within poricidal anthers (Harris 1905) which arearranged in the centre of the flower sometimes forming a coneThe most efficient way to release pollen is through vibrationsapplied to the anthers (Bowers 1975) Solanum is a classic systemfor the study of buzz pollination (Buchmann et al 1977 De Lucaand Vallejo-Mariacuten 2013) Here we studied two annual species inthe monophyletic clade Solanum Section Androceras (Whalen1979) S rostratum Dunal and S citrullifolium (A Braun) NieuwlThese two species depart form the classical Solanum-type flowerby having heterantherous flowers in which the anthers aredifferentiated into two functional types (Vallejo-Marin et al2014) (Fig 1) One type the four feeding anthers is yellow andplaced at the centre of the flower and the second type the singlepollinating anther is usually differently coloured (yellow to reddishbrown or light violet in the species studied here) and displaced toeither the right- or left-hand side of the centre of the flower (Vallejo-Marin et al 2014) At anthesis the flowers are orientated with theanthers parallel to the ground Bees usually manipulate the feedinganthers (Fig 1) and only rarely directly manipulate the pollinatinganther (Vallejo-Marin et al 2009) Plants were grown from seed atthe University of Stirling research glasshouses following theprotocol described in Vallejo-Marin et al (2014) Seeds fromS rostratum were collected in the field from open-pollinated fruitsfrom three separate individuals (accessions 10s34 10s36 and10s39) in Puerto del Aire Veracruz Mexico (1874degN 9753degW)Seeds from S citrullifolium were obtained from self-fertilised fruitsof three individuals (accession 199) grown from seeds originallyobtained from the Radboud University Solanaceae collection(accession 894750197) Fresh flowers were transported from theglasshouse to the laboratory in open containers by cutting entireinflorescences and placing the stems in water-soaked Ideal FloralFoam (Oasis Floral Products Washington UK) Inflorescences

3

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

were then kept in wet floral foam and stored at room temperatureuntil used for experiments

Bumblebee colony conditioningBumblebee colony conditioning was performed in a122 cmtimes100 cmtimes37 cm wooden flight arena with a Perspex lidThe arena was divided in half with a wooden panel to allow up totwo colonies to be connected at the same time The arena wasilluminated with an LED light panel (595 cmtimes595 cm 48 WDaylight Opus Lighting Technology Birmingham UK) on eachside To encourage bees to forage during the experiment a pollenand a nectar feeder (Russell and Papaj 2016) were placed in thecentre of the flight arena A single colony was attached to each sideof the arena and bees were allowed to enter the arena and return tothe colony freely for at least 1 day Following this all colonies wereexposed to a bouquet of S citrullifolium and S rostratum placed inthe flight arena After this period we considered the colonylsquoconditionedrsquo and used it in buzz pollination trials

Buzz pollination trialsWe conducted the buzz pollination trials in a 60 cmtimes60 cmtimes60 cmlsquoRussellrsquo flight arena made of wooden panels and a UV-transparentPerspex lid illuminated with an LED light panel on the top(A Russell personal communication) The flight arena had twoopenings in the front side of the box covered with clear plastic toallow direct observations into the arena A single bumblebee colonywas connected at any one time to the flight arena using plastic tubeswith sliding plasticmetal doors to control bee access to the arena Ineach buzz pollination trial a single bee was allowed to enter theflight arena and forage for a maximum of 10 min If the bee did notvisit the flower attached to the accelerometer (see below) during thattime we returned it to the colony and allowed the next bee to enterthe arena Bees that vibrated any flower were allowed to continuevisiting the experimental array for up to 15 min to enable datacollection and subsequently captured and marked on the thoraxwith an individual ID using coloured pens (POSCA MitsubishiMilton Keynes UK) Marked bees were returned to the colony withtheir pollen baskets intact Between trials we replaced any flowerthat had been visited more than 3 times with fresh flowersIn each buzz pollination trial we placed an inflorescence with

1ndash3 flowers within the flight arena and recorded floral vibrations ona single focal flower using a calibrated miniature lightweight(08 g) uniaxial accelerometer (Model 352A24 PCB PiezotronicsInc Depew NY YSA) The accelerometer was attached to thefocal flower through a small (5times035 mm) metal pin glued withLoctite 454 to the accelerometer The pin was then inserted intothe base of the flower (receptacle) at a 90 deg angle to the longestaxis of the stamens (Fig S1) We chose to attach the accelerometerusing a pin as gluing the accelerometer to the flower would haveresulted in additional mass loading to the flower and introducedmass variation among flowers Although the vibrations producedby the bee and transmitted to the flower probably result inmovement of the flower in three dimensions (Gibson and Cocroft2018) our accelerometer only records vibrations along a singleaxis We chose to measure vibrations along this axis as theoreticalmodels of buzz pollination suggest that pollen ejection is afunction of the vibration magnitude along an axis perpendicularto the longest axis of the stamen (Buchmann and Hurley 1978Vallejo-Mariacuten 2019) The accelerometer was connected to abattery-powered signal conditioner (480C02 PCB PiezotronicsInc) and the signal was digitised using a digital storageoscilloscope (TBS1032B Tektronix UK Ltd Bracknell UK)

at a sampling rate of 2500ndash25000 Hz (mean 6745 Hz median5000 Hz) and stored as a text file

Estimating vibrational propertiesAccelerometerAccelerometer recordings were imported into R v351 (httpwwwR-projectorg) and stored as time series objects We selected1ndash4 recordings per bee and foraging bout Sampling frequency andprobe attenuation were extracted from each accelerometer file Thevoltage readings of the accelerometer recorded in the oscilloscopewere transformed into acceleration units (m sminus2) using the productrsquoscalibration (102 mV msminus2) Time series were centred on zero andanalysed using seewave (Sueur 2018) In order to minimise low-frequency noise we used a high-pass filter of 80 Hz (which is wellbelow the fundamental frequency of bumblebee floral vibrationsDe Luca and Vallejo-Mariacuten 2013) with a 512 Hamming Fouriertransformwindow using the function fir We used the timer functionto identify individual vibrations not broken by silence A singlevibration from each accelerometer file was randomly selected forextracting frequency and acceleration information (Fig 2A) Wecalculated root mean square (RMS) acceleration from the time-series data using the rms function (Fig 2B) To calculate peakacceleration we first obtained the amplitude envelope of the timewave (absolute amplitude) smoothed it with a window size of 2 andan overlap of 75 and then obtained the maximum observedacceleration (Sueur 2018) (Fig 2C) This approach should providea conservative estimate of peak acceleration that is more robust tovoltage spikes in the time wave Fundamental frequency wascalculated over the duration of the vibration using the maximumpossible window length (128ndash2048 samples median 512) withinthe limit imposed by the total number of samples of each vibration(window lengthlenumber of samples) Larger window lengthsimprove frequency resolution (Δf=sampling frequencywindowlength) (Sueur 2018 Vallejo-Mariacuten 2019) We estimated thefundamental frequency using the function fund with a maximum of500 Hz (De Luca and Vallejo-Mariacuten 2013) and a window overlapof 75 When the recording length allowed for multiple estimatesof fundamental frequency for the same vibration we calculated themean and used this for analysis

Coupling factorThe transmission of mechanical vibrations from the bee to the floweris probably influenced by the structural andmaterial properties of theflower itself (Vallejo-Mariacuten 2019) The vibrations we measured atthe base of the flower are therefore potentially different from thevibrations experienced at the anthers Moreover interspecificdifferences in floral architecture make it likely that thetransmission of vibrations from the bee to the floral organs variesbetween plant species (Vallejo-Mariacuten 2019) King (1993) suggestscalculating a lsquocoupling factorrsquo to quantify the change in magnitudebetween the vibrations produced at the anther level and thosemeasured at a distance from the anthers Following King (1993) weused a Bruel amp Kjaer 4294 Calibration Exciter (Naerum Denmark)to produce vibrations of known frequency and acceleration(15915 Hz plusmn002 10 m sminus2 plusmn3 RMS acceleration) Weapplied the calibration exciter vibrations to the flower by firmlycontacting the feeding anthers arranged in their natural position (FigS1B) We recorded the vibrations experienced by the accelerometerattached to the flowerrsquos calyx and calculated the coupling factor iethe ratio of RMS acceleration between the anther and calyxvibrations (coupling factor=RMScalibration exciterRMSaccelerometer)Coupling factor values greater than 1 indicate that the amplitude of

4

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

substrate-borne vibrations produced by different species of bee in thesame flower and among different flowers are needed to betterunderstand how bee species and individual characteristics influencethe mechanical vibrations transmitted during buzz pollinationOne of the largest gaps in our current knowledge on buzz

pollination is the effect of plant characteristics on the transmission ofbee vibrations Characteristics of the flower such as mass stiffnessgeometry and other material properties of anthers and associatedfloral structures are expected to affect the transmission of vibrations(Michelsen et al 1982) Studies of animal communication usingplant-borne substrates have shown that plant architecture (eg branchdiameter number of internodes between caller and receiver) affectsthe transmission of insect vibrations (Gibson and Cocroft 2018)However it is less clear whether plant characteristics influence thevibrations transmitted over short distances within a single flower Todate only one study has quantitatively compared the transmission offloral vibrations among different plant species King (1993) studiedthe differences in amplitude of floral vibrations measured as peakacceleration amplitude produced by bumblebees visiting flowers ofSymphytum sp (comfrey) andActinidia deliciosa (kiwi)King (1993)calculated the ratio between the peak acceleration of vibrations ofknown amplitude applied to the anthers using a shaker table with thepeak acceleration of vibrations detected in the stem adjacent to theflowermeasuredwith an accelerometer The ratiowhich he called thelsquocoupling factorrsquo provides an estimate of the attenuation observedbetween the source of the vibration and the accelerometer and isexpected to be a function of the filtering and attenuating properties ofthe flower and intervening plant tissue (Cocroft andRodriguez 2005Mortimer 2017) King (1993) showed that the coupling factor differsbetween specieswith radically distinct floralmorphologies (comfrey264 kiwi 182) The results from this pioneering study raise twosimple but key points First plant identity and floral characteristicsmediate the transmission of vibrations through the flower Secondsignal attenuation must be taken into account when trying to infer thecharacteristics of vibrations experienced by the anthers frommeasurements taken in other parts of the flower (eg stem pedicelcalyx petals)Here we investigated the characteristics of substrate-borne floral

vibrations produced by four closely related bumblebee taxa(Bombus ss) (Cameron et al 2007) on two buzz-pollinatedplants in the genus Solanum L Section Androceras (Solanaceae)We used an accelerometer to collect measurements of accelerationand frequency from three taxa in the Bombus terrestris speciesaggregate (B terrestris B terrestris ssp audax and B terrestris sspcanariensis) as well as in B ignitus while visiting flowers ofSolanum rostratum and Solanum citrullifolium We used these datato address four questions (1) Do the properties of substrate-bornefloral vibrations differ between bumblebee taxa when visitingflowers of the same plant species (2) Do floral vibrations producedby the same bee taxon depend on the species of flower visited (3)Do closely related plant species differ in their transmissionproperties of substrate-borne vibrations as estimated using Kingrsquoscoupling factor (4) Does bee size andor floral biomass influencethe characteristics of floral-borne vibrations transmitted to theflower during buzz pollination

MATERIALS AND METHODSBee materialWe studied four closely related bumblebee taxa in the Bombussubgenus Bombus Latreille (Hymenoptera Apidae) All beecolonies used were reared and supplied by Biobest (WesterloBelgium) Three taxa belong to the B terrestris L species

aggregate two of these are B terrestris ssp audax (Harris 1776)(hereafter B audax) and B terrestris ssp canariensis Peacuterez 1895(hereafter B canariensis) (Rasmont et al 2008) These two taxahave different distributions with B audax native to the British Islesand B canariensis native to the Canary Islands (Rasmont et al2008) The third taxon belongs to B terrestris but its exact originremains uncertain (hereafter B terrestris) it is probably B terrestrisssp dalmatinusDalla Torre 1882 (Lecocq et al 2016 Velthuis andVan Doorn 2006) which has a European continental distributionDifferences among subspecies within the B terrestris aggregateinclude coloration and behavioural traits (Rasmont et al 2008) Thefourth taxon studied was B ignitus (Smith 1869) which can befound in China Korea and Japan (Shao et al 2004)

We used five colonies from each the four taxa (six for B audax)When the colonies were received (3 July 2018) each colonyconsisted of a single queen and approximately 20ndash40 workers Anadditional colony of B audax was added to the experiment onAugust 2018 Upon receipt the colonies were placed underenclosed laboratory conditions at room temperature (20ndash23degC)and fed ad libitum with sugar solution from a plastic containerplaced under the colony (Biogluc Biobest) Additionally weprovided each colony with 2 g of ground honeybee-collected pollenevery week throughout the experiment Bee experiments wereconducted with approval from the Ethics Committee of theUniversity of Stirling

Plant materialWe studied two closely related species of buzz-pollinated plants inthe genus Solanum (Solanaceae) Solanum flowers are nectarlessand offer pollen as the main reward to attract pollinators Pollen iscontained within poricidal anthers (Harris 1905) which arearranged in the centre of the flower sometimes forming a coneThe most efficient way to release pollen is through vibrationsapplied to the anthers (Bowers 1975) Solanum is a classic systemfor the study of buzz pollination (Buchmann et al 1977 De Lucaand Vallejo-Mariacuten 2013) Here we studied two annual species inthe monophyletic clade Solanum Section Androceras (Whalen1979) S rostratum Dunal and S citrullifolium (A Braun) NieuwlThese two species depart form the classical Solanum-type flowerby having heterantherous flowers in which the anthers aredifferentiated into two functional types (Vallejo-Marin et al2014) (Fig 1) One type the four feeding anthers is yellow andplaced at the centre of the flower and the second type the singlepollinating anther is usually differently coloured (yellow to reddishbrown or light violet in the species studied here) and displaced toeither the right- or left-hand side of the centre of the flower (Vallejo-Marin et al 2014) At anthesis the flowers are orientated with theanthers parallel to the ground Bees usually manipulate the feedinganthers (Fig 1) and only rarely directly manipulate the pollinatinganther (Vallejo-Marin et al 2009) Plants were grown from seed atthe University of Stirling research glasshouses following theprotocol described in Vallejo-Marin et al (2014) Seeds fromS rostratum were collected in the field from open-pollinated fruitsfrom three separate individuals (accessions 10s34 10s36 and10s39) in Puerto del Aire Veracruz Mexico (1874degN 9753degW)Seeds from S citrullifolium were obtained from self-fertilised fruitsof three individuals (accession 199) grown from seeds originallyobtained from the Radboud University Solanaceae collection(accession 894750197) Fresh flowers were transported from theglasshouse to the laboratory in open containers by cutting entireinflorescences and placing the stems in water-soaked Ideal FloralFoam (Oasis Floral Products Washington UK) Inflorescences

3

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

were then kept in wet floral foam and stored at room temperatureuntil used for experiments

Bumblebee colony conditioningBumblebee colony conditioning was performed in a122 cmtimes100 cmtimes37 cm wooden flight arena with a Perspex lidThe arena was divided in half with a wooden panel to allow up totwo colonies to be connected at the same time The arena wasilluminated with an LED light panel (595 cmtimes595 cm 48 WDaylight Opus Lighting Technology Birmingham UK) on eachside To encourage bees to forage during the experiment a pollenand a nectar feeder (Russell and Papaj 2016) were placed in thecentre of the flight arena A single colony was attached to each sideof the arena and bees were allowed to enter the arena and return tothe colony freely for at least 1 day Following this all colonies wereexposed to a bouquet of S citrullifolium and S rostratum placed inthe flight arena After this period we considered the colonylsquoconditionedrsquo and used it in buzz pollination trials

Buzz pollination trialsWe conducted the buzz pollination trials in a 60 cmtimes60 cmtimes60 cmlsquoRussellrsquo flight arena made of wooden panels and a UV-transparentPerspex lid illuminated with an LED light panel on the top(A Russell personal communication) The flight arena had twoopenings in the front side of the box covered with clear plastic toallow direct observations into the arena A single bumblebee colonywas connected at any one time to the flight arena using plastic tubeswith sliding plasticmetal doors to control bee access to the arena Ineach buzz pollination trial a single bee was allowed to enter theflight arena and forage for a maximum of 10 min If the bee did notvisit the flower attached to the accelerometer (see below) during thattime we returned it to the colony and allowed the next bee to enterthe arena Bees that vibrated any flower were allowed to continuevisiting the experimental array for up to 15 min to enable datacollection and subsequently captured and marked on the thoraxwith an individual ID using coloured pens (POSCA MitsubishiMilton Keynes UK) Marked bees were returned to the colony withtheir pollen baskets intact Between trials we replaced any flowerthat had been visited more than 3 times with fresh flowersIn each buzz pollination trial we placed an inflorescence with

1ndash3 flowers within the flight arena and recorded floral vibrations ona single focal flower using a calibrated miniature lightweight(08 g) uniaxial accelerometer (Model 352A24 PCB PiezotronicsInc Depew NY YSA) The accelerometer was attached to thefocal flower through a small (5times035 mm) metal pin glued withLoctite 454 to the accelerometer The pin was then inserted intothe base of the flower (receptacle) at a 90 deg angle to the longestaxis of the stamens (Fig S1) We chose to attach the accelerometerusing a pin as gluing the accelerometer to the flower would haveresulted in additional mass loading to the flower and introducedmass variation among flowers Although the vibrations producedby the bee and transmitted to the flower probably result inmovement of the flower in three dimensions (Gibson and Cocroft2018) our accelerometer only records vibrations along a singleaxis We chose to measure vibrations along this axis as theoreticalmodels of buzz pollination suggest that pollen ejection is afunction of the vibration magnitude along an axis perpendicularto the longest axis of the stamen (Buchmann and Hurley 1978Vallejo-Mariacuten 2019) The accelerometer was connected to abattery-powered signal conditioner (480C02 PCB PiezotronicsInc) and the signal was digitised using a digital storageoscilloscope (TBS1032B Tektronix UK Ltd Bracknell UK)

at a sampling rate of 2500ndash25000 Hz (mean 6745 Hz median5000 Hz) and stored as a text file

Estimating vibrational propertiesAccelerometerAccelerometer recordings were imported into R v351 (httpwwwR-projectorg) and stored as time series objects We selected1ndash4 recordings per bee and foraging bout Sampling frequency andprobe attenuation were extracted from each accelerometer file Thevoltage readings of the accelerometer recorded in the oscilloscopewere transformed into acceleration units (m sminus2) using the productrsquoscalibration (102 mV msminus2) Time series were centred on zero andanalysed using seewave (Sueur 2018) In order to minimise low-frequency noise we used a high-pass filter of 80 Hz (which is wellbelow the fundamental frequency of bumblebee floral vibrationsDe Luca and Vallejo-Mariacuten 2013) with a 512 Hamming Fouriertransformwindow using the function fir We used the timer functionto identify individual vibrations not broken by silence A singlevibration from each accelerometer file was randomly selected forextracting frequency and acceleration information (Fig 2A) Wecalculated root mean square (RMS) acceleration from the time-series data using the rms function (Fig 2B) To calculate peakacceleration we first obtained the amplitude envelope of the timewave (absolute amplitude) smoothed it with a window size of 2 andan overlap of 75 and then obtained the maximum observedacceleration (Sueur 2018) (Fig 2C) This approach should providea conservative estimate of peak acceleration that is more robust tovoltage spikes in the time wave Fundamental frequency wascalculated over the duration of the vibration using the maximumpossible window length (128ndash2048 samples median 512) withinthe limit imposed by the total number of samples of each vibration(window lengthlenumber of samples) Larger window lengthsimprove frequency resolution (Δf=sampling frequencywindowlength) (Sueur 2018 Vallejo-Mariacuten 2019) We estimated thefundamental frequency using the function fund with a maximum of500 Hz (De Luca and Vallejo-Mariacuten 2013) and a window overlapof 75 When the recording length allowed for multiple estimatesof fundamental frequency for the same vibration we calculated themean and used this for analysis

Coupling factorThe transmission of mechanical vibrations from the bee to the floweris probably influenced by the structural andmaterial properties of theflower itself (Vallejo-Mariacuten 2019) The vibrations we measured atthe base of the flower are therefore potentially different from thevibrations experienced at the anthers Moreover interspecificdifferences in floral architecture make it likely that thetransmission of vibrations from the bee to the floral organs variesbetween plant species (Vallejo-Mariacuten 2019) King (1993) suggestscalculating a lsquocoupling factorrsquo to quantify the change in magnitudebetween the vibrations produced at the anther level and thosemeasured at a distance from the anthers Following King (1993) weused a Bruel amp Kjaer 4294 Calibration Exciter (Naerum Denmark)to produce vibrations of known frequency and acceleration(15915 Hz plusmn002 10 m sminus2 plusmn3 RMS acceleration) Weapplied the calibration exciter vibrations to the flower by firmlycontacting the feeding anthers arranged in their natural position (FigS1B) We recorded the vibrations experienced by the accelerometerattached to the flowerrsquos calyx and calculated the coupling factor iethe ratio of RMS acceleration between the anther and calyxvibrations (coupling factor=RMScalibration exciterRMSaccelerometer)Coupling factor values greater than 1 indicate that the amplitude of

4

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

were then kept in wet floral foam and stored at room temperatureuntil used for experiments

Bumblebee colony conditioningBumblebee colony conditioning was performed in a122 cmtimes100 cmtimes37 cm wooden flight arena with a Perspex lidThe arena was divided in half with a wooden panel to allow up totwo colonies to be connected at the same time The arena wasilluminated with an LED light panel (595 cmtimes595 cm 48 WDaylight Opus Lighting Technology Birmingham UK) on eachside To encourage bees to forage during the experiment a pollenand a nectar feeder (Russell and Papaj 2016) were placed in thecentre of the flight arena A single colony was attached to each sideof the arena and bees were allowed to enter the arena and return tothe colony freely for at least 1 day Following this all colonies wereexposed to a bouquet of S citrullifolium and S rostratum placed inthe flight arena After this period we considered the colonylsquoconditionedrsquo and used it in buzz pollination trials

Buzz pollination trialsWe conducted the buzz pollination trials in a 60 cmtimes60 cmtimes60 cmlsquoRussellrsquo flight arena made of wooden panels and a UV-transparentPerspex lid illuminated with an LED light panel on the top(A Russell personal communication) The flight arena had twoopenings in the front side of the box covered with clear plastic toallow direct observations into the arena A single bumblebee colonywas connected at any one time to the flight arena using plastic tubeswith sliding plasticmetal doors to control bee access to the arena Ineach buzz pollination trial a single bee was allowed to enter theflight arena and forage for a maximum of 10 min If the bee did notvisit the flower attached to the accelerometer (see below) during thattime we returned it to the colony and allowed the next bee to enterthe arena Bees that vibrated any flower were allowed to continuevisiting the experimental array for up to 15 min to enable datacollection and subsequently captured and marked on the thoraxwith an individual ID using coloured pens (POSCA MitsubishiMilton Keynes UK) Marked bees were returned to the colony withtheir pollen baskets intact Between trials we replaced any flowerthat had been visited more than 3 times with fresh flowersIn each buzz pollination trial we placed an inflorescence with

1ndash3 flowers within the flight arena and recorded floral vibrations ona single focal flower using a calibrated miniature lightweight(08 g) uniaxial accelerometer (Model 352A24 PCB PiezotronicsInc Depew NY YSA) The accelerometer was attached to thefocal flower through a small (5times035 mm) metal pin glued withLoctite 454 to the accelerometer The pin was then inserted intothe base of the flower (receptacle) at a 90 deg angle to the longestaxis of the stamens (Fig S1) We chose to attach the accelerometerusing a pin as gluing the accelerometer to the flower would haveresulted in additional mass loading to the flower and introducedmass variation among flowers Although the vibrations producedby the bee and transmitted to the flower probably result inmovement of the flower in three dimensions (Gibson and Cocroft2018) our accelerometer only records vibrations along a singleaxis We chose to measure vibrations along this axis as theoreticalmodels of buzz pollination suggest that pollen ejection is afunction of the vibration magnitude along an axis perpendicularto the longest axis of the stamen (Buchmann and Hurley 1978Vallejo-Mariacuten 2019) The accelerometer was connected to abattery-powered signal conditioner (480C02 PCB PiezotronicsInc) and the signal was digitised using a digital storageoscilloscope (TBS1032B Tektronix UK Ltd Bracknell UK)

at a sampling rate of 2500ndash25000 Hz (mean 6745 Hz median5000 Hz) and stored as a text file

Estimating vibrational propertiesAccelerometerAccelerometer recordings were imported into R v351 (httpwwwR-projectorg) and stored as time series objects We selected1ndash4 recordings per bee and foraging bout Sampling frequency andprobe attenuation were extracted from each accelerometer file Thevoltage readings of the accelerometer recorded in the oscilloscopewere transformed into acceleration units (m sminus2) using the productrsquoscalibration (102 mV msminus2) Time series were centred on zero andanalysed using seewave (Sueur 2018) In order to minimise low-frequency noise we used a high-pass filter of 80 Hz (which is wellbelow the fundamental frequency of bumblebee floral vibrationsDe Luca and Vallejo-Mariacuten 2013) with a 512 Hamming Fouriertransformwindow using the function fir We used the timer functionto identify individual vibrations not broken by silence A singlevibration from each accelerometer file was randomly selected forextracting frequency and acceleration information (Fig 2A) Wecalculated root mean square (RMS) acceleration from the time-series data using the rms function (Fig 2B) To calculate peakacceleration we first obtained the amplitude envelope of the timewave (absolute amplitude) smoothed it with a window size of 2 andan overlap of 75 and then obtained the maximum observedacceleration (Sueur 2018) (Fig 2C) This approach should providea conservative estimate of peak acceleration that is more robust tovoltage spikes in the time wave Fundamental frequency wascalculated over the duration of the vibration using the maximumpossible window length (128ndash2048 samples median 512) withinthe limit imposed by the total number of samples of each vibration(window lengthlenumber of samples) Larger window lengthsimprove frequency resolution (Δf=sampling frequencywindowlength) (Sueur 2018 Vallejo-Mariacuten 2019) We estimated thefundamental frequency using the function fund with a maximum of500 Hz (De Luca and Vallejo-Mariacuten 2013) and a window overlapof 75 When the recording length allowed for multiple estimatesof fundamental frequency for the same vibration we calculated themean and used this for analysis

Coupling factorThe transmission of mechanical vibrations from the bee to the floweris probably influenced by the structural andmaterial properties of theflower itself (Vallejo-Mariacuten 2019) The vibrations we measured atthe base of the flower are therefore potentially different from thevibrations experienced at the anthers Moreover interspecificdifferences in floral architecture make it likely that thetransmission of vibrations from the bee to the floral organs variesbetween plant species (Vallejo-Mariacuten 2019) King (1993) suggestscalculating a lsquocoupling factorrsquo to quantify the change in magnitudebetween the vibrations produced at the anther level and thosemeasured at a distance from the anthers Following King (1993) weused a Bruel amp Kjaer 4294 Calibration Exciter (Naerum Denmark)to produce vibrations of known frequency and acceleration(15915 Hz plusmn002 10 m sminus2 plusmn3 RMS acceleration) Weapplied the calibration exciter vibrations to the flower by firmlycontacting the feeding anthers arranged in their natural position (FigS1B) We recorded the vibrations experienced by the accelerometerattached to the flowerrsquos calyx and calculated the coupling factor iethe ratio of RMS acceleration between the anther and calyxvibrations (coupling factor=RMScalibration exciterRMSaccelerometer)Coupling factor values greater than 1 indicate that the amplitude of

4

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

the vibrations is reduced RMSacceleration valueswere calculated asdescribed in the lsquoEstimating vibrational propertiesrsquo section aboveWe estimated the coupling factor for each plant species using 30flowers of S rostratum (from 5 individuals and 3 accessions) and 35flowers of S citrullifolium (from 3 individuals and 3 accessions)Weanalysed an average of 11 recordings per flower (N=365 for Scitrullifolium and N=370 for S rostratum) RMS acceleration wascalculated as above

Peak displacementWe estimated peak displacement using fundamental frequency andpeak acceleration assuming simple sinusoidal motion using theequation Dpeak=Apeak(2πF )

2 where Dpeak is peak displacementApeak is peak acceleration amplitude and F is fundamental frequency(see King and Buchmann 1996) To estimate peak displacementincorporating the coupling factor we used the value of Apeak

multiplied by the species-specific coupling factor calculatedas above

Flower massAs the mass of the flower could influence the magnitude of thevibrations transmitted from the body of the bee to the floralreceptacle we estimated flower mass of each flower used in the buzzpollination trials In order to estimate flower mass before visitation(and hence before pollen removal) we used an indirect approach Werandomly sampled 100 flowers of each plant species (including thepedicel) and used them to estimate a species-specific correlationbetween flower morphological measurements and dry biomass (mg)To do this we measured maximum corolla length the length ofthe feeding anthers and the pollinating anther length These threemorphological measurements were summarised using a principalcomponent analysis and the first principal component (PC1) wasused as an overall estimate of flower size Then each flowerwas dried at 60degC for 36 h and weighed with an analytical balancewith 001 mg resolution (Analytical Plus AP250D OhausGreifensee Switzerland) We then calculated a correlation betweenPC1 and flower mass (mg) and used this association to calculatethe dry biomass of experimental flowers from morphologicalmeasurements

Bee sizeAs an estimate of bee size we used intertegular distance (ITD)which has been shown to be a good approximation of overall beesize (Cane 1987) and is regularly used in buzz pollination studies(P A De Luca S L Buchman C Galen A Mason and MV-Munpublished) At the end of the behavioural trials (approximatelythe end of August) we killed bees by placing the entire colony in aminus80degC freezer for 48 h All marked bees and an additional randomsample of workers all identifiable males and queens weremeasured (100 bees per species 20 per colony) Bees (workersonly) were measured with digital callipers (Mitutoyo AndoverUK) to the nearest 01 mm In total we measured 557 worker beesfrom 21 colonies of four taxa

Statistical analysisTo compare differences in size (ITD) between workers of differentbee species we used a linear mixed-effects model with ITD as aresponse variable species as an explanatory variable (fixed effect)and colony as a random effect (21 coloniesN=557 bees) We used aTukey test for pairwise comparisons between bee taxa To comparethe size of bees that were observed producing floral vibrations(buzzers) versus a random sample of workers from each colony

(all bees) we selected colonies that included both categories of bees(8 colonies N=282 bees) In this analysis ITD was used as aresponse variable bee species and bee type (buzzer versus all bees)as fixed effects and colony as a random effect The interactionbetween species and bee type was excluded during preliminaryanalyses as it was not significant All bees included worker bees thatleft the colony and did not buzz as well as bees that never left thecolony and may have included some bees that buzzed but were notobserved and marked

To test for differences in the vibrational properties betweenbumblebee species and between flowers of the two Solanumspecies we also used linear mixed-effects models The responsevariables were RMS acceleration peak acceleration fundamentalfrequency or peak displacement Bee size (ITD) flower massbumblebee species and plant species were included as fixed effectsAs we had several vibrational measures for most bumblebeeindividuals we used bumblebee individual as a random effectwhich accounts for the statistical associations among vibrationsproduced by the same individual All analyses were done in Rv351 (httpwwwR-projectorg) using packages lme4 (httpCRANR-projectorgpackage=lme4) for parameter estimationand lmerTest (httpsCRANR-projectorgpackage=lmerTest) forassessing statistical significance

RESULTSWe found that size (ITD) of individual worker bees wassignificantly different among species (Plt001 Figs S2 and S3)Specifically Bombus audax was smaller than the other three taxawhich did not differ significantly in size from each other (Fig S2)A comparison of the size of bees observed vibrating Solanumflowers against all worker bees sampled within each colony (usingcolony as a random effect and bee species as a fixed effect) showedthat buzz-pollinating bees were on average larger (coefficient for allworkers=minus0286 Plt0001) (Fig 3)

We obtained a total of 230 recordings of floral vibrations from 79different individual bumblebees from eight colonies of fourdifferent taxa (Table 1) All colonies and bee taxa were recordedon S rostratum while three of these colonies (two Bombus audaxand one B canariensis colony) were recorded on both S rostratumand S citrullifolium (102 recordings on S rostratum and 82 onS citrullifolium) Analysis of the calibrated vibrations applied tothe anthers and measured with the accelerometer at the base of theflower showed that the two studied plant species differed in thedamping of vibrations transmitted through the flower The couplingfactor for S rostratum was 884plusmn023 (meanplusmnsem) and that forS citrullifolium was 1221plusmn029 indicating that flowers ofS citrullifolium dampen floral vibrations more strongly than thoseof its congener The observed differences in the coupling factor ofeach plant species were statistically significant as assessed with alinear model with species as the explanatory variable [speciescoefficient (S rostratum)=minus3367 Plt0001]

Differences between bumblebee taxa on S rostratumWe found significant differences among bumblebee species in peakacceleration RMS acceleration (with and without coupling factor)and fundamental frequency (Table 2 Fig 4) Bombus audax andB terrestris had floral vibrations with the highest fundamentalfrequency followed by B canariensis and B ignitus (Fig 4A)Bombus audax had a RMS acceleration higher than that of allother taxa as well as one of the highest peak accelerations (withB ignitus) (Fig 4BC) The other three taxa achieved similar levelsof both RMS and peak acceleration ITD and flower biomass were

5

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

not significantly associated with either measurement of accelerationor with frequency (Table 2) Peak displacement calculated usingfundamental frequency and peak acceleration was significantlydifferent among bee species Peak displacement was higher forB ignitus and B audax and lower for B terrestris (Table 2Fig 4B)

Differences between bee taxa visiting two plant speciesWe found a statistically significant effect of both bee taxon(B audax versus B canariensis) and plant species (S rostratumversus S citrullifolium) on peak and RMS acceleration of floral

vibrations (Table 3 Fig 5) In contrast when we multiplied RMSacceleration by the coupling factor calculated for each Solanumspecies bee species still contributed significantly to explainingacceleration values but the effect of plant species on accelerationdisappeared Fundamental frequency was also significantlydifferent between bee taxa but not between plant species(Table 3 Fig 5) Neither bee size nor flower biomass wassignificantly associated with the acceleration of floral vibrations(Table 3) Unexpectedly bee size (ITD) was positively associatedwith the fundamental frequency of floral vibrations after statisticallyaccounting for differences in size between the two bumblebee taxa(Table 3) Peak displacement was significantly different betweenplant species but not between B audax and B terrestris Howeverthe difference in peak displacement between plants disappearedwhen accounting for the plantrsquos coupling factor (Table 3 Fig 6)

DISCUSSIONOur study represents one of the few available investigations on themechanical properties of floral-borne vibrations produced by beesduring buzz pollination and one of the first to systematicallycompare substrate-borne vibrations of multiple bee taxa on the sameflower and of the same bee taxon on different plant species Wefound that floral vibrations on the same plant species (S rostratum)differ between bee taxa in both fundamental frequency andacceleration amplitude (both RMS acceleration and peakacceleration) Interestingly although neither individual bee sizenor floral biomass explained variation in floral vibration propertiesB audax the bee taxon with the smallest average size producedfloral vibrations that combined both relatively high frequency andacceleration When measurements of frequency and accelerationwere used to calculate peak displacement (the maximum distancetravelled by the oscillating structure from its resting position)we found statistically significant differences between bee taxawith both B audax and B ignitus achieving larger displacements

Table 1 Sample size for experiments measuring floral vibrations

Bee taxon Colony Plant speciesNo ofrecordings

No ofbees

B terrestris sspaudax A3 S citrullifolium 22 10S rostratum 10 5

A6 S citrullifolium 23 10S rostratum 33 10

B terrestris sspcanariensis

C1 S citrullifolium 37 10S rostratum 42 10

C2 S rostratum 17 7B ignitus I1 S rostratum 5 2

I2 13 5B terrestris T1 2 1

T3 26 9Total 8 230 79

Bee taxon comprises Bombus terrestris ssp audax (B audax) Bombusterrestris ssp canariensis (B canariensis) Bombus ignitus and Bombusterrestris Plant species were Solanum citrullifolium and Solanum rostratumThe number of recordings represents the number of accelerometer recordingsthat were used in the final analysis The number of bees indicates thenumber of individuals used in each plant speciesbee colony combinationA few of the same individuals were recorded in both plant species and thusthe total number of different bees across the whole experiment was 72

Buzzers Allworkers

Buzzers Allworkers

Buzzers Allworkers

Buzzers Allworkers

3

4

5

Buzzing behaviour

ITD

(mm

) B audax

B canariensis

B terrestris

B ignitus

Fig 3 Size of bees observed buzzing during the experiment (lsquobuzzersrsquo) versus a random sample of workers in the colony (lsquoall workersrsquo) Size isgiven as intertegular distance (ITD) for all species studied B audax Bombus terrestris ssp canariensis (hereafter B canariensis) Bombus terrestris andBombus ignitus This analysis included only colonies in which both buzzers and all workers were observed and measured during the experiment (8 coloniestwo per taxon N=282 bees)

6

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

than B terrestris A comparison of floral vibrations produced byB audax and B canariensis visiting the same two Solanumplant species showed that acceleration (RMS and peak) and peakdisplacement measured at the base of the flower but not frequency

were significantly different between plants Comparison of theplant-specific coupling factors clearly indicated that flowers of evenclosely related species differ significantly in their capacity todampen substrate-borne vibrations We found that when these

Table 2 Effect of bee taxon bee size (intertegular distance ITD) and floral mass on the mechanical properties of floral vibrations produced by fourbee taxa visiting the same plant species (Solanum rostratum)

Parameter

Peak acceleration RMS accelerationRMS acceleration withflower coupling factor Fundamental frequency Peak displacement

Estimate(m sminus2)

se(m sminus2) P

Estimate(m sminus2)

se(m sminus2) P

Estimate(m sminus2)

se(m sminus2) P

Estimate(Hz)

se(Hz) P

Estimate(microm)

se(microm) P

Intercept(B audax)

1908 7689 1411 4953 12477 43788 421686 44406 minus0969 1729

ITD (mm) 1032 1597 0521 0938 1079 0389 8294 9538 0389 minus1768 8818 0842 0390 0345 0265Floral drymass (mg)

0392 0241 0106 0127 0139 0361 1125 1227 0361 minus2696 1482 0071 0109 0057 0058

Bee taxon 0015 0003 0003 lt00001 lt0001B canariensis minus5291 1556 minus4059 1064 minus35886 9409 minus40565 8460 minus0707 0332B terrestris minus5348 1995 minus4506 1355 minus39831 11976 minus3418 10992 minus1288 0430B ignitus minus4268 2016 minus4140 1382 minus36594 12219 minus93305 10990 0622 0431

Floral vibrations produced by four bumblebee taxa on flowers of buzz-pollinated Solanum rostratum were recorded with an accelerometer attached to thebase of the flower Peak acceleration root mean square (RMS) acceleration and fundamental frequency were calculated for a subset of randomly selected floralvibrations Parameter estimates and standard errors for individual coefficients were obtained from a linear mixed-effects model with bee individual as a random effectP-values were calculated for each explanatory variable using a Type III analysis of variance with Satterthwaitersquos method bold indicates statistical significance

a b a c

100

200

300

400

500

Fund

amen

tal f

requ

ency

(Hz)

A

a b b ab

0

10

20

30

Pea

k ac

cele

ratio

n (m

sndash2

)

B

a b b b

0

5

10

15

20

Ba Bc Bt Bi

RM

S a

ccel

erat

ion

(m s

ndash2)

C

ac ab b c

0

2

4

6

Ba Bc Bt Bi

Pea

k di

spla

cem

ent (microm

)

D

Bee taxon

Fig 4 Comparison of themechanical properties of floral vibrations from four taxa of bumblebees on flowers of buzz-pollinated S rostratumVibrationswere recorded using an accelerometer attached to the calyx of the flower using a metal pin (A) Fundamental frequency (B) Peak acceleration (C) RMSacceleration (D) Peak displacement calculated with RMS acceleration Within each panel different letters indicate statistically different mean valuesassessed by a Tukey test of pairwise comparisons with a confidence level of 95 N=148 floral vibrations from 49 bees from eight colonies of four taxaBa B audax Bc B canariensis Bi B ignitus Bt B terrestris

7

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

differences in transmission properties of flowers were taken intoaccount plant-specific differences in acceleration and displacementdisappeared Together our results suggest that both B audax andB canariensis produce vibrations at the anther level that achievesimilar acceleration amplitude and displacement irrespective of theplant they visit The main findings of our study can be summarisedin three points (1) Even closely related and morphologicallysimilar bumblebee taxa differ in the mechanical properties of thevibrations transmitted to flowers during buzz pollination (2) Floraltype (ie plant species) significantly influences the accelerationamplitude of vibrations transmitted over floral structures (3)Bombus audax and B canariensis two closely related subspecieswithin the B terrestris species aggregate produce approximatelysimilar acceleration and displacement while vibrating the anthers oftwo different buzz-pollinated plants

Differences in floral vibrations between closely related beesPrevious studies have shown that different species of bees producevibrations with different properties when sonicating flowers of thesame plant species (eg Burkart et al 2011 Corbet and Huang2014 King and Buchmann 2003) Most of these previous studieshave focused on the properties of the acoustic (airborne)component of floral vibrations particularly frequency in partbecause of the ease and accessibility of recording acousticvibrations (De Luca et al 2018) Differences in the frequencyof floral vibrations might be expected given the size variationamong buzz-pollinating bees It is well established thatwingbeat frequency in insects is correlated with morphologicalcharacteristics including wing length and body mass (Byrne et al1988) and similarly body mass is associated with frequency ininsects communicating using airborne sound (Cocroft and DeLuca 2006) However floral vibrations seem to be less stronglyassociated with specific morphological traits in bees such as bodysize (Burkart et al 2011 Nunes-Silva et al 2013 Rosi-Denadaiet al 2018 P A De Luca S L Buchman C Galen A Masonand MV-M unpublished)

Here we found that although different bee species producefloral vibrations with different properties body size (ITD) didnot consistently explain variation in either the frequency or theamplitude of substrate-borne vibrations In the first experimentcomparing multiple bee species on the same plant we found noassociation between bee size and frequency or amplitude In thesecond experiment comparing vibrations of two bee species inS rostratum and S citrullifolium we found that bee size (ITD)was positively correlated with frequency but not with amplitudeA positive association between body size and frequency iscontrary to the negative association usually found betweenbody size and wingbeat frequency (P A De Luca S LBuchman C Galen A Mason and MV-M unpublished) Thelack of a consistent relationship between body size and frequency isparalleled by previous studies of buzz-pollinating Bombusimpatiens (Nunes-Silva et al 2013 Switzer and Combes 2017)For instance Switzer and Combes (2017) found that higher(acoustic) frequency during floral vibrations was positivelyassociated with bee body mass but negatively associated withITD In a study of the characteristics of substrate-borne vibrationsused in insect communication Cocroft and De Luca (2006) found anegative relationship between body size and signal frequency across51 species in the Family Membracidae (Hemiptera) but nocorrelation when analysing within-population variation in twospecies of treehoppers Importantly they also found that theassociation between body size and signal frequency across speciesTa

ble3

Effec

tofb

eetaxo

nbe

esize

(ITD)plan

tspe

cies

andflo

ralm

asson

themec

hanica

lprope

rtiesof

floralv

ibratio

nsdu

ring

buzz

pollina

tionof

twoplan

tspe

cies

(Sros

tratum

andS

citrullifolium)v

isite

dby

twobe

etaxa

(Ba

udax

andBc

anariens

is)

Param

eter

Pea

kac

celeratio

nRMSac

celeratio

nRMSac

celeratio

ntimesflo

wer

coup

lingfactor

Fun

damen

talfrequ

ency

Pea

kdisp

lace

men

tPea

kdisp

lace

men

t(w

ithco

upling)

Dagger

Estim

ate

(msminus

2)

se

(msminus

2)

PEstim

ate

(msminus

2)

se

(msminus

2)

PEstim

ate

(msminus

2)

se

(msminus

2)

PEstim

ate

(Hz)

se

(Hz)

PEstim

ate

(microm)

se

(microm)

PEstim

ate

(microm)

se

(microm)

P

Intercep

t(Ba

udax

Sc

itrullifolium)

231

3714

0026

9496

6568

748

049

30022

744

813

097

6153

016

535

15185

ITD

(mm)

281

7162

3008

9195

8114

9009

420

852

11081

1006

525

047

986

1001

6034

1034

2032

5337

0339

1032

5Florald

rymas

s(m

g)minus026

6016

3010

3minus016

2010

0010

6minus161

6099

3010

5minus229

9119

1005

5minus003

3003

8038

5minus049

0038

0019

9Bee

taxo

n(Bc

anariens

is)

minus393

7124

5000

3minus317

0089

4lt000

1minus30

963

860

4lt000

01minus48

592

725

7lt000

01minus032

8025

9021

2minus296

0255

0025

3

Plant

spec

ies

(Sros

tratum)

308

7072

7lt000

01146

9046

2000

2minus056

9456

1090

1280

4501

7057

7064

9016

4lt000

1minus008

2164

4096

0

Floralvibratio

nsprod

uced

byfour

bumbleb

eetaxa

onflo

wersof

buzz-pollinated

Solan

umrostratum

werereco

rded

with

anac

celerometer

attach

edto

theba

seof

theflo

werP

eakac

celeratio

nroot

mea

nsq

uare

(RMS)ac

celeratio

nan

dfund

amen

talfrequ

ency

wereca

lculated

forasu

bset

ofrand

omlyse

lected

floralvibratio

nsP

aram

eter

estim

ates

andstan

dard

errors

forindividu

alco

efficientswereob

tained

from

alinea

rmixed

-effe

ctsmod

elwith

beeindividu

alas

arand

omeffectP

-value

swereca

lculated

fore

achex

plan

atoryva

riableus

ingaTyp

eIII

analysisof

varia

ncewith

Satterthw

aitersquosmetho

dbo

ldindica

tes

sign

ifica

nce

Inthis

analysisthe

resp

onse

varia

bleispe

akac

celeratio

nmultip

liedby

asp

ecies-sp

ecificflo

wer

coup

lingfactorF

lower

coup

lingfactorS

ros

tratum

=856

Sc

itrullifolium=11

30

DaggerHere

peak

disp

lace

men

twas

calculated

from

acce

leratio

nmultip

liedby

theco

uplingfactor

(see

Materials

andMetho

ds)

8

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

of different orders varied with the function of those signals (egalarm or mating) Our results can be interpreted as indicating thatafter accounting for the effect of bee species size variation withinspecies does not correlate with the amplitude of floral vibrationsalthough it may be positively associated with frequency in someplants This has some interesting implications for buzz pollinationas it suggests that within the range of size variation seen amongbuzz-pollinating worker bees of a given species different-sizedworkers are capable of producing similar floral vibrations Thiscould be achieved either passively (floral vibrations do not dependstrongly on bee traits related to size Nunes-Silva et al 2013) oractively (bees may be able to adjust their vibrations to achieve agiven outcome) Given that the properties of floral vibrations affectpollen removal (De Luca et al 2013) this should also influence theamount of pollen obtained by workers Buzz-pollinating bees arecapable of altering several aspects of their behaviour in response topollen availability including visit duration buzz number andduration and level of pollen grooming (Buchmann and Cane 1989Nunes-Silva et al 2013 Russell et al 2016) In contrast evidenceof whether bumblebees can adjust the mechanical properties ofbuzzes in response to pollen availability is mixed For example onthe one hand a study of B impatiens vibrating flowers of tomato(Solanum lycopersicum) found that frequency and velocityamplitude (m sminus1) do not differ in consecutive visits (wherepollen is presumably being depleted) or in comparisons between

flowers with full pollen loads and flowers in which pollen access hasbeen restricted by gluing the anther pore shut (Nunes-Silva et al2013) On the other hand Russell et al (2016) found that Bimpatiens produced significantly higher amplitudes for the acousticcomponent (dB) of floral vibrations when visiting flowers ofSolanum houstonii in which the anther pores were glued shut thanwhen visiting unmanipulated flowers More research is needed toestablish the morphological physiological and behaviouraldeterminants of floral vibrations and the question as to why evenclosely related species differ in their floral vibrations remainsunanswered

Do bees adjust their vibrations depending on the plantspeciesPrevious studies indicated that the characteristics of floral vibrationsof the same bee species on different plant species can vary(reviewed in Switzer and Combes 2017) For example in agreenhouse study of B impatiens visiting three species of Solanum(S carolinense S dulcamara and S lycopersicum) Switzer andCombes (2017) showed that the same individual bee can producefloral vibrations of different frequency and duration on differentplant species We found that the acceleration amplitude but notfundamental frequency experienced at the base of the flower whenbuzz pollinated by the same bee species varied depending on planttype In contrast to previous studies measuring the frequency of

250

300

350

400

450Fu

ndam

enta

l fre

quen

cy (H

z)A

0

10

20

30

Pea

k ac

cele

ratio

n (m

sndash2

)

B

0

5

10

15

20

S citrullifolium S rostratum

RM

S a

ccel

erat

ion

(m s

ndash2)

C

0

50

100

150

S citrullifolium S rostratum

RM

S a

ccel

erat

ion

coup

ling

fact

or (m

sndash2

) D

Plant species

Fig 5 Comparison of the mechanical properties of floral vibrations from two bumblebee taxa on flowers of two buzz-pollinated plants Data arefrom B audax (purple) and B canariensis (blue) on Solanum citrullifolium and S rostratum Vibrations were recorded using an accelerometer attached tothe calyx of the flower using a metal pin (A) Fundamental frequency (B) Peak acceleration (C) RMS acceleration (D) RMS acceleration multiplied bythe species-specific flower coupling factor (see Materials and Methods) N=184 floral vibrations from 53 bees from four colonies of two taxa

9

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

airborne vibrations during buzz pollination in the current study wefound that two species of Bombus produced substrate-bornevibrations of similar frequency while visiting two closely relatedspecies of Solanum In our study the difference in floral vibrationsamong plants was restricted to changes in vibration amplitudespecifically acceleration amplitudeThe observed difference in vibration characteristics detected at

the base of the flower of two plant species could have at least twoexplanations (Switzer and Combes 2017) First the bee may beactively adjusting its floral vibrations depending on the plantspecies Individual bees seem to be capable of modifying thefrequency of their floral vibrations on a given plant species (Morganet al 2016) and it has been suggested that differences in floralcharacteristics including morphology and pollen availability maycause bees to change the vibrations produced by the same bee ondifferent plants (Switzer and Combes 2017) Second differences inthe characteristics of the substrate-borne vibrations measured herecould be due to intrinsic differences in the mechanical properties ofthe flowers of the two species of Solanum studied Our experimentaldesign and analysis allowed us to calculate the mechanical changesto known vibrational signals as they travelled through the flower (thecoupling factor) The analysis of floral vibrations after accountingfor these plant species-specific differences indicate that a given beespecies maintains similar types of vibrations regardless of the plantbeing visited Thus our study suggests that the differences in floralvibrations observed here are due to differences in the mechanicalproperties of the flowers rather than to behavioural changes in theproduction of vibrations by bees It remains a distinct possibility thatthe observed lack of amplitude differences in floral vibrations ondifferent plants results from the fact that bees are vibrating flowersat or close to the maximum amplitude they can generate using theirthoracic musclesIn nature the coupling between the bee and the rest of the flower

is influenced by both the characteristics of the flower and the way inwhich a bee holds onto floral parts while buzz pollinating (King1993) Our experiment could only account for differences in thetransmission of floral vibrations due to the mechanical properties offlowers as the calibrated vibrometer used here cannot accommodate

subtle variation in how a bee holds a flower during buzz pollinationMore work is needed to determine how the way in which a beemanipulates a flower eg how lsquotightlyrsquo the bee holds the anthers orwhether a bee lsquobitesrsquo the anthers (eg Russell et al 2016) affectsthe transmission of vibrations Different floral morphologies maycause changes in the way in which a bee manipulates and vibratesa flower and therefore on the transmission of those vibrationsThe two species of Solanum studied here have similar floralmorphologies and were manipulated in the same manner bybumblebees but comparisons of floral species with more disparatemorphologies may reveal whether and to what extent changes inbee behaviour affect the transmission of floral vibrations includingvibration amplitude and potentially pollen release

Effect of floral properties on substrate-borne floral vibrationsThe field of animal communication using substrate-borne vibrations(biotremology) has long recognised the importance of consideringplant characteristics in the functional ecology and evolution ofvibrational communication (Cocroft and Rodriguez 2005Michelsen et al 1982 Mortimer 2017) In buzz pollination theimportance of plant traits in the transmission of vibrations isacknowledged (eg Buchmann and Hurley 1978 King 1993) buthas rarely been explicitly incorporated into mechanistic ecologicalor evolutionary studies

Characteristics of the floral structures with which bees interactduring buzz pollination eg the androecium should have an effecton the transmission of vibrational energy from the bee to the flower(King and Buchmann 1996) Bees often vibrate only some of theanthers on a flower but their vibrations can be transmitted to otherfloral parts including other stamens not directly in contact with thebee resulting in pollen release (MV-M personal observation)Biotremology studies show that plants can cause significantdistortion of insectsrsquo vibration signals (Cocroft and Rodriguez2005 Michelsen et al 1982) The characteristics of vegetative andflower tissues are therefore likely to cause alterations in thetransmission of vibrations (King and Buchmann 1996) Our resultssuggest that mass cannot explain differences between flowers fromtwo Solanum species when being vibrated by the same Bombus

2

4

6

S citrullifolium S rostratum

Pea

k di

spla

cem

ent (microm

)A

0

20

40

60

S citrullifolium S rostratum

Pea

k di

spla

cem

ent

coup

ling

fact

or (micro

m)

B

Plant species

Fig 6 Estimated peak displacement of floral vibrations produced by two bumblebee taxa while visiting flowers of two plant species Data arefrom B audax (purple) and B canariensis (blue) on S citrullifolium and S rostratum (A) Estimated peak displacement at the point of measurement ie thebase of the flower (B) Estimated peak displacement at the anthers obtained by multiplying peak acceleration by the plant species-specific coupling factorN=184 floral vibrations from 53 bees from four colonies of two taxa

10

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

species However we showed that S rostratum reduced theamplitude of the vibrations transmitted through the flower to alesser extent than S citrullifolium The transmission of vibrationsthrough plant tissue depends not only on the distance from thesource but also on the characteristics and vibration modes of theplant (Michelsen et al 1982) Even fine structural differences suchas the presence of vascular tissue and vein size can affect thedamping of vibrations (Michelsen et al 1982) Thus floralproperties such as the morphology geometry and materialproperties of the flower are expected to contribute to vibrationaldifferences For example the stiffness of the filament holding theanther and whether the anther cones are loosely arranged or tightlypacked into a cone (Glover et al 2004) are likely to affect thetransmission of vibrations away from its source (the bee) Thevibrational properties of flowers particularly in an evolutionarycontext are not well understood The convergence of disparate plantgroups onto similar buzz-pollinated floral morphologies (eg theSolanum-type flower has evolved in more than 21 plant familiesDe Luca and Vallejo-Mariacuten 2013) provides an opportunity toinvestigate whether morphological similarity is associated withsimilar vibrational properties

ConclusionsTogether with previous work our results suggest that bees differ intheir capacity to transmit vibrations to flowers Because thecharacteristics of vibrations in particular their amplitude areassociated with pollen release (De Luca et al 2013 Vallejo-Mariacuten2019) visitation by different species of bees may affect patterns ofpollen dispensing and ultimately plant fitness (Harder and Barclay1994 Harder and Wilson 1994) Therefore from a plantrsquosperspective it may be selectively advantageous to favour visitationby specific species Similarly from a beersquos perspective the choiceof a given plant species may determine the extent to which theirfloral vibrations translate into sufficient vibrational energy to elicitpollen release (Buchmann and Hurley 1978) In particular in caseswhere bees cannot adjust their floral vibrations depending on plantspecies selectively foraging in plants in which vibrations aretransmitted with low attenuation (eg with low coupling factors)may allow bees to maximise pollen release per buzz effortMoreover bee traits that maximise the beersquos transmission ofvibrations (eg the location and manner in which a bee holds aflower) should also be favoured as mechanisms to transmitvibrations more efficiently potentially removing more pollen

AcknowledgementsWe thank M Pozo and Biobest for providing the bumblebee colonies for thisexperiment A Russell kindly shared his protocols and advice on flight arenaconstruction and bee experiments in general J Gibson and P De Luca providedkey advice on the analysis and interpretation of substrate-borne vibrations and onbuzz pollination We thank J Weir for help with flight arena construction Membersof the Vallejo-Marin lab in particular Lucy Nevard provided help with beemaintenance

Competing interestsThe authors declare no competing or financial interests

Author contributionsConceptualization BA-C MV-M Methodology BA-C CEB MV-M Formalanalysis BA-C MV-M Investigation BA-C CEB MV-M ResourcesMV-M Writing - original draft BA-C MV-M Writing - review amp editingBA-C MV-M Visualization BA-C MV-M Supervision MV-M Projectadministration MV-M Funding acquisition BA-C MV-M

FundingThis study was partially funded by a SPARK grant (University of Stirling) to MV-Mand an ERASMUS internship to BA-C

Data availabilityData have been deposited in the Dryad digital repository (Arroyo-Correa et al2019) dryad3q0b1r1

Supplementary informationSupplementary information available online athttpjebbiologistsorglookupdoi101242jeb198176supplemental

ReferencesArroyo-Correa B Beattie C E andVallejo-MarinM (2019) Data from Bee and

floral traits affect the characteristics of the vibrations experienced by flowers duringbuzz-pollination Dryad Digital Repository

Bowers K A W (1975) The pollination ecology of Solanum rostratum(Solanaceae) Am J Bot 62 633ndash638

Buchmann S L (1983) Buzz pollination in angiosperms In Handbook ofExperimental Pollination Biology (ed C E Jones andR J Little) pp 73ndash113 NYScientific and Academic Editions

Buchmann S L and Cane J H (1989) Bees assess pollen returns whilesonicating Solanum flowers Oecologia 81 289ndash294

Buchmann S L and Hurley J P (1978) Biophysical model for buzz pollination inangiosperms J Theor Biol 72 639ndash657

Buchmann S L Jones C E and Colin L J (1977) Vibratile pollination ofSolanum douglasii and Solanum xanti (Solanaceae) in Southern California USAWasmann J Biol 35 1ndash25

Burkart A Lunau K and Schlindwein C (2011) Comparative bioacousticalstudies on flight and buzzing of neotropical bees J Pollin Ecol 6 118ndash124

Byrne D N Buchmann S L and Spangler H G (1988) Relationship betweenwing loading wingbeat frequency and body mass in homopterous insects J ExpBiol 135 9ndash23

Cameron S A Hines H M and Williams P H (2007) A comprehensivephylogeny of the bumble bees (Bombus) Biol J Linn Soc 91 161ndash188

Cane J H (1987) Estimation of bee size using intertegular span (Apoidea)J Kans Entomol Soc 60 145ndash147

Cardinal S Buchmann S L and Russell A L (2018) The evolution of floralsonication a pollen foraging behavior used by bees (Anthophila) Evolution 72590ndash600

Cocroft R B and De Luca P (2006) Size-frequency relationships in insectvibratory signals In Insect Sounds and Communication Physiology BehaviorEcology and Evolution (ed S Drosopoulos and M F Claridge) pp 99ndash110New York CRC

Cocroft R B and Rodriguez R L (2005) The behavioral ecology of insectvibrational communication Bioscience 55 323ndash334

Corbet S A and Huang S-Q (2014) Buzz pollination in eight bumblebee-pollinated Pedicularis species does it involve vibration-induced triboelectriccharging of pollen grains Ann Bot 114 1665ndash1674

De Luca P A and Vallejo-Marın M (2013) Whatrsquos the lsquobuzzrsquo about The ecologyand evolutionary significance of buzz-pollination Curr Opin Plant Biol 16429ndash435

De Luca P A Bussiere L F Souto-Vilaros D Goulson D Mason A C andVallejo-Marın M (2013) Variability in bumblebee pollination buzzes affects thequantity of pollen released from flowers Oecologia 172 805ndash816

De Luca P A Cox D A and Vallejo-Marın M (2014) Comparison of pollinationand defensive buzzes in bumblebees indicates species-specific and context-dependent vibrations Naturwissenschaften 101 331ndash338

De Luca P A Giebink N Mason A C Papaj D and Buchmann S L(2018) How well do acoustic recordings characterize properties of bee(Anthophila) floral sonication vibrations Bioacoustics Int J Anim SoundRecord 1ndash14

Gibson J S and Cocroft R B (2018) Vibration-guided mate searching intreehoppers directional accuracy and sampling strategies in a complex sensoryenvironment J Exp Biol 221 jeb175083

Glover B J Bunnewell S andMartin C (2004) Convergent evolution within thegenus Solanum the specialised anther cone develops through alternativepathways Gene 331 1ndash7

Harder L D and Barclay R M R (1994) The functional significance of poricidalanthers and buzz pollination controlled pollen removal from Dodecatheon FunctEcol 8 509ndash517

Harder L D and Wilson W G (1994) Floral evolution and male reproductivesuccess optimal dispensing schedules for pollen dispersal by animal-pollinatedplants Evol Ecol 8 542ndash559

Harris J A (1905) The dehiscence of anthers by apical pores Missouri BotGarden Annu Rep 1905 167ndash257

Kawai Y and Kudo G (2008) Effectiveness of buzz pollination inPedicularis chamissonis significance of multiple visits by bumblebees EcolRes 24 215

King M J (1993) Buzz foraging mechanism of bumble bees J Apic Res32 41ndash49

King M J and Buchmann S L (1996) Sonication dispensing of pollen fromSolanum laciniatum flowers Funct Ecol 10 449ndash456

11

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

than B terrestris A comparison of floral vibrations produced byB audax and B canariensis visiting the same two Solanumplant species showed that acceleration (RMS and peak) and peakdisplacement measured at the base of the flower but not frequency

were significantly different between plants Comparison of theplant-specific coupling factors clearly indicated that flowers of evenclosely related species differ significantly in their capacity todampen substrate-borne vibrations We found that when these

Table 2 Effect of bee taxon bee size (intertegular distance ITD) and floral mass on the mechanical properties of floral vibrations produced by fourbee taxa visiting the same plant species (Solanum rostratum)

Parameter

Peak acceleration RMS accelerationRMS acceleration withflower coupling factor Fundamental frequency Peak displacement

Estimate(m sminus2)

se(m sminus2) P

Estimate(m sminus2)

se(m sminus2) P

Estimate(m sminus2)

se(m sminus2) P

Estimate(Hz)

se(Hz) P

Estimate(microm)

se(microm) P

Intercept(B audax)

1908 7689 1411 4953 12477 43788 421686 44406 minus0969 1729

ITD (mm) 1032 1597 0521 0938 1079 0389 8294 9538 0389 minus1768 8818 0842 0390 0345 0265Floral drymass (mg)

0392 0241 0106 0127 0139 0361 1125 1227 0361 minus2696 1482 0071 0109 0057 0058

Bee taxon 0015 0003 0003 lt00001 lt0001B canariensis minus5291 1556 minus4059 1064 minus35886 9409 minus40565 8460 minus0707 0332B terrestris minus5348 1995 minus4506 1355 minus39831 11976 minus3418 10992 minus1288 0430B ignitus minus4268 2016 minus4140 1382 minus36594 12219 minus93305 10990 0622 0431

Floral vibrations produced by four bumblebee taxa on flowers of buzz-pollinated Solanum rostratum were recorded with an accelerometer attached to thebase of the flower Peak acceleration root mean square (RMS) acceleration and fundamental frequency were calculated for a subset of randomly selected floralvibrations Parameter estimates and standard errors for individual coefficients were obtained from a linear mixed-effects model with bee individual as a random effectP-values were calculated for each explanatory variable using a Type III analysis of variance with Satterthwaitersquos method bold indicates statistical significance

a b a c

100

200

300

400

500

Fund

amen

tal f

requ

ency

(Hz)

A

a b b ab

0

10

20

30

Pea

k ac

cele

ratio

n (m

sndash2

)

B

a b b b

0

5

10

15

20

Ba Bc Bt Bi

RM

S a

ccel

erat

ion

(m s

ndash2)

C

ac ab b c

0

2

4

6

Ba Bc Bt Bi

Pea

k di

spla

cem

ent (microm

)

D

Bee taxon

Fig 4 Comparison of themechanical properties of floral vibrations from four taxa of bumblebees on flowers of buzz-pollinated S rostratumVibrationswere recorded using an accelerometer attached to the calyx of the flower using a metal pin (A) Fundamental frequency (B) Peak acceleration (C) RMSacceleration (D) Peak displacement calculated with RMS acceleration Within each panel different letters indicate statistically different mean valuesassessed by a Tukey test of pairwise comparisons with a confidence level of 95 N=148 floral vibrations from 49 bees from eight colonies of four taxaBa B audax Bc B canariensis Bi B ignitus Bt B terrestris

7

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

differences in transmission properties of flowers were taken intoaccount plant-specific differences in acceleration and displacementdisappeared Together our results suggest that both B audax andB canariensis produce vibrations at the anther level that achievesimilar acceleration amplitude and displacement irrespective of theplant they visit The main findings of our study can be summarisedin three points (1) Even closely related and morphologicallysimilar bumblebee taxa differ in the mechanical properties of thevibrations transmitted to flowers during buzz pollination (2) Floraltype (ie plant species) significantly influences the accelerationamplitude of vibrations transmitted over floral structures (3)Bombus audax and B canariensis two closely related subspecieswithin the B terrestris species aggregate produce approximatelysimilar acceleration and displacement while vibrating the anthers oftwo different buzz-pollinated plants

Differences in floral vibrations between closely related beesPrevious studies have shown that different species of bees producevibrations with different properties when sonicating flowers of thesame plant species (eg Burkart et al 2011 Corbet and Huang2014 King and Buchmann 2003) Most of these previous studieshave focused on the properties of the acoustic (airborne)component of floral vibrations particularly frequency in partbecause of the ease and accessibility of recording acousticvibrations (De Luca et al 2018) Differences in the frequencyof floral vibrations might be expected given the size variationamong buzz-pollinating bees It is well established thatwingbeat frequency in insects is correlated with morphologicalcharacteristics including wing length and body mass (Byrne et al1988) and similarly body mass is associated with frequency ininsects communicating using airborne sound (Cocroft and DeLuca 2006) However floral vibrations seem to be less stronglyassociated with specific morphological traits in bees such as bodysize (Burkart et al 2011 Nunes-Silva et al 2013 Rosi-Denadaiet al 2018 P A De Luca S L Buchman C Galen A Masonand MV-M unpublished)

Here we found that although different bee species producefloral vibrations with different properties body size (ITD) didnot consistently explain variation in either the frequency or theamplitude of substrate-borne vibrations In the first experimentcomparing multiple bee species on the same plant we found noassociation between bee size and frequency or amplitude In thesecond experiment comparing vibrations of two bee species inS rostratum and S citrullifolium we found that bee size (ITD)was positively correlated with frequency but not with amplitudeA positive association between body size and frequency iscontrary to the negative association usually found betweenbody size and wingbeat frequency (P A De Luca S LBuchman C Galen A Mason and MV-M unpublished) Thelack of a consistent relationship between body size and frequency isparalleled by previous studies of buzz-pollinating Bombusimpatiens (Nunes-Silva et al 2013 Switzer and Combes 2017)For instance Switzer and Combes (2017) found that higher(acoustic) frequency during floral vibrations was positivelyassociated with bee body mass but negatively associated withITD In a study of the characteristics of substrate-borne vibrationsused in insect communication Cocroft and De Luca (2006) found anegative relationship between body size and signal frequency across51 species in the Family Membracidae (Hemiptera) but nocorrelation when analysing within-population variation in twospecies of treehoppers Importantly they also found that theassociation between body size and signal frequency across speciesTa

ble3

Effec

tofb

eetaxo

nbe

esize

(ITD)plan

tspe

cies

andflo

ralm

asson

themec

hanica

lprope

rtiesof

floralv

ibratio

nsdu

ring

buzz

pollina

tionof

twoplan

tspe

cies

(Sros

tratum

andS

citrullifolium)v

isite

dby

twobe

etaxa

(Ba

udax

andBc

anariens

is)

Param

eter

Pea

kac

celeratio

nRMSac

celeratio

nRMSac

celeratio

ntimesflo

wer

coup

lingfactor

Fun

damen

talfrequ

ency

Pea

kdisp

lace

men

tPea

kdisp

lace

men

t(w

ithco

upling)

Dagger

Estim

ate

(msminus

2)

se

(msminus

2)

PEstim

ate

(msminus

2)

se

(msminus

2)

PEstim

ate

(msminus

2)

se

(msminus

2)

PEstim

ate

(Hz)

se

(Hz)

PEstim

ate

(microm)

se

(microm)

PEstim

ate

(microm)

se

(microm)

P

Intercep

t(Ba

udax

Sc

itrullifolium)

231

3714

0026

9496

6568

748

049

30022

744

813

097

6153

016

535

15185

ITD

(mm)

281

7162

3008

9195

8114

9009

420

852

11081

1006

525

047

986

1001

6034

1034

2032

5337

0339

1032

5Florald

rymas

s(m

g)minus026

6016

3010

3minus016

2010

0010

6minus161

6099

3010

5minus229

9119

1005

5minus003

3003

8038

5minus049

0038

0019

9Bee

taxo

n(Bc

anariens

is)

minus393

7124

5000

3minus317

0089

4lt000

1minus30

963

860

4lt000

01minus48

592

725

7lt000

01minus032

8025

9021

2minus296

0255

0025

3

Plant

spec

ies

(Sros

tratum)

308

7072

7lt000

01146

9046

2000

2minus056

9456

1090

1280

4501

7057

7064

9016

4lt000

1minus008

2164

4096

0

Floralvibratio

nsprod

uced

byfour

bumbleb

eetaxa

onflo

wersof

buzz-pollinated

Solan

umrostratum

werereco

rded

with

anac

celerometer

attach

edto

theba

seof

theflo

werP

eakac

celeratio

nroot

mea

nsq

uare

(RMS)ac

celeratio

nan

dfund

amen

talfrequ

ency

wereca

lculated

forasu

bset

ofrand

omlyse

lected

floralvibratio

nsP

aram

eter

estim

ates

andstan

dard

errors

forindividu

alco

efficientswereob

tained

from

alinea

rmixed

-effe

ctsmod

elwith

beeindividu

alas

arand

omeffectP

-value

swereca

lculated

fore

achex

plan

atoryva

riableus

ingaTyp

eIII

analysisof

varia

ncewith

Satterthw

aitersquosmetho

dbo

ldindica

tes

sign

ifica

nce

Inthis

analysisthe

resp

onse

varia

bleispe

akac

celeratio

nmultip

liedby

asp

ecies-sp

ecificflo

wer

coup

lingfactorF

lower

coup

lingfactorS

ros

tratum

=856

Sc

itrullifolium=11

30

DaggerHere

peak

disp

lace

men

twas

calculated

from

acce

leratio

nmultip

liedby

theco

uplingfactor

(see

Materials

andMetho

ds)

8

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

of different orders varied with the function of those signals (egalarm or mating) Our results can be interpreted as indicating thatafter accounting for the effect of bee species size variation withinspecies does not correlate with the amplitude of floral vibrationsalthough it may be positively associated with frequency in someplants This has some interesting implications for buzz pollinationas it suggests that within the range of size variation seen amongbuzz-pollinating worker bees of a given species different-sizedworkers are capable of producing similar floral vibrations Thiscould be achieved either passively (floral vibrations do not dependstrongly on bee traits related to size Nunes-Silva et al 2013) oractively (bees may be able to adjust their vibrations to achieve agiven outcome) Given that the properties of floral vibrations affectpollen removal (De Luca et al 2013) this should also influence theamount of pollen obtained by workers Buzz-pollinating bees arecapable of altering several aspects of their behaviour in response topollen availability including visit duration buzz number andduration and level of pollen grooming (Buchmann and Cane 1989Nunes-Silva et al 2013 Russell et al 2016) In contrast evidenceof whether bumblebees can adjust the mechanical properties ofbuzzes in response to pollen availability is mixed For example onthe one hand a study of B impatiens vibrating flowers of tomato(Solanum lycopersicum) found that frequency and velocityamplitude (m sminus1) do not differ in consecutive visits (wherepollen is presumably being depleted) or in comparisons between

flowers with full pollen loads and flowers in which pollen access hasbeen restricted by gluing the anther pore shut (Nunes-Silva et al2013) On the other hand Russell et al (2016) found that Bimpatiens produced significantly higher amplitudes for the acousticcomponent (dB) of floral vibrations when visiting flowers ofSolanum houstonii in which the anther pores were glued shut thanwhen visiting unmanipulated flowers More research is needed toestablish the morphological physiological and behaviouraldeterminants of floral vibrations and the question as to why evenclosely related species differ in their floral vibrations remainsunanswered

Do bees adjust their vibrations depending on the plantspeciesPrevious studies indicated that the characteristics of floral vibrationsof the same bee species on different plant species can vary(reviewed in Switzer and Combes 2017) For example in agreenhouse study of B impatiens visiting three species of Solanum(S carolinense S dulcamara and S lycopersicum) Switzer andCombes (2017) showed that the same individual bee can producefloral vibrations of different frequency and duration on differentplant species We found that the acceleration amplitude but notfundamental frequency experienced at the base of the flower whenbuzz pollinated by the same bee species varied depending on planttype In contrast to previous studies measuring the frequency of

250

300

350

400

450Fu

ndam

enta

l fre

quen

cy (H

z)A

0

10

20

30

Pea

k ac

cele

ratio

n (m

sndash2

)

B

0

5

10

15

20

S citrullifolium S rostratum

RM

S a

ccel

erat

ion

(m s

ndash2)

C

0

50

100

150

S citrullifolium S rostratum

RM

S a

ccel

erat

ion

coup

ling

fact

or (m

sndash2

) D

Plant species

Fig 5 Comparison of the mechanical properties of floral vibrations from two bumblebee taxa on flowers of two buzz-pollinated plants Data arefrom B audax (purple) and B canariensis (blue) on Solanum citrullifolium and S rostratum Vibrations were recorded using an accelerometer attached tothe calyx of the flower using a metal pin (A) Fundamental frequency (B) Peak acceleration (C) RMS acceleration (D) RMS acceleration multiplied bythe species-specific flower coupling factor (see Materials and Methods) N=184 floral vibrations from 53 bees from four colonies of two taxa

9

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

airborne vibrations during buzz pollination in the current study wefound that two species of Bombus produced substrate-bornevibrations of similar frequency while visiting two closely relatedspecies of Solanum In our study the difference in floral vibrationsamong plants was restricted to changes in vibration amplitudespecifically acceleration amplitudeThe observed difference in vibration characteristics detected at

the base of the flower of two plant species could have at least twoexplanations (Switzer and Combes 2017) First the bee may beactively adjusting its floral vibrations depending on the plantspecies Individual bees seem to be capable of modifying thefrequency of their floral vibrations on a given plant species (Morganet al 2016) and it has been suggested that differences in floralcharacteristics including morphology and pollen availability maycause bees to change the vibrations produced by the same bee ondifferent plants (Switzer and Combes 2017) Second differences inthe characteristics of the substrate-borne vibrations measured herecould be due to intrinsic differences in the mechanical properties ofthe flowers of the two species of Solanum studied Our experimentaldesign and analysis allowed us to calculate the mechanical changesto known vibrational signals as they travelled through the flower (thecoupling factor) The analysis of floral vibrations after accountingfor these plant species-specific differences indicate that a given beespecies maintains similar types of vibrations regardless of the plantbeing visited Thus our study suggests that the differences in floralvibrations observed here are due to differences in the mechanicalproperties of the flowers rather than to behavioural changes in theproduction of vibrations by bees It remains a distinct possibility thatthe observed lack of amplitude differences in floral vibrations ondifferent plants results from the fact that bees are vibrating flowersat or close to the maximum amplitude they can generate using theirthoracic musclesIn nature the coupling between the bee and the rest of the flower

is influenced by both the characteristics of the flower and the way inwhich a bee holds onto floral parts while buzz pollinating (King1993) Our experiment could only account for differences in thetransmission of floral vibrations due to the mechanical properties offlowers as the calibrated vibrometer used here cannot accommodate

subtle variation in how a bee holds a flower during buzz pollinationMore work is needed to determine how the way in which a beemanipulates a flower eg how lsquotightlyrsquo the bee holds the anthers orwhether a bee lsquobitesrsquo the anthers (eg Russell et al 2016) affectsthe transmission of vibrations Different floral morphologies maycause changes in the way in which a bee manipulates and vibratesa flower and therefore on the transmission of those vibrationsThe two species of Solanum studied here have similar floralmorphologies and were manipulated in the same manner bybumblebees but comparisons of floral species with more disparatemorphologies may reveal whether and to what extent changes inbee behaviour affect the transmission of floral vibrations includingvibration amplitude and potentially pollen release

Effect of floral properties on substrate-borne floral vibrationsThe field of animal communication using substrate-borne vibrations(biotremology) has long recognised the importance of consideringplant characteristics in the functional ecology and evolution ofvibrational communication (Cocroft and Rodriguez 2005Michelsen et al 1982 Mortimer 2017) In buzz pollination theimportance of plant traits in the transmission of vibrations isacknowledged (eg Buchmann and Hurley 1978 King 1993) buthas rarely been explicitly incorporated into mechanistic ecologicalor evolutionary studies

Characteristics of the floral structures with which bees interactduring buzz pollination eg the androecium should have an effecton the transmission of vibrational energy from the bee to the flower(King and Buchmann 1996) Bees often vibrate only some of theanthers on a flower but their vibrations can be transmitted to otherfloral parts including other stamens not directly in contact with thebee resulting in pollen release (MV-M personal observation)Biotremology studies show that plants can cause significantdistortion of insectsrsquo vibration signals (Cocroft and Rodriguez2005 Michelsen et al 1982) The characteristics of vegetative andflower tissues are therefore likely to cause alterations in thetransmission of vibrations (King and Buchmann 1996) Our resultssuggest that mass cannot explain differences between flowers fromtwo Solanum species when being vibrated by the same Bombus

2

4

6

S citrullifolium S rostratum

Pea

k di

spla

cem

ent (microm

)A

0

20

40

60

S citrullifolium S rostratum

Pea

k di

spla

cem

ent

coup

ling

fact

or (micro

m)

B

Plant species

Fig 6 Estimated peak displacement of floral vibrations produced by two bumblebee taxa while visiting flowers of two plant species Data arefrom B audax (purple) and B canariensis (blue) on S citrullifolium and S rostratum (A) Estimated peak displacement at the point of measurement ie thebase of the flower (B) Estimated peak displacement at the anthers obtained by multiplying peak acceleration by the plant species-specific coupling factorN=184 floral vibrations from 53 bees from four colonies of two taxa

10

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

species However we showed that S rostratum reduced theamplitude of the vibrations transmitted through the flower to alesser extent than S citrullifolium The transmission of vibrationsthrough plant tissue depends not only on the distance from thesource but also on the characteristics and vibration modes of theplant (Michelsen et al 1982) Even fine structural differences suchas the presence of vascular tissue and vein size can affect thedamping of vibrations (Michelsen et al 1982) Thus floralproperties such as the morphology geometry and materialproperties of the flower are expected to contribute to vibrationaldifferences For example the stiffness of the filament holding theanther and whether the anther cones are loosely arranged or tightlypacked into a cone (Glover et al 2004) are likely to affect thetransmission of vibrations away from its source (the bee) Thevibrational properties of flowers particularly in an evolutionarycontext are not well understood The convergence of disparate plantgroups onto similar buzz-pollinated floral morphologies (eg theSolanum-type flower has evolved in more than 21 plant familiesDe Luca and Vallejo-Mariacuten 2013) provides an opportunity toinvestigate whether morphological similarity is associated withsimilar vibrational properties

ConclusionsTogether with previous work our results suggest that bees differ intheir capacity to transmit vibrations to flowers Because thecharacteristics of vibrations in particular their amplitude areassociated with pollen release (De Luca et al 2013 Vallejo-Mariacuten2019) visitation by different species of bees may affect patterns ofpollen dispensing and ultimately plant fitness (Harder and Barclay1994 Harder and Wilson 1994) Therefore from a plantrsquosperspective it may be selectively advantageous to favour visitationby specific species Similarly from a beersquos perspective the choiceof a given plant species may determine the extent to which theirfloral vibrations translate into sufficient vibrational energy to elicitpollen release (Buchmann and Hurley 1978) In particular in caseswhere bees cannot adjust their floral vibrations depending on plantspecies selectively foraging in plants in which vibrations aretransmitted with low attenuation (eg with low coupling factors)may allow bees to maximise pollen release per buzz effortMoreover bee traits that maximise the beersquos transmission ofvibrations (eg the location and manner in which a bee holds aflower) should also be favoured as mechanisms to transmitvibrations more efficiently potentially removing more pollen

AcknowledgementsWe thank M Pozo and Biobest for providing the bumblebee colonies for thisexperiment A Russell kindly shared his protocols and advice on flight arenaconstruction and bee experiments in general J Gibson and P De Luca providedkey advice on the analysis and interpretation of substrate-borne vibrations and onbuzz pollination We thank J Weir for help with flight arena construction Membersof the Vallejo-Marin lab in particular Lucy Nevard provided help with beemaintenance

Competing interestsThe authors declare no competing or financial interests

Author contributionsConceptualization BA-C MV-M Methodology BA-C CEB MV-M Formalanalysis BA-C MV-M Investigation BA-C CEB MV-M ResourcesMV-M Writing - original draft BA-C MV-M Writing - review amp editingBA-C MV-M Visualization BA-C MV-M Supervision MV-M Projectadministration MV-M Funding acquisition BA-C MV-M

FundingThis study was partially funded by a SPARK grant (University of Stirling) to MV-Mand an ERASMUS internship to BA-C

Data availabilityData have been deposited in the Dryad digital repository (Arroyo-Correa et al2019) dryad3q0b1r1

Supplementary informationSupplementary information available online athttpjebbiologistsorglookupdoi101242jeb198176supplemental

ReferencesArroyo-Correa B Beattie C E andVallejo-MarinM (2019) Data from Bee and

floral traits affect the characteristics of the vibrations experienced by flowers duringbuzz-pollination Dryad Digital Repository

Bowers K A W (1975) The pollination ecology of Solanum rostratum(Solanaceae) Am J Bot 62 633ndash638

Buchmann S L (1983) Buzz pollination in angiosperms In Handbook ofExperimental Pollination Biology (ed C E Jones andR J Little) pp 73ndash113 NYScientific and Academic Editions

Buchmann S L and Cane J H (1989) Bees assess pollen returns whilesonicating Solanum flowers Oecologia 81 289ndash294

Buchmann S L and Hurley J P (1978) Biophysical model for buzz pollination inangiosperms J Theor Biol 72 639ndash657

Buchmann S L Jones C E and Colin L J (1977) Vibratile pollination ofSolanum douglasii and Solanum xanti (Solanaceae) in Southern California USAWasmann J Biol 35 1ndash25

Burkart A Lunau K and Schlindwein C (2011) Comparative bioacousticalstudies on flight and buzzing of neotropical bees J Pollin Ecol 6 118ndash124

Byrne D N Buchmann S L and Spangler H G (1988) Relationship betweenwing loading wingbeat frequency and body mass in homopterous insects J ExpBiol 135 9ndash23

Cameron S A Hines H M and Williams P H (2007) A comprehensivephylogeny of the bumble bees (Bombus) Biol J Linn Soc 91 161ndash188

Cane J H (1987) Estimation of bee size using intertegular span (Apoidea)J Kans Entomol Soc 60 145ndash147

Cardinal S Buchmann S L and Russell A L (2018) The evolution of floralsonication a pollen foraging behavior used by bees (Anthophila) Evolution 72590ndash600

Cocroft R B and De Luca P (2006) Size-frequency relationships in insectvibratory signals In Insect Sounds and Communication Physiology BehaviorEcology and Evolution (ed S Drosopoulos and M F Claridge) pp 99ndash110New York CRC

Cocroft R B and Rodriguez R L (2005) The behavioral ecology of insectvibrational communication Bioscience 55 323ndash334

Corbet S A and Huang S-Q (2014) Buzz pollination in eight bumblebee-pollinated Pedicularis species does it involve vibration-induced triboelectriccharging of pollen grains Ann Bot 114 1665ndash1674

De Luca P A and Vallejo-Marın M (2013) Whatrsquos the lsquobuzzrsquo about The ecologyand evolutionary significance of buzz-pollination Curr Opin Plant Biol 16429ndash435

De Luca P A Bussiere L F Souto-Vilaros D Goulson D Mason A C andVallejo-Marın M (2013) Variability in bumblebee pollination buzzes affects thequantity of pollen released from flowers Oecologia 172 805ndash816

De Luca P A Cox D A and Vallejo-Marın M (2014) Comparison of pollinationand defensive buzzes in bumblebees indicates species-specific and context-dependent vibrations Naturwissenschaften 101 331ndash338

De Luca P A Giebink N Mason A C Papaj D and Buchmann S L(2018) How well do acoustic recordings characterize properties of bee(Anthophila) floral sonication vibrations Bioacoustics Int J Anim SoundRecord 1ndash14

Gibson J S and Cocroft R B (2018) Vibration-guided mate searching intreehoppers directional accuracy and sampling strategies in a complex sensoryenvironment J Exp Biol 221 jeb175083

Glover B J Bunnewell S andMartin C (2004) Convergent evolution within thegenus Solanum the specialised anther cone develops through alternativepathways Gene 331 1ndash7

Harder L D and Barclay R M R (1994) The functional significance of poricidalanthers and buzz pollination controlled pollen removal from Dodecatheon FunctEcol 8 509ndash517

Harder L D and Wilson W G (1994) Floral evolution and male reproductivesuccess optimal dispensing schedules for pollen dispersal by animal-pollinatedplants Evol Ecol 8 542ndash559

Harris J A (1905) The dehiscence of anthers by apical pores Missouri BotGarden Annu Rep 1905 167ndash257

Kawai Y and Kudo G (2008) Effectiveness of buzz pollination inPedicularis chamissonis significance of multiple visits by bumblebees EcolRes 24 215

King M J (1993) Buzz foraging mechanism of bumble bees J Apic Res32 41ndash49

King M J and Buchmann S L (1996) Sonication dispensing of pollen fromSolanum laciniatum flowers Funct Ecol 10 449ndash456

11

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

differences in transmission properties of flowers were taken intoaccount plant-specific differences in acceleration and displacementdisappeared Together our results suggest that both B audax andB canariensis produce vibrations at the anther level that achievesimilar acceleration amplitude and displacement irrespective of theplant they visit The main findings of our study can be summarisedin three points (1) Even closely related and morphologicallysimilar bumblebee taxa differ in the mechanical properties of thevibrations transmitted to flowers during buzz pollination (2) Floraltype (ie plant species) significantly influences the accelerationamplitude of vibrations transmitted over floral structures (3)Bombus audax and B canariensis two closely related subspecieswithin the B terrestris species aggregate produce approximatelysimilar acceleration and displacement while vibrating the anthers oftwo different buzz-pollinated plants

Differences in floral vibrations between closely related beesPrevious studies have shown that different species of bees producevibrations with different properties when sonicating flowers of thesame plant species (eg Burkart et al 2011 Corbet and Huang2014 King and Buchmann 2003) Most of these previous studieshave focused on the properties of the acoustic (airborne)component of floral vibrations particularly frequency in partbecause of the ease and accessibility of recording acousticvibrations (De Luca et al 2018) Differences in the frequencyof floral vibrations might be expected given the size variationamong buzz-pollinating bees It is well established thatwingbeat frequency in insects is correlated with morphologicalcharacteristics including wing length and body mass (Byrne et al1988) and similarly body mass is associated with frequency ininsects communicating using airborne sound (Cocroft and DeLuca 2006) However floral vibrations seem to be less stronglyassociated with specific morphological traits in bees such as bodysize (Burkart et al 2011 Nunes-Silva et al 2013 Rosi-Denadaiet al 2018 P A De Luca S L Buchman C Galen A Masonand MV-M unpublished)

Here we found that although different bee species producefloral vibrations with different properties body size (ITD) didnot consistently explain variation in either the frequency or theamplitude of substrate-borne vibrations In the first experimentcomparing multiple bee species on the same plant we found noassociation between bee size and frequency or amplitude In thesecond experiment comparing vibrations of two bee species inS rostratum and S citrullifolium we found that bee size (ITD)was positively correlated with frequency but not with amplitudeA positive association between body size and frequency iscontrary to the negative association usually found betweenbody size and wingbeat frequency (P A De Luca S LBuchman C Galen A Mason and MV-M unpublished) Thelack of a consistent relationship between body size and frequency isparalleled by previous studies of buzz-pollinating Bombusimpatiens (Nunes-Silva et al 2013 Switzer and Combes 2017)For instance Switzer and Combes (2017) found that higher(acoustic) frequency during floral vibrations was positivelyassociated with bee body mass but negatively associated withITD In a study of the characteristics of substrate-borne vibrationsused in insect communication Cocroft and De Luca (2006) found anegative relationship between body size and signal frequency across51 species in the Family Membracidae (Hemiptera) but nocorrelation when analysing within-population variation in twospecies of treehoppers Importantly they also found that theassociation between body size and signal frequency across speciesTa

ble3

Effec

tofb

eetaxo

nbe

esize

(ITD)plan

tspe

cies

andflo

ralm

asson

themec

hanica

lprope

rtiesof

floralv

ibratio

nsdu

ring

buzz

pollina

tionof

twoplan

tspe

cies

(Sros

tratum

andS

citrullifolium)v

isite

dby

twobe

etaxa

(Ba

udax

andBc

anariens

is)

Param

eter

Pea

kac

celeratio

nRMSac

celeratio

nRMSac

celeratio

ntimesflo

wer

coup

lingfactor

Fun

damen

talfrequ

ency

Pea

kdisp

lace

men

tPea

kdisp

lace

men

t(w

ithco

upling)

Dagger

Estim

ate

(msminus

2)

se

(msminus

2)

PEstim

ate

(msminus

2)

se

(msminus

2)

PEstim

ate

(msminus

2)

se

(msminus

2)

PEstim

ate

(Hz)

se

(Hz)

PEstim

ate

(microm)

se

(microm)

PEstim

ate

(microm)

se

(microm)

P

Intercep

t(Ba

udax

Sc

itrullifolium)

231

3714

0026

9496

6568

748

049

30022

744

813

097

6153

016

535

15185

ITD

(mm)

281

7162

3008

9195

8114

9009

420

852

11081

1006

525

047

986

1001

6034

1034

2032

5337

0339

1032

5Florald

rymas

s(m

g)minus026

6016

3010

3minus016

2010

0010

6minus161

6099

3010

5minus229

9119

1005

5minus003

3003

8038

5minus049

0038

0019

9Bee

taxo

n(Bc

anariens

is)

minus393

7124

5000

3minus317

0089

4lt000

1minus30

963

860

4lt000

01minus48

592

725

7lt000

01minus032

8025

9021

2minus296

0255

0025

3

Plant

spec

ies

(Sros

tratum)

308

7072

7lt000

01146

9046

2000

2minus056

9456

1090

1280

4501

7057

7064

9016

4lt000

1minus008

2164

4096

0

Floralvibratio

nsprod

uced

byfour

bumbleb

eetaxa

onflo

wersof

buzz-pollinated

Solan

umrostratum

werereco

rded

with

anac

celerometer

attach

edto

theba

seof

theflo

werP

eakac

celeratio

nroot

mea

nsq

uare

(RMS)ac

celeratio

nan

dfund

amen

talfrequ

ency

wereca

lculated

forasu

bset

ofrand

omlyse

lected

floralvibratio

nsP

aram

eter

estim

ates

andstan

dard

errors

forindividu

alco

efficientswereob

tained

from

alinea

rmixed

-effe

ctsmod

elwith

beeindividu

alas

arand

omeffectP

-value

swereca

lculated

fore

achex

plan

atoryva

riableus

ingaTyp

eIII

analysisof

varia

ncewith

Satterthw

aitersquosmetho

dbo

ldindica

tes

sign

ifica

nce

Inthis

analysisthe

resp

onse

varia

bleispe

akac

celeratio

nmultip

liedby

asp

ecies-sp

ecificflo

wer

coup

lingfactorF

lower

coup

lingfactorS

ros

tratum

=856

Sc

itrullifolium=11

30

DaggerHere

peak

disp

lace

men

twas

calculated

from

acce

leratio

nmultip

liedby

theco

uplingfactor

(see

Materials

andMetho

ds)

8

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

of different orders varied with the function of those signals (egalarm or mating) Our results can be interpreted as indicating thatafter accounting for the effect of bee species size variation withinspecies does not correlate with the amplitude of floral vibrationsalthough it may be positively associated with frequency in someplants This has some interesting implications for buzz pollinationas it suggests that within the range of size variation seen amongbuzz-pollinating worker bees of a given species different-sizedworkers are capable of producing similar floral vibrations Thiscould be achieved either passively (floral vibrations do not dependstrongly on bee traits related to size Nunes-Silva et al 2013) oractively (bees may be able to adjust their vibrations to achieve agiven outcome) Given that the properties of floral vibrations affectpollen removal (De Luca et al 2013) this should also influence theamount of pollen obtained by workers Buzz-pollinating bees arecapable of altering several aspects of their behaviour in response topollen availability including visit duration buzz number andduration and level of pollen grooming (Buchmann and Cane 1989Nunes-Silva et al 2013 Russell et al 2016) In contrast evidenceof whether bumblebees can adjust the mechanical properties ofbuzzes in response to pollen availability is mixed For example onthe one hand a study of B impatiens vibrating flowers of tomato(Solanum lycopersicum) found that frequency and velocityamplitude (m sminus1) do not differ in consecutive visits (wherepollen is presumably being depleted) or in comparisons between

flowers with full pollen loads and flowers in which pollen access hasbeen restricted by gluing the anther pore shut (Nunes-Silva et al2013) On the other hand Russell et al (2016) found that Bimpatiens produced significantly higher amplitudes for the acousticcomponent (dB) of floral vibrations when visiting flowers ofSolanum houstonii in which the anther pores were glued shut thanwhen visiting unmanipulated flowers More research is needed toestablish the morphological physiological and behaviouraldeterminants of floral vibrations and the question as to why evenclosely related species differ in their floral vibrations remainsunanswered

Do bees adjust their vibrations depending on the plantspeciesPrevious studies indicated that the characteristics of floral vibrationsof the same bee species on different plant species can vary(reviewed in Switzer and Combes 2017) For example in agreenhouse study of B impatiens visiting three species of Solanum(S carolinense S dulcamara and S lycopersicum) Switzer andCombes (2017) showed that the same individual bee can producefloral vibrations of different frequency and duration on differentplant species We found that the acceleration amplitude but notfundamental frequency experienced at the base of the flower whenbuzz pollinated by the same bee species varied depending on planttype In contrast to previous studies measuring the frequency of

250

300

350

400

450Fu

ndam

enta

l fre

quen

cy (H

z)A

0

10

20

30

Pea

k ac

cele

ratio

n (m

sndash2

)

B

0

5

10

15

20

S citrullifolium S rostratum

RM

S a

ccel

erat

ion

(m s

ndash2)

C

0

50

100

150

S citrullifolium S rostratum

RM

S a

ccel

erat

ion

coup

ling

fact

or (m

sndash2

) D

Plant species

Fig 5 Comparison of the mechanical properties of floral vibrations from two bumblebee taxa on flowers of two buzz-pollinated plants Data arefrom B audax (purple) and B canariensis (blue) on Solanum citrullifolium and S rostratum Vibrations were recorded using an accelerometer attached tothe calyx of the flower using a metal pin (A) Fundamental frequency (B) Peak acceleration (C) RMS acceleration (D) RMS acceleration multiplied bythe species-specific flower coupling factor (see Materials and Methods) N=184 floral vibrations from 53 bees from four colonies of two taxa

9

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

airborne vibrations during buzz pollination in the current study wefound that two species of Bombus produced substrate-bornevibrations of similar frequency while visiting two closely relatedspecies of Solanum In our study the difference in floral vibrationsamong plants was restricted to changes in vibration amplitudespecifically acceleration amplitudeThe observed difference in vibration characteristics detected at

the base of the flower of two plant species could have at least twoexplanations (Switzer and Combes 2017) First the bee may beactively adjusting its floral vibrations depending on the plantspecies Individual bees seem to be capable of modifying thefrequency of their floral vibrations on a given plant species (Morganet al 2016) and it has been suggested that differences in floralcharacteristics including morphology and pollen availability maycause bees to change the vibrations produced by the same bee ondifferent plants (Switzer and Combes 2017) Second differences inthe characteristics of the substrate-borne vibrations measured herecould be due to intrinsic differences in the mechanical properties ofthe flowers of the two species of Solanum studied Our experimentaldesign and analysis allowed us to calculate the mechanical changesto known vibrational signals as they travelled through the flower (thecoupling factor) The analysis of floral vibrations after accountingfor these plant species-specific differences indicate that a given beespecies maintains similar types of vibrations regardless of the plantbeing visited Thus our study suggests that the differences in floralvibrations observed here are due to differences in the mechanicalproperties of the flowers rather than to behavioural changes in theproduction of vibrations by bees It remains a distinct possibility thatthe observed lack of amplitude differences in floral vibrations ondifferent plants results from the fact that bees are vibrating flowersat or close to the maximum amplitude they can generate using theirthoracic musclesIn nature the coupling between the bee and the rest of the flower

is influenced by both the characteristics of the flower and the way inwhich a bee holds onto floral parts while buzz pollinating (King1993) Our experiment could only account for differences in thetransmission of floral vibrations due to the mechanical properties offlowers as the calibrated vibrometer used here cannot accommodate

subtle variation in how a bee holds a flower during buzz pollinationMore work is needed to determine how the way in which a beemanipulates a flower eg how lsquotightlyrsquo the bee holds the anthers orwhether a bee lsquobitesrsquo the anthers (eg Russell et al 2016) affectsthe transmission of vibrations Different floral morphologies maycause changes in the way in which a bee manipulates and vibratesa flower and therefore on the transmission of those vibrationsThe two species of Solanum studied here have similar floralmorphologies and were manipulated in the same manner bybumblebees but comparisons of floral species with more disparatemorphologies may reveal whether and to what extent changes inbee behaviour affect the transmission of floral vibrations includingvibration amplitude and potentially pollen release

Effect of floral properties on substrate-borne floral vibrationsThe field of animal communication using substrate-borne vibrations(biotremology) has long recognised the importance of consideringplant characteristics in the functional ecology and evolution ofvibrational communication (Cocroft and Rodriguez 2005Michelsen et al 1982 Mortimer 2017) In buzz pollination theimportance of plant traits in the transmission of vibrations isacknowledged (eg Buchmann and Hurley 1978 King 1993) buthas rarely been explicitly incorporated into mechanistic ecologicalor evolutionary studies

Characteristics of the floral structures with which bees interactduring buzz pollination eg the androecium should have an effecton the transmission of vibrational energy from the bee to the flower(King and Buchmann 1996) Bees often vibrate only some of theanthers on a flower but their vibrations can be transmitted to otherfloral parts including other stamens not directly in contact with thebee resulting in pollen release (MV-M personal observation)Biotremology studies show that plants can cause significantdistortion of insectsrsquo vibration signals (Cocroft and Rodriguez2005 Michelsen et al 1982) The characteristics of vegetative andflower tissues are therefore likely to cause alterations in thetransmission of vibrations (King and Buchmann 1996) Our resultssuggest that mass cannot explain differences between flowers fromtwo Solanum species when being vibrated by the same Bombus

2

4

6

S citrullifolium S rostratum

Pea

k di

spla

cem

ent (microm

)A

0

20

40

60

S citrullifolium S rostratum

Pea

k di

spla

cem

ent

coup

ling

fact

or (micro

m)

B

Plant species

Fig 6 Estimated peak displacement of floral vibrations produced by two bumblebee taxa while visiting flowers of two plant species Data arefrom B audax (purple) and B canariensis (blue) on S citrullifolium and S rostratum (A) Estimated peak displacement at the point of measurement ie thebase of the flower (B) Estimated peak displacement at the anthers obtained by multiplying peak acceleration by the plant species-specific coupling factorN=184 floral vibrations from 53 bees from four colonies of two taxa

10

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

species However we showed that S rostratum reduced theamplitude of the vibrations transmitted through the flower to alesser extent than S citrullifolium The transmission of vibrationsthrough plant tissue depends not only on the distance from thesource but also on the characteristics and vibration modes of theplant (Michelsen et al 1982) Even fine structural differences suchas the presence of vascular tissue and vein size can affect thedamping of vibrations (Michelsen et al 1982) Thus floralproperties such as the morphology geometry and materialproperties of the flower are expected to contribute to vibrationaldifferences For example the stiffness of the filament holding theanther and whether the anther cones are loosely arranged or tightlypacked into a cone (Glover et al 2004) are likely to affect thetransmission of vibrations away from its source (the bee) Thevibrational properties of flowers particularly in an evolutionarycontext are not well understood The convergence of disparate plantgroups onto similar buzz-pollinated floral morphologies (eg theSolanum-type flower has evolved in more than 21 plant familiesDe Luca and Vallejo-Mariacuten 2013) provides an opportunity toinvestigate whether morphological similarity is associated withsimilar vibrational properties

ConclusionsTogether with previous work our results suggest that bees differ intheir capacity to transmit vibrations to flowers Because thecharacteristics of vibrations in particular their amplitude areassociated with pollen release (De Luca et al 2013 Vallejo-Mariacuten2019) visitation by different species of bees may affect patterns ofpollen dispensing and ultimately plant fitness (Harder and Barclay1994 Harder and Wilson 1994) Therefore from a plantrsquosperspective it may be selectively advantageous to favour visitationby specific species Similarly from a beersquos perspective the choiceof a given plant species may determine the extent to which theirfloral vibrations translate into sufficient vibrational energy to elicitpollen release (Buchmann and Hurley 1978) In particular in caseswhere bees cannot adjust their floral vibrations depending on plantspecies selectively foraging in plants in which vibrations aretransmitted with low attenuation (eg with low coupling factors)may allow bees to maximise pollen release per buzz effortMoreover bee traits that maximise the beersquos transmission ofvibrations (eg the location and manner in which a bee holds aflower) should also be favoured as mechanisms to transmitvibrations more efficiently potentially removing more pollen

AcknowledgementsWe thank M Pozo and Biobest for providing the bumblebee colonies for thisexperiment A Russell kindly shared his protocols and advice on flight arenaconstruction and bee experiments in general J Gibson and P De Luca providedkey advice on the analysis and interpretation of substrate-borne vibrations and onbuzz pollination We thank J Weir for help with flight arena construction Membersof the Vallejo-Marin lab in particular Lucy Nevard provided help with beemaintenance

Competing interestsThe authors declare no competing or financial interests

Author contributionsConceptualization BA-C MV-M Methodology BA-C CEB MV-M Formalanalysis BA-C MV-M Investigation BA-C CEB MV-M ResourcesMV-M Writing - original draft BA-C MV-M Writing - review amp editingBA-C MV-M Visualization BA-C MV-M Supervision MV-M Projectadministration MV-M Funding acquisition BA-C MV-M

FundingThis study was partially funded by a SPARK grant (University of Stirling) to MV-Mand an ERASMUS internship to BA-C

Data availabilityData have been deposited in the Dryad digital repository (Arroyo-Correa et al2019) dryad3q0b1r1

Supplementary informationSupplementary information available online athttpjebbiologistsorglookupdoi101242jeb198176supplemental

ReferencesArroyo-Correa B Beattie C E andVallejo-MarinM (2019) Data from Bee and

floral traits affect the characteristics of the vibrations experienced by flowers duringbuzz-pollination Dryad Digital Repository

Bowers K A W (1975) The pollination ecology of Solanum rostratum(Solanaceae) Am J Bot 62 633ndash638

Buchmann S L (1983) Buzz pollination in angiosperms In Handbook ofExperimental Pollination Biology (ed C E Jones andR J Little) pp 73ndash113 NYScientific and Academic Editions

Buchmann S L and Cane J H (1989) Bees assess pollen returns whilesonicating Solanum flowers Oecologia 81 289ndash294

Buchmann S L and Hurley J P (1978) Biophysical model for buzz pollination inangiosperms J Theor Biol 72 639ndash657

Buchmann S L Jones C E and Colin L J (1977) Vibratile pollination ofSolanum douglasii and Solanum xanti (Solanaceae) in Southern California USAWasmann J Biol 35 1ndash25

Burkart A Lunau K and Schlindwein C (2011) Comparative bioacousticalstudies on flight and buzzing of neotropical bees J Pollin Ecol 6 118ndash124

Byrne D N Buchmann S L and Spangler H G (1988) Relationship betweenwing loading wingbeat frequency and body mass in homopterous insects J ExpBiol 135 9ndash23

Cameron S A Hines H M and Williams P H (2007) A comprehensivephylogeny of the bumble bees (Bombus) Biol J Linn Soc 91 161ndash188

Cane J H (1987) Estimation of bee size using intertegular span (Apoidea)J Kans Entomol Soc 60 145ndash147

Cardinal S Buchmann S L and Russell A L (2018) The evolution of floralsonication a pollen foraging behavior used by bees (Anthophila) Evolution 72590ndash600

Cocroft R B and De Luca P (2006) Size-frequency relationships in insectvibratory signals In Insect Sounds and Communication Physiology BehaviorEcology and Evolution (ed S Drosopoulos and M F Claridge) pp 99ndash110New York CRC

Cocroft R B and Rodriguez R L (2005) The behavioral ecology of insectvibrational communication Bioscience 55 323ndash334

Corbet S A and Huang S-Q (2014) Buzz pollination in eight bumblebee-pollinated Pedicularis species does it involve vibration-induced triboelectriccharging of pollen grains Ann Bot 114 1665ndash1674

De Luca P A and Vallejo-Marın M (2013) Whatrsquos the lsquobuzzrsquo about The ecologyand evolutionary significance of buzz-pollination Curr Opin Plant Biol 16429ndash435

De Luca P A Bussiere L F Souto-Vilaros D Goulson D Mason A C andVallejo-Marın M (2013) Variability in bumblebee pollination buzzes affects thequantity of pollen released from flowers Oecologia 172 805ndash816

De Luca P A Cox D A and Vallejo-Marın M (2014) Comparison of pollinationand defensive buzzes in bumblebees indicates species-specific and context-dependent vibrations Naturwissenschaften 101 331ndash338

De Luca P A Giebink N Mason A C Papaj D and Buchmann S L(2018) How well do acoustic recordings characterize properties of bee(Anthophila) floral sonication vibrations Bioacoustics Int J Anim SoundRecord 1ndash14

Gibson J S and Cocroft R B (2018) Vibration-guided mate searching intreehoppers directional accuracy and sampling strategies in a complex sensoryenvironment J Exp Biol 221 jeb175083

Glover B J Bunnewell S andMartin C (2004) Convergent evolution within thegenus Solanum the specialised anther cone develops through alternativepathways Gene 331 1ndash7

Harder L D and Barclay R M R (1994) The functional significance of poricidalanthers and buzz pollination controlled pollen removal from Dodecatheon FunctEcol 8 509ndash517

Harder L D and Wilson W G (1994) Floral evolution and male reproductivesuccess optimal dispensing schedules for pollen dispersal by animal-pollinatedplants Evol Ecol 8 542ndash559

Harris J A (1905) The dehiscence of anthers by apical pores Missouri BotGarden Annu Rep 1905 167ndash257

Kawai Y and Kudo G (2008) Effectiveness of buzz pollination inPedicularis chamissonis significance of multiple visits by bumblebees EcolRes 24 215

King M J (1993) Buzz foraging mechanism of bumble bees J Apic Res32 41ndash49

King M J and Buchmann S L (1996) Sonication dispensing of pollen fromSolanum laciniatum flowers Funct Ecol 10 449ndash456

11

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

of different orders varied with the function of those signals (egalarm or mating) Our results can be interpreted as indicating thatafter accounting for the effect of bee species size variation withinspecies does not correlate with the amplitude of floral vibrationsalthough it may be positively associated with frequency in someplants This has some interesting implications for buzz pollinationas it suggests that within the range of size variation seen amongbuzz-pollinating worker bees of a given species different-sizedworkers are capable of producing similar floral vibrations Thiscould be achieved either passively (floral vibrations do not dependstrongly on bee traits related to size Nunes-Silva et al 2013) oractively (bees may be able to adjust their vibrations to achieve agiven outcome) Given that the properties of floral vibrations affectpollen removal (De Luca et al 2013) this should also influence theamount of pollen obtained by workers Buzz-pollinating bees arecapable of altering several aspects of their behaviour in response topollen availability including visit duration buzz number andduration and level of pollen grooming (Buchmann and Cane 1989Nunes-Silva et al 2013 Russell et al 2016) In contrast evidenceof whether bumblebees can adjust the mechanical properties ofbuzzes in response to pollen availability is mixed For example onthe one hand a study of B impatiens vibrating flowers of tomato(Solanum lycopersicum) found that frequency and velocityamplitude (m sminus1) do not differ in consecutive visits (wherepollen is presumably being depleted) or in comparisons between

flowers with full pollen loads and flowers in which pollen access hasbeen restricted by gluing the anther pore shut (Nunes-Silva et al2013) On the other hand Russell et al (2016) found that Bimpatiens produced significantly higher amplitudes for the acousticcomponent (dB) of floral vibrations when visiting flowers ofSolanum houstonii in which the anther pores were glued shut thanwhen visiting unmanipulated flowers More research is needed toestablish the morphological physiological and behaviouraldeterminants of floral vibrations and the question as to why evenclosely related species differ in their floral vibrations remainsunanswered

Do bees adjust their vibrations depending on the plantspeciesPrevious studies indicated that the characteristics of floral vibrationsof the same bee species on different plant species can vary(reviewed in Switzer and Combes 2017) For example in agreenhouse study of B impatiens visiting three species of Solanum(S carolinense S dulcamara and S lycopersicum) Switzer andCombes (2017) showed that the same individual bee can producefloral vibrations of different frequency and duration on differentplant species We found that the acceleration amplitude but notfundamental frequency experienced at the base of the flower whenbuzz pollinated by the same bee species varied depending on planttype In contrast to previous studies measuring the frequency of

250

300

350

400

450Fu

ndam

enta

l fre

quen

cy (H

z)A

0

10

20

30

Pea

k ac

cele

ratio

n (m

sndash2

)

B

0

5

10

15

20

S citrullifolium S rostratum

RM

S a

ccel

erat

ion

(m s

ndash2)

C

0

50

100

150

S citrullifolium S rostratum

RM

S a

ccel

erat

ion

coup

ling

fact

or (m

sndash2

) D

Plant species

Fig 5 Comparison of the mechanical properties of floral vibrations from two bumblebee taxa on flowers of two buzz-pollinated plants Data arefrom B audax (purple) and B canariensis (blue) on Solanum citrullifolium and S rostratum Vibrations were recorded using an accelerometer attached tothe calyx of the flower using a metal pin (A) Fundamental frequency (B) Peak acceleration (C) RMS acceleration (D) RMS acceleration multiplied bythe species-specific flower coupling factor (see Materials and Methods) N=184 floral vibrations from 53 bees from four colonies of two taxa

9

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

airborne vibrations during buzz pollination in the current study wefound that two species of Bombus produced substrate-bornevibrations of similar frequency while visiting two closely relatedspecies of Solanum In our study the difference in floral vibrationsamong plants was restricted to changes in vibration amplitudespecifically acceleration amplitudeThe observed difference in vibration characteristics detected at

the base of the flower of two plant species could have at least twoexplanations (Switzer and Combes 2017) First the bee may beactively adjusting its floral vibrations depending on the plantspecies Individual bees seem to be capable of modifying thefrequency of their floral vibrations on a given plant species (Morganet al 2016) and it has been suggested that differences in floralcharacteristics including morphology and pollen availability maycause bees to change the vibrations produced by the same bee ondifferent plants (Switzer and Combes 2017) Second differences inthe characteristics of the substrate-borne vibrations measured herecould be due to intrinsic differences in the mechanical properties ofthe flowers of the two species of Solanum studied Our experimentaldesign and analysis allowed us to calculate the mechanical changesto known vibrational signals as they travelled through the flower (thecoupling factor) The analysis of floral vibrations after accountingfor these plant species-specific differences indicate that a given beespecies maintains similar types of vibrations regardless of the plantbeing visited Thus our study suggests that the differences in floralvibrations observed here are due to differences in the mechanicalproperties of the flowers rather than to behavioural changes in theproduction of vibrations by bees It remains a distinct possibility thatthe observed lack of amplitude differences in floral vibrations ondifferent plants results from the fact that bees are vibrating flowersat or close to the maximum amplitude they can generate using theirthoracic musclesIn nature the coupling between the bee and the rest of the flower

is influenced by both the characteristics of the flower and the way inwhich a bee holds onto floral parts while buzz pollinating (King1993) Our experiment could only account for differences in thetransmission of floral vibrations due to the mechanical properties offlowers as the calibrated vibrometer used here cannot accommodate

subtle variation in how a bee holds a flower during buzz pollinationMore work is needed to determine how the way in which a beemanipulates a flower eg how lsquotightlyrsquo the bee holds the anthers orwhether a bee lsquobitesrsquo the anthers (eg Russell et al 2016) affectsthe transmission of vibrations Different floral morphologies maycause changes in the way in which a bee manipulates and vibratesa flower and therefore on the transmission of those vibrationsThe two species of Solanum studied here have similar floralmorphologies and were manipulated in the same manner bybumblebees but comparisons of floral species with more disparatemorphologies may reveal whether and to what extent changes inbee behaviour affect the transmission of floral vibrations includingvibration amplitude and potentially pollen release

Effect of floral properties on substrate-borne floral vibrationsThe field of animal communication using substrate-borne vibrations(biotremology) has long recognised the importance of consideringplant characteristics in the functional ecology and evolution ofvibrational communication (Cocroft and Rodriguez 2005Michelsen et al 1982 Mortimer 2017) In buzz pollination theimportance of plant traits in the transmission of vibrations isacknowledged (eg Buchmann and Hurley 1978 King 1993) buthas rarely been explicitly incorporated into mechanistic ecologicalor evolutionary studies

Characteristics of the floral structures with which bees interactduring buzz pollination eg the androecium should have an effecton the transmission of vibrational energy from the bee to the flower(King and Buchmann 1996) Bees often vibrate only some of theanthers on a flower but their vibrations can be transmitted to otherfloral parts including other stamens not directly in contact with thebee resulting in pollen release (MV-M personal observation)Biotremology studies show that plants can cause significantdistortion of insectsrsquo vibration signals (Cocroft and Rodriguez2005 Michelsen et al 1982) The characteristics of vegetative andflower tissues are therefore likely to cause alterations in thetransmission of vibrations (King and Buchmann 1996) Our resultssuggest that mass cannot explain differences between flowers fromtwo Solanum species when being vibrated by the same Bombus

2

4

6

S citrullifolium S rostratum

Pea

k di

spla

cem

ent (microm

)A

0

20

40

60

S citrullifolium S rostratum

Pea

k di

spla

cem

ent

coup

ling

fact

or (micro

m)

B

Plant species

Fig 6 Estimated peak displacement of floral vibrations produced by two bumblebee taxa while visiting flowers of two plant species Data arefrom B audax (purple) and B canariensis (blue) on S citrullifolium and S rostratum (A) Estimated peak displacement at the point of measurement ie thebase of the flower (B) Estimated peak displacement at the anthers obtained by multiplying peak acceleration by the plant species-specific coupling factorN=184 floral vibrations from 53 bees from four colonies of two taxa

10

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

species However we showed that S rostratum reduced theamplitude of the vibrations transmitted through the flower to alesser extent than S citrullifolium The transmission of vibrationsthrough plant tissue depends not only on the distance from thesource but also on the characteristics and vibration modes of theplant (Michelsen et al 1982) Even fine structural differences suchas the presence of vascular tissue and vein size can affect thedamping of vibrations (Michelsen et al 1982) Thus floralproperties such as the morphology geometry and materialproperties of the flower are expected to contribute to vibrationaldifferences For example the stiffness of the filament holding theanther and whether the anther cones are loosely arranged or tightlypacked into a cone (Glover et al 2004) are likely to affect thetransmission of vibrations away from its source (the bee) Thevibrational properties of flowers particularly in an evolutionarycontext are not well understood The convergence of disparate plantgroups onto similar buzz-pollinated floral morphologies (eg theSolanum-type flower has evolved in more than 21 plant familiesDe Luca and Vallejo-Mariacuten 2013) provides an opportunity toinvestigate whether morphological similarity is associated withsimilar vibrational properties

ConclusionsTogether with previous work our results suggest that bees differ intheir capacity to transmit vibrations to flowers Because thecharacteristics of vibrations in particular their amplitude areassociated with pollen release (De Luca et al 2013 Vallejo-Mariacuten2019) visitation by different species of bees may affect patterns ofpollen dispensing and ultimately plant fitness (Harder and Barclay1994 Harder and Wilson 1994) Therefore from a plantrsquosperspective it may be selectively advantageous to favour visitationby specific species Similarly from a beersquos perspective the choiceof a given plant species may determine the extent to which theirfloral vibrations translate into sufficient vibrational energy to elicitpollen release (Buchmann and Hurley 1978) In particular in caseswhere bees cannot adjust their floral vibrations depending on plantspecies selectively foraging in plants in which vibrations aretransmitted with low attenuation (eg with low coupling factors)may allow bees to maximise pollen release per buzz effortMoreover bee traits that maximise the beersquos transmission ofvibrations (eg the location and manner in which a bee holds aflower) should also be favoured as mechanisms to transmitvibrations more efficiently potentially removing more pollen

AcknowledgementsWe thank M Pozo and Biobest for providing the bumblebee colonies for thisexperiment A Russell kindly shared his protocols and advice on flight arenaconstruction and bee experiments in general J Gibson and P De Luca providedkey advice on the analysis and interpretation of substrate-borne vibrations and onbuzz pollination We thank J Weir for help with flight arena construction Membersof the Vallejo-Marin lab in particular Lucy Nevard provided help with beemaintenance

Competing interestsThe authors declare no competing or financial interests

Author contributionsConceptualization BA-C MV-M Methodology BA-C CEB MV-M Formalanalysis BA-C MV-M Investigation BA-C CEB MV-M ResourcesMV-M Writing - original draft BA-C MV-M Writing - review amp editingBA-C MV-M Visualization BA-C MV-M Supervision MV-M Projectadministration MV-M Funding acquisition BA-C MV-M

FundingThis study was partially funded by a SPARK grant (University of Stirling) to MV-Mand an ERASMUS internship to BA-C

Data availabilityData have been deposited in the Dryad digital repository (Arroyo-Correa et al2019) dryad3q0b1r1

Supplementary informationSupplementary information available online athttpjebbiologistsorglookupdoi101242jeb198176supplemental

ReferencesArroyo-Correa B Beattie C E andVallejo-MarinM (2019) Data from Bee and

floral traits affect the characteristics of the vibrations experienced by flowers duringbuzz-pollination Dryad Digital Repository

Bowers K A W (1975) The pollination ecology of Solanum rostratum(Solanaceae) Am J Bot 62 633ndash638

Buchmann S L (1983) Buzz pollination in angiosperms In Handbook ofExperimental Pollination Biology (ed C E Jones andR J Little) pp 73ndash113 NYScientific and Academic Editions

Buchmann S L and Cane J H (1989) Bees assess pollen returns whilesonicating Solanum flowers Oecologia 81 289ndash294

Buchmann S L and Hurley J P (1978) Biophysical model for buzz pollination inangiosperms J Theor Biol 72 639ndash657

Buchmann S L Jones C E and Colin L J (1977) Vibratile pollination ofSolanum douglasii and Solanum xanti (Solanaceae) in Southern California USAWasmann J Biol 35 1ndash25

Burkart A Lunau K and Schlindwein C (2011) Comparative bioacousticalstudies on flight and buzzing of neotropical bees J Pollin Ecol 6 118ndash124

Byrne D N Buchmann S L and Spangler H G (1988) Relationship betweenwing loading wingbeat frequency and body mass in homopterous insects J ExpBiol 135 9ndash23

Cameron S A Hines H M and Williams P H (2007) A comprehensivephylogeny of the bumble bees (Bombus) Biol J Linn Soc 91 161ndash188

Cane J H (1987) Estimation of bee size using intertegular span (Apoidea)J Kans Entomol Soc 60 145ndash147

Cardinal S Buchmann S L and Russell A L (2018) The evolution of floralsonication a pollen foraging behavior used by bees (Anthophila) Evolution 72590ndash600

Cocroft R B and De Luca P (2006) Size-frequency relationships in insectvibratory signals In Insect Sounds and Communication Physiology BehaviorEcology and Evolution (ed S Drosopoulos and M F Claridge) pp 99ndash110New York CRC

Cocroft R B and Rodriguez R L (2005) The behavioral ecology of insectvibrational communication Bioscience 55 323ndash334

Corbet S A and Huang S-Q (2014) Buzz pollination in eight bumblebee-pollinated Pedicularis species does it involve vibration-induced triboelectriccharging of pollen grains Ann Bot 114 1665ndash1674

De Luca P A and Vallejo-Marın M (2013) Whatrsquos the lsquobuzzrsquo about The ecologyand evolutionary significance of buzz-pollination Curr Opin Plant Biol 16429ndash435

De Luca P A Bussiere L F Souto-Vilaros D Goulson D Mason A C andVallejo-Marın M (2013) Variability in bumblebee pollination buzzes affects thequantity of pollen released from flowers Oecologia 172 805ndash816

De Luca P A Cox D A and Vallejo-Marın M (2014) Comparison of pollinationand defensive buzzes in bumblebees indicates species-specific and context-dependent vibrations Naturwissenschaften 101 331ndash338

De Luca P A Giebink N Mason A C Papaj D and Buchmann S L(2018) How well do acoustic recordings characterize properties of bee(Anthophila) floral sonication vibrations Bioacoustics Int J Anim SoundRecord 1ndash14

Gibson J S and Cocroft R B (2018) Vibration-guided mate searching intreehoppers directional accuracy and sampling strategies in a complex sensoryenvironment J Exp Biol 221 jeb175083

Glover B J Bunnewell S andMartin C (2004) Convergent evolution within thegenus Solanum the specialised anther cone develops through alternativepathways Gene 331 1ndash7

Harder L D and Barclay R M R (1994) The functional significance of poricidalanthers and buzz pollination controlled pollen removal from Dodecatheon FunctEcol 8 509ndash517

Harder L D and Wilson W G (1994) Floral evolution and male reproductivesuccess optimal dispensing schedules for pollen dispersal by animal-pollinatedplants Evol Ecol 8 542ndash559

Harris J A (1905) The dehiscence of anthers by apical pores Missouri BotGarden Annu Rep 1905 167ndash257

Kawai Y and Kudo G (2008) Effectiveness of buzz pollination inPedicularis chamissonis significance of multiple visits by bumblebees EcolRes 24 215

King M J (1993) Buzz foraging mechanism of bumble bees J Apic Res32 41ndash49

King M J and Buchmann S L (1996) Sonication dispensing of pollen fromSolanum laciniatum flowers Funct Ecol 10 449ndash456

11

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

airborne vibrations during buzz pollination in the current study wefound that two species of Bombus produced substrate-bornevibrations of similar frequency while visiting two closely relatedspecies of Solanum In our study the difference in floral vibrationsamong plants was restricted to changes in vibration amplitudespecifically acceleration amplitudeThe observed difference in vibration characteristics detected at

the base of the flower of two plant species could have at least twoexplanations (Switzer and Combes 2017) First the bee may beactively adjusting its floral vibrations depending on the plantspecies Individual bees seem to be capable of modifying thefrequency of their floral vibrations on a given plant species (Morganet al 2016) and it has been suggested that differences in floralcharacteristics including morphology and pollen availability maycause bees to change the vibrations produced by the same bee ondifferent plants (Switzer and Combes 2017) Second differences inthe characteristics of the substrate-borne vibrations measured herecould be due to intrinsic differences in the mechanical properties ofthe flowers of the two species of Solanum studied Our experimentaldesign and analysis allowed us to calculate the mechanical changesto known vibrational signals as they travelled through the flower (thecoupling factor) The analysis of floral vibrations after accountingfor these plant species-specific differences indicate that a given beespecies maintains similar types of vibrations regardless of the plantbeing visited Thus our study suggests that the differences in floralvibrations observed here are due to differences in the mechanicalproperties of the flowers rather than to behavioural changes in theproduction of vibrations by bees It remains a distinct possibility thatthe observed lack of amplitude differences in floral vibrations ondifferent plants results from the fact that bees are vibrating flowersat or close to the maximum amplitude they can generate using theirthoracic musclesIn nature the coupling between the bee and the rest of the flower

is influenced by both the characteristics of the flower and the way inwhich a bee holds onto floral parts while buzz pollinating (King1993) Our experiment could only account for differences in thetransmission of floral vibrations due to the mechanical properties offlowers as the calibrated vibrometer used here cannot accommodate

subtle variation in how a bee holds a flower during buzz pollinationMore work is needed to determine how the way in which a beemanipulates a flower eg how lsquotightlyrsquo the bee holds the anthers orwhether a bee lsquobitesrsquo the anthers (eg Russell et al 2016) affectsthe transmission of vibrations Different floral morphologies maycause changes in the way in which a bee manipulates and vibratesa flower and therefore on the transmission of those vibrationsThe two species of Solanum studied here have similar floralmorphologies and were manipulated in the same manner bybumblebees but comparisons of floral species with more disparatemorphologies may reveal whether and to what extent changes inbee behaviour affect the transmission of floral vibrations includingvibration amplitude and potentially pollen release

Effect of floral properties on substrate-borne floral vibrationsThe field of animal communication using substrate-borne vibrations(biotremology) has long recognised the importance of consideringplant characteristics in the functional ecology and evolution ofvibrational communication (Cocroft and Rodriguez 2005Michelsen et al 1982 Mortimer 2017) In buzz pollination theimportance of plant traits in the transmission of vibrations isacknowledged (eg Buchmann and Hurley 1978 King 1993) buthas rarely been explicitly incorporated into mechanistic ecologicalor evolutionary studies

Characteristics of the floral structures with which bees interactduring buzz pollination eg the androecium should have an effecton the transmission of vibrational energy from the bee to the flower(King and Buchmann 1996) Bees often vibrate only some of theanthers on a flower but their vibrations can be transmitted to otherfloral parts including other stamens not directly in contact with thebee resulting in pollen release (MV-M personal observation)Biotremology studies show that plants can cause significantdistortion of insectsrsquo vibration signals (Cocroft and Rodriguez2005 Michelsen et al 1982) The characteristics of vegetative andflower tissues are therefore likely to cause alterations in thetransmission of vibrations (King and Buchmann 1996) Our resultssuggest that mass cannot explain differences between flowers fromtwo Solanum species when being vibrated by the same Bombus

2

4

6

S citrullifolium S rostratum

Pea

k di

spla

cem

ent (microm

)A

0

20

40

60

S citrullifolium S rostratum

Pea

k di

spla

cem

ent

coup

ling

fact

or (micro

m)

B

Plant species

Fig 6 Estimated peak displacement of floral vibrations produced by two bumblebee taxa while visiting flowers of two plant species Data arefrom B audax (purple) and B canariensis (blue) on S citrullifolium and S rostratum (A) Estimated peak displacement at the point of measurement ie thebase of the flower (B) Estimated peak displacement at the anthers obtained by multiplying peak acceleration by the plant species-specific coupling factorN=184 floral vibrations from 53 bees from four colonies of two taxa

10

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

species However we showed that S rostratum reduced theamplitude of the vibrations transmitted through the flower to alesser extent than S citrullifolium The transmission of vibrationsthrough plant tissue depends not only on the distance from thesource but also on the characteristics and vibration modes of theplant (Michelsen et al 1982) Even fine structural differences suchas the presence of vascular tissue and vein size can affect thedamping of vibrations (Michelsen et al 1982) Thus floralproperties such as the morphology geometry and materialproperties of the flower are expected to contribute to vibrationaldifferences For example the stiffness of the filament holding theanther and whether the anther cones are loosely arranged or tightlypacked into a cone (Glover et al 2004) are likely to affect thetransmission of vibrations away from its source (the bee) Thevibrational properties of flowers particularly in an evolutionarycontext are not well understood The convergence of disparate plantgroups onto similar buzz-pollinated floral morphologies (eg theSolanum-type flower has evolved in more than 21 plant familiesDe Luca and Vallejo-Mariacuten 2013) provides an opportunity toinvestigate whether morphological similarity is associated withsimilar vibrational properties

ConclusionsTogether with previous work our results suggest that bees differ intheir capacity to transmit vibrations to flowers Because thecharacteristics of vibrations in particular their amplitude areassociated with pollen release (De Luca et al 2013 Vallejo-Mariacuten2019) visitation by different species of bees may affect patterns ofpollen dispensing and ultimately plant fitness (Harder and Barclay1994 Harder and Wilson 1994) Therefore from a plantrsquosperspective it may be selectively advantageous to favour visitationby specific species Similarly from a beersquos perspective the choiceof a given plant species may determine the extent to which theirfloral vibrations translate into sufficient vibrational energy to elicitpollen release (Buchmann and Hurley 1978) In particular in caseswhere bees cannot adjust their floral vibrations depending on plantspecies selectively foraging in plants in which vibrations aretransmitted with low attenuation (eg with low coupling factors)may allow bees to maximise pollen release per buzz effortMoreover bee traits that maximise the beersquos transmission ofvibrations (eg the location and manner in which a bee holds aflower) should also be favoured as mechanisms to transmitvibrations more efficiently potentially removing more pollen

AcknowledgementsWe thank M Pozo and Biobest for providing the bumblebee colonies for thisexperiment A Russell kindly shared his protocols and advice on flight arenaconstruction and bee experiments in general J Gibson and P De Luca providedkey advice on the analysis and interpretation of substrate-borne vibrations and onbuzz pollination We thank J Weir for help with flight arena construction Membersof the Vallejo-Marin lab in particular Lucy Nevard provided help with beemaintenance

Competing interestsThe authors declare no competing or financial interests

Author contributionsConceptualization BA-C MV-M Methodology BA-C CEB MV-M Formalanalysis BA-C MV-M Investigation BA-C CEB MV-M ResourcesMV-M Writing - original draft BA-C MV-M Writing - review amp editingBA-C MV-M Visualization BA-C MV-M Supervision MV-M Projectadministration MV-M Funding acquisition BA-C MV-M

FundingThis study was partially funded by a SPARK grant (University of Stirling) to MV-Mand an ERASMUS internship to BA-C

Data availabilityData have been deposited in the Dryad digital repository (Arroyo-Correa et al2019) dryad3q0b1r1

Supplementary informationSupplementary information available online athttpjebbiologistsorglookupdoi101242jeb198176supplemental

ReferencesArroyo-Correa B Beattie C E andVallejo-MarinM (2019) Data from Bee and

floral traits affect the characteristics of the vibrations experienced by flowers duringbuzz-pollination Dryad Digital Repository

Bowers K A W (1975) The pollination ecology of Solanum rostratum(Solanaceae) Am J Bot 62 633ndash638

Buchmann S L (1983) Buzz pollination in angiosperms In Handbook ofExperimental Pollination Biology (ed C E Jones andR J Little) pp 73ndash113 NYScientific and Academic Editions

Buchmann S L and Cane J H (1989) Bees assess pollen returns whilesonicating Solanum flowers Oecologia 81 289ndash294

Buchmann S L and Hurley J P (1978) Biophysical model for buzz pollination inangiosperms J Theor Biol 72 639ndash657

Buchmann S L Jones C E and Colin L J (1977) Vibratile pollination ofSolanum douglasii and Solanum xanti (Solanaceae) in Southern California USAWasmann J Biol 35 1ndash25

Burkart A Lunau K and Schlindwein C (2011) Comparative bioacousticalstudies on flight and buzzing of neotropical bees J Pollin Ecol 6 118ndash124

Byrne D N Buchmann S L and Spangler H G (1988) Relationship betweenwing loading wingbeat frequency and body mass in homopterous insects J ExpBiol 135 9ndash23

Cameron S A Hines H M and Williams P H (2007) A comprehensivephylogeny of the bumble bees (Bombus) Biol J Linn Soc 91 161ndash188

Cane J H (1987) Estimation of bee size using intertegular span (Apoidea)J Kans Entomol Soc 60 145ndash147

Cardinal S Buchmann S L and Russell A L (2018) The evolution of floralsonication a pollen foraging behavior used by bees (Anthophila) Evolution 72590ndash600

Cocroft R B and De Luca P (2006) Size-frequency relationships in insectvibratory signals In Insect Sounds and Communication Physiology BehaviorEcology and Evolution (ed S Drosopoulos and M F Claridge) pp 99ndash110New York CRC

Cocroft R B and Rodriguez R L (2005) The behavioral ecology of insectvibrational communication Bioscience 55 323ndash334

Corbet S A and Huang S-Q (2014) Buzz pollination in eight bumblebee-pollinated Pedicularis species does it involve vibration-induced triboelectriccharging of pollen grains Ann Bot 114 1665ndash1674

De Luca P A and Vallejo-Marın M (2013) Whatrsquos the lsquobuzzrsquo about The ecologyand evolutionary significance of buzz-pollination Curr Opin Plant Biol 16429ndash435

De Luca P A Bussiere L F Souto-Vilaros D Goulson D Mason A C andVallejo-Marın M (2013) Variability in bumblebee pollination buzzes affects thequantity of pollen released from flowers Oecologia 172 805ndash816

De Luca P A Cox D A and Vallejo-Marın M (2014) Comparison of pollinationand defensive buzzes in bumblebees indicates species-specific and context-dependent vibrations Naturwissenschaften 101 331ndash338

De Luca P A Giebink N Mason A C Papaj D and Buchmann S L(2018) How well do acoustic recordings characterize properties of bee(Anthophila) floral sonication vibrations Bioacoustics Int J Anim SoundRecord 1ndash14

Gibson J S and Cocroft R B (2018) Vibration-guided mate searching intreehoppers directional accuracy and sampling strategies in a complex sensoryenvironment J Exp Biol 221 jeb175083

Glover B J Bunnewell S andMartin C (2004) Convergent evolution within thegenus Solanum the specialised anther cone develops through alternativepathways Gene 331 1ndash7

Harder L D and Barclay R M R (1994) The functional significance of poricidalanthers and buzz pollination controlled pollen removal from Dodecatheon FunctEcol 8 509ndash517

Harder L D and Wilson W G (1994) Floral evolution and male reproductivesuccess optimal dispensing schedules for pollen dispersal by animal-pollinatedplants Evol Ecol 8 542ndash559

Harris J A (1905) The dehiscence of anthers by apical pores Missouri BotGarden Annu Rep 1905 167ndash257

Kawai Y and Kudo G (2008) Effectiveness of buzz pollination inPedicularis chamissonis significance of multiple visits by bumblebees EcolRes 24 215

King M J (1993) Buzz foraging mechanism of bumble bees J Apic Res32 41ndash49

King M J and Buchmann S L (1996) Sonication dispensing of pollen fromSolanum laciniatum flowers Funct Ecol 10 449ndash456

11

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

species However we showed that S rostratum reduced theamplitude of the vibrations transmitted through the flower to alesser extent than S citrullifolium The transmission of vibrationsthrough plant tissue depends not only on the distance from thesource but also on the characteristics and vibration modes of theplant (Michelsen et al 1982) Even fine structural differences suchas the presence of vascular tissue and vein size can affect thedamping of vibrations (Michelsen et al 1982) Thus floralproperties such as the morphology geometry and materialproperties of the flower are expected to contribute to vibrationaldifferences For example the stiffness of the filament holding theanther and whether the anther cones are loosely arranged or tightlypacked into a cone (Glover et al 2004) are likely to affect thetransmission of vibrations away from its source (the bee) Thevibrational properties of flowers particularly in an evolutionarycontext are not well understood The convergence of disparate plantgroups onto similar buzz-pollinated floral morphologies (eg theSolanum-type flower has evolved in more than 21 plant familiesDe Luca and Vallejo-Mariacuten 2013) provides an opportunity toinvestigate whether morphological similarity is associated withsimilar vibrational properties

ConclusionsTogether with previous work our results suggest that bees differ intheir capacity to transmit vibrations to flowers Because thecharacteristics of vibrations in particular their amplitude areassociated with pollen release (De Luca et al 2013 Vallejo-Mariacuten2019) visitation by different species of bees may affect patterns ofpollen dispensing and ultimately plant fitness (Harder and Barclay1994 Harder and Wilson 1994) Therefore from a plantrsquosperspective it may be selectively advantageous to favour visitationby specific species Similarly from a beersquos perspective the choiceof a given plant species may determine the extent to which theirfloral vibrations translate into sufficient vibrational energy to elicitpollen release (Buchmann and Hurley 1978) In particular in caseswhere bees cannot adjust their floral vibrations depending on plantspecies selectively foraging in plants in which vibrations aretransmitted with low attenuation (eg with low coupling factors)may allow bees to maximise pollen release per buzz effortMoreover bee traits that maximise the beersquos transmission ofvibrations (eg the location and manner in which a bee holds aflower) should also be favoured as mechanisms to transmitvibrations more efficiently potentially removing more pollen

AcknowledgementsWe thank M Pozo and Biobest for providing the bumblebee colonies for thisexperiment A Russell kindly shared his protocols and advice on flight arenaconstruction and bee experiments in general J Gibson and P De Luca providedkey advice on the analysis and interpretation of substrate-borne vibrations and onbuzz pollination We thank J Weir for help with flight arena construction Membersof the Vallejo-Marin lab in particular Lucy Nevard provided help with beemaintenance

Competing interestsThe authors declare no competing or financial interests

Author contributionsConceptualization BA-C MV-M Methodology BA-C CEB MV-M Formalanalysis BA-C MV-M Investigation BA-C CEB MV-M ResourcesMV-M Writing - original draft BA-C MV-M Writing - review amp editingBA-C MV-M Visualization BA-C MV-M Supervision MV-M Projectadministration MV-M Funding acquisition BA-C MV-M

FundingThis study was partially funded by a SPARK grant (University of Stirling) to MV-Mand an ERASMUS internship to BA-C

Data availabilityData have been deposited in the Dryad digital repository (Arroyo-Correa et al2019) dryad3q0b1r1

Supplementary informationSupplementary information available online athttpjebbiologistsorglookupdoi101242jeb198176supplemental

ReferencesArroyo-Correa B Beattie C E andVallejo-MarinM (2019) Data from Bee and

floral traits affect the characteristics of the vibrations experienced by flowers duringbuzz-pollination Dryad Digital Repository

Bowers K A W (1975) The pollination ecology of Solanum rostratum(Solanaceae) Am J Bot 62 633ndash638

Buchmann S L (1983) Buzz pollination in angiosperms In Handbook ofExperimental Pollination Biology (ed C E Jones andR J Little) pp 73ndash113 NYScientific and Academic Editions

Buchmann S L and Cane J H (1989) Bees assess pollen returns whilesonicating Solanum flowers Oecologia 81 289ndash294

Buchmann S L and Hurley J P (1978) Biophysical model for buzz pollination inangiosperms J Theor Biol 72 639ndash657

Buchmann S L Jones C E and Colin L J (1977) Vibratile pollination ofSolanum douglasii and Solanum xanti (Solanaceae) in Southern California USAWasmann J Biol 35 1ndash25

Burkart A Lunau K and Schlindwein C (2011) Comparative bioacousticalstudies on flight and buzzing of neotropical bees J Pollin Ecol 6 118ndash124

Byrne D N Buchmann S L and Spangler H G (1988) Relationship betweenwing loading wingbeat frequency and body mass in homopterous insects J ExpBiol 135 9ndash23

Cameron S A Hines H M and Williams P H (2007) A comprehensivephylogeny of the bumble bees (Bombus) Biol J Linn Soc 91 161ndash188

Cane J H (1987) Estimation of bee size using intertegular span (Apoidea)J Kans Entomol Soc 60 145ndash147

Cardinal S Buchmann S L and Russell A L (2018) The evolution of floralsonication a pollen foraging behavior used by bees (Anthophila) Evolution 72590ndash600

Cocroft R B and De Luca P (2006) Size-frequency relationships in insectvibratory signals In Insect Sounds and Communication Physiology BehaviorEcology and Evolution (ed S Drosopoulos and M F Claridge) pp 99ndash110New York CRC

Cocroft R B and Rodriguez R L (2005) The behavioral ecology of insectvibrational communication Bioscience 55 323ndash334

Corbet S A and Huang S-Q (2014) Buzz pollination in eight bumblebee-pollinated Pedicularis species does it involve vibration-induced triboelectriccharging of pollen grains Ann Bot 114 1665ndash1674

De Luca P A and Vallejo-Marın M (2013) Whatrsquos the lsquobuzzrsquo about The ecologyand evolutionary significance of buzz-pollination Curr Opin Plant Biol 16429ndash435

De Luca P A Bussiere L F Souto-Vilaros D Goulson D Mason A C andVallejo-Marın M (2013) Variability in bumblebee pollination buzzes affects thequantity of pollen released from flowers Oecologia 172 805ndash816

De Luca P A Cox D A and Vallejo-Marın M (2014) Comparison of pollinationand defensive buzzes in bumblebees indicates species-specific and context-dependent vibrations Naturwissenschaften 101 331ndash338

De Luca P A Giebink N Mason A C Papaj D and Buchmann S L(2018) How well do acoustic recordings characterize properties of bee(Anthophila) floral sonication vibrations Bioacoustics Int J Anim SoundRecord 1ndash14

Gibson J S and Cocroft R B (2018) Vibration-guided mate searching intreehoppers directional accuracy and sampling strategies in a complex sensoryenvironment J Exp Biol 221 jeb175083

Glover B J Bunnewell S andMartin C (2004) Convergent evolution within thegenus Solanum the specialised anther cone develops through alternativepathways Gene 331 1ndash7

Harder L D and Barclay R M R (1994) The functional significance of poricidalanthers and buzz pollination controlled pollen removal from Dodecatheon FunctEcol 8 509ndash517

Harder L D and Wilson W G (1994) Floral evolution and male reproductivesuccess optimal dispensing schedules for pollen dispersal by animal-pollinatedplants Evol Ecol 8 542ndash559

Harris J A (1905) The dehiscence of anthers by apical pores Missouri BotGarden Annu Rep 1905 167ndash257

Kawai Y and Kudo G (2008) Effectiveness of buzz pollination inPedicularis chamissonis significance of multiple visits by bumblebees EcolRes 24 215

King M J (1993) Buzz foraging mechanism of bumble bees J Apic Res32 41ndash49

King M J and Buchmann S L (1996) Sonication dispensing of pollen fromSolanum laciniatum flowers Funct Ecol 10 449ndash456

11

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology

King M J and Buchmann S L (2003) Floral sonication by bees mesosomalvibration by Bombus and Xylocopa but not Apis (Hymenoptera Apidae) ejectspollen from poricidal anthers J Kans Entomol Soc 76 295ndash305

Lecocq T Coppee A Michez D Brasero N Rasplus J-Y Valterova I andRasmont P (2016) The alienrsquos identity consequences of taxonomic status forthe international bumblebee trade regulations Biol Conserv 195 169ndash176

Lye G C Jennings S N Osborne J L andGoulson D (2011) Impacts of theuse of nonnative commercial bumble bees for pollinator supplementation inraspberry J Econ Entomol 104 107ndash114

Macior L W (1964) An experimental study of the floral ecology of Dodecatheonmeadia Am J Bot 51 96ndash108

Macior L W (1968) Pollination adaptation in Pedicularis groenlandica Am J Bot55 927ndash932

Mesquita-Neto J N Bluthgen N and Schlindwein C (2018) Flowers withporicidal anthers and their complex interaction networks-Disentangling legitimatepollinators and illegitimate visitors Funct Ecol 32 2321ndash2332

Michelsen A Fink F Gogala M and Traue D (1982) Plants as transmissionchannels for insect vibrational songs Behav Ecol Sociobiol 11 269ndash281

Michener C D (2000) The Bees of the World Baltimore USA John HopkinsUniversity Press

Morandin L A Laverty T M and Kevan P G (2001) Effect of bumble bee(Hymenoptera Apidae) pollination intensity on the quality of greenhousetomatoes J Econ Entomol 94 172ndash179

Morgan T Whitehorn P Lye G C and Vallejo-Marın M (2016) Floralsonication is an innate behaviour in bumblebees that can be fine-tuned withexperience in manipulating flowers J Insect Behav 29 233ndash241

Mortimer B (2017) Biotremology do physical constraints limit the propagation ofvibrational information Anim Behav 130 165ndash174

Nunes-Silva P Hnrcir M Shipp L Imperatriz-Fonseca V L andKevan P G(2013) The behaviour of Bombus impatiens (Apidae Bombini) on tomato(Lycopersicon esculentum Mill Solanaceae) flowers pollination and rewardperception J Pollin Ecol 11 33ndash40

Rasmont P Coppee A Michez D and De Meulemeester T (2008) Anoverview of the Bombus terrestris (L 1758) subspecies (Hymenoptera Apidae)Ann Soc Entomol Fr 44 243ndash250

Rosi-Denadai C A Araujo P C S Campos L A D O Cosme L Jr andGuedes R N C (2018) Buzz-pollination in Neotropical bees genus-dependentfrequencies and lack of optimal frequency for pollen release Insect Sci 1ndash10

Russell A L and Papaj D R (2016) Artificial pollen dispensing flowers andfeeders for bee behaviour experiments J Pollin Ecol 18 13ndash22

Russell A L Leonard A S Gillette H D and Papaj D R (2016) Concealedfloral rewards and the role of experience in floral sonication by bees Anim Behav120 83ndash91

Shao Z-Y Mao H-X Fu W-J Ono M Wang D-S Bonizzoni M andZhang Y-P (2004) Genetic structure of Asian populations of Bombus ignitus(Hymenoptera Apidae) J Hered 95 46ndash52

Sueur J (2018) Sound Analysis and Synthesis with R The Netherlands SpringerSwitzer C M and Combes S A (2017) Bumblebee sonication behavior

changes with plant species and environmental conditions Apidologie 48223ndash233

Vallejo-Marın M (2019) Buzz pollination studying bee vibrations on flowers NewPhytol

Vallejo-Marın M Manson J S Thomson J D and Barrett S C H (2009)Division of labour within flowers heteranthery a floral strategy to reconcilecontrasting pollen fates J Evol Biol 22 828ndash839

Vallejo-Marin M Walker C Friston-Reilly P Solis-Montero L and Igic B(2014) Recurrent modification of floral morphology in heterantherous Solanumreveals a parallel shift in reproductive strategy Philos Trans R Soc B Biol Sci369 20130256

Velthuis H H W and Van Doorn A (2006) A century of advances in bumblebeedomestication and the economic and environmental aspects of itscommercialization for pollination Apidologie 37 421ndash451

Whalen M D (1979) Taxonomy of Solanum section Androceras GentesHerbarum 11 359ndash426

12

RESEARCH ARTICLE Journal of Experimental Biology (2019) 222 jeb198176 doi101242jeb198176

Journal

ofEx

perim

entalB

iology