Flowers respond to pollinator sound within minutes by increasing nectar sugar concentration

Abstract

Can plants sense natural airborne sounds and respond to them rapidly? We show that Oenothera drummondii flowers, exposed to playback sound of a flying bee or to synthetic sound signals at similar frequencies, produce sweeter nectar within 3 min, potentially increasing the chances of cross pollination. We found that the flowers vibrated mechanically in response to these sounds, suggesting a plausible mechanism where the flower serves as an auditory sensory organ. Both the vibration and the nectar response were frequency‐specific: the flowers responded and vibrated to pollinator sounds, but not to higher frequency sound. Our results document for the first time that plants can rapidly respond to pollinator sounds in an ecologically relevant way. Potential implications include plant resource allocation, the evolution of flower shape and the evolution of pollinators sound. Finally, our results suggest that plants may be affected by other sounds as well, including anthropogenic ones.

Full text

LETTER Flowers respond to pollinator sound within minutes by
increasing nectar sugar concentration
Marine Veits,
1†
Itzhak Khait,
1†
Uri Obolski,
1
Eyal Zinger,
1
Arjan Boonman,
2
Aya Goldshtein,
2
Kfir Saban,
1
Rya Seltzer,
2
Udi Ben-Dor,
1
Paz Estlein,
1
Areej Kabat,
1
Dor Peretz,
1
Ittai Ratzersdorfer,
1
Slava Krylov,
3
Daniel Chamovitz,
1
Yuval Sapir,
1§
Yossi Yovel,
2§
and
Lilach Hadany
1
*
§
Abstract
Can plants sense natural airborne sounds and respond to them rapidly? We show that Oenothera
drummondii flowers, exposed to playback sound of a flying bee or to synthetic sound signals at
similar frequencies, produce sweeter nectar within 3 min, potentially increasing the chances of
cross pollination. We found that the flowers vibrated mechanically in response to these sounds,
suggesting a plausible mechanism where the flower serves as an auditory sensory organ. Both the
vibration and the nectar response were frequency-specific: the flowers responded and vibrated to
pollinator sounds, but not to higher frequency sound. Our results document for the first time that
plants can rapidly respond to pollinator sounds in an ecologically relevant way. Potential implica-
tions include plant resource allocation, the evolution of flower shape and the evolution of pollina-
tors sound. Finally, our results suggest that plants may be affected by other sounds as well,
including anthropogenic ones.
Keywords
Communication, nectar, plant bioacoustics, plant–pollinator interactions, pollination, signalling,
vibration.
Ecology Letters (2019) 22: 1483–1492
INTRODUCTION
Plants’ ability to sense their environment and respond to it is
critical to their survival. Plants responses to light (Jiao et al.
2007; Chory 2010), volatile chemicals (Arimura et al. 2000;
Baldwin et al. 2006; Heil & Bueno 2007; Karban et al. 2014;
Karban 2015) and different forms of touch, including continu-
ous (Darwin 1892; Slack 2000; Braam 2005; Monshausen &
Haswell 2013) and vibrating (De Luca & Vallejo-Mar
ın 2013;
Appel & Cocroft 2014) are well documented. However, the
ability of plants to sense and respond to airborne sound –one
of the most widely used communication modalities in the ani-
mal kingdom –has hardly been investigated (Chamovitz 2012;
Gagliano et al. 2012; Hassanien et al. 2014). Recent studies
demonstrated slow responses, such as changes in the growth
rate of plants, after exposure to artificial acoustic stimuli last-
ing hours or days (Takahashi et al. 1991; Xiujuan et al. 2003;
Yi et al. 2003; Bochu et al. 2004; Ghosh et al. 2016; Choi
et al. 2017; Gagliano et al., 2017; Ghosh et al. 2017; Kim
et al. 2017; L
opez-Ribera & Vicient 2017; Jung et al. 2018).
Furthermore, plant tissues have been shown to vibrate to a
range of sounds (Telewski 2006; Rebar et al. 2012; Davis
et al., 2014). In contrast, to the best of our knowledge, a
rapid reaction to airborne sound has never been reported for
plants; neither has the biological function of any plant
response to airborne sound been identified. In this work, we
aimed to test rapid plant responses to airborne sound in the
context of plant–pollinator interactions.
The great majority (87.5%) of flowering plants rely on ani-
mal pollinators for reproduction (Ollerton et al. 2011). In
these plants, attracting pollinators can increase plant fitness
and is achieved using signals such as colour, odour and shape,
and by food rewards of nectar and pollen (Willmer 2011).
Increased reward quality or quantity can result in longer polli-
nator visits or in a higher likelihood that a pollinator will visit
another flower of the same species in the near future, poten-
tially increasing the flower’s fitness by increasing the chances
of pollination and reproduction (Faegri & Van Der Pijl 1979;
Pappers et al. 1999; Stout & Goulson 2002). Producing an
enhanced reward can be expensive (Pleasants & Chaplin 1983;
Southwick 1984; Pyke 1991; Ordano & Ornelas 2005; Ornelas
& Lara 2009; Galetto et al. 2018) and standing crop of nectar
is subjected to degradation by microbes (Herrera et al. 2008;
Vannette et al., 2013) as well as to robbery (Irwin et al. 2010),
including silent robbers like ants (Galen 1999). Thus, a mech-
anism for timing the production of enhanced reward to a time
when pollinators are likely to be present could be highly bene-
ficial to the plant. Here we suggest that a response of plants
to the sound of a pollinator can serve as such a timing mecha-
nism. Specifically, we hypothesise that plants could respond to
the sound of a flying pollinator by increasing the reward in a
way that would increase the probability of pollination and
reproduction by the same or similar pollinators.
The wingbeats of flying pollinators, including insects, birds
and bats, produce sound waves that travel rapidly through
air. If plants were able to receive such sounds and react to
1
School of Plant Sciences and Food Security, Tel-Aviv University, Tel-Aviv,
Israel
2
School of Zoology, Tel-Aviv University, Tel-Aviv, Israel
3
School of Mechanical Engineering, Tel-Aviv University, Tel-Aviv, Israel
*Correspondence: E-mail: lilach.hadany@gmail.com
†
These authors contributed equally.
§
These authors contributed equally.
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and
distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
Ecology Letters, (2019) 22: 1483–1492 doi: 10.1111/ele.13331

them rapidly, they could temporarily increase their advertise-
ment and/or reward when pollinators are likely to be present,
resulting in improved resource allocation. A possible plant
organ that could relay the airborne acoustic signal into a
response is the flower itself, especially in flowers with ‘bowl’
shape. If this is the case, we expect that part of the flower (or
the entire flower) would vibrate physically in response to the
airborne sound of a potential pollinator. We further predict
that nectar sugar concentration would increase in response to
the sound. None of these predictions have been tested before.
To test these predictions we used the beach evening primrose,
Oenothera drummondii, whose major pollinators are hawk-
moths (at night and early morning) and bees (at dusk and
morning) (Eisikowitch & Lazar 1987). We measured petal
vibration and nectar sugar concentration in response to
sounds. We analysed the effect of different sound frequencies,
including both pollinator recordings and synthetic sounds at
similar and different frequencies. We show that pollinator
sounds, and synthetic sound signals at similar frequencies,
cause vibration of the petals and evoke a rapid response –an
increase in the plant’s nectar sugar concentration.
MATERIALS AND METHODS
General
We exposed Oenothera drummondii plants to different sound
playbacks (see below) and measured the concentration of
sugar in their nectar. We compared plants’ response to differ-
ent sounds including pollinator recordings, synthetic sounds
in pollinator frequencies and in much higher frequencies, and
silence. To determine whether the playback sounds result in
physical vibration of the flower petals, we used laser vibrome-
try. To evaluate pollinator temporal distribution in the field,
we performed field observations.
Experimental setup: Measuring plant nectar response under
different treatments
The nectar response was tested in four different experiments
(see Table S1 for summary): Experiment 1a (n=90 flowers),
where the plants were grown outdoors in a natural environ-
ment, exposed to natural acoustic conditions in the summer.
The response was tested to the acoustic treatments (see
Sound signals and playbacks for details): ‘Silence’ –no
sound playback, ‘Low’ –playback of a low-frequency sound
signal with energy between 50 and 1000 Hz, covering the
range of pollinator wingbeat frequencies, and ‘High’ –play-
back of a high-frequency sound signal with energy between
158 and 160 kHz. This treatment served as a control for the
potential effect of the speaker’s electromagnetic field, which
was absent in the ‘Silence’ treatment. Experiment 1b
(n=167 flowers), where the plants were grown indoors in
the summer, and the response was measured to the previous
three stimuli plus a ‘Bee’ stimulus –playback of the record-
ings of a single hovering honeybee with a peak frequency of
200–500 Hz; Experiment 2 (n=298 flowers), where the
plants were grown indoors in the fall, and response was
tested for ‘Low’, ‘High’, and ‘Intermediate’ stimuli –
playback of a sound signal with energy between 34 and
35 kHz. To test the role of the flower itself (rather than
other parts of the plant exposed to sound) in the response,
the ‘Low’ and ‘High’ treatments were also tested for flowers
contained in glass jars ‘Low in Jar’ and ‘High in Jar’; and
Experiment 3 (n=112 flowers), where the plants were grown
indoors in the spring, and response was tested for ‘Low’ and
‘High’ stimuli.
In each experiment, plants were numbered, randomly
assigned to treatments and tested at a random order, alternat-
ing between the different treatments (see Table S1 for num-
bers per treatment). We sometimes used different flowers of
the same plant in more than one treatment; however, never in
the same day, nor in the same treatment. To measure a flow-
er’s response, it was emptied of nectar, and immediately
exposed to one of the treatments above. Its newly produced
nectar was extracted 3 min after the beginning of the treat-
ment (we had to wait 3 min for the amount of nectar accumu-
lated to be measurable by refractormeter). Sugar
concentration and nectar volume were quantified before and
after the treatment (for details see Nectar measurements meth-
ods, Fig. S1).
In the jar manipulation, we used six identical 1 L sound
proof glass jars, padded with acoustically isolating foam (see
Fig. S2). The jar’s ability to block sound was tested by posi-
tioning a calibrated microphone (GRAS, 40DP) inside it and
playing the ‘Low’ playback from a 10-cm distance (as in the
experiment). This measurement confirmed that jars reduced
sound intensity by 14 dB.
Sound signals and playbacks
In the nectar experiments, we used five signals, including bee
recordings, three artificial sound stimuli and silence. The arti-
ficial sound stimuli were generated using acoustic software
(Avisoft, Saslablite). The ‘Low’ frequency stimulus consisted
of a 10-s frequency modulated (FM) sound signal sweeping
from 1000 Hz to 50 Hz, covering the frequency range of the
wingbeat of natural pollinators. The ‘Intermediate’ frequency
stimulus consisted of a 10-s frequency modulated sound signal
sweeping from 35 to 34 kHz. The ‘High’ frequency stimulus
consisted of a 10-s frequency modulated sound signal sweep-
ing from 160 to 158 kHz, a frequency that is clearly out of
range for pollinator wingbeat. The ‘Bee’ stimulus was
recorded by positioning a calibrated microphone (GRAS,
40DP) and recording an individual honey bee (Apis mellifera)
from a distance of 10cm. The ‘Silence’ control treatment con-
sisted of no playback.
Acoustic playbacks were performed using a Vifa speaker
(XT25sc90-04, Vifa). D/A converter (Player 216-2, Avisoft
Bioacoustics) at a sampling rate of 500 kHz. All signals were
recorded using a calibrated microphone before playback to
validate their intensity. Playback intensity in the indoors
groups ‘Low’, ‘Bee’ and ‘Intermediate’ were set to resemble
the intensity of a bee hovering 10 cm above the plant, with a
peak sound pressure level of ca. 75 dBSPL relative to 20 µPa
at a distance of 10 cm. ‘Low’ playbacks in the outdoor group
had a peak pressure of c. 95 dB SPL (relative to 20 µPa at
10 cm).
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
1484 M. Veits et al. Letter

For control, we used either ‘Silence’, where no sound was
played, or the ‘High’ playback which had a weak intensity
(ca. 55dBSPL) but served as an additional ‘Silence’-like condi-
tion controlling for the electromagnetic field, absent in the
‘Silence’ control. All playbacks were played continuously for
3 min in all treatments, including the silent control. Each
playback was played to a group of 5–6 flowers, hovering over
each of them with a speaker for a period of 10 s each, return-
ing to the first flower at the end. The speakers were moved
from plant to plant for 3 min at a distance of c. 10 cm from
the nearest flower, mimicking a pollinator hovering around a
bush. Thus each flower was exposed to direct sound for
33.8 0.3 s on average (we validated that the number of
flowers per group had no effect on the significance of the
results, see Results). Such movement of the speakers was done
also in the ‘Silence’ treatment. In both ‘grown indoors’ and
‘grown outdoors’ experiments, all playbacks were performed
indoors: the flowers were brought into a silent room and were
treated there. To control for the accuracy of our playback sys-
tem, we recorded the playback of the bee sound using a cali-
brated microphone (GRASS 40DP). The recording obtained
that way (Fig. S3) was nearly identical to the original record-
ing used in the experiment.
The vibration experiments were performed with playbacks
of a bee (peak energy at 250–500 Hz) and a moth (peak
energy at c. 100 Hz and no energy above 400 Hz), and pure
tones at the peak frequencies of the signals described above:
‘Low’ (1 kHz), ‘Intermediate’ (35 kHz) and ‘High’ (160 kHz).
A control for the bee playback was performed with a live bee,
held by its legs with tweezers and hovering at several centime-
tres from the flower.
Nectar measurements
Nectar was extracted from all the flowers before treatments
using PTFE (Teflon) tubes (external diameter =0.9 mm, inter-
nal diameter =0.6 mm), followed by disposable 1 µL
glass capillaries for the nectar remaining after emptying by
the Teflon tubes. The treatments were applied immediately
after extraction. To avoid differences resulting from variation
in emptying times, we left a capillary inside the first emptied
flowers to assure that no new nectar has accumulated. When
the last flower was emptied, all capillaries were removed and
the treatment (‘High’, ‘Low’, ‘Bee’, ‘Intermediate’ or ‘Silence’)
started. Three minutes later, after the treatment ended, nectar
was drawn again from all the flowers. Sugar concentration in
each flower was measured by calibrated Bellingham-Stanley
low-volume Eclipse refractometers (0–50 Brix), where concen-
tration measurements are accurate in volumes as low as
0.2 µL. Three minutes allowed for enough nectar to accumu-
late in each flower (see Fig. S4, presenting nectar quantities)
sufficient for refractometer measurement.
Measuring petal vibration using laser vibrometry
To determine whether the playback sounds result in physical
vibration of the flower petals, we used laser vibrometry. This
method allows measuring minute physical vibrations through
Doppler shifts of a laser beam reflected from a vibrating
surface. To this end, the flowers were positioned on a wafer
prober (Karl Suss PSM6, Mitutoyo FS70L-S microscope) and
operated in ambient air. The motion of the petals was regis-
tered using a laser Doppler vibrometer (Polytec LDV, OFV-
5000 controller). The vibrometer was operated in the velocity
acquisition mode using VD-02 Velocity Output Decoder, (up
to 1.5 MHz bandwidth). The laser beam was focused on the
base of the petal (see Fig. S5) using the 95 long working dis-
tance lens of the microscope.
Signals from the LDV were fed into the oscilloscope KEY-
SIGHT DSOX2004A (70MHz, 1Mpts memory). We compared
flower vibrations in response to different playback frequencies
and in the absence of playback (‘Silence’) in a paired experimen-
tal design (within the same plant). To validate that the presence
of petals was crucial to the vibration, we also compared petal
vibration in intact flowers to vibration in intact petal of flowers
where some of the petals were removed (see Figs S5). To mea-
sure the actual vibration amplitude (i.e. displacement), we sub-
divided the measured velocity by 2
pf, where fis the frequency of
the oscillation. We used vibration models of objects with similar
shapes [both a beam and a circular thin plate (Blevins & Plun-
kett 1980)] to estimate the flower’s resonance vibration fre-
quency. The resonance frequency of an object is dictated by the
material properties, geometry and boundary conditions. For
flower size of c. 6 cm and thickness of c. 0.4 mm, we estimated
a fundamental mode frequency to be in the range of 100–
500 Hz. A measured density of c. 230 kg/m3 and the Young’s
modulus of c. 1 MPa adopted from Watanabe & Ziegler 2013
were used in calculations.
Monitoring pollinator temporal distribution in the field
In order to assess the temporal distribution of pollinators
around the plants in the field, two sets of field observations
were done on the Tel Aviv beach: (a) To test whether the
presence of a pollinator can indicate the vicinity of additional
pollinators, we videoed Oenothera drummondii plants during
the night. Seventeen plants were videoed over two nights for
4 h after sunset in summer 2017, using IR video cameras (Full
Spectrum POV Cam, GhostStop USA, resolution
1920 91080, 30 fps). Cameras were positioned at a distance
of 1–1.5 m from the plant. The videos were scrutinised manu-
ally using Matlab R2016a and VLC media player 2.2.4. A
moth passing within a distance of 1 m from a plant was
defined as ‘near the plant’. We then analysed the distribution
of intervals between these events (see Results). (b) To estimate
the time that a single pollinator spends close to an Oenothera
drummondii plant, the plants were visually observed during
the day, when it was possible to track the same individual
over time. Six plants were observed over 4 days for 3 h in
each day. A bee passing within a distance of 10 cm from a
plant was defined as ‘adjacent to the plant’ and the time it
spent within this distance was estimated.
Plants and growth conditions
Oenothera drummondii plants were propagated from grafts of
plants taken from Bet-Yanai coast, Israel. In all experiments,
irrespective of the plants growth conditions, the response of the
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
Letter Plants hear their pollinators 1485

plants to sound playback was tested indoors, in a quiet room.
For Experiment 1a (‘grown outdoors, summer 2014’), 200
plants were placed in 3 L pots and grown in the Tel Aviv
University Botanical Gardens in an outdoor setting. Flower
buds were covered with nets a day before the experiment, to
avoid pollination and nectar withdrawal by pollinators. For
Experiment 1b (‘grown indoors, summer 2015’), 100 plants were
placed in 0.5 L pots. We originally expected a difference in
sugar concentration between plants grown outdoors and
indoors, since a pilot experiment revealed that the nectar vol-
ume was dramatically different between these groups, with
more nectar in the flowers of plants grown outdoors. For
Experiment 2 (‘grown indoors, fall 2016’), 400 plants were
placed in 1.1 L pots, and for Experiment 3 (‘grown indoors,
spring 2016’), 200 plants were placed in 0.5 L pots. Experiments
1b, 2 and 3 used indoor-grown plants only, as outdoor plants
flower only in the summer. For all indoor experiments, the
plants were grown in a controlled growth room, at 27–28
degrees centigrade, with 16 h of artificial daylight, about
1 month prior to the beginning of the experiment. Altogether,
more than 650 flowers from these 900 plants were used in the
nectar experiments, and another c. 200 flowers in the laser
experiments (taken from the plants of Experiments 2 and 3). In
each experiment, only plants of the same age, season and pot
size were tested. See Table S1 for summary of the experiments.
Statistical analysis
Experiment 1: We performed a two-way ANOVA on log (sugar
concentration), including the treatment (‘Silence’, ‘High’, ‘Low’
or ‘Bee’) and group (1a, ‘grown indoors’ or 1b, ‘grown out-
doors’) variables. The group variable was not found to have a
significant effect (P=0.793). Therefore, data from both groups
were combined and the sugar concentration and nectar volume
between different treatments were compared. Shapiro–Wilks
test concluded significant deviation from normality (P<0.05)
in some of the cases (nectar volume data), so Wilcoxon rank-
sum was used for comparison. Within groups, the reported P-
values were adjusted for multiple comparisons using the Holm–
Bonferroni method.). Experiment 2: nectar traits (sugar concen-
tration and nectar volume) under different treatments were
compared using Wilcoxon rank-sum. To test the effect of
hydration status (days since watering), number of flowers in
group and time of day on our results, we used analysis of vari-
ance (ANOVA) model. We used log (sugar concentration) as the
dependent variable, and hydration status (or number of flowers
in group or time of day), treatment group, and their interaction
as predictors. Post-hoc P-values were calculated using a Tukey
HSD test. Constant variance assumption was corroborated
using Levene’s Test. All petal vibration levels were compared
using paired Wilcoxon test (comparing vibration levels of the
same flower under different treatments or petal removal). Polli-
nator temporal distribution in the field was compared using
paired Wilcoxon test as well.
RESULTS
We found that Oenothera drummondii flowers produced nectar
with significantly increased sugar concentration after exposure
to the playback of the natural sound of bee wingbeats (‘Bee’
treatment) in comparison with flowers exposed to either high
frequency sounds (‘High’) or no sound at all (‘silence’, Wil-
coxon P<0.01 for both comparisons, Fig. 1a and b). The
same result was obtained for artificial sounds with bee-like
frequencies (the ‘Low’ treatments, Fig. 1b middle and right).
The average sugar concentration was increased by a factor of
1.2 in flowers exposed to pollinator-like frequencies (‘Bee’ and
‘Low’ sound signals), in comparison with flowers exposed to
‘Silence’ or ‘High’, while no difference was observed between
flowers exposed to ‘High’ frequencies and flowers exposed to
the ‘Silence’ treatment. No difference in sugar concentration
was observed between experimental groups before the treat-
ment, and the volume of the nectar produced by the flowers
did not change significantly in the ‘Bee’ and ‘Low’ treatments
(Fig. S4), showing that the increase in sugar concentration in
these groups could not be attributed to a decrease in water
volume. Analysing the data using Student’s t-test of log-trans-
formed data resulted in similar significant results (P<0.001
for each of the comparisons between treatment (‘Low’ and
‘Bee’) and control (‘High’ and ‘Silence’).
To determine whether the pollinator sounds result in physi-
cal vibrations of the flower, we used laser vibrometry (see
Methods). Oenothera drummondii flowers vibrated mechani-
cally in response to the airborne sounds of a bee or a moth
recording (Fig. 2a, and Fig. S6 for moth sound spectra), oscil-
lating in velocities that have already been shown to elicit a
defence response in a plant that was mechanically moved in
such velocities (Appel & Cocroft 2014). The flowers also
vibrated in response to the hovering of a live bee similar to
their vibration in response to the bee’s playback (Fig. S7).
The amplitude of the mechanical vibrations (which reached
0.1 mm) depended on the presence of intact petals, and signif-
icantly decreased upon removal of petals (Fig. 2b, P<0.0005,
see Fig. S5 for details), suggesting that the petals either
directly receive, or serve to enhance the received signal.
To test the frequency specificity of both the physical vibra-
tion and the nectar response, we performed another indoors
experiment (Experiment 2, in the fall) in which we repeated
the use of the previous sound stimuli (Low and High) and
introduced another ‘Intermediate’ sound signal with a peak
frequency of 35 kHz (240 new flowers were used in this exper-
iment, in the fall, see Table S1). The flowers showed some fre-
quency specificity, both functionally and mechanically: they
vibrated significantly (paired Wilcoxon P<0.0001, n=21) in
response to sound signals of the ‘Low’ signal, 1 kHz, but not
in response to the peak frequency of an ‘Intermediate’ signal,
35 kHz (P>0.9, n=23), or the ‘High’ signal, 160 kHz
(P>0.9, n=21, see Fig. 2c and d black line). Similarly, the
flowers increased sugar concentration in response to ‘Low’
sound signals significantly (P<0.002, 2D red dotted line) in
comparison to the ‘Intermediate-’ or ‘High’-treated flowers.
‘Low’ sounds resulted in significantly higher sugar concentra-
tion than all other treatments (High, Intermediate, High in
jar, Low in jar) also when accounting for hydration status,
number of flowers in group, or the time of day (P<0.03).
We cannot report how the flower responds between 1 kHz
and 35 kHz. Differences between the four other treatment
groups were not significant. The ratios of post-treatment to
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
1486 M. Veits et al. Letter

pre-treatment concentration and vibration, per plant, revealed
an identical pattern: the ratios were significantly higher
(P<0.002 for concentration ratio, see Fig. S8A, P<e-07 for
vibration ratio) in plants exposed to ‘low’ sounds in compar-
ison with plants exposed to ‘high’ or ‘intermediate’ sounds
(Table S2). The volume of the nectar produced by the flowers
did not decrease in response to the ‘Low’ treatment
(Fig. S8B), so the increase in sugar concentration in this
group could not be attributed to a decrease in water volume.
In another experiment (Experiment 3, n=112 flowers, in
spring) where only ‘Low’ and ‘High’ stimuli were tested, there
was again significant increase in the sugar concentration in
response to the ‘Low’ stimuli (see Table S3). The plants
showed different flowering phenotype in different seasons
probably reflecting the season experienced before entering the
growth room: summer plants (Experiment 1) had larger flow-
ers with higher sugar concentration before treatment in com-
parison with either fall plants (Experiment 2) or spring (see
Table S4). Regardless, the major pattern –an increase in nec-
tar sugar concentration in response to pollinator sound play-
backs –was highly significant in all seasons (Fig. 2D,
1ATable S3).
To test one potential advantage of increasing reward within
minutes after a pollinator’s sound, we video-monitored the
distribution of pollinators near Oenothera drummondii flowers
in the field over two nights. We found that one pollinator fly-
ing in the vicinity of the plant –and producing sound in the
process –is a strong indication that another or same individ-
ual may be in the plant vicinity within a few minutes. Specifi-
cally, a pollinator was >9 times more common near the plant
if a pollinator was near the plant in the preceding 6 min, than
if no pollinator was around in the preceding 6 min (see
Fig. S9 and Monitoring pollinator visitations methods). This
activity pattern of the pollinators suggests that a response of
the plant within minutes of the sound could more often be rel-
evant to pollinators than a response not preceded by a sound.
We further quantified the time pollinators tend to stay next to
Oenothera drummondii flowers in the field (Methods). Two
species of bees were observed around the flowers, and the
observed time spent within a distance that would allow the
bee sound to generate flower vibrations (‘adjacent to the
plant’) were 27.8 7.7 s for honey bees (n=44), and
38.9 11.8 sec for carpenter bees (n=23), see Fig. S10. In
reality, plants may of course be exposed to longer sound stim-
uli due to multiple bee pass one after the other. Notably, as
our playback lasted 3 min and we had six plants at each ses-
sion, each plant was exposed to 30 s of direct playback, on
average.
Figure 1 Flowers respond rapidly to pollinator sounds by producing sweeter nectar (a). Mean sugar concentration under the different treatments in plants
grown outdoors (dashed black) and indoors (dotted red). Mean sugar concentration across both indoors and outdoors groups differed significantly
(P<0.01) between flowers exposed to frequencies below 1 kHz (sugar concentration 19.8% 0.6, n= 72 and 19.1% 0.7, n= 42 for ‘Low’ and ‘Bee’
after 3 min respectively), compared to flowers exposed to ‘Silence’ or ‘High’ frequency sound (16.3% 0.5, n= 71, and 16.0% 0.4, n=72
respectively). Insert shows a flower of Oenothera drummondii. (b) Spectra (frequency content) of the playback signals used in the experiment. Both ‘Bee’
and ‘Low’ signals contain most energy below 1000 Hz, while the ‘High’ control peaked at c. 159 000 Hz.
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
Letter Plants hear their pollinators 1487

Finally, to validate the importance of the flower itself as an
organ responsible for reception of pollinator sounds, we ran
another experiment. When the flowers (but not the stem or
leaves) were covered with glass jars that blocked sound (see
Fig. S2), then the ‘Low’ playback had no effect on the sugar
concentration: For flowers enclosed in jars, there was no sig-
nificant difference between exposure to ‘Low’ treatment and
exposure to ‘High’ (P>0.64, for n=58 and 59 flowers
respectively), and none of these groups differed significantly
from the no jar ‘High’ treatment, that served as a control
Figure 2 (a) Flowers vibrate mechanically in response to airborne sound of a pollinator. Top: Left –time signal of a honey bee sound signal (airborne
signal recorded using a microphone). Right –time signal of a flying Plodia interpunctella male moth (the signal’s spectrum peaks at c. 100 Hz, see Fig. S6).
Bottom: Mechanical vibration recorded in an Oenothera drummondii flower in response to the playback of the bee (left) and moth (right) sound signals. (b)
Vibration velocity in response to the bee signal depended on the presence of petals: a significantly stronger vibration was recorded when all four petals
were intact in comparison to when flowers were trimmed and had only 1 or 0.5 petals (paired Wilcoxon, P<0.0005 for the comparison between four and
one petal and P<0.005 for the comparison between 4 and 0.5 petals). (c) Flowers vibrated in response to playback of low frequencies around 1 kHz (left)
while they did not vibrate above background noise to playbacks at higher frequencies of c. 35 kHz (right). Top: the time that the playback was ‘on’.
Bottom: Vibration time signals of the flowers. (d) Frequency specificity in both vibration and sugar concentration response. The flowers vibrated (dashed
black) significantly more than background noise in response to sound signals in low frequencies around 1 kHz (paired Wilcoxon P<0.0001, n= 21) but
not in response to high frequencies around 160 kHz (P>0.6, n= 23) or to intermediate frequencies around 35 kHz (P>0.9, n= 21); The flowers also
increased sugar concentration (dotted red line) in response to ‘Low’ signals significantly more than in response to the ‘Intermediate’ signal presented in the
inset (P<0.002), or to the ‘High’ signal serving as control (P<0.0001). (Sugar concentration 15.9% 0.57, n= 81, 12.8% 0.7, n= 49, and
12.3% 0.77, n= 51, for Low, High and Intermediate respectively). Inset shows the spectrum of the ‘Intermediate’ playback signal used in the nectar
experiment. (e) Summary of experimental results. Flowers vibrate in response to airborne sound at pollinator’s frequency range, and increase nectar sugar
concentration (right panel). Glass covered flowers do not respond (middle), suggesting that the flower serves as the plant’s ‘ear’. The flowers responseis
frequency specific, and they do not vibrate or respond to frequencies around 35 kHz (left).
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
1488 M. Veits et al. Letter

(P>0.49, n=49 flowers, see Table S2). Mean nectar volume
did not differ significantly between any of the groups
(Table S2). These results suggest that flowers are important
for hearing pollinators, but we cannot exclude the possibility
that other parts of the plant may also respond to pollinator
sounds, resulting in nectar response later than 3 min, or that
other parts of the plant may serve as sensory organs for
sounds at other frequencies.
DISCUSSION
We found that plants respond rapidly to specific airborne
sound frequencies (Figs 1, 2d) in a way that could potentially
increase their chances of pollination, and that flowers can
serve as sound sensing organs (Fig. 2). Consistent results were
obtained in four independent experiments (Table S1) with
over 650 flowers in total. The flowers responded similar to bee
wingbeat sounds and to artificial sound waves that were simi-
lar in their frequency spectrum but differed greatly in their
temporal pattern, suggesting that the frequency of the sound
is sufficient to elicit a response. The flowers responded
rapidly, within 3 min. The concentration of sugar in the nec-
tar produced following the exposure to sound increased by a
factor of 1.2 on average.
Bees have been shown to be capable of perceiving differ-
ences in sugar concentration, as small as 1-3% (Afik et al.
2006; Whitney et al. 2008). Thus, even if the new sugar-rich
nectar is diluted by lower concentration nectar already present
in the flower, the bees would be able to detect the difference
in many cases.
Increased sugar concentration can enhance the learning pro-
cess of the pollinators, and facilitate the pollinator constancy
–the tendency to visit flowers from the same species (Cnaani
et al. 2006) –thus increasing the effectiveness of pollination.
Enhanced reward can also increase visit duration, further
enhancing pollination efficiency (Manetas & Petropoulou
2000; Brandenburg et al. 2012). This is not without caveats:
too high sugar concentration could result in too viscous nec-
tar for some pollinators, but the values measured here are
below the optimum for both bees and moths (Josens & Farina
2001; Krenn 2010; Kim et al. 2011), suggesting that the polli-
nators can benefit from the increased concentration. It may
also result in a higher number of flowers visited per plant,
possibly leading to geitonogamous selfing (Klinkhamer & de
Jong 1993; Hodges 1995; Dafni et al. 2005). Yet, if only part
of the flowers in the plant carry enhanced rewards –for exam-
ple due to depletion –then the response could result in
increased variation in nectar standing crop within the plant,
encouraging the pollinators to move to the next plant and
facilitating outcrossing (Ott et al. 1985; Biernaskie & Cartar
2004; Pyke 2016).
A response within 3 min is advantageous when pollinators
move between nearby flowers, or when the presence of one
pollinator is a good predictor of other nearby pollinators,
such as in bees (Goulson 1999; Slaa et al. 2003) and in moths
according to our field observations (Fig. S9). Such a response
would allow the plant to identify the beginning and intensity
of pollinator activity which can differ from day-to-day due to
various factors such as weather conditions (Corbet et al.
1993). The plant could then switch to an increased sugar pro-
duction mode, in order to reward the first actual visitors.
Rapidly increasing nectar sugar concentration would be
advantageous also in the case of a sporadic pollinator remain-
ing in the area of the plant for a long time. Note that in a
plant like the evening primrose, characterised by multiple
flowers (dozens of flowers in a mature bush), the response to
the sound of a nearby pollinator could be beneficial even if
the pollinator avoids visiting the specific flowers that had
recently been visited (Giurfa & N
u~
nez 1992; Goulson et al.
1998), since it can still visit other flowers of the same plant.
Other pollinators actually prefer occupied or recently occupied
food sources (Schmidt et al. 2003; Kawaguchi et al. 2006;
Lihoreau et al. 2016), and might especially benefit from
enhanced refilling.
The plants responded to sound frequencies characteristic of
pollinators’ wingbeat (Figs 1, 2). How could frequency speci-
ficity be attained? We estimated the resonance frequency of
the evening primrose petal to be a few hundred Hz based on
vibration models developed for objects with similar shapes
(Blevins & Plunkett 1980). This is close to the sound frequen-
cies typically generated by bee and moth wingbeat. Further-
more, the flower should vibrate mostly around the resonance
frequency, and vibrate less in response to higher or lower fre-
quencies. Indeed, we observed that the flower-filtered frequen-
cies above 350 Hz produced by the hovering bee, responding
less to these frequencies (Figure S7). This could explain how
the flower increased sugar concentration in its nectar only in
response to low frequencies. Moreover, this frequency speci-
ficity might also explain, in theory, how the flower filters
wind-induced vibrations, which are typically at lower frequen-
cies (Appel & Cocroft 2014).
The current work is the first step in a new field, and can be
extended in several ways. First, the response to sound can be
further studied in the wild, on the background of other natu-
ral sounds. Second, all our nectar measurements were per-
formed by first emptying the flower and then measuring
refilled nectar. Testing the response to sound without prior
manipulation will be more realistic (Corbet 2003), but would
require large sample sizes due to the high variation in the nec-
tar standing crop present in the model species. Third, the
actual functionality of the response has yet to be tested –that
is, do pollinators indeed prefer plants exposed to sound, and
to what extent? Fourth, we tested the response to sound in a
single plant species. Additional species might reveal different
responses according to their specific ecologies (e.g. bat polli-
nated plants may respond to different frequencies).
The petal vibrations that we measure could be picked up by
mechanoreceptors, which are common in plants (Monshausen
& Gilroy 2009), and have been shown to respond to vibra-
tions with similar velocities (Appel & Cocroft 2014). We
hypothesise that the flower serves as an external ‘ear’ in terms
of receiving pollinator airborne sounds by the plant. We posit
that the petals of other flowering species could have evolved
to detect sound, similar to our findings in Oenothera drum-
mondii. The resonance frequency of a flower will be dictated
by its mechanical parameters: size, shape and density, which
could be under natural selection. If plant responses to air-
borne acoustic signals are indeed adaptive in the context of
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
Letter Plants hear their pollinators 1489

pollination, we expect plants with ‘noisy’ pollinators –such as
bees, moths and birds –to have evolved large ear-like flowers
with proper mechanical parameters making them sensitive to
the sounds of their pollinators.
Much is known about the response of pollinators to plant sig-
nalling from a distance (Patiny 2011; Schaefer & Ruxton 2011).
In contrast, the response of plants to pollinators from a distance
has never been demonstrated. The implications of such a
response to the ecological system might be far reaching, since
pollination is critical to the survival of many plant species,
including many agriculturally important crops (Kremen et al.
2002; Fægri & Van der Pijl 2013). Plant response to sound could
allow bidirectional feedback between pollinators and plants,
which can improve the synchronisation between them, lowering
nectar waste and potentially improving the efficiency of pollina-
tion in changing environments. These advantages can be dimin-
ished in very noisy environments, suggesting possible sensitivity
of pollination to external noises, including antropogenic ones.
Finally, plants’ ability to sense airborne sounds has implications
way beyond pollination: plants could potentially sense and
respond to herbivores’ airborne sounds, other animals, and pos-
sibly other plants (Khait et al., 2018).
ACKNOWLEDGEMENTS
We thank Prof. Dan Eisikowitch, Prof. Amram Eshel, Dr.
Iftach Vaknin and Dr. Yael Mendelik for contributing nectar
measuring equipment; Stella Lulinski for help with the laser
experiments; Stav Hen, Dorin Cohn, Oren Rabinowitz and
Ran Perelman for help with the nectar experiments; Prof. Nir
Ohad, Dr. Tuvik Beker and Prof. Judith Berman for com-
ments on the manuscript. The research has been supported in
part by Bikura 2308/16 (LH, YS, YY), Bikura 2064/18 (LH,
YS, YY), ISF 1568/13 (LH), by the Smaller Winnikow fellow-
ship (MV) and by the Manna Center Program for Food
Safety and Security fellowships (IK, UO).
AUTHOR CONTRIBUTIONS
LH conceived the idea. LH, YY, YS and DAC designed the
research. MV, IK, UBD, PE, AB, AK, DP, IR, AG, KS, RS,
and EZ performed the experiments. IK, UO, MV, YY and
LH analysed the data. YY LH and SK supervised the acous-
tic experiments. YS and LH supervised the nectar experi-
ments. LH, YY and YS contributed equally to the study. All
the authors discussed the results and took part in writing the
manuscript.
DATA ACCESSABILITY STATEMENT
Data available from the Dryad Digital Repository: https://
doi.org/10.5061/dryad.6n5h0pb.
REFERENCES
Afik, O., Dag, A., Kerem, Z. & Shafir, S. (2006). Analyses of avocado
(Persea americana) nectar properties and their perception by honey
bees (Apis mellifera). J. Chem. Ecol., 32, 1949–1963.
Appel, H. & Cocroft, R. (2014). Plants respond to leaf vibrations caused
by insect herbivore chewing. Oecologia, 175, 1257–1266.
Arimura, G.-I., Ozawa, R., Shimoda, T., Nishioka, T., Boland, W. &
Takabayashi, J. (2000). Herbivory-induced volatiles elicit defence genes
in lima bean leaves. Nature, 406, 512–515.
Baldwin, I.T., Halitschke, R., Paschold, A., von Dahl, C.C. & Preston,
C.A. (2006). Volatile signaling in plant-plant interactions: "Talking
Trees" in the genomics Era. Science, 311, 812–815.
Biernaskie, J. & Cartar, R. (2004). Variation in rate of nectar production
depends on floral display size: a pollinator manipulation hypothesis.
Funct. Ecol., 18, 125–129.
Blevins, R.D. & Plunkett, R. (1980). Formulas for natural frequency and
mode shape. American Society of Mechanical Engineers, 47, 461–462.
Bochu, W., Jiping, S., Biao, L., Jie, L. & Chuanren, D. (2004). Soundwave
stimulation triggers the content change of the endogenous hormone of
the Chrysanthemum mature callus. Colloids Surf., B, 37, 107–112.
Braam, J. (2005). In touch: plant responses to mechanical stimuli. New
Phytol., 165, 373–389.
Brandenburg, A., Kuhlemeier, C. & Bshary, R. (2012). Hawkmoth
pollinators decrease seed set of a low-nectar Petunia axillaris line
through reduced probing time. Curr. Biol., 22, 1635–1639.
Chamovitz, D. (2012). What a plant knows: A field guide to the senses of
your garden-and beyond. Scientific American/Farrar, Straus and Giroux,
New York.
Choi, B., Ghosh, R., Gururani, M.A., Shanmugam, G., Jeon, J., Kim, J.,
et al. (2017). Positive regulatory role of sound vibration treatment in
Arabidopsis thaliana against Botrytis cinerea infection. Sci. Rep.,7,2527.
Chory, J. (2010). Light signal transduction: an infinite spectrum of
possibilities. Plant J., 61, 982–991.
Cnaani, J., Thomson, J.D. & Papaj, D.R. (2006). Flower choice and
learning in foraging bumblebees: effects of variation in nectar volume
and concentration. Ethology, 112, 278–285.
Corbet, S.A. (2003). Nectar sugar content: estimating standing crop and
secretion rate in the field. Apidologie, 34, 1–10.
Corbet, S.A., Fussell, M., Ake, R., Fraser, A., Gunson, C., Savage, A.,
et al. (1993). Temperature and the pollinating activity of social bees.
Ecol. Entomol., 18, 17–30.
Dafni, A., Kevan, P.G. & Husband, B.C. (2005). Practical pollination
biology. Enviroquest Ltd, Cambridge.
Darwin, C. (1892). The formation of vegetable mould through the action
of worms: with observations on their habits. Appleton.
Davis, A., Rubinstein, M., Wadhwa, N., Mysore, G.J., Durand, F. &
Freeman, W.T. (2014). The visual microphone: passive recovery of
sound from video. ACM Trans. Graph., 33, 79.
Eisikowitch, D. & Lazar, Z. (1987). Flower change in Oenothera
drummondii Hooker as a response to pollinators’ visits. Bot. J. Linn.
Soc., 95, 101–111.
Faegri, K. & Van Der Pijl, L. (1979). The principles of pollination ecology.
Pergamon Press, New York.
Fægri, K. & Van der Pijl, L. (2013). Principles of pollination ecology.
Elsevier, London.
Gagliano, M., Mancuso, S. & Robert, D. (2012). Towards understanding
plant bioacoustics. Trends Plant Sci., 17, 323–325.
Gagliano, M., Grimonprez, M., Depczynski, M. & Renton, M. (2017).
Tuned in: plant roots use sound to locate water. Oecologia, 184(1),
151–160.
Galen, C. (1999). Flowers and enemies: predation by nectar-thieving ants
in relation to variation in floral form of an alpine wildflower,
Polemonium viscosum. Oikos, 85, 426–434.
Galetto, L., Araujo, F.P., Grilli, G., Amarilla, L.D., Torres, C. & Sazima,
M. (2018). Flower trade-offs derived from nectar investment in female
reproduction of two Nicotiana species (Solanaceae). Acta Botanica
Brasilica, 32, 473–478.
Ghosh, R., Mishra, R.C., Choi, B., Kwon, Y.S., Bae, D.W., Park, S.-C.,
et al. (2016). Exposure to sound vibrations lead to transcriptomic,
proteomic and hormonal changes in Arabidopsis. Sci. Rep., 6, 33370.
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
1490 M. Veits et al. Letter

Ghosh, R., Gururani, M.A., Ponpandian, L.N., Mishra, R.C., Park, S.-
C., Jeong, M.-J., et al. (2017). Expression analysis of sound vibration-
regulated genes by touch treatment in Arabidopsis. Frontiers in Plant
Science, 8, 100.
Giurfa, M. & N
u~
nez, J.A. (1992). Honeybees mark with scent and reject
recently visited flowers. Oecologia, 89, 113–117.
Goulson, D. (1999). Foraging strategies of insects for gathering nectar
and pollen, and implications for plant ecology and evolution. Perspect.
Plant Ecol. Evol. Syst., 2, 185–209.
Goulson, D., Hawson, S.A. & Stout, J.C. (1998). Foraging bumblebees
avoid flowers already visited by conspecifics or by other bumblebee
species. Anim. Behav., 55, 199–206.
Hassanien, R.H.E., Hou, T.-Z., Li, Y.-F. & Li, B.-M. (2014). Advances in
effects of sound waves on plants. Journal of Integrative Agriculture, 13,
335–348.
Heil, M. & Bueno, J.C.S. (2007). Within-plant signaling by volatiles leads
to induction and priming of an indirect plant defense in nature. Proc.
Natl Acad. Sci., 104, 5467–5472.
Herrera, C.M., Garc
ıa, I.M. & P
erez, R. (2008). Invisible floral larcenies:
microbial communities degrade floral nectar of bumble bee-pollinated
plants. Ecology, 89, 2369–2376.
Hodges, S.A. (1995). The influence of nectar production on hawkmoth
behavior, self pollination, and seed production in Mirabilis multiflora
(Nyctaginaceae). Am. J. Bot., 82, 197–204.
Irwin, R.E., Bronstein, J.L., Manson, J.S. & Richardson, L. (2010).
Nectar robbing: ecological and evolutionary perspectives. Annu. Rev.
Ecol. Evol. Syst., 41, 271–292.
Jiao, Y., Lau, O.S. & Deng, X.W. (2007). Light-regulated transcriptional
networks in higher plants. Nat. Rev. Genet., 8, 217–230.
Josens, R.B. & Farina, W.M. (2001). Nectar feeding by the hovering
hawk moth Macroglossum stellatarum: intake rate as a function of
viscosity and concentration of sucrose solutions. J. Comp. Physiol. A.,
187, 661–665.
Jung, J., Kim, S.-K., Kim, J.Y., Jeong, M.-J. & Ryu, C.-M. (2018).
Beyond chemical triggers: evidence for sound-evoked physiological
reactions in plants. Front. Plant Sci., 9, 25.
Karban, R. (2015). Plant sensing and communication. University of
Chicago Press, Chicago.
Karban, R., Yang, L.H. & Edwards, K.F. (2014). Volatile communication
between plants that affects herbivory: a meta-analysis. Ecol. Lett., 17,
44–52.
Kawaguchi, L., Ohashi, K. & Toquenaga, Y. (2006). Do bumble bees
save time when choosing novel flowers by following conspecifics? Funct.
Ecol., 20, 239–244.
Khait, I., Sharon, R., Perelman, R., Boonman, A., Yovel, Y. & Hadany,
L. (2018). Plants emit remotely detectable ultrasounds that can reveal
plant stress bioRxiv/2018/507590.
Kim, W., Gilet, T. & Bush, J.W. (2011). Optimal concentrations in nectar
feeding. Proc. Natl Acad. Sci.,108, 16618–16621
Kim, J.Y., Lee, S.I., Kim, J.A., Park, S.-C. & Jeong, M.-J. (2017). Sound
waves increases the ascorbic acid content of alfalfa sprouts by affecting
the expression of ascorbic acid biosynthesis-related genes. Plant
Biotechnol. Rep., 11, 355–364.
Klinkhamer, P.G. & de Jong, T.J. (1993). Attractiveness to pollinators: a
plant’s dilemma. Oikos, 180–184.
Kremen, C., Williams, N.M. & Thorp, R.W. (2002). Crop pollination
from native bees at risk from agricultural intensification. Proc. Natl
Acad. Sci., 99, 16812–16816.
Krenn, H.W. (2010). Feeding mechanisms of adult Lepidoptera: structure,
function, and evolution of the mouthparts. Annu. Rev. Entomol., 55,
307–327.
Lihoreau, M., Chittka, L. & Raine, N.E. (2016). Monitoring flower
visitation networks and interactions between pairs of bumble bees in a
large outdoor flight cage. PLoS ONE, 11, e0150844.
L
opez-Ribera, I. & Vicient, C.M. (2017). Drought tolerance induced by
sound in Arabidopsis plants. Plant Signaling & Behav., 12, e1368938.
De Luca, P.A. & Vallejo-Mar

ın, M. (2013). What’s the ‘buzz’ about? The
ecology and evolutionary significance of buzz-pollination. Curr. Opin.
Plant Biol., 16, 429–435.
Manetas, Y. & Petropoulou, Y. (2000). Nectar amount, pollinator visit
duration and pollination success in the mediterranean shrub Cistus
creticus.Ann. Bot., 86, 815–820.
Monshausen, G.B. & Gilroy, S. (2009). Feeling green: mechanosensing in
plants. Trends Cell Biol., 19, 228–235.
Monshausen, G.B. & Haswell, E.S. (2013). A force of nature: molecular
mechanisms of mechanoperception in plants. J. Exp. Bot., 64, 4663–
4680.
Ollerton, J., Winfree, R. & Tarrant, S. (2011). How many flowering
plants are pollinated by animals? Oikos, 120, 321–326.
Ordano, M. & Ornelas, J.F. (2005). The cost of nectar replenishment in
two epiphytic bromeliads. J. Trop. Ecol., 21, 541–547.
Ornelas, J.F. & Lara, C. (2009). Nectar replenishment and pollen receipt
interact in their effects on seed production of Penstemon roseus.
Oecologia, 160, 675–685.
Ott, J.R., Real, L.A. & Silverfine, E.M. (1985). The effect of nectar
variance on bumblebee patterns of movement and potential gene
dispersal. Oikos, 45, 333–340.
Pappers, S.M., de Jong, T.J., Klinkhamer, P.G. & Meelis, E. (1999).
Effects of nectar content on the number of bumblebee approaches and
the length of visitation sequences in Echium vulgare (Boraginaceae).
Oikos, 87, 580–586.
Patiny, S. (2011). Evolution of plant-pollinator relationships. Cambridge
University Press, Cambridge.
Pleasants, J.M. & Chaplin, S.J. (1983). Nectar production rates of
Asclepias quadrifolia: causes and consequences of individual variation.
Oecologia, 59, 232–238.
Pyke, G.H. (1991). What does it cost a plant to produce floral nectar?
Nature, 350, 58–59.
Pyke, G.H. (2016). Floral nectar: Pollinator attraction or manipulation?
Trends Ecol. Evol., 31, 339–341.
Rebar, D., H€
obel, G. & Rodr
ıguez, R.L. (2012). Vibrational playback by
means of airborne stimuli. J. Exp. Biol., 215, 3513–3518.
Schaefer, H.M. & Ruxton, G.D. (2011). Plant-animal communication.
Oxford University Press, Oxford
Schmidt, V.M., Zucchi, R. & Barth, F.G. (2003). A stingless bee marks
the feeding site in addition to the scent path (Scaptotrigona aff.
depilis). Apidologie, 34, 237–248.
Slaa, E.J., Wassenberg, J. & Biesmeijer, J.C. (2003). The use of field–
based social information in eusocial foragers: local enhancement among
nestmates and heterospecifics in stingless bees. Ecol. Entomol., 28, 369–
379.
Slack, A. (2000). Carnivorous plants. The MIT Press, Cambridge, MA.
Southwick, E. (1984). Photosynthate allocation to floral nectar: a
neglected energy investment. Ecology, 65, 1775–1779.
Stout, J.C. & Goulson, D. (2002). The influence of nectar secretion rates
on the responses of bumblebees (Bombus spp.) to previously visited
flowers. Behav. Ecol. Sociobiol., 52, 239–246.
Takahashi, H., Suge, H. & Kato, T. (1991). Growth promotion by
vibration at 50 Hz in rice and cucumber seedlings. Plant Cell Physiol.,
32, 729–732.
Telewski, F.W. (2006). A unified hypothesis of mechanoperception in
plants. Am. J. Bot., 93, 1466–1476.
Vannette, R.L., Gauthier, M.-P.L. & Fukami, T. (2013). Nectar bacteria,
but not yeast, weaken a plant–pollinator mutualism. Biol. Sci., 280,
20122601.
Watanabe, K. & Ziegler, F. (2013). Dynamics of advanced materials and
smart structures. Springer Science & Business Media, Dordrecht.
Whitney, H.M., Dyer, A., Chittka, L., Rands, S.A. & Glover, B.J. (2008).
The interaction of temperature and sucrose concentration on foraging
preferences in bumblebees. Naturwissenschaften, 95, 845–850.
Willmer, P. (2011). Pollination and floral ecology. Princeton University
Press, Princeton, NJ.
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
Letter Plants hear their pollinators 1491

Xiujuan, W., Bochu, W., Yi, J., Chuanren, D. & Sakanishi, A. (2003).
Effect of sound wave on the synthesis of nucleic acid and protein in
chrysanthemum. Colloids Surf., B, 29, 99–102.
Yi, J., Bochu, W., Xiujuan, W., Daohong, W., Chuanren, D., Toyama,
Y., et al. (2003). Effect of sound wave on the metabolism of
chrysanthemum roots. Colloids Surf., B, 29, 115–118.
SUPPORTING INFORMATION
Additional supporting information may be found online in
the Supporting Information section at the end of the article.
Editor, Christoph Scherber
Manuscript received 2 December 2018
First decision made 7 January 2019
Second decision made 26 March 2019
Manuscript accepted 10 April 2019
©2019 The Authors Ecology Letters published by CNRS and John Wiley & Sons Ltd
1492 M. Veits et al. Letter