Shakers and head bangers: differences in sonication behavior between Australian Amegilla murrayensis (blue-banded bees) and North American Bombus impatiens (bumblebees)

Abstract

Many bees collect pollen by grasping the anthers of a flower and vibrating their flight muscles at high frequencies—a behavior termed sonication, or buzz-pollination. Here we compare buzz-pollination on Solanum lycopersicum (cherry tomatoes) by two bees that fill similar niches on different continents—in Australia, Amegilla murrayensis (blue-banded bee), and in North America, Bombus impatiens (bumblebee). We collected audio recordings of buzz-pollination and quantified the frequency and length of buzzes, as well as the total time spent per flower. We found that A. murrayensis buzzes at significantly higher frequencies (~350 Hz) than B. impatiens (~240 Hz) and flaps its wings at higher frequencies during flight. There was no difference in the length of a single buzz, but A. murrayensis spent less time on each flower, as B. impatiens buzzed the flower several times before departing, whereas A. murrayensis typically buzzed the flower only once. High-speed videos of A. murrayensis during buzz-pollination revealed that its physical interaction with the flower differs markedly from the mechanism described for Bombus and other bees previously examined. Rather than grasping the anther cone with its mandibles and shaking, A. murrayensis taps the anther cone with its head at the high buzzing frequencies generated by its flight muscles. This unique behavior, combined with its higher buzzing frequency and reduced flower visit duration, suggests that A. murrayensis may be able to extract pollen more quickly than B. impatiens, and points to the need for further studies directly comparing the pollination effectiveness of these species.

Full text

ORIGINAL PAPER
Shakers and head bangers: differences in sonication behavior
between Australian Amegilla murrayensis (blue-banded bees)
and North American Bombus impatiens (bumblebees)
Callin M. Switzer
1
•Katja Hogendoorn
2
•Sridhar Ravi
3
•Stacey A. Combes
4
Received: 2 November 2015 / Accepted: 20 November 2015 / Published online: 1 December 2015
Springer Science+Business Media Dordrecht 2015
Abstract Many bees collect pollen by grasping the
anthers of a flower and vibrating their flight muscles at high
frequencies—a behavior termed sonication, or buzz-polli-
nation. Here we compare buzz-pollination on Solanum
lycopersicum (cherry tomatoes) by two bees that fill similar
niches on different continents—in Australia, Amegilla
murrayensis (blue-banded bee), and in North America,
Bombus impatiens (bumblebee). We collected audio
recordings of buzz-pollination and quantified the frequency
and length of buzzes, as well as the total time spent per
flower. We found that A. murrayensis buzzes at signifi-
cantly higher frequencies (*350 Hz) than B. impatiens
(*240 Hz) and flaps its wings at higher frequencies during
flight. There was no difference in the length of a single
buzz, but A. murrayensis spent less time on each flower, as
B. impatiens buzzed the flower several times before
departing, whereas A. murrayensis typically buzzed the
flower only once. High-speed videos of A. murrayensis
during buzz-pollination revealed that its physical interac-
tion with the flower differs markedly from the mechanism
described for Bombus and other bees previously examined.
Rather than grasping the anther cone with its mandibles
and shaking, A. murrayensis taps the anther cone with its
head at the high buzzing frequencies generated by its flight
muscles. This unique behavior, combined with its higher
buzzing frequency and reduced flower visit duration, sug-
gests that A. murrayensis may be able to extract pollen
more quickly than B. impatiens, and points to the need for
further studies directly comparing the pollination effec-
tiveness of these species.
Keywords Sonication Solanum Vibration 
Pollination Native bees
Introduction
Over 200,000 plant species depend on insects for pollina-
tion (Buchmann 1983). Pollinating insects often consume
both nectar and pollen, and they transfer pollen grains
among plants as they travel from flower to flower, an
essential step in the reproduction of many plants. Under-
standing the physical interactions between plants and insect
pollinators (primarily bees) can provide insight into the
requirements and evolution of these critical plant–pollina-
tor relationships.
Although considered a mutualistic relationship, the
interaction between plants and pollinating insects is not
entirely without conflict. The conflict arises because the
ideal behavior of the pollinator is different from the pol-
linator’s perspective versus the plant’s perspective (Gegear
and Laverty 2001). Bees attempt to expend the least pos-
sible energy for the greatest reward; bumblebees forage for
Handling Editor: Heikki Hokkanen.
Electronic supplementary material The online version of this
article (doi:10.1007/s11829-015-9407-7) contains supplementary
material, which is available to authorized users.
&Callin M. Switzer
callin.switzer@gmail.com
1
Department of Organismic and Evolutionary Biology,
Harvard University, Cambridge, MA, USA
2
School of Agriculture, Food and Wine, The University of
Adelaide, Adelaide, SA, Australia
3
School of Aerospace, Mechanical, and Manufacturing
Engineering, RMIT University, Melbourne, VIC, Australia
4
Department of Neurobiology, Physiology, and Behavior,
University of California, Davis, Davis, CA, USA
123
Arthropod-Plant Interactions (2016) 10:1–8
DOI 10.1007/s11829-015-9407-7

pollen in a manner that increases their probability of
maximizing their net energy intake (Zimmerman 1982).
Plants, on the other hand, would benefit most if pollinators
moved sequentially among flower of the same species—a
strategy that an optimally foraging pollinator would rarely
use (Gegear and Laverty 2001). One evolutionary ‘‘strat-
egy’’ for increasing a plant’s reproductive success is to
dispense only a little pollen at a time, ensuring that its
flowers are visited multiple times and that pollinators must
visit multiple flowers to obtain sufficient pollen (Harder
and Thomson 1989). Plants may also benefit from ‘‘messy’’
bees that cannot clean all the pollen off their bodies, since
this excess pollen is not consumed by the pollinator, and is
more likely to be transferred to different flowers that the
bee visits subsequently; outcrossing plants require a polli-
nator that accumulates pollen on its body where it has a
high chance of fertilizing a conspecific plant ovule (Gegear
and Laverty 2001).
These evolutionary strategies are particularly evident in
the approximately 20,000 insect-pollinated plants (*8%
of angiosperms) that have evolved poricidal anthers
(Buchmann 1983)—anthers with only small pores through
which pollen is released. Poricidal anthers restrict direct
access to pollen (De Luca and Vallejo-Marı
´n2013; Harder
and Thomson 1989), helping to limit the amount of pollen
that bees can collect during a visit and depositing pollen in
locations on the bees’ bodies that are poorly groomed.
Throughout areas with temperate climates, Bombus spp.
(bumblebees) play a vital role in pollinating plants with
poricidal anthers, as they are capable of performing soni-
cation, or buzz-pollination, to release pollen that is largely
inaccessible to insects that do not perform this behavior
(e.g., honeybees) (King and Buchmann 2003). In warmer
areas, species belonging to other taxa, e.g., Xylocopa
(Hogendoorn et al. 2000) and Amegilla (Hogendoorn et al.
2006), perform buzz-pollination.
Buzz-pollination has been well described in Bombus
spp.: The bee lands on a flower, curls her abdomen around
the anther tips while grasping the anthers with her mand-
ibles, and then uses her flight muscles to vibrate her body
without flapping the wings (King et al. 2006). These
vibrations are transmitted through the head and body to the
flower, and pollen is released from the pores onto the bee’s
body (De Luca and Vallejo-Marı
´n2013; Harder and Bar-
clay 1994; King and Buchmann 2003). Due to the bee’s
position on the anther during sonication, pollen is deposited
onto her ventral body surface, and although she collects
some of the pollen grains, several locations on the ventral
body surface are poorly groomed (Buchmann 1983;
Michener et al. 1978), which facilitates the transfer of
pollen to other flowers. Both species of bees groom the
pollen from their bodies and place it onto specialized
carrying structures (Michener et al. 1978). Bombus has
basket-like corbiculae, while Amegilla has brush-like sco-
pae for holding pollen on the hind legs (Michener 2000).
Bombus moistens the pollen with nectar before packing it
into the corbicula (Michener 2000; Michener et al. 1978).
Amegilla packs relatively dry pollen among the hairs of the
scopa (Anderson and Symon 1988).
The asynchronous flight muscles that drive the wings
form part of a resonant system, whose vibration frequency
depends on the mass it is driving (i.e., the mass of the
wings) (Josephson et al. 2000). Thus, when the wings are
disengaged during sonication, the vibration frequency of
the flight muscles is higher than the bees’ flapping fre-
quency during flight (King et al. 1996).
Although much of the previous work on buzz-pollina-
tion has focused on Bombus spp. (Asada and Ono 1996;
Buchmann and Hurley 1978; De Luca et al. 2013; Harder
1990; King 1993; King and Buchmann 2003; Morandin
et al. 2001), many other bee genera perform buzz-polli-
nation, including Protandrena: Andrenidae (Cane and
Buchmann 1989), Megachile: Megachilidae (Neff and
Simpson 1988), Augochloropsis: Halictidae (Thorp and
Estes 1975), Xylocopa: Apidae (Hogendoorn et al. 2000;
King and Buchmann 2003), Nomia: Apidae (Anderson and
Symon 1988), and Amegilla: Apidae (Hogendoorn et al.
2006). Information on the mechanics of buzz-pollination in
these genera is far more limited, and comparative studies of
buzz-pollination mechanisms among different groups of
bees are scarce.
Buzz-pollination is known to be critical for many
endangered plants, such as Dianella longifolia in Australia
(listed on the Advisory List of Rare or Threatened Plants in
Victoria in 2014 and the Northern Territory Threatened
Species list), which can reproduce only through buzz-pol-
lination. In addition, the economic value of buzz-pollina-
tion is very high, as it contributes to increased yields in
crops ranging from tomatoes (Asada and Ono 1996;
Hogendoorn et al. 2006) to blueberries (Javorek et al.
2002) and cranberries (MacKenzie 1994).
In mainland Australia, Bombus spp. are not present, and
multiple native bees perform buzz-pollination. The poten-
tial introduction of Bombus spp. to the Australian mainland
for tomato pollination (Hogendoorn et al. 2006) is being
debated intensively, as Bombus spp. have been commer-
cialized in other parts of the world, and their effectiveness
at pollinating crops in greenhouses is well established
(King 1993). Native Australian bees, like Amegilla spp.,
have not been commercialized to the same degree, but
research suggests that they also present a viable method of
pollinating tomatoes in greenhouses (Bell et al. 2006;
Hogendoorn et al. 2006). However, few studies have
compared the mechanisms by which native Australian bees
and Bombus spp. extract pollen via sonication, and buzz-
pollination by Amegilla spp. has not been quantified.
2 C. M. Switzer et al.
123

Here we compare buzz-pollination on Solanum lycop-
ersicum (cherry tomatoes) by two bees that fill similar
niches on different continents—in Australia, Amegilla
murrayensis (blue-banded bee), and in North America,
Bombus impatiens (common Eastern bumblebee).
To determine whether these species pollinating the same
flower perform buzz-pollination in the same way, we col-
lected audio recordings of buzz-pollination and quantified
the frequency and length of individual buzzes, as well as
the total time spent on a single flower (which may
encompass multiple buzzes). We also recorded bees during
flight, to compare sonication frequency to flight frequency.
Finally, we filmed A. murrayensis during buzz-pollination
using high-speed video, to compare its physical interaction
with the flower to the well-described sonication behavior of
B. impatiens.
Materials and methods
Study species and locations
We collected audio recordings of pollination buzzes by
Bombus impatiens (bumblebees) on Solanum lycopersicum
‘‘Sweet 100’’ (cherry tomatoes) and by A. murrayensis
(blue-banded bees) on S. lycopersicum ‘‘ Heirloom Roma
Cherry’’ and S. lycopersicum ‘‘ Tommy Toe’’ (cherry
tomatoes) (Fig. 1). Although the varieties of cherry toma-
toes (S. lycopersicum) used were different, the flowers are
very similar in size and morphology, and thus, we do not
expect that the tomato variety significantly affected buzz-
pollination characteristics (Online Resource 1, Tables VII
and IX). Recordings of B. impatiens were collected in a
community garden in Carlisle, Massachusetts, USA
(42520N; 71320W), and those of A. murrayensis at the
Adelaide Botanic Garden, Adelaide, Australia (34920S;
138610E). In Australia, S. lycopersicum ‘‘ Tommy Toe’’
plants growing in the garden were supplemented with
potted tomato plants (S. lycopersicum ‘‘ Heirloom Roma
Cherry’’) to provide additional flowers. For the potted
plants, we recorded if the flower had been previously
visited.
Audio recordings and analysis
We collected audio recordings with a shotgun microphone
(SGM-1X, Azden, Tokyo, Japan) attached to a digital
recorder (DR-100mkII, Tascam, Montebello, California),
held within 3 cm of the bees’ bodies. We attempted to
position the microphone pointed at each bee’s thorax,
approximately orthogonal to the bee’s frontal plane. We
were not able to maintain that position for all recordings;
however, we have no evidence that recording from dif-
ferent angles affects the analysis of sonication frequency or
duration. We recorded bees while landing, buzzing flowers,
and flying away, to analyze audio characteristics of both
flight and buzz-pollination.
Because some bees perform multiple buzzes on a single
flower with pauses in-between, we recorded the time of
landing and takeoff to calculate the total visit duration.
This was considered a suitable estimate of the time spent
on a single blossom, since these bees generally did not
crawl between tomato flowers. When audio recordings did
not span the entire length of a flower visit, we excluded
them from the analysis of visit duration.
After collecting audio recordings of landing, buzz-pol-
lination, and takeoff flight, we captured bees with a net and
noted the time, temperature, and relative humidity. To
ensure independent samples, we either marked bees after
the first capture (and excluded recaptured bees from the
analysis) or collected the bees and pinned them as speci-
mens. We measured intertegular (IT) span with digital
calipers on bees that were released and with ImageJ (http://
imagej.nih.gov/ij/) on photographs of pinned specimens to
obtain the average size of each species (Online Resource 1,
Table I). When bees performed multiple buzzes while
visiting a single flower, the frequency and duration of these
buzzes were averaged for statistical analysis.
We played recordings in Audacity (http://audacity.sour
ceforge.net/) and identified the start and end of each buzz
aurally and visually to determine the buzz length. We
defined buzzes that had breaks of less than about 0.1 s to be
single buzzes. Figure 2a shows an oscillogram from a
series of buzzes by B. impatiens, with a single buzz
expanded in Fig. 2b. We calculated buzz frequencies in R
(R Core Team 2012), using the ‘‘seewave’’ (Sueur et al.
2008) and ‘‘signal’’ (Signal Developers 2013) packages.
We first filtered recordings to remove low-frequency noise
and then calculated fundamental frequencies within sliding
Fig. 1 Photographs of bees used in this study. Bombus impatiens
workers (left) are typically larger than Amegilla murrayensis (right;
Online Resource 1, Table I). Both bees are shown on Solanum
lycopersicum flowers. Black bars indicate approx. 1 cm. Bumblebee
photograph credit: Tim Stanley/Native Beeology
Shakers and head bangers: differences in sonication behavior between Australian Amegilla…3
123

windows of 2048 points, with 80 % overlap. Recordings
typically contained small number points that were clearly
outliers (Fig. 2c)—single data points at frequencies more
than one standard deviation beyond the median funda-
mental frequency. These outliers were most likely artifacts
caused by using a relatively small sliding window and/or
collecting recordings in noisy, outdoor environments. We
removed outliers and then calculated the median of the
trimmed distribution to determine the frequency of each
buzz (Fig. 2c).
We calculated wing beat (flight) frequency using the
same method as for buzz frequency—identifying flights
aurally and visually in Audacity, and then using R to cal-
culate fundamental frequency (Online Resource 1,
Table II). For one recording of a pollination buzz by B.
impatiens and four recordings of flight by A. murrayensis,
we were unable to obtain an accurate frequency using
seewave, so we analyzed these recordings manually, by
performing a fast Fourier transform (FFT) with the spec-
trum function in Audacity, using a Hanning window of
2048 points. We then listened to the recording and matched
the sound with one of the peaks from the FFT spectrum.
Video recordings
We collected videos of A. murrayensis performing buzz-
pollination on S. lycopersicum (cherry tomato) flowers in
the Adelaide Botanic Garden, using a high-speed camera
(TS3, Fastec Imaging, San Diego, California) recording at
2000 fps. We recorded a total of nine videos, four of which
are known to be of unique individuals, because we were
able to capture these bees after filming.
Statistical tests
All statistical tests were performed in R (R Core Team
2012). We used multiple linear regression to compare flight
frequency, average buzz-pollination frequency, average
buzz length, and visit duration between A. murrayensis and
B. impatiens; this method allowed us to compare the two
species of bees while accounting for environmental vari-
ables: temperature, time of day, and relative humidity. To
fit the assumptions of linear regression, we squared buzz-
pollination frequency, square-root-transformed flight fre-
quency, and log-transformed buzz length and visit dura-
tion. We also used paired ttests to compare flight versus
pollination buzz frequency for individuals within each
species.
We used multiple linear regression to compare buzz-
pollination characteristics for A. murrayensis on different
tomato varieties and on virgin versus nonvirgin flowers
(Online Resource 1, Tables VII–IX).
We adjusted significance level using Bonferroni cor-
rection, to account for performing multiple comparisons
with the same individuals. Because we performed four
multiple regressions and one ttest, we adjusted our sig-
nificance level to 0.05 divided by 5, or 0.01. We did not
adjust the significance level to include the four covariates
in each of the multiple regressions, because the only
Fig. 2 Audio recordings of buzz-pollination. aOscillogram showing
four pollination buzzes by Bombus impatiens (bumblebee) on a flower
of Solanum lycopersicum ‘‘Sweet 100’’ (cherry tomato). Shaded
region indicates a single buzz. bExpanded oscillogram of the single
buzz shaded in a.cFundamental frequency calculated over the course
of the buzz shown in b.Dots represents the fundamental frequencies
calculated from overlapping windows of 2048 data points. Frequency
measurements that were identified as outliers and removed are
indicated by a plus symbol
4 C. M. Switzer et al.
123

variable of interest was the bee species. This correction is
overly conservative, but using a less-conservative adjust-
ment would not have changed our conclusions. Graphs
were made with ggplot2 (Wickham 2009).
Results
None of the buzz-pollination or flight characteristics
quantified were associated with environmental variables
(temperature, relative humidity, time of day). We found
that buzz frequency was significantly higher for Amegilla
than for Bombus (t
(70)
=8.452, pvalue 0.001; Fig. 3a;
Online Resource 1, Table III), and wing beat frequency
was also higher for Amegilla than for Bombus
(t
(71)
=13.372, pvalue 0.001; Fig. 3a; Online Resource
1, Table IV). Within each bee species, the wing beat fre-
quency during flight was significantly lower than the buzz-
pollination frequency (Amegilla t
(21)
=24.67, pvalue
0.001; Bombus t
(52)
=16.59, pvalue 0.001).
There was no significant difference between the two bee
species in the length of an individual pollination buzz
(Fig. 4b; t
(70)
=1.124, pvalue [0.2; Online Resource 1,
Table V), but B. impatiens spent more time on a single
flower than A. murrayensis (Fig. 3c; t
(53)
=3.974, pvalue
\0.005; Online Resource 1, Table VI). We found no sig-
nificant differences in buzz characteristics of A. mur-
rayensis when pollinating the two different varieties of S.
lycopersicum or when pollinating unvisited versus previ-
ously visited flowers (Online Resource 1, Tables VII–IX).
The high-speed videos revealed that A. murrayensis
differs markedly from B. impatiens (and many other buzz-
pollinating bees described thus far) in how it physically
interacts with the flower during buzz-pollination. While B.
impatiens and other bees grasp the flower’s anthers with
their mandibles as well as their legs, A. murrayensis does
not. All videos we collected showed A. murrayensis
grabbing the anther with only its legs and repeatedly tap-
ping the anther with its head at the high buzzing frequency
that is likely generated by its flight muscles (Fig. 4, Online
Fig. 3 Buzz-pollination and
flight characteristics of A.
murrayensis and B. impatiens.
aPollination buzz frequency of
A. murrayensis (white,n=22)
versus B. impatiens (gray,
n=53), and flight (wing beat)
frequency of A. murrayensis
(n=23) versus B. impatiens
(n=53). bAverage buzz
length of A. murrayensis
(n=22) versus B. impatiens
(n=53). cFlower visit
duration of A. murrayensis
(n=22) versus B. impatiens
(n=36). Double asterisks and
single asterisks indicate a
significant difference at
p\0.0001 and 0.0005,
respectively
Shakers and head bangers: differences in sonication behavior between Australian Amegilla…5
123

Resource 2). Research indicates that the flight muscles are
used during sonication in Bombus occidentalis (King et al.
1996), and our high-speed videos show that the mesosoma
of A. murrayensis is deforming with each tap of the head
(Online Resource 2)—similar to the way that bumblebees’
mesosomas deform, while they buzz (Online Resource 3).
In particular, the videos for both bumblebees and blue-
banded bees show the first segment of the mesosoma,
called the pronotum, moving at the same frequency as the
head. The mesopleuron can also be seen oscillating during
buzz-pollination. In two recordings, we saw a bee briefly
grasp the anther with its mandibles, but it quickly switched
to the head-tapping behavior.
We noticed that A. murrayensis left brown marks on the
anther cone—these ‘‘bee kisses’’ are interpreted by com-
mercial tomato growers as a sign that bees have visited the
flowers (Buchmann and Nabhan 1996). A. murrayensis
may be damaging the anthers with impact forces, but the
resulting ‘‘bee kisses’’ are similar to those left by B.
impatiens.
Discussion
We found that Amegilla murrayensis (blue-banded bees)
buzz cherry tomato flowers at significantly higher fre-
quencies (*350 Hz) than B. impatiens (*240 Hz; Online
Resource 1, Table II), while accounting for environmental
variables. The flight (wing beat) frequencies of both species
are lower than their buzz-pollination frequencies. This is
likely due to the properties of asynchronous muscles, which
are part of a resonant system (Josephson 2006). When the
mass of the wings is reduced in this system, the wing beat
frequency increases (Roberts and Cartar 2015). Likewise,
when the mass on the wings is increased or simply moved
further from axis of rotation, the frequency of the resonant
system should decrease. The lower frequency during flap-
ping flight is likely because the wings are extended during
flapping, but held close to the body during buzz-pollination.
A. murrayensis’ flight frequency is significantly higher than
that of B. impatiens, which is not surprising due to its
smaller body size (Burkart et al. 2011). Previous studies
have suggested that the amount of pollen released from
poricidal anthers increases with buzz frequency (Harder and
Barclay 1994) or with buzz frequency and displacement (De
Luca et al. 2013; King and Buchmann 1996); thus, the
higher buzzing frequency of A. murrayensis may be more
effective at releasing pollen from the anthers.
Despite the large difference in buzzing frequency
between species, the average length of a single buzz was the
same in A. murrayensis and B. impatiens, with both bees
buzzing in bouts lasting approximately 1 s—similar to the
duration required to eject most of the pollen from Solanum
laciniatum flowers (King and Buchmann 1996). However,
B. impatiens spent significantly more time on a single
flower (approximately 3.7 s), as compared to A. mur-
rayensis, which departed after *1 s. This difference is due
to the fact that B. impatiens typically buzzed an individual
flower several times (often gathering and cleaning pollen
from its body in-between buzzes) before departing, whereas
A. murrayensis typically buzzed a flower once and then flew
away to clean pollen from its body. Occasionally, A. mur-
rayensis returned to the same flower after cleaning, but most
often it moved onto a new flower.
The fact that A. murrayensis spent significantly less time
on each flower, typically buzzing the flower only once,
Fig. 4 Image sequence of Amegilla murrayensis during buzz-polli-
nation. Rather than grasping the anther firmly with its mandibles like
other buzz-pollinating bees, A. murrayensis taps its head against the
anther of a Solanum lycopersicum (cherry tomato) flower at the high
buzzing frequency generated by its flight muscles (approximately
350 Hz; Online Resource 1, Table II). The interval between images
(from Online Resource 2) is 1/1000 of a second. The dark marks on
the anthers were made with ink, to help visualize the movement of the
anther
6 C. M. Switzer et al.
123

combined with its higher (and possibly more effective)
buzzing frequency, suggests that A. murrayensis may be
able to extract pollen from flowers more quickly than B.
impatiens. An alternative explanation for the difference
between bee species in the amount of time spent on each
flower is the possibility that tomatoes in Australia could
provide different amounts of pollen during buzzing than
tomatoes in the USA (due to potential differences in tomato
varieties, local environment, visitation rates by local bees,
etc.). However, because we recorded A. murrayensis per-
forming only a single buzz on both virgin and previously
buzzed flowers—which are known to release less pollen
(King and Buchmann 1996)—we do not believe that A.
murrayensis is adjusting the number of buzzes it performs
based on the pollen reward. Whether or not A. murrayensis
is in fact obtaining more pollen than B. impatiens with a
single buzz, the behavior of often moving onto the next
flower after performing only one buzz appears to be typical
for this species, at least when foraging from cherry tomato
plants.
We also found that A. murrayensis interacts with the
flower in a unique way during buzz-pollination. Rather
than grasping the anthers firmly with its mandibles and
shaking (as described for Bombus and many other bee
genera previously studied (Buchmann 1983; Buchmann
and Hurley 1978; Crobet and Huang 2014; Jesson and
Barrett 2005; King 1993; Online Resource 3), A. mur-
rayensis taps the anthers with its head repeatedly, at the
high frequencies most likely generated by its flight mus-
cles. This ‘‘head-banging’’ behavior may be intentional, or
it may be a side effect of the bees being unable to grasp the
anthers firmly enough with their mandibles while sonicat-
ing, possibly due to their small size or insufficient grip
strength—although other, smaller bee species, such as
Lipotriches (Halictidae), have been recorded grasping
anthers firmly with their mandibles during buzz-pollination
(Online Resource 4), and we observed A. murrayensis
grasping onto leaves with its mandibles while grooming its
body.
The mechanical features of a buzz that have been pro-
posed to determine how much pollen is dislodged include
the length of a buzz and the maximum acceleration (often
called amplitude) (De Luca et al. 2013) or velocity of the
buzz (Corbet and Huang 2014). Acceleration and velocity
both increase with buzz frequency and with displacement
(Corbet and Huang 2014; De Luca et al. 2013). Impact
forces, which occur when two objects collide, cause sharp
changes in velocity, which in turn produce large spikes in
acceleration. Thus, the head tapping observed in A. mur-
rayensis, in which the head collides and then disengages
with the anther repeatedly at high frequencies, may pro-
duce higher accelerations than grasping the anthers firmly
and shaking. High accelerations produced by collisions
with the anther could lead to higher pollen release rates or
could help break up clumps of wet pollen. These
hypotheses could be tested in future studies by manipu-
lating flowers with mechanical shakers that either grasp
firmly or collide repeatedly with the anthers, and quanti-
fying the amount of pollen released.
If A. murrayensis is capable of removing more pollen
with a single buzz (due to its higher buzzing frequency and/
or head-tapping behavior), this could be detrimental from
the plant’s perspective, since a single forager is removing a
large portion of its pollen. On the other hand, some
researchers have suggested that high pollen removal from
poricidal anthers is associated with higher pollen deposi-
tion onto stigmas (Harder 1990; Harder and Thomson
1989), although this is not always the case for plants with
nonporicidal anthers (Wilson and Thomson 1991). In
addition, if A. murrayensis routinely spends less time on a
single flower, then it may move onto another flower more
rapidly than B. impatiens, which could lead to higher
pollination rates.
The relative effectiveness of A. murrayensis and B.
impatiens as pollinators may also be affected by other
aspects of their behavior—for instance, how well each
species grooms the pollen from its body (decreasing the
probability of transferring pollen between flowers), and
whether the behavior of one species brings its body closer
to the stigma (increasing the chances of depositing pollen).
A definitive answer to the question of pollination effec-
tiveness will ultimately require controlled experiments
comparing the yield of tomato plants buzz-pollinated by A.
murrayensis versus B. impatiens.
Although further work is required to make any claims
about the pollination effectiveness of these different bees,
our work shows that at least one native Australian bee—
Amegilla murrayensis—differs significantly from Bombus
spp. in several aspects of buzz-pollination, including its
buzzing frequency and the amount of time spent per
flower. Furthermore, our observation that A. murrayensis
interacts with flowers in a unique way during buzz-pol-
lination—by ‘‘head-butting’’ rather than ‘‘shaking’’ the
anthers—leads to further questions about the behavioral,
mechanistic, and evolutionary roots of this method of
buzz-pollination.
Acknowledgments The authors thank Christopher De Ieso and
Remko Leijs for help in collecting data in the Adelaide Botanic
Garden, as well as the staff of the Adelaide Botanic Garden, notably
Carolyn Sawtell and Robert Hatcher, for allowing us to conduct
research with their plants. We would also like to thank Robin Hopkins
for helpful suggestions on the manuscript. This project was funded by
a Putnam Expedition Grant to C.S. from the Harvard Museum of
Comparative Zoology and by the National Science Foundation
(CAREER IOS-1253677) to S.C.
Shakers and head bangers: differences in sonication behavior between Australian Amegilla…7
123

References
Anderson GJ, Symon D (1988) Insect foragers on Solanum flowers in
Australia. Ann Mo Bot Gard 75:842–852
Asada S, Ono M (1996) Crop pollination by Japanese bumblebees,
Bombus spp. (Hymenoptera: Apidae): tomato foraging behavior
and pollination efficiency. Appl Entomol Zool 31(4):581–586
Bell M, Spooner-Hart R, Haigh A (2006) Pollination of greenhouse
tomatoes by the Australian blue banded bee Amegilla (Zon-
amegilla) holmesi (Hymenoptera: Apidae). J Econ Entomol
99(2):437–442
Buchmann SL (1983) Buzz pollination in angiosperms. In: Jones CE,
Little RJ (eds) Handbook of experimental pollination biology.
van Nostrand Reinhold Company Inc, New York, pp 73–113
Buchmann SL, Hurley JP (1978) A biophysical model for buzz
pollination in angiosperms. J Theor Biol 72(4):639–657
Buchmann SL, Nabhan GP (1996) The forgotten pollinators. Island
Press, Washington, DC
Burkart A, Lunau K, Schlindwein C (2011) Comparative bioacous-
tical studies on flight and buzzing of neotropical bees. J Pollinat
Ecol 6(16):118–124
Cane JH, Buchmann SL (1989) Novel pollen-harvesting behavior by
the bee Protandrena mexicanorum (Hymenoptera: Andrenidae).
J Insect Behav 2(3):431–436
Corbet SA, Huang SQ (2014) Buzz pollination in eight bumblebee-
pollinated Pedicularis species: Does it involve vibration-induced
triboelectric charging of pollen grains? Ann Bot 114(8):1665–
1674
Core Team R (2012) R: a language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna
De Luca PA, Vallejo-Marı
´n M (2013) What’s the ‘buzz’about? The
ecology and evolutionary significance of buzz-pollination. Curr
Opin Plant Biol 16(4):429–435
De Luca PA, Bussiere LF, Souto-Vilaros D, Goulson D, Mason AC,
Vallejo-Marı
´n M (2013) Variability in bumblebee pollination
buzzes affects the quantity of pollen released from flowers.
Oecologia 172(3):805–816
Gegear RJ, Laverty TM (2001) The effect of variation among floral
traits on the flower constancy of pollinators. In: Chittka L,
Thomson JD (eds) Cognitive ecology of pollination. Animal
behaviour and floral evolution. Cambridge University Press,
Cambridge, pp 1–20
Harder LD (1990) Pollen removal by bumble bees and its implica-
tions for pollen dispersal. Ecology 71:1110–1125
Harder L, Barclay R (1994) The functional significance of poricidal
anthers and buzz pollination: controlled pollen removal from
Dodecatheon. Funct Ecol 8:509–517
Harder LD, Thomson JD (1989) Evolutionary options for maximizing
pollen dispersal of animal-pollinated plants. Am Nat 133:
323–344
Hogendoorn K, Steen Z, Schwarz MP (2000) Native Australian
carpenter bees as a potential alternative to introducing bumble
bees for tomato pollination in greenhouses. J Apic Res 39:67–
74
Hogendoorn K, Gross CL, Sedgley M, Keller MA (2006) Increased
tomato yield through pollination by native Australian Amegilla
chlorocyanea (Hymenoptera: Anthophoridae). J Econ Entomol
99(3):828–833
Javorek S, Mackenzie K, Vander Kloet S (2002) Comparative
pollination effectiveness among bees (Hymenoptera: Apoidea)
on lowbush blueberry (Ericaceae: Vaccinium angustifolium).
Ann Entomol Soc Am 95(3):345–351
Jesson LK, Barrett SC (2005) Experimental tests of the function of
mirror-image flowers. Biol J Linn Soc 85(2):167–179
Josephson RK, Malamud JG, Stokes DR (2000) Asynchronous
muscle: a primer. J Exp Biol 203(18):2713–2722
King MJ (1993) Buzz foraging mechanism of bumble bees. J Apic
Res 32:41–49
King MJ, Buchmann SL (1996) Sonication dispensing of pollen from
Solanum laciniatum flowers. Funct Ecol 10:449–456
King MJ, Buchmann SL (2003) Floral sonication by bees: mesosomal
vibration by Bombus and Xylocopa, but not Apis (Hymenoptera:
Apidae), ejects pollen from poricidal anthers. J Kans Entomol
Soc 76:295–305
King MJ, Buchmann SL, Spangler H (1996) Activity of asynchronous
flight muscle from two bee families during sonication (buzzing).
J Exp Biol 199(10):2317–2321
MacKenzie K (1994) The foraging behavior of honey bees (Apis
mellifera L) and bumble bees (Bombus spp) on cranberry
(Vaccinium macrocarpon Ait). Apidologie 25:375–383
Michener CD (2000) The bees of the world. Johns Hopkins University
Press, Baltimore
Michener CD, Winston ML, Jander R (1978) Pollen manipulation and
selected activities and structures in the family Apidae. Univ
Kans Sci Bull 51:575–601
Morandin L, Laverty T, Kevan P (2001) Bumble bee (Hymenoptera:
Apidae) activity and pollination levels in commercial tomato
greenhouses. J Econ Entomol 94(2):462–467
Neff JL, Simpson BB (1988) Vibratile pollen-harvesting by Me-
gachile mendica Cresson (Hymenoptera, Megachilidae). J Kans
Entomol Soc 61:242–244
Roberts J, Cartar RV (2015) Shape of wing wear fails to affect load
lifting in bumble bees with experimental wing wear. Can J Zool
93(7):531–537
Signal Developers (2013) Signal: signal processing. http://r-forge.r-
project.org/projects/signal/last. Accessed Oct 2015
Sueur J, Aubin T, Simonis C (2008) Equipment review: seewave, a
free modular tool for sound analysis and synthesis. Bioacoustics
18(2):213–226
Thorp RW, Estes JR (1975) Intrafloral behavior of bees on flowers of
Cassia fasciculata. J Kans Entomol Soc 48:175–184
Wickham H (2009) ggplot2: elegant graphics for data analysis.
Springer, New York
Wilson P, Thomson JD (1991) Heterogeneity among floral visitors
leads to discordance between removal and deposition of pollen.
Ecology 72:1503–1507
Zimmerman M (1982) Optimal foraging: random movement by
pollen collecting bumblebees. Oecologia 53(3):394–398
8 C. M. Switzer et al.
123