Enhancing pollination by attracting and retaining leafcutting bees (Megachile rotundata) in
alfalfa seed production fields
Johanne Brunet1 and Zainulabeuddin Syed2
1. USDA Agricultural Research Service, VCRU, Madison, Wisconsin
2. Department of Biological Sciences University of Notre-Dame, Notre-Dame, Indiana
The alfalfa leafcutting bee, Megachile rotundata (F.), has become an important managed
pollinator of alfalfa, Medicago sativa L. One problem when using alfalfa leafcutting bees as
managed pollinator, is the dispersal of many females upon release, even when adequate nesting
sites are present. While dispersal of female bees from the site of emergence may represent a
successful evolutionary adaptation to avoid inbreeding (Gandon 1999; Guillaume and Perrin
2006), this behavior is problematic when these pollinators are used as managed pollinators of
alfalfa. Reducing the dispersal of females upon release into the fields, would facilitate the
maintenance of viable commercial populations of alfalfa leafcutting bees. Previous work
indicated M. rotundata females were more likely to initiate nests in used as opposed to new nesting
boards (Fairey and Lieverse 1986, Fairey and Lefkovitch 1993). However, the use of old nesting
boards is not a viable option because it facilitates disease propagation and increases bee mortality
(Vandenberg and Stephen 1982). One proposed explanation for the preference of bees for the old as
opposed to new nesting boards is attraction at a short distance to odors of old nest contents (Stephen
and Torchio 1961). While old bee boards cannot be used to retain bees, chemical attractants represent a
viable alternative. Early attempts to identify compounds involved in such attraction were not
successful (Buttery et al. 1981; Parker et al. 1983), but Pitts-Singer (2007) demonstrated a
preference of alfalfa leafcutting bee females for nest cells or fecal rings relative to blanks; leaf
pieces compared to fecal rings and finally leaf piece extracts relative to control (solvent), when
performing behavioral assays in Y-tubes. Such results are promising and justify pursuing the
question whether chemicals extracted from alfalfa leafcutting bee cells could be used to prevent
female alfalfa leafcutting bees from dispersing from a site upon release.
The objectives of this study are to extract, isolate, and identify biologically active constituents
from alfalfa leafcutting bee cells (where eggs were laid and bees developed) and to quantify the
attraction of these chemicals to the bees. We conventionally divide the chemosensory behavior
of ACLBs into long range direct insect movement towards the odor cue (induced by Long range
attractants (LRA)), and slowing down/arrestment behavior in the vicinity of the cues (induced by
additional short range arrestants (SRA)). We isolate LRAs and SRAs from empty bee cells and
perform behavioral tests to quantify attraction of the different extracts to alfalfa leafcutting bees.
The ultimate goal of this research is to develop field deployable attractive baits using the
attracting chemicals in order to facilitate the maintenance of alfalfa leafcutting bees in seed-
production fields and improve alfalfa pollination.
Isolation of biologically active constituents from alfalfa leafcutting bee cells
Empty cocoons were shipped in brown paper bags from Madison, WI and stored at -4°C until
use. The isolation of biologically active material was performed based on the prior knowledge
that bee cells, following bee emergence, potentially produce long range and short range
chemosensory cues, hereafter referred to, respectively, as long range attractants (LRA) and short
range arrestants (SRA).
Long range attractants (LRAs)
Two hundred grams of empty bee cells were placed in an airtight glass chamber (Rad Lab Glass
Shop Notre Dame, IN) fitted with Teflon® lined inlet and outlet tubes to collect airborne odors.
This set up was used to sample and analyze “Long range odors” following previously described
methods (Scheidler et al 2015). Briefly, we used Solid phase micro-extraction (SPME) fiber
(50/30 um DVB/CAR/PDMS Stableflex 23 Ga; Supelco, US) method wherein a pre-cleaned
SPME fiber was exposed for an hour to the bee cells confined in the glass chamber before being
injected for mass spectrometry analysis. The second method was collection of the bee cell
headspace odors onto an adsorbent and eluting in an organic solvent to compare and contrast
with the SPME profile. Charcoal-filtered air (Pall Life Sciences, US) was pushed through the
inlet @400 ml/min, and the outlet was connected to a glass cartridge containing Super-Q
adsorbent (Alltech USA). An electric vacuum (Whisper AP-150, Tetra, US) pulled air over the
bee cells on to the adsorbent for 48 hours. Volatiles were desorbed in glass distilled hexane
(Fischer, ≥98.5% purity) and stored at -80 oC until they were shipped on dry ice to Madison,
Short range arrestants (SRAs)
The Headspace method above collected lighter volatile constituents (LRA). To collect heavier
volatile constituents (SRA), bee cells were immersed for 5 min in glass-distilled non-polar
hexane or a polar solvent methanol (Fischer, ≥98.5% purity) in a Wheaton® high recovery
NextGen™ V-Vial attached with PTFE lined septa caps (Wheaton, Millville, NJ). After
removing the bee cells from their respective solution, they were subjected to two additional
rinses. Extracts were pooled, evaporated under a gentle stream of helium to increase their
concentration before being reconstituted in 10 ml µl of their respective solution. Both polar and
non-polar samples were shipped on dry ice to Madison, Wisconsin for behavioral assays.
Chemical analyses were performed on a 7890A GC system (Agilent Technologies, Santa Clara,
CA) coupled with a 5975C Agilent Technologies mass spectrophotometer (inert XL MSD with a
triple-Axis Detector). The exposed SMPE fiber, or the extracts (LRA or SRA) where an extract
was injected in a split-splitless injector operating under splitless mode at 250°C in an Agilent
HP-5 capillary column (30 m, 0.32 mm ID, 0.25 µm phase thickness). Helium (Ultra High
Purity 5.0 Grade; Airgas, USA) was used as the carrier gas at a constant flow rate of 1 ml/min.
The column was held isothermally at 50°C for 1 minute, then programmed to increase at a rate of
10°C per minute until 300°C, with a final hold of 5 minutes. The Mass Spectrometer (MS) was
operated at 70 eV. Data recording and quantification was performed using the Agilent MSD
ChemStation software (E.02.02.1431). Initial chemical identity was determined using the NIST
2011 MS library and compounds with a 80% match or greater, using this library, are reported
here. We are in the process of further confirming these initial chemical identities by : 1)
comparing the Retention Index (RI) calculated for each of the compounds to known RIs from the
published literature; and 2) confirming retention times and their mass spectra by comparing them
to synthetic standards.
Choice trials were performed in a cage, 2.44 m x 1.83 m x 1.83 m (L x W x H), set up in a room
at the DC Smith greenhouse at the University of Wisconsin-Madison. Temperature was set at 26-
28°C during the day and 21°C at night. The cage faced southeast. Flowering alfalfa plants were
kept in the cage to feed the bees.
In the cage, we set up two 30.5 cm x 44.0 cm cardboards, separated by 1.3 meters, and attached
using small binder clips on the north wall of the cage (facing south), 0.8 m above the ground. On
each cardboard, we set up six 6 cm x 4.5 cm styrofoam
nesting blocks, forming three rows of two blocks with each
block separated by 10 cm and attached using mounting
tape (Fig. 1). Each bee block contained 20 (5 x 4) holes
and a fresh paper tube was inserted into each hole prior to
One cardboard served as control, where a solvent was
added to some of the blocks on the cardboard, while a
chemical attractant was added to the blocks on the other
cardboard. A one cm2 piece of Fisher brand filter paper P5
was placed with a stick pin on top of each of the four
corner blocks on a cardboard (Fig. 1). Twenty five µl of
the test chemical or solvent (control) was added to each
filter paper using a Hamilton 250 ul glass syringe for a
total of 100 µl per cardboard. The test chemicals and
Fig. 1 Experimental set up. One cardboard attached to the north facing wall of a large cage with 6 small
bee boards set up 10 cm apart. Filter papers were attached with a pin at the top of four of the small bee
solvents were mixed 1:3 with paraffin oil to slow down evaporation. Fresh chemical and solvent
were added at the beginning of each trial. Solvents and chemicals were stored in the freezer.
Trials were run in the mornings, typically between 11:00- 12:00 during months with standard
time and between 10:00- 11:00 a.m. during months with daylight saving time. The position of the
treatment and control cardboards in a cage was switched each day of the experiment. A trial
typically lasted one hour. Females were used in the experiment and the following behaviors were
recorded: approach, defined as flying within 4 inches of a block; landing, where a bee landed on
the block; and entering, where a bee entered a tube. The length of time the bee remained on the
block or in the tube was recorded together with time of day, treatment, and bee block position
within a cardboard. Bees that landed or entered were removed from the cage and kept in a small
cage (61 cm x 61 cm x 61 cm) outside the large mesh cage for one full day or until the next trial
(if longer than one day) when they were returned to the large cage. A tube visited by a bee was
immediately removed and replaced with a fresh tube. If a bee had not left a tube after 5 min, both
the tube and the bee were removed. The small bee cage contained cut alfalfa flowers to feed the
bees and flowers within the small cages were replaced daily.
The experiments performed to date include one set of experiments to determine preference of
alfalfa leafcutting bees for the Long Range chemicals relative to the control (hexane) and a
second set of experiments to examine preference for the short range arrestants extracted in non-
polar hexane relative to the control (hexane). Differences in the number of total visits between
either the long range or the short range chemical and the control were contrasted using Chi-
square tests. Differences were also examined separately, also using Chi-square tests, for the
number of approaches, number of landings or number of enters within each experiment when
sample sizes were sufficient with a minimum of 5 visits in each category.
Biologically active constituents from alfalfa leafcutting bee cells
From the odors collected either by SPME or Super-Q adsorbent, we identified 22 volatile
compounds (Table 1; Fig. 2). The most abundant compound collected by either method was a
monoterpene, α- pinene, followed by a common green leaf volatile (GLV), 3-hexen-1-ol (Z) (Fig.
2B). Interestingly the
SPME protocol and
resulted in qualitatively
thus only the Total Ion
Chromatogram (TIC) of
the Super-Q extract is
displayed here in the
analyzed by GC-MS
were subsequently used
for behavioral studies.
Table 1. List of odorant compounds collected from the bee cell headspace. Odors were collected
on Super-Q and desorbed into hexane and analyzed on a high resolution capillary column.
Fig. 2. Odor profile of the bee cell headspace. A- Total Ion Chromatogram and B- Identified
compounds arranged by decreasing abundance.
Behavioral response to chemicals
Long range attractant
Over all visits, we observed no statistically
significant differences between the
number of visits to control and long range
attractants (df= 1, χ2= 1.8, p = 0.18). Bees
visited the boards with the control 7 times
and boards with the long range attractant
Table 2. Behavioral assay for preference of alfalfa leafcutting bees to long range attractants.
Number of approaches, landings and entries and total visits for control or chemical attractant
13 times (Table 2). While the number of approaches to the long range attractant
were more common than to the control (Table 2), numbers are small and we are currently
increasing the sample sizes to confirm the pattern. Chi-square tests were not presented on
individual behaviors because the minimum of 5 visits per cell was not reached.
Short range arrestants
Over all visits, bees did not visit bee boards with short range arrestants more often than the
control boards (df = 1, χ2 = 0.67, P = 0.41). Bees made 30 visits to the control boards and 24
visits to the boards with the short range arrestants (Table 3). When each behavior was examined
separately, bees did not make more approaches (df = 1, χ2 = 0.36, P = 0.55), more landings (df =
1, χ2 = 1.67, P = 0.197) or more entry (df =
1, χ2 = 0.29, P = 0.59) to the boards with
short range arrestants relative to the
control boards (Table 3).
Table 3. Behavioral assay for preference of alfalfa leafcutting bees to short range arrestants.
Number of approaches, landings and entries and total visits for control or chemical attractant
Our GC-MS analysis of bee cell headspace odors revealed a series of monoterpenes,
sesquiterpenes and GLVs. In addition, presence of common floral and leaf odors such as
benzaldehyde and straight chain aldehydes (heptanal, nonanal and 1-octen-3ol) reflect the
presence of leaves in the bee cell material. While our behavioral data, to date, do not indicate a
preference by ALCBs for either long range attractants or short range arrestants, more trials are
needed to increase sample sizes and test different settings. The fact ALCBs are known to use
chemical odors for nest recognition (Guedot et al., 2013) and another solitary bee, Osmia
lignaria, has been shown to respond to some of the chemical constituents of empty cocoons
(Teresa Pitts-Singer, pers. comm.), further justifies pursuing the behavioral tests. We will
perform more behavioral assays to determine if biologically active material (LSA and SRA)
from the bee cells are attractive to ALCBs. Once the identity is established and quantified, we
will run these active extracts on a Gas Chromatography system (GC) linked to an ALCB antenna
that serves as a biological detector. We can then identify the set of compounds eliciting olfactory
physiological response in ALCBs and use behavioral tests to confirm the preference of ALBCs
to these compounds, individually and in different combinations. Ultimately, the goal of this
research is to develop field deployable attractive baits using the attracting chemicals to facilitate
the maintenance of alfalfa leafcutting bees upon release and improve alfalfa pollination.
Because dispersal of females at the site of release in solitary bees may well represent an
adaptation to reduce inbreeding in natural populations (Gandon 1999; Guillaume and Perrin
2006), mating females prior to release in alfalfa seed production fields may significantly reduce
their probability of dispersing. One recommendation to bee managers and alfalfa seed growers
would be to implement bee management practices that increase the chances of females mating
prior to release while maintaining adequate nesting sites in order to limit the loss of female
ALCBs upon release.
We thank Kari Steiger for collecting the behavior data and Nicole Scheidler for help with
extractions. Theresa Pitts-Singer provided useful suggestions for the experimental setup.
Bohart, G.E. 1962. How to manage the alfalfa leaf-cutting bee (Megachile rotundata Fabr.) for
alfalfa pollination. Utah Agricultural Experiment Station, Logan, UT.
Buttery, R.G., F.D. Parker, R. Teranishi, R.M. Thomas, and L.C. Ling. 1981. Volatile
components of alfalfa leafcutter bee cells. J. Agric. Food Chem. 29: 955- 958.
Fairey, D.T., and J.A.C. Lieverse. 1986. Cell production by the alfalfa leafcutting bee
(Megachile rotundata F.) in new and used wood and polystyrene nesting materials. J. Appl.
Entomol. 102: 148-153.
Fairey, D.T., and L.P. Lefkovitch. 1993. Arrangement of new and used nesting materials in
leafcutting bee, Megachile rotundata (F.) (Hym., Megachilidae), shelters to maximize cell
production. J. Appl. Entomol. 115: 62-65.
Gandon, S. 1999. Kin competition, the cost of inbreeding and the evolution of dispersal. J.
Theor. Biol. 200: 345-364.
Guédot, C., J.S. Buckner, M.M. Hagen, J. Bosch, W.P. Kemp, and T.L. Pitts-Singer. 2013. Nest
marking behavior and chemical composition of olfactory cues involved in nest recognition in
Megachile rotundata. Environ. Entomol. 42: 779-789.
Guillaume, F. and N. Perrin. 2006. Joint evolution of dispersal and inbreeding load. Genetics 13:
Parker, F.D.R., R. Teranishi, and A. C. Olson. 1983. Influence of attractants on nest
establishment by the alfalfa leafcutting bee (Hymenoptera: Megachilidae) in styrofoam and
rolled paper. J. Kansas Entomol. Soc. 56: 477-482.
Pitts-Singer, T.L. 2007. Olfactory response of Megachilid bees, Osmia lignaria, Megachile
rotundata and M. pugnata, to individual cues from old nest cavities. Environmental Entomology
Scheidler, N. H., C. Liu, K. A. Hamby, F. G. Zalom, and Z. Syed. 2015. Volatile codes:
Correlation of olfactory signals and reception in Drosophila-yeast chemical communication.
Scientific reports 5: 14059.
Stephen, W. P., and P. F. Torchio. 1961. Biological notes on the leaf-cutter bee, Megachile
(Eutricharaea) rotundata (Fabricius). Pan-Pac. Entomol. 32: 85-92.
Vandenberg J.D. Stephen W.P. 1982. Etiology and symptomology of chalkbrood in the alfalfa
leafcutting bee, Megachile rotundata. J. Invertebr. Pathol . 39: 133–137.