Content uploaded by Daniel Aureliano Newman
Author content
All content in this area was uploaded by Daniel Aureliano Newman on Dec 06, 2018
Content may be subject to copyright.
Effects of nectar robbing on nectar dynamics and bumblebee foraging
strategies in Linaria vulgaris (Scrophulariaceae)
Daniel A. Newman and James D. Thomson
Newman, D. A. and Thomson, J. D. 2005. Effects of nectar robbing on nectar dynamics
and bumblebee foraging strategies in Linaria vulgaris (Scrophulariaceae). /Oikos 110:
309 /320.
Differences in morphology among bumblebee species sharing a nectar resource may
lead to variation in foraging behaviour and efficiency. Less efficient bumblebees might
opportunistically switch foraging strategies from legitimate visitation to secondary
robbing when hole-biting primary robbers are present. We observed various aspects of
pollination and nectar robbing ecology of Linaria vulgaris in the Colorado Rocky
Mountains, with emphasis on the role of bumblebee proboscis length. Bees can extract
nectar from a nectar spur legitimately, by entering the front of the flower, or
illegitimately, by biting or reusing holes in the spur. Although L. vulgaris flowers are
apparently adapted for pollination by long-tongued bees, short-tongued bees visited
them legitimately for trace amounts of nectar but switched to secondary robbing in the
presence of primary robbers. Longer-tongued bees removed more nectar in less time
than did shorter-tongued bees, and were less likely to switch to secondary robbing even
when
/100% of flowers had been pierced. As the proportion of robbed flowers in the
population increased, the relative number of legitimate visits decreased while the
relative number of robbing visits increased. Robbing decreased nectar standing crop
and increased the proportion of empty flowers per inflorescence. Despite these
potentially detrimental effects of robbers, differences in inflorescence use among
robbers and pollinators, and the placement of holes made by primary robbers, may
mitigate negative effects of nectar robbing in L. vulgaris. We discuss some of the
reasons that L. vulgaris pollination ecology and growth form might temper the
potentially negative effect of nectar robbing.
D. A. Newman and J. D. Thomson, Dept of Zoology, Univ. of Toronto, 25 Harbord
Street, Toronto, Ontario, Canada, M5G 3G5 (danewman@zoo.utotonto.ca) and Rocky
Mountain Biological Laboratory, Gothic, CO 81224, USA.
Floral adaptations that promote pollen transport by
pollinators, which include some of the most striking
examples of plant/animal coevolution, are often inter-
preted as evidence of specialization to a specific polli-
nator type (Castellanos et al. 2003, 2004). Specialized
floral morphology may limit access to nectar and pollen
to visitors that effectively transport pollen, by excluding
inefficient pollinators. However, specialized flowers are
vulnerable to exploitation by parasites of plant/polli-
nator mutualisms (sensu Mainero and del Rio 1985)
that, by removing nectar without pollinating, may have
detrimental consequences for plant fitness (Roubik et al.
1985, Irwin and Brody 1998, 1999, 2000, Traveset et al.
1998, Navarro 2001).
Nectar robbers are animals that pierce flowers to
obtain nectar, usually without effecting pollination
(primary robbing). Other nectar-seeking animals, includ-
ing corruptible pollinators, can then exploit robber-made
holes (secondary robbing). Changes in pollinator beha-
viour due to nectar robbing may have positive, neutral,
or negative effects on plant fitness (Maloof and Inouye
2000, Irwin et al. 2001), although most authors consider
Accepted 18 January 2005
Copyright #OIKOS 2005
ISSN 0030-1299
OIKOS 110: 309 /320, 2005
OIKOS 110:2 (2005) 309
robbers to be detrimental (Proctor et al. 1996, Yu 2001,
Anderson and Midgley 2002, Stanton 2003). Negative
effects are typically attributed to pollinator avoidance of
robbed plants with reduced nectar rewards (Irwin and
Brody 1999, Irwin 2000, Navarro 2001); however,
pollinators that switch from legitimate visitation to
secondary robbing may be more problematic for plants
than pollinator avoidance. For example, bees are unlikely
to avoid robbed flowers, as they are not able to
discriminate between rewarding and unrewarding flow-
ers without sampling (Maloof and Inouye 2000). Bee-
pollinated flowers are therefore likely to receive at least
some legitimate visits per bee, provided not all visitors
switch to secondary robbing. If bees are numerous,
female fitness should not suffer much (Irwin et al.
2001) and male fitness may even increase (Zimmerman
and Cook 1985, Richardson 2004, but see Irwin 2003). If
avoidance of robbed flowers does not occur, switches to
secondary robbing will divert pollinators from serving
the plants; this may prove costly to the robbed plants if
pollinators in the population are limiting. If avoidance of
robbed flowers does occur, secondary robbers will
compound its costs by depleting the remaining nectar
rewards that could entice less wary pollinators. Despite
these potential impacts of nectar robbers on plants, we
know of no ecological studies investigating the role of
secondary robbers. It is noteworthy that agricultural
papers on nectar robbing typically refer not to the
problems of pollinator deterrence, but to secondary
robbing (Free 1962, 1965, Eaton and Stewart 1969,
Poulsen 1973 cited in Goulson 2003, Dedej and
Delaplane 2003, 2004). Indeed, Eaton and Stewart
(1969, p. 150) were sufficiently alarmed by the rate
at which honeybees were abandoning legitimate visita-
tion of Vaccinium corymbosum crops to use holes made
by B. occidentalis occidentalis that they recommended
‘‘eliminating this subspecies from areas of high-
bush blueberry plantings’’. Their concerns may
have been warranted: Dedej and Delaplane (2004) found
that secondary robbing of the related V. ashei by
honeybees caused significant reductions in seed set and
fruit size.
Maloof and Inouye (2000), in the hope of spurring
research into finding generalities about nectar robbing,
listed four factors that may determine whether its effects
are negative, neutral or positive. (1) The identity of the
pollinators. Differences among pollinators may cause
species to react differentially to the activities of nectar
robbers; some may become secondary robbers, others
may avoid plants altogether, and others still may not
respond at all. A small change in the cost/benefit
balance of the interaction /such as the appearance of
an easier route to nectar /may cause some visitors to
switch from mutualistic pollinators to opportunistic
robbers. Therefore, the effects of nectar robbing on
plants may depend on how different pollinator species
respond to the presence of robbed flowers, on how
effective each pollinator is at transporting pollen (Wilson
and Thomson 1991, Thomson and Thomson 1992), and
on the relative abundances and visitation rates of each of
these species (Herrera 1987, Pellmyr and Thompson
1996). For example, if a plant’s most effective pollinator
is not deterred from robbed plants and does not become
a secondary robber, the fitness effects for the plant will
most likely be minimal; on the other hand, if only
mediocre or rare pollinators remain, robbed plants will
suffer relative to unrobbed neighbours. Even taxonomi-
cally similar pollinator species may vary: although some
authors have treated bumblebee species as a behaviou-
rally homogeneous group, Goulson et al. (1998) found
that species responded differently to floral display size,
and cautioned against treating all of them as equivalent
despite superficial similarities in foraging behaviour.
(2) The amount of nectar that robbers remove is also
an important factor in determining the impacts of nectar
robbing. Reductions in nectar standing crop can nega-
tively affect plants by discouraging pollinators (Irwin
2000), and increased variability in nectar volume can
positively affect plants by promoting outcrossing
(Harder and Real 1987, Biernaskie et al. 2002). The
effects of nectar robbers on nectar standing crop may
also be related to the morphology of the plant’s flowers.
In flowers that accumulate nectar at, or close to, the site
of nectar production (the nectaries), nectar robbers can
easily pierce corollas at that site, and therefore drain all
available nectar (e.g. Ipomopsis aggregata; Irwin 2000).
In flowers whose nectar is presented apart from its
production site, as is the case in some spurred flowers
(Vogel 1998), primary robbers may face a choice of
where to bite holes; they could either pierce flowers near
the nectaries to drain nectar as it is produced, or pierce
the spurs to obtain nectar accumulation. If primary
robbers pierce flowers far from the location of the
standing crop, secondary robbers may concentrate on
nectar easily accessible from the holes (e.g. the nectaries,
or the nectar ducts, Vogel 1998), while legitimate visitors
may still obtain nectar from the tip of the spur. Such an
effect might reduce the potential negative effects of
nectar robbers on nectar standing crop. (3) The plant’s
growth form may affect pollinator and robber behaviour
in ways that could attenuate or amplify the effects of
nectar robbing. Plant with many flowers, rich nectar
rewards, or clonal growth may suffer from extensive
receipt of self-pollen by pollinators using area-restricted
searches (Pyke 1978, Ohashi 2002); nectar robbers,
by reducing rewards or by creating increased variability
in nectar standing crop, might benefit plants by
encouraging pollinators to leave clonal patches.
Some plant architectures may also directly influence
of nectar robbers and their impacts on plants. The
degree of flower congestion on an inflorescence could
increase nectar robbers’ handling time, and has been
310 OIKOS 110:2 (2005)
hypothesized to dissuade nectar robbers by making
spurs or nectaries difficult to reach (Fogg 1950, Inouye
1983). (4) If alternate nectar resources are abundant,
pollinators faced with a nectar-robbed food source can
easily switch to other plant species. If such alternatives
are scarce, however, they may be forced to remain
constant (Irwin et al. 2001), and pollination will not be
strongly affected.
In the Colorado Rocky Mountains, we observed a
morphologically variable group of bumblebee species
foraging on a single, apparently highly specialized, plant
species. We paid specific attention to the differences in
tongue-length among species and to the role that nectar-
robbing bumblebees play in the relationships between a
plant, Linaria vulgaris, and its pollinators. We used the
first three of Maloof and Inouye’s (2000) aforemen-
tioned factors as guidelines for the following five
questions: (1) does the gradual increase in robbing
activity in L. vulgaris patches over a season affect
pollinator and secondary robber assemblages? (2) are
shorter-tongued bumblebees most likely to become
secondary robbers? (3) does bumblebee foraging effi-
ciency vary among species with different proboscis
lengths? (4) how does robbing affect nectar rewards
(standing crop) in L. vulgaris? and how might these
effects affect bumblebee behaviour? and (5) do legitimate
and robbing visitors differ in their use of L. vulgaris
inflorescences?
Methods
Study site and system
We conducted the study between July 24 and August
16, 2003, at the Rocky Mountain Biological Labo-
ratory (RMBL) in Gothic, Colorado (106859?15ƒN;
38857?30ƒW; 2900 m). Weather during this period was
consistently sunny in the mornings and more or less
cloudy in the afternoons; rain was rare and intermittent
late in the day, but became increasingly frequent and
persistent by the second week of August. In all, we
observed plants in ten patches within the town of
Gothic; all of these were located in disturbed soils on
roadsides or near human housing, and contained
between approximately 50 to several thousand individual
ramets.
Linaria vulgaris Mill. (Scrophulariaceae) is a long-
lived perennial herb introduced to New England in the
1700s, and established in the North American west since
at least the early 1900s. It is a noxious weed in pastures
and other agricultural lands, due in large part to its
aggressive clonal propagation and prolific production of
small, winged seeds (Saner et al. 1995). Ramets have
several racemose inflorescences with numerous yellow
flowers (per inflorescence mean9
/SD in this population:
9.459
/3.64; n/155; range/1/39) that produce nectar
at the base of the ovaries, but present it in the end of long
spurs (mean9
/SD in this population: 13.379/3.04 mm;
n
/673). Older flowers are generally found lower on the
inflorescence. The vestibular flowers are kept closed,
and, as such, limit the availability of nectar rewards to
pollinators that are strong enough to pry open the
corolla lips (bumblebees and other robust bees). Very
small insects, and hummingbirds and hawkmoths whose
mouthparts are thin and long enough to fit between the
corolla lips can also obtain nectar, although they
probably do not pollinate (D. A. Newman, pers. obs.).
Linaria vulgaris flowers conform to the classic bumble-
bee syndrome, and authors typically describe them as
being adapted to pollination by long-tongued bumble-
bees, stating that shorter-tongued bees are unable to
reach the nectar legitimately (Hill 1909, Proctor et al.
1996, Stout et al. 1998, 2000, Corbet et al. 2001). Linaria
vulgaris is particularly useful for foraging studies
because its nectar spurs are translucent, a feature that
allowed us to measure the height of nectar (which
correlates strongly with volume; r
S
/0.98, p B/0.01;
Nepi et al. 2002) without removing it or otherwise
damaging the plant.
Of nine bumblebee species observed at the RMBL in
2003, five of them were common enough on L. vulgaris
for inclusion in statistical analyses; of those, only
three are included in all experimental and observational
studies described below. Bombus flavifrons Cresson,
a species with an intermediate tongue length (7.3 mm,
Pyke 1982; tongue lengths listed here refer only
to workers) was the most common. Bombus bifarius
Cresson, a short-tongued (5.8 mm, Pyke 1982)
species and B. appositus Cresson, a large long-tongued
(10.5 mm, Pyke 1982) species, were also abundant.
Bombus occidentalis Greene, a large short-tongued
(5.7 mm, Pyke 1982) species well known for its ability
to rob flowers, was rare early in the season but became
more common by the end of July; it does not figure in all
studies in this paper due to its rarity in the early summer.
Bombus occidentalis robs L. vulgaris flowers by chewing
holes in the spur (Corbet et al. 1981, Stout et al. 1998,
2000, Irwin et al. 2001, Irwin and Maloof 2002). Bombus
californicus Smith, another large long-tongued species,
was only occasionally common, and then only in some
sites; like B. occidentalis, it does not figure in every study
for this reason. These five species are very easy to
identify, even in flight, due to marked differences in size
and pile colour patterns. Because size variation may be
partially responsible for differences in foraging efficiency
among species, we measured the radial cell length
(Harder 1982) of worker bumblebees of the four species
listed above. We measured the radial cell on each bee’s
right wing with callipers. These bees were either caught
wild near the RMBL in 2002 and 2003, or borrowed
from the RMBL insect collection.
OIKOS 110:2 (2005) 311
Study 1. Bumblebee abundances and foraging
strategies
On nine sunny days between July 17 and August 14, we
haphazardly chose a cluster of five ramets in bloom (i.e.
close neighbours), and observed bumblebee activity for
five minutes. We repeated this protocol twice a day (once
in the mid morning and again in the mid afternoon) at
ten sites in the Gothic townsite, except in mid July when
only two patches were blooming. During these observa-
tion periods, we identified visiting bumblebees to species,
and recorded each individual’s foraging strategy (legit-
imate versus robbing).
Study 2. Bumblebee foraging efficiency
Nectar removal and handling time (experimental)
In order to understand why some pollinator species
switch to secondary robbing, we compared the foraging
efficiency of bumblebee species with different proboscis
lengths. We collected inflorescences and kept them in
water overnight to allow nectar to accumulate in the
spurs. The following day, we cut off flowers until each
inflorescence had ten flowers, and individually numbered
them with permanent marker. We then measured nectar
height in the spurs with a digital calliper. Immediately
after this, we put out the inflorescence next to a natural
L. vulgaris patch, and waited for the first bumblebee
visit. We videotaped the visits with a handheld digital
video camera in order to measure handling times for
individual bees, which we did by watching the video
sequences in slow motion (5
/real time). We measured
handling time (access time
/probing time) from the
moment the bumblebee aimed its head downward to
open the flower until it pulled out of the flower and
groomed its head and thorax (both the starting and
ending behaviours are highly stereotypical, and similar
across species). After the bee left the inflorescence, we
re-measured the height of nectar in the spurs. We
conducted this experiment for legitimate visitors, using
unrobbed inflorescences, and for robbers using pre-
viously robbed inflorescences.
Handling time (observational)
Since we had allowed unusually large amounts of nectar
to accumulate in the experiment described above, hand-
ling times by bumblebees were artificially long. For a
more natural account of handling time differences
among species, we also followed wild bumblebees fora-
ging on unmanipulated L. vulgaris plants and video-
taped them on as many flowers as possible. Later, we
viewed the videocassettes and measured handling time as
above.
Study 3. Nectar in robbed and unrobbed flowers
Nectar standing crop in robbed and unrobbed flowers
On July 28 and August 2, we randomly collected flowers
from L. vulgaris in five patches in the townsite of
Gothic. We only removed one flower per raceme. We
scored each flower as robbed or unrobbed, and measured
the height of nectar in the spur. On the morning of
August 15, we covered the spurs of 46 haphazardly
chosen flowers with the tapered end of disposable
micropipette tips, which we stuck on with Minwax
TM
Polycrylic finish (at this time, we had to actively protect
spurs since robbers were biting almost all available
flowers). As a sham control, we also glued half-pipette
tips onto the spurs of 46 other flowers in a way that
would allow robbers to pierce or reuse holes. After six
hours, we collected the flowers and measured the nectar
height as described above.
Location of bites on robbed flowers
On August 14 and 17, to investigate whether robbers bite
holes at specific location on the nectar spur, we excluded
bumblebees from several racemes to allow nectar accu-
mulation, measured nectar standing crop and spur
length on 100 marked flowers, and placed the inflor-
escences into natural L. vulgaris patches with high
bumblebee activity. Four hours later, we collected the
marked flowers and measured the location of the robber
holes (i.e. the distance between the tip of the spur and
the bottom of the hole).
Study 4. Vertical patterns of nectar standing crop
and bumblebee behaviour on L. vulgaris
inflorescences
Bumblebee visitation patterns on inflorescences: location
of flower visits on inflorescences
Bumblebee species may visit the vertical inflorescences
of L. vulgaris differently, with potential consequences
to pollen transport and the effects of nectar robbing.
On August 12/13, and 15 /16, we randomly chose
an inflorescence (typically one that had more than
10 flowers) and observed the behaviour of the first
bumblebee visitor. We scored the visit as occurring
on the bottom, top, or both halves of the inflorescence,
and identified the bumblebee species and its foraging
strategy.
Bumblebee visitation patterns on inflorescences: effect of
flower congestion
Along with the observations above, we scored the flower
density in of each inflorescence into three categories:
sparse (no overlap between flowers and the spurs
of flowers above them), intermediate (some overlap
between flowers, but spurs of the upper flowers are
312 OIKOS 110:2 (2005)
clearly visible), and dense (flowers almost completely
hide the spurs of the upper flowers).
Nectar standing crop, nectar accumulation, and flower
location on inflorescences
We collected 112 inflorescences between August 3 and 8,
and, recording the vertical location of flowers along the
inflorescence, measured nectar standing crop and the
presence or absence of robber holes. To compare nectar
standing crop (which is at least partially determined by
the activity of floral visitors) with actual nectar produc-
tion rates, we also covered patches of inflorescences with
metal cages covered with bridal veil (0.5 m in length and
width, 1.0 m tall) to prevent bee visitation. Twenty-four
hours later, we collected inflorescences and took the
same measurements as described above.
Results
Floral visitors
Seven bumblebee species visited L. vulgaris. In addition
to the five species listed above, we occasionally saw
B. mixtus, a short-tongued species (all secondarily
robbing) and B. nevadensis, a long-tongued species (all
legitimate except once, by a particularly small worker).
In late June 2003, bumblebee numbers were unusually
low at the RMBL (N. M. Waser and D. W. Inouye, pers.
comm.), but numbers grew rapidly in early July; by late
July, bumblebees were so numerous in all L. vulgaris
patches that we very frequently observed more than one
bee per inflorescence. Anthophora furcata-terminalis
Cresson was an uncommon but regular legitimate
visitor throughout the summer. Aside from bumblebees,
secondary robbers included ants (Formica fusca ,
F. lasioides,Tapinoma spp) and unidentified lepidoptera
and diptera. Long-tongued bumblebee species (B. appo-
situs,B. californicus and B. nevadensis) only visited
legitimately, with a single aforementioned exception. We
never saw B. occidentalis visiting legitimately.
Bumblebee species collected at Gothic, Colorado,
varied significantly in radial cell length (Kruskal/Wallis
test for workers only: H
/42.167, df/3, pB/0.0001,
Fig. 1). Larger species, with the exception of B.
occidentalis, have longer tongues.
Study 1. Bumblebee species abundances and
foraging strategies
Before B. occidentalis began foraging on L. vulgaris
around July 24, all flower visits to the plant were
legitimate. Bombus flavifrons and B. appositus made
most of these visits, although the short-tongued
B. bifarius was also quite common (Fig. 2a). Only a
few days after we observed the first nectar robbers, some
bees had already begun to secondarily rob (Fig. 2b). By
the second week of August, secondarily robbing
B. flavifrons and B. bifarius were making almost all of
the observed visits, although a small proportion of
individuals from both species remained legitimate
throughout the study. Bombus occidentalis , although
active enough to pierce nearly 100% of the flowers at
Fig. 1. Radial cell length differences among the four common
bumblebees species used in this study. Bombus bifarius and
B. occidentalis have short proboscides, B. flavifrons has an
intermediate proboscis length, and B. appositus has a long
proboscis. Bombus californicus is excluded due to lack of
specimens available for measurement.
Fig. 2. Mean number of (a) legitimately visiting and (b)
primary and secondary robbing bumblebees during five-minute
observations of L. vulgaris patches. Dates refer to days in July
and August 2003; note that this axis is not to scale. B. appositus
and B. californicus are shown together under ‘‘long-tongued’’,
although B. appositus represents the great majority of observa-
tions in this category. The ‘‘others’’ category includes rare
bumblebees (B. mixtus,B. nevadensis ) and non-Bombus insects.
OIKOS 110:2 (2005) 313
the RMBL, only made up a small proportion of the total
observed visits to L. vulgaris. Although all bumblebee
populations increased throughout the summer, the
absolute number of legitimate visits by all species did
not increase significantly (r
S
/0.330, p /0.368, n /9),
and the proportion of legitimate visits by all species
declined significantly (r
S
//0.905, p/0.0008, n/9),
as did the proportion of visits by long-tongued
species (r
S
//0.867, p/0.002, n/9; most of these
are B. appositus).
Study 2. Foraging efficiency of bumblebee species
The long-tongued B. appositus was much more efficient
than the other species in both handling time and nectar
removal (Table 1). Bombus flavifrons was intermediately
efficient. Bombus bifarius workers have such short
proboscides that they could very rarely reach the nectar
standing crop legitimately; therefore the nectar removal
data appear to show zero reward for this species.
However, these workers licked the back of spurs along
the duct that channels nectar (Vogel 1998), and were still
getting a small, if immeasurable, reward. Although some
B. flavifrons workers could reach up to half of the
available standing crop, others only licked the nectar
duct. Similarly, all three species of primary and second-
ary robbers did not drink from the nectar standing crop
while robbing, but licked upward along the nectar duct
to probe the base of the ovary where the nectaries are
located (Fig. 3). Very few robbers visited the inflor-
escences that we set out for them; this is puzzling,
because robbers were much more common than legit-
imate visitors at the time of the experiment.
Study 3. Effects of robbing on nectar standing crop
Nectar standing crop in robbed and unrobbed flowers
Out of 68 flowers randomly collected on July 28, 38
(56%) were robbed. Unrobbed flowers contained sig-
nificantly more nectar than robbed flowers (mean9
/SE:
1.1829
/0.303 mm and 0.6329/0.222 mm, respectively;
Mann/Whitney U test; Z/2.379, p/0.017), and were
significantly more likely to contain at least some nectar
(x
2
/15.583, df /1, p /0.017). Every one of the 135
flowers collected on August 2 was robbed, and therefore
no nectar comparison was possible. Out of 46 ‘‘sham-
protected’’ flowers on August 15, 45 (98%) were robbed
after six hours. In this case, flowers that we had
protected from robbers had significantly more nectar
than those whose spurs were only partly covered by
pipette tips (mean9
/SE: 1.139/0.227 mm and 0.3199/
0.070 mm, respectively; Mann/Whitney U test;
Z
/2.067, p/0.039).
Location of bites on robbed flowers
In unmanipulated flowers, most of the holes made by
B. occidentalis were located just below the level of the
ovary, and therefore any nectar in the spurs was out of
tongue’s reach for most visitors; indeed, almost all
primary and secondary robbing visits involved bees
licking upwards along the back of the spur (in the
‘‘nectar duct’’, Vogel 1998; Fig. 3a). Bombus flavifrons
workers occasionally licked downwards into the spur,
and their relatively long proboscides could probably
reach some of the standing crop.
Out of the 100 flowers that we left out for four hours,
98 were robbed. The locations of holes relative to the
height of nectar in the spurs suggest that B. occidentalis
makes a decision of where to bite based on the nectar
Table 1. Differences in foraging behaviours and efficiency among three bumblebee species visiting Linaria vulgaris, using legitimate
(L) and robbing (R) strategies. Means and standard errors are shown. Analyses are Kruskal/Wallis H tests. Unsuccessful visits
occurred when bees attempted to open flowers, but could not. Bombus occidentalis and B. californicus visited too rarely to be
included in analyses. Sample sizes in the experimental study are: B. bifarius (L
/5, R/5), B. flavifrons (L/34, R/3) and
B. appositus (L
/22). In the observational study (*), sample sizes are: B. bifarius (L/8, R/22), B. flavifrons (L/9, R/13) and
B. appositus (L
/19).
Variable Strategy bifarius flavifrons appositus Hp
Access time (s)* L 2.94 (0.45) 3.01 (0.30) 2.05 (0.25) 8.324 0.015
R 1.76 (0.26) 1.47 (0.35) /1.593 0.451
Access time (s) L 9.07 (2.10) 9.28 (1.11) 4.91 (0.61) 12.31 0.002
R 2.57 (0.26) 4.35 (2.37) /0.125 0.724
Flowers visited L 3.33 (0.49) 6.47 (0.87) 7.81 (1.03) 4.716 0.095
R 3.50 (0.65) 7.00 (4.51) /0/
Unsuccessful visits L 1.33 (0.62) 0.73 (0.19) 0.10 (0.07) 9.653 0.080
R 0.50 (0.29) 0 (0) /1.800 0.180
Nectar removal (0.1 mm) L 0.43 (0.04) 32.80 (8.50) 147.34 (28.35) 18.80 B
/0.001
R 3.08 (2.30) 0.24 (0.12) /1.559 0.212
Foraging efficiency** L 0.07 (0.04) 3.62 (0.93) 30.49 (4.84) 23.64 B
/0.001
R 1.06 (0.70) 0.06 (0.04) /1.559 0.212
*Results from bumblebees visiting natural L. vulgaris inflorescences (study 1 /observational). All other results are from the
experimental runs with manipulated inflorescences (study 2-experimental). Sample sizes for this experiment are shown in superscript
after the standard errors.
**Foraging efficiency
/(nectar removal)/(handling time).
314 OIKOS 110:2 (2005)
standing crop. The location of the bite along the spur
was negatively correlated with the original level of nectar
in the spur (r
S
//0.467, n/98, pB/0.00001); in flowers
with small quantities of nectar, primary robbers bit the
upper part of spurs, far from the standing crop. In
flowers with greater quantities of nectar, robbers bit
holes lower on the spurs, within reach of the standing
crop. This result corroborates our observations that
robbers drink from the accumulated nectar when stand-
ing crop is high, but directly from the nectaries at the
base of the ovaries when it is low. Robbers rarely made
holes lower than halfway down the spur; only 12.6% of
holes were below the middle of the spur, despite the
presence of collected nectar in the tip. There was no
apparent sidedness to the location of holes; most were
either double slits (one on the right and one of the left
sides of the spur) or single holes in the centre of the spur.
Study 4. Inflorescence partitioning among
bumblebee species, and its effects on nectar
dynamics
Bumblebee visitation patterns on inflorescences: location
of flower visits on inflorescences
Bumblebee species visited L. vulgaris inflorescences
differently (Fig. 4). Smaller species (B. bifarius and
B. flavifrons), those most likely to be secondary
robbers, visited primarily the bottom half of inflores-
cences, while the opposite was true for the larger, long-
tongued, and exclusively legitimate, species (B. appositus
and B. californicus). Long-tongued bees were more likely
than short- and intermediate-tongued bees to visit the
top half of the raceme (x
2
/54.86, df /2, p B/0.000001).
Short- and intermediate-tongued bees did not visit
inflorescences differently (x
2
/1.587, df /1, p /0.208).
Bombus occidentalis tended to prefer the bottom half
Fig. 3. (a) Bombus bifarius
worker licking the nectary (N)
at the base of a Linaria vulgaris
ovary (O) (flower is shown in
partial cross-section) through a
hole made by B. occidentalis
near the ovary; the nectar
standing crop is in the bottom
of the nectar spur (S) where
secondary robbers cannot reach
it. (b) Bombus flavifrons
probing nectar legitimately
from L. vulgaris. The bee uses
its weight and sometimes, as is
illustrated here, its foreleg to
hold the flower lips open.
Drawings by D. A. Newman
adapted from photographs by
D. A. Newman.
Fig. 4. Proportion of visits by five bumblebee species to the
lower (white bars) and upper (grey bars) flowers on L. vulgaris
inflorescences. Species are arranged from short-tongued on the
left to long-tongued on the right. Some proportions exceed
100% because some individuals visited both lower and upper
halves of the inflorescences.
OIKOS 110:2 (2005) 315
of the inflorescence, although less so than the
smaller species (x
2
/3.429, df /1, p /0.064), and
most often visited both top and bottom halves. Regard-
less of species, robbing individuals were more likely
to visit the bottom half of inflorescences while the
opposite was the case for legitimate visitors (x
2
/48.86,
df
/1, pB/0.000001). Unlike Corbet et al. (1981),
we saw no robbers working upside-down on the inflor-
escences.
Bumblebee visitation patterns on inflorescences: effect of
flower congestion
Flower congestion significantly affected inflorescence
use by nectar robbers. Robbers only visited upper flowers
on dense inflorescences in 5.26% of the observations,
compared to 16.27% in intermediate and 25.0% in sparse
inflorescences (x
2
/6.40, df /2, p /0.041). There was
no such effect for legitimate visitors (x
2
/0.220, df /2,
p
/0.896).
Nectar standing crop, nectar accumulation, and flower
location on inflorescences
Empty flowers were frequent and flowers with nectar
occurred along the entire length of the inflorescence.
However, nectar standing crop was significantly higher
in the upper half of the inflorescences, and significantly
more flowers in the lower half were empty (Table 2).
In addition, significantly more flowers in the lower
half of the inflorescences were robbed than in the top
half (Table 2). In contrast, in the 24-hour nectar
accumulation treatment, there were no differences in
either nectar accumulation or in percent empty flowers
between the bottom and top halves of inflorescences.
Not surprisingly, the significant difference in percent
robbed flowers between the lower and upper flowers
remained (Table 2).
Discussion
Nectar robbing in Linaria vulgaris significantly reduced
nectar standing crop and visitation rates by legitimate
pollinators. These changes might imply that robbing has
negative fitness consequences for the plant (Maloof and
Inouye 2000), but two European (Stout et al. 2000, Nepi
et al. 2002) and one North American (Irwin and Maloof
2002) studies on the effects of nectar robbing on female
fitness on L. vulgaris have found no reductions in seed
set. This ability to tolerate the potentially damaging
effects of nectar robbing may be due to aspects of its
growth form and to primary robber behaviour. In this
study, we found that inflorescence structure and patterns
of flower maturation, as well as the spatial separation of
nectar accumulation from the nectaries, might act as
‘tolerance traits’ that neutralize the impacts of nectar
robbing.
Robbed flowers had less nectar than unrobbed flowers
under both natural and experimental conditions. This
finding is consistent with Stout et al. (2000) for the same
plant species in England, although the differences they
found were not significant due to high variability in
nectar standing crop. We, like Stout et al. (2000), also
found that robbed flowers were more often empty than
unrobbed flowers. Most flowers in the population were
robbed, but it is conceivable that local patches unaf-
fected by robbers might benefit from higher nectar
standing crops, and increased legitimate visitation,
than those of other plants (Irwin and Brody 1998,
1999, 2000).
Primary robbers might be considered ‘‘resource
engineers’’ that make a food source available or easier
to obtain to other species. As B. occidentalis workers
pierced an increasing number of flowers in the popula-
tion, secondary robbing made up a greater number
and proportion of the total visits to L. vulgaris. Indeed,
even if it is faster and easier than visiting legitimately,
secondary robbing will only be worthwhile if individuals
Table 2. Results from Wilcoxon matched pairs tests comparing the lower and upper halves of Linaria vulgaris inflorescences in
nectar standing crop and nectar accumulation variables. Nectar standing crops were measured immediately after unmanipulated
inflorescences were collected. Nectar accumulation was measured after 24 hour pollinator exclusions. Note that since the standing
crop and accumulation data were collected from different plants, dates, and patches, standing crop and accumulation values should
not be compared quantitatively.
Variable Inflorescence location Mean (SE) Z-value n p-value
Standing crop nectar height (mm) bottom half 0.36 (0.05) 7.62 112 B
/0.00001
top half 0.93 (0.08)
% empty flowers bottom half 67.99 (2.41) 6.72 112 B
/0.00001
top half 45.70 (2.314)
% robbed flowers bottom half 98.32 (0.62) 3.33 112 B
/0.001
top half 94.42 (1.31)
Accumulation nectar height bottom half 0.576 (0.064) 0.69 31 0.491
top half 0.506 (0.053)
% empty flowers bottom half 59.87 (15.79) 0.26 31 0.795
top half 60.87 (16.82)
% robbed flowers bottom half 100.0 (0.0) 2.95 31 0.003
top half 92.009 (2.218)
316 OIKOS 110:2 (2005)
do not have to waste much time seeking flowers
with holes (Inouye 1983). Secondarily robbing bumble-
bees appear to find holes in nectar spurs while crawling
from flower to flower on racemes; following the
discovery, there is a period during which individuals
alternate between visiting legitimately and illegally
(D. A. Newman, pers. obs.). Provided there are enough
holes in flowers, secondary robbing seems to become a
fixed strategy for most individuals. In a flight cage
experiment, short-tongued B. rufocinctus workers
trained to visit L. vulgaris legitimately quickly (within
one or two foraging trips) switched to secondary
robbing, but then resisted switching back to legitimate
visitation when we removed robbed plants. Instead,
individuals flew around inflorescences for several min-
utes, presumably to look for holes, before returning
to the colony, and only switched back to legitimate
visitation after more than 24 hours (D. A. Newman,
pers. obs.). Not far from the RMBL, B. bifarius and
B. flavifrons workers only began to secondarily rob
Penstemon strictus when the frequency of robbed flowers
approached 100% (K. Ohashi, pers. comm.).
Bumblebee species used L. vulgaris inflorescences
differently. Secondary robbers preferentially used the
lower half of inflorescences while legitimate visitors
used the top half. This is likely due to the increased
probability that older flowers, which are located lower on
the inflorescences, will have been pierced by B. occiden-
talis (Corbet et al. 1981), and perhaps because lower
flowers (whose spurs are exposed) are easier to rob when
inflorescences are congested. While almost all of the
flowers in the population we studied were robbed,
flowers in the top half of inflorescences were more likely
to be unrobbed, and had more nectar than flowers on
the bottom half. Young flowers, whose corollas were not
yet fully developed, had more nectar on average than
mature flowers, although the difference was not signifi-
cant, and were less often robbed than their mature
counterparts (x
2
/18.331, df /1, p B/0.0001). Bombus
appositus and occasionally B. flavifrons preferred these
younger flowers, spending considerable amounts of time
trying, often unsuccessfully, to open them. The parti-
tioning of flower use among bumblebees using different
foraging strategies may reduce or neutralize potential
negative effects of nectar robbing. Linaria vulgaris
flowers begin anthesis before they open (Corbet et al.
1981), and most anthers in the open flowers we observed
were already empty (D. A. Newman, pers. obs.). There-
fore, if legitimate pollinators are visiting mostly young
flowers on the upper end of inflorescences, they are likely
importing enough outcross pollen to guarantee full seed
set, and exporting all of the pollen produced by
individual flowers, even before primary robbers have
attacked all of the flowers on the plant, or before
secondary robbers have the time to deplete nectar
rewards. Although not a defensive trait, the L. vulgaris
inflorescence arrangement, in combination with its
patterns of flower development and anthesis, might act
as a tolerance mechanism that reduces or eliminates any
detrimental effect of nectar robbers. Linaria vulgaris
appears to keep its older (lower) flowers long after they
are useful for reproduction. It is likely that such a
pattern of flower longevity might benefit the plant by
increasing attractiveness to pollinators, as Ishii and
Sakai (2001) showed in Narthecium asiaticum. This
characteristic has the additional benefit of keeping
most robbing visits away from the younger flowers.
Our observations of wild bees also suggested that the
degree of flower congestion along an inflorescence
(mean9
/SD in this population/3.009/0.869 flowers
per cm; range
/1.05 to 5.94) might also reduce the
impact of secondary robbing by making the spurs
difficult to reach; although the robust B. occidentalis
can pry flowers apart to bite the spurs, the smaller
B. bifarius and B. flavifrons appeared to put consider-
able effort into reaching them, and often left inflores-
cences after several unsuccessful attempts. The
disproportionate preference by secondary robbers for
lower flowers in dense inflorescences supports this
observation. Although the notion that congested inflor-
escences might act as protection from nectar robbers is
not new (Fogg 1950, Inouye 1983), these are to our
knowledge the first reported observations and results
supporting the possibility that such protection might
occur in the field. Variation in flower congestion among
plants within and among patches may make some plants
less vulnerable to primary or secondary robbing.
Due to the location of the holes in most spurs (Fig. 3),
most short-tongued bees (B. bifarius,B. occidentalis and
even B. flavifrons) did not reach the nectar standing crop
even when they robbed; instead, they usually probed
upwards along the back of the nectar duct to the
nectaries. The location of holes far from the site of
nectar collection allowed any accumulation (i.e. in the
morning, or after a day with low bumblebee activity) to
be effectively ‘‘reserved’’ for legitimate long-tongued
visitors. When visitation was low or artificially pre-
vented, and nectar was allowed to accumulate, primary
robbers made their holes lower on the spurs, and were
able to drain flowers completely. Bombus occidentalis
workers probably find it difficult to bite holes near the
tapered ends of nectar spurs. Therefore, it appears that
the abundance of legitimate visitors can influence the
robbers’ choice of where to make holes in the flower. If
legitimate visitors are common, and are able to keep
nectar standing crops low, robbers appear to prefer
making holes closer to the nectaries; therefore any
residual nectar accumulation is protected from second-
ary robbers and more legitimate visits are likely. As
mentioned above, reductions in nectar standing crop due
to robbing are expected to affect plants negatively
(Maloof and Inouye 2000). However, we found that as
OIKOS 110:2 (2005) 317
long as nectar standing crop was low, B. occidentalis
made holes closer to the nectaries, and therefore left
nectar in the spur where only long-tongued pollinators
could reach it. This behaviour by hole-biting bees, in
plants with separate nectar production and nectar
storage sites, may reduce the likelihood that a flower
will be completely empty unless it has just been visited by
a legitimate visitor; it is likely that long-tongued bees will
continue to visit flowers that contain some nectar.
Although primary and secondary robbers might severely
or completely stem the flow of nectar into L. vulgaris ’
spurs on a daily basis, resulting in a higher percentage of
empty flowers following nectar depletion by a long-
tongued visitor, continuous nectar production would
replenish a standing crop in each flower overnight,
thereby increasing its odds of successful pollen deposi-
tion and removal.
Bumblebee species varied in foraging efficiency on
L. vulgaris flowers. The larger long-tongued species
could open flowers more rapidly, and removed a much
greater volume of nectar than their shorter-tongued
counterparts. Shorter-tongued bees removed only trace
amounts of nectar by visiting flowers legitimately, and
often could not reach the standing crop. Pollinating
bumblebee species also differed in their preference for
secondary robbing; shorter-tongued species, less efficient
legitimate visitors than longer-tongued species, were
more likely to exploit the holes made by B. occidentalis.
Bees with longer tongues can reach more of the available
nectar, so that their absolute reward is greater; in
addition, manoeuvring a long proboscis into a small
hole may make secondary robbing difficult or impossible
for long-tongued species (Ranta and Lundberg 1980,
Plowright and Plowright 1997). Furthermore, since
longer-tongued bees are also generally larger (Heinrich
1976, Harder 1983), obtaining nectar legitimately is
easier for these species, probably because their weight
aids in opening the flower. A plausible consequence of
this advantage is a more stereotypical visitation techni-
que relative to smaller species, which could result in
more precise pollen transfer from anthers to the stigmas
of other flowers.
Nectar robbers may radically alter the relationship
between L. vulgaris and its visitors because they provide
a new route to the nectar. Shorter-tongued bees will thus
find their access to the nectar made easier. Short-
tongued bees most likely benefit, since nectar is easier
to obtain in the presence of holes. Long-tongued bees
probably suffer from increased competition with a larger
community of nectar foragers due to robbing (Mainero
and del Rio 1985, Irwin and Brody 1998). In the
presence of holes, it is in the best interest of short-
tongued bumblebees to switch to secondary robbing,
provided the frequency of holes is high enough to make
the strategy sustainable. The reason that some B. bifarius
visited L. vulgaris legitimately despite the presence of
holes is most likely that they were collecting pollen
(Macior 1967), either exclusively or in addition to nectar.
Bombus flavifrons observed preference for secondary
robbing is puzzling in light of our foraging efficiency
measurements (Table 1). However, despite the apparent
trade off between handling time and reward that
bumblebees, especially B. flavifrons, face when they
secondarily rob instead of visiting legitimately, cheating
probably has other benefits that we did not measure. An
unmeasured energetic cost incurred by B. flavifrons
visiting L. vulgaris legitimately probably accounts
for the species’ preference for secondary robbing.
Indeed, it is commonly assumed that bees expend
much less energy by robbing than by opening closed
flowers (Macior 1966, Inouye 1983), and this seems
apparent based on our observations of smaller bees
struggling to enter L. vulgaris flowers.
The nectar robbing literature concentrates on the
effects of nectar-parasitism on plant fitness. However,
results have proved so variable among systems that
conclusions about the phenomenon elude generalization.
In L. vulgaris, nectar robbing is apparently not detri-
mental (Stout et al. 2000, Irwin and Maloof 2002, Nepi
et al. 2002), despite reductions in nectar standing crop
and in legitimate pollinator visitation. Most of the
reduction in legitimate visitation in this study appeared
to result from pollinators switching to secondary rob-
bing. In a year or area with a small long-tongued
bumblebee population, it is likely that nectar robbing
would have more severe effects, since most pollinators
visiting L. vulgaris would be short-tongued and therefore
prone to switch to robbing.
Our findings point towards some of the plant
traits that may mitigate the effects of nectar robbing.
Plants with racemose inflorescences whose flowers
mature along the vertical axis, and whose female and
male functions are easily fulfilled (L. vulgaris has sticky
pollen that is very effectively picked up in large
quantities by bumblebees), may not suffer from the
pollen limitation associated with nectar robbing, as do
some other species (e.g. Ipomopsis aggregata; Irwin and
Brody 1998, 1999, 2000). Spatial separation of nectaries
from the site of nectar accumulation, found in several
plant families (e.g. Scrophulariaceae, Violaceae, Solana-
ceae, Alliaceae, among others; Vogel 1998), might also
help plants tolerate nectar robbing if it prevents robbers
from completely draining the flowers they attack.
Although it may be impossible to find generalizations
about the impacts of nectar robbers on plants, identify-
ing traits such as these may help determine in what plant
species to expect negative or neutral effects of nectar
robbing.
Acknowledgements /The authors thank M. Danielczyk
and R. J. Smith for equipment, D. C. Darling, K. Ohashi,
F. H. Rodd and J. S. Thaler for useful discussion. We
are indebted to the Rocky Mountain Biological Labora-
tory for use of its hymenopteran collection. Funding was
318 OIKOS 110:2 (2005)
granted to D. A. Newman by the Natural Sciences and
Engineering Research Council of Canada, and by the
University of Toronto.
References
Anderson, B. and Midgley, J. J. 2002. It takes two to tango but
three is a tangle: mutualists and cheaters on the carnivorous
plant Roridula ./Oecologia 132: 369 /373.
Biernaskie, J. M., Cartar, R. V. and Hurley, T. A. 2002. Risk-
averse inflorescence departure in hummingbirds and bumble
bees: could plants benefit from variable nectar rewards?
/Oikos 98: 98 /104.
Castellanos, M. C., Wilson, P. and Thomson, J. D. 2003. Pollen
transfer by hummingbirds and bumblebees, and the diver-
gence of pollinator modes in Penstemon./Evolution 57:
2742 /2752.
Castellanos, M. C., Wilson, P. and Thomson, J. D. 2004.
‘Anti-bee’ and ‘pro-bird’ changes during the evolution of
hummingbird pollination in Penstemon./J. Evol. Biol. 17:
876 /885.
Corbet, S. A., Cuthill, I., Fallows, M. et al. 1981. Why do
nectar-foraging bees and wasps work upwards on inflor-
escences? /Oecologia 51: 79/83.
Corbet, S. A., Bee, J., Dasmahapatra, K. et al. 2001. Native or
exotic? Double or single? Evaluating plants for pollinator-
friendly gardens. /Ann. Bot. 87: 219 /232.
Dedej, S. and Delaplane, K. S. 2003. Honey bee (Hymenoptera:
Apidae) pollination of rabbiteye blueberry Vaccinium ashei
var. ‘Climax’ is pollinator density-dependent. /J. Econ.
Entomol. 96: 1215 /1220.
Dedej, S. and Delaplane, K. S. 2004. Nectar-robbing carpenter
bees reduce seed-setting capability of honey bees (Hyme-
noptera: Apidae) in rabbiteye blueberry, Vaccinium ashei ,
‘Climax’. /Environ. Entomol. 33: 100 /106.
Eaton, G. W. and Stewart, M. G. 1969. Blueberry blossom
damage caused by bumblebees. /Can. Entomol. 101: 149 /
150.
Fogg, G. E. 1950. Biological flora of the British Isles. Sinapis
arvensis L. /J. Ecol. 38: 415 /439.
Free, J. B. 1962. The behaviour of honeybees visiting field beans
(Vicia faba ). /J. Anim. Ecol. 31: 497 /502.
Free, J. B. 1965. The ability of bumblebees and honeybees to
pollinate red clover. /J. Appl. Ecol. 2: 289 /294.
Goulson, D. 2003. Bumblebees: behaviour and eco-
logy. /Oxford Univ. Press.
Goulson, D., Stout, J. C., Hawson, S. A. et al. 1998. Floral
display size in comfrey, Symphytum officinale L. (Boragina-
ceae): relationships with visitation by three bumblebee
species and subsequent seed set. /Oecologia 113: 502 /508.
Harder, L. D. 1982. Measurement and estimation of functional
proboscis length in bumblebees (Hymenoptera: Apidae).
/Can. J. Zool. 60: 1073 /1079.
Harder, L. D. 1983. Flower handling efficiency of bumble bees:
morphological aspects of probing time. /Oecologica 57:
274 /280.
Harder, L. D. and Real, L. A. 1987. Why are bumble bees risk
averse? /Ecology 68: 1104 /1108.
Heinrich, B. 1976. Resource partitioning among some eusocial
insects: bumblebees. /Ecology 57: 874 /889.
Herrera, C. M. 1987. Components of pollinator ‘‘quality’’:
comparative analysis of a diverse insect assemblage. /Oikos
50: 79 /90.
Hill, E. J. 1909. Pollination of Linaria with special reference to
cleistogamy. /Bot. Gaz. 47: 454 /466.
Inouye, D. W. 1983. The ecology of nectar robbing. /In:
Bentley, B. and Elias, T. (eds), The biology of nectaries.
Columbia Univ. Press, pp. 153/173.
Irwin, R. E. 2000. Hummingbird avoidance of nectar-robbed
plants: spatial location or visual cues. /Oikos 91: 499 /506.
Irwin, R. E. 2003. Impact of nectar robbing on estimates of
pollen flow: conceptual predictions and empirical outcomes.
/Ecology 84: 485 /495.
Irwin, R. E. and Brody, A. K. 1998. Nectar robbing in
Ipomopsis aggregata: effects on pollinator behaviour and
plant fitness. /Oecologia 116: 519 /527.
Irwin, R. E. and Brody, A. K. 1999. Nectar-robbing bumble
bees reduce the fitness of Ipomopsis aggregata (Polemonia-
ceae). /Ecology 80: 1703/1712.
Irwin, R. E. and Brody, A. K. 2000. Consequences of nectar
robbing for realized male function in a hummingbird-
pollinated plant. /Ecology 81: 2637 /2643.
Irwin, R. E. and Maloof, J. E. 2002. Variation in nectar robbing
over time, space, and species. /Oecologia 133: 525/533.
Irwin, R. E., Brody, A. K and Waser, N. M. 2001. The impact of
floral larceny on individuals, populations, and communities.
/Oecologia 129: 161 /168.
Ishii, H. S. and Sakai, S. 2001. Effects of display size and
position on individual flower longevity in racemes of
Narthecium asiaticum (Liliaceae). /Funct. Ecol. 15: 396 /
405.
Macior, L. W. 1966. Foraging behaviour of Bombus (Hyme-
noptera: Apidae) in relation to Aquilegia pollination. /Am.
J. Bot. 53: 302 /309.
Macior, L. W. 1967. Pollen-foraging behaviour of Bombus in
relation to pollination of nototribic flowers. /Am. J. Bot 54:
359 /364.
Mainero, J. S. and del Rio, C. M. 1985. Cheating and taking
advantage in mutualistic associations. /In: Boucher, D. H.
(ed.), The biology of mutualism: ecology and evolution.
Croom Helm, London, pp. 192/216.
Maloof, J. E. and Inouye, D. W. 2000. Are nectar robbers
cheaters or mutualists? /Ecology 81: 2651/2661.
Navarro, L. 2001. Reproductive biology and effect of nectar
robbing on fruit production in Macleania bullata (Erica-
ceae). /Plant Ecol. 152: 59/65.
Nepi, M., Pacini, E., Nencini, C. et al. 2002. Variability in
nectar production and composition in Linaria vulgaris (L.)
Mill. (Scrophulariaceae). /Plant Syst. Evol. 238: 109 /118.
Ohashi, K. 2002. Consequences of floral complexity for
bumblebee-mediated geitonogamous self-pollination in Sal-
via nipponica Miq. (Labiatae). /Evolution 56: 2414 /2423.
Pellmyr, O. and Thompson, J. N. 1996. Sources of variation in
pollinator contribution within a guild: the effects of plant
and pollinator factors. /Oecologia 107: 595 /604.
Plowright, C. M. S. and Plowright, R. C. 1997. The advantage
of short tongues in bumble bees (Bombus )/analyses of
species distributions according to flower corolla depth, and
of working speeds on white clover. /Can. Entomol. 129:
51 /59.
Poulsen, M. H. 1973. The frequency and foraging behaviour of
honeybees and bumblebees on field beans in Denmark. /J.
Apic. Res. 12: 75 /80.
Proctor, M., Yeo, P. and Lack, A. 1996. The natural history of
pollination. /Timber Press, Oregon.
Pyke, G. H. 1978. Optimal foraging in bumblebees: patterns of
movements between inflorescences. /Theor. Popul. Biol. 13:
72 /98.
Pyke, G. H. 1982. Local geographic distributions of bumblebees
near Crested Butte, Colorado: competition and community
structure. /Ecology 63: 555 /573.
Ranta, E. and Lundberg, H. 1980. Resource partitioning in
bumblebees: the significance of differences in proboscis
length. /Oikos 35: 298 /302.
Richardson, S. C. 2004. Are nectar-robbers mutualists or
antagonists? /Oecologia 139: 246 /254.
Roubik, D. W., Holbrook, N. M. and Parra, G. 1985. Roles of
nectar robbers in reproduction of the tropical treelet Quassia
amara (Simaroubaceae). /Oecologia 66: 161/167.
Saner, M. A., Clements, D. R., Hall, M. R. et al. 1995. The
biology of Canadian weeds. 105. Linaria vulgaris Mill.
/Can. J. Plant. Sci. 75: 525 /537.
OIKOS 110:2 (2005) 319
Stanton, M. L. 2003. Interacting guilds: moving beyond the
pairwise perspective on mutualistms. /Am. Nat. 162: S10/
S23.
Stout, J. C., Allen, J. A. and Goulson, D. 1998. The influence of
relative plant density and floral morphological complexity
on the behaviour of bumblebees. /Oecologia 117: 543 /550.
Stout, J. C., Allen, J. A. and Goulson, D. 2000. Nectar robbing,
forager efficiency and seed set: bumblebees foraging on the
self incompatible plant Linaria vulgaris (Scrophulariaceae).
/Acta Oecol. 21: 277 /283.
Thomson, J. D. and Thomson, B. A. 1992. Pollen presentation
and viability schedules in animal-pollinated plants: conse-
quences for reproductive success. /In: Wyatt, R. (ed.),
Ecology and evolution of plant reproduction. Chapman &
Hall, pp. 1 /24.
Traveset, A., Willson, M. F. and Sabag, C. 1998. Effect of
nectar-robbing birds on fruit set of Fuchsia magellanica in
Tierra Del Fuego: a disrupted mutualism. /Funct. Ecol. 12:
459 /464.
Vogel, S. 1998. Remarkable nectaries: structure, ecology,
organophyletic perspectives III. Nectar ducts. /Flora 193:
113 /131.
Wilson, P. and Thomson, J. D. 1991. Heterogeneity among
floral visitors leads to discordance between removal and
deposition of pollen. /Ecology 72: 1503 /1507.
Yu, D. W. 2001. Parasites of mutualisms. /Biol. J. Linn. Soc.
72: 529 /546.
Zimmerman, M. and Cook, S. 1985. Pollinator foraging,
experimental nectar-robbing and plant fitness in Impatiens
capensis./Am. Midl. Nat. 113: 84/91.
Subject Editor : Esa Ranta
320 OIKOS 110:2 (2005)
