Current Biology 22, 1635–1639, September 11, 2012 ª2012 Elsevier Ltd All rights reservedhttp://dx.doi.org/10.1016/j.cub.2012.06.058
Hawkmoth Pollinators Decrease Seed Set
of a Low-Nectar Petunia axillaris Line
through Reduced Probing Time
Anna Brandenburg,1,2,* Cris Kuhlemeier,1
and Redouan Bshary2
1Institute of Plant Sciences, University of Bern, Altenbergrain
21, 3013 Bern, Switzerland
2Institute of Biology, University of Neucha ˆtel, Emile Argand 13,
2009 Neucha ˆtel, Switzerland
Although deception of floral pollinators is well known
among orchids [1, 2], the majority of animal-pollinated
plants secure pollination by nectar rewards. The costs
and benefits of nectar production remain poorly under-
stood [3–5]. Here, we developed a crossing design to intro-
gress a low-nectar-volume locus of Petunia integrifolia into
the genetic background of P. axillaris. The resulting intro-
gression line resembled P. axillaris but produced only
one-third of the nectar volume. When exposed simulta-
neously to low-nectar and wild-type P. axillaris plants,
hawkmoth pollinators reduced their probing duration on
low-nectar plants but otherwise did not show any signs
of discrimination against these plants. However, reduced
probing duration resulted in reduced seed production in
the low-nectar plants despite their higher reproductive
potential as evidenced by hand pollination. In line with
this interpretation, we found a positive correlation between
probing duration and seed set, and hawkmoth pollination
of low-nectar plants that were manually supplemented
with nectar to parental levels yielded seed sets similar to
hand pollination. Thus, a simple self-serving pollinator
behavior—the adjustment of probing time in response to
nectar volume—may select against reducing nectar and
protect many plant-pollinator mutualisms against a drift
Results and Discussion
Mutualisms are cooperative interactions between individuals
from two or more species in which all individuals gain a net
benefit . They appear in many biological systems and are
believed to be critical in shaping nearly every existing eco-
system . Most of these interactions involve investments
(where cooperating leads to a reduction of the actor’s imme-
diate payoffs relative to individuals that do not cooperate)
by at least one partner [6–9]. The existence of investments
raises the question of which factors stabilize mutualisms
and prevent cheaters from spreading in the population. To
answer this question, one would ideally measure the fitness
consequences of an individual with reduced investment
relative to the wild-type. Here, we provide such data for a
plant-pollinator mutualism involving Petunia axillaris ssp.
axillaris N, a common laboratory strain of P. axillaris, and one
of its main pollinators, the tobacco hornworm moth Manduca
sexta, a nocturnal hawkmoth species.
Pollination mutualisms are usually asymmetrical interac-
tions in the sense that only the plant invests in the production
of a reward (typically floral nectar), whereas the pollinator
ensures reproductive success of the plant by cross-fertiliza-
tion as a by-product of self-serving foraging decisions. The
vast majority of flowering plant species provide nectar
rewards, suggesting that rewards are typically under positive
selection (deceptive orchids being a well-known exception;
for overview see [1, 2]). Nevertheless, there is disagreement
concerning the actual costs and benefits of nectar production
[10–12]. Optimal nectar volumes will depend strongly on polli-
nator behavior. Higher nectar volumes will make a plant more
attractive, but such benefits must be balanced against the
costs of production and the effects on flower visitation: if polli-
nators visit more flowers on the same plant and fewer plants
overall, this will increase inbreeding and thus be under nega-
tive selection [13, 14].
Manipulation of nectar quantities has been instrumental in
deducing pollinator decision rules with respect to nectar pro-
duction that affect the fitness of plants [4, 15–17]. In addition,
the reproductive success of deceptive versus nectar-
providing orchid species has been compared [5,18]. However,
these approaches measure either the costs or the benefits
of manual nectar manipulation. Here, we use a different
approach in which net fitness consequences for individual
plants with reduced nectar investment relative to conspecifics
may be studied as a function of pollinator behavior. We ex-
ploited the interspecific fertility between two Petunia species,
one with high nectar volumes (P. axillaris ssp. axillaris N, here-
after referred to as P. axillaris) and one producing low nectar
volumes (P. integrifolia ssp. inflata, hereafter referred to as
P. integrifolia). We introgressed a low-nectar locus from
P. integrifolia into the genetic background of P. axillaris by
repetitive backcrossing. First, we assessed all phenotypic
traits known to affect pollinator behavior. Next, we studied
the behavioral responses of tobacco hornworm moths to
both our introgression line and wild-type plants. We were
particularly interested in behaviors that may affect the fitness
of a plant [3, 19], namely whether pollinators avoid low-nectar
nectar volume. Finally, we tested whether probing duration
may affect female reproductive success, measured as seed
set (number of mature seeds per capsule) per flower after
one pollination event.
Phenotypic Comparison of Wild-Type P. axillaris
and Low-Nectar Introgression Line F25
Our aim was to produce an introgression line that would be
similar to P. axillaris in all phenotypic traits except for nectar
production. Our low-nectar introgression line (F25) largely
fulfilled these criteria (Table 1; see also Figure S1 available
online). F25 and control plants differed significantly in nectar
volumes, with F25 containing on average only 30% of the
nectar volume of P. axillaris (Table 1). The reduced nectar
volume in F25 plants was accompanied by half the total nectar
sugar compared to parental plants. Other phenotypic traits
including flower anatomy and pollen/ovule production were
similar, with the exception of methyl benzoate (MB) emission.
In F25, MB emission was nearly three times as high as in
P. axillaris (Table 1). Consistent with our backcross design,
the phenotype and genotype of our F25 line was much more
like P. axillaris than its other parent, P. integrifolia (Figure S1;
Table 1). For example, 65 genetic markers spread over
the seven chromosomes of Petunia were homozygous for
P. axillaris, whereas only one simple sequence repeat marker
retained the P. integrifolia genotype.
Pollinator Behavior during Choice Experiments
We simultaneously exposed one P. axillaris and one low-
nectar F25 plant to one hawkmoth at a time. We measured
first choice, number of flowers visited per plant, and probing
duration for each flower. There was no significant initial pre-
ference for either F25 (15 first approaches) or P. axillaris (21
first approaches) (n = 37, c2= 0.7, df = 1, p = 0.4). Similarly,
there was no difference in first choice when moths were
offered F25 supplemented with 15 ml of nectar, bringing F25
up to P. axillaris nectar levels (7 first approaches) and
P. axillaris (7 first approaches; n = 14).
The number of flowers visited was not significantly different
between F25 (median 3 flowers) and control plants (median 3
flowers) (Wilcoxon signed-rank test; n = 25, Z = 21.37, p =
0.171) (Figure 1A). There was also no significant difference
in the number of flowers visited between P. axillaris and F25
when the latter was supplemented with 15 ml of nectar
(Wilcoxon signed-rank test; n = 14, Z = 20.184, p = 0.854). In
this respect, our negative results differ from a previous study
on Petunia in which hawkmoths visited fewer flowers on
plants where nectar was manually removed .
Probing time per flower on F25 was significantly decreased
compared to control plants (Wilcoxon signed-rank test; n = 25,
Z = 24.39, p < 0.001). The mean probing time per flower on
F25 was 6.93 s, compared to 11.14 s on P. axillaris (Figure 1B),
representing a feeding time reduction of 47%. The difference
in probing time was restored when F25 was supplemented
with 15 ml of nectar (mean probing time 13.3 s) and tested
against P. axillaris (mean probing time 14.3 s; Figure 1B)
(Wilcoxon signed-rank test; n = 14, Z = 24.89, p = 0.625).
Taking these results together, we found no evidence for
pollinator discrimination between the two plant lines. The
only observed pollinator behavior that could potentially act
as a control mechanism against reduced nectar production
was probing duration. Therefore, next we analyzed how
probing duration affected seed set in our two Petunia lines.
The Effects of Probing Duration on Seed Set
We compared seed set in wild-type and F25 plants both after
after one hawkmoth visit. There was a significant interaction
between plant line and pollination method (Table S1). Post
hoc tests revealed that hand-pollinated seed set was sig-
nificantly higher (+20%) in F25 (n = 57, mean number of seeds
480 6 20 SE) than in P. axillaris (n = 56, mean number of seeds
383 6 20) (Tukey’s honestly significant difference [HSD] test;
p = 0.009). However, when both lines were pollinated by
moths, F25 produced significantly fewer seeds (n = 52, mean
number of seeds 328 6 22) than P. axillaris (n = 40, mean
number of seeds 424 6 25) (223%) (Tukey’s HSD; p = 0.03).
Additionally, moth-pollinated F25 plants showed a significant
decrease in seed set compared to hand-pollinated F25 plants
(232%) (Tukey’s HSD; p < 0.0001). The reproductive potential
of moth-pollinated F25 plants was restored when nectar
volumes were supplemented with 15 ml of nectar (n = 26,
mean number of seeds 468 6 30) (Tukey’s HSD; p = 0.997)
(Figure 2A). The number of seeds per flower was positively
correlated with hawkmoth probing time on P. axillaris and
F25 plants (n = 15, linear regression, R2= 0.336, p < 0.01;
Game theoretic models propose that in cases in which
payoffs are positively correlated with the duration of an inter-
action, an early ending in response to cheating by a partner
may indeed be a suitable mechanism to diminish the payoffs
for cheaters , as long as cheaters cannot easily find new
partners. This condition is apparently fulfilled in our system
because hawkmoth visits are low in wild P. axillaris popula-
tions , and several studies provide evidence that hawk-
moth-pollinated plants are actually pollinator limited [24, 25].
Table 1. Comparison of Phenotypic Traits
P. axillaris-F25P. integrifolia
Tube length D1 (cm)
Tube length D2 (cm)
Corolla diameter (cm)
Corolla UV reflectance
Median MB emission (pptv)
Number of pollen/flower
Number of ovules/ovary
Mean ovary weight (mg)
Nectar volume (ml)
Nectar sugar proportion (%)
Nectar sugar concentration (%w/v)
Stigma surface area (cm2)
2.1 6 0.1
1.2 6 0.1
5.4 6 0.3
185,360 6 12,990
207.2 6 30.6
5.2 6 0.6
34.7 6 6.8
2.1 6 0.1
1.3 6 0.1
5.6 6 0.2
196,289 6 29,961
196.2 6 31
5.3 6 0.9
10.4 6 4.1
p = 0.35
p = 0.35
p = 0.28
0.2 6 0.02
1.6 6 0.2
3.1 6 0.18
p < 0.001
p < 0.001
p < 0.001
p < 0.001p < 0.00117
p = 0.78
p = 0.485
p = 0.766
p < 0.001
101.35 6 0.47p < 0.001
16.53 6 1.5
30.75 6 4.3
p < 0.02p = 0.2626
p = 0.24310
Statisticalanalysisofaveragephenotypic parametersofP.axillaris,F25,andP.integrifolia.Allvaluesaregivenasaverages 6standarddeviation,exceptfor
methyl benzoate (MB) emission, given as median (pptv, particles per trillion volumetric). For an illustration of the similarities in morphological parameters
between P. axillaris and F25, see Figure S1.
Current Biology Vol 22 No 17
Therefore, it is reasonable to assume that one pollinator visit
per flower reflects natural conditions in our study system.
ductive potential of a flower is a function of pollinator probing
duration, which in turn is a function of nectar volume. Thus,
a very simple self-serving foraging rule of pollinators may
select against a reduction of nectar volume in flowering plants.
This mechanism is similar to host sanctions in legume-mycor-
rhiza interactions, where plants self-servingly grow roots in
areas where nitrogen fixation is highest, which is typically
where the most cooperative mycorrhiza lines live in the plant
nodules [26, 27].
Previous studies on the relationship between nectar pro-
duction and pollinator behavior have been limited to testing
the effect of nectar depletion or addition on pollinator
behavior. Although this approach is useful for testing changes
in pollinator behavior or either costs or benefits, a simulta-
neous measurement of costs and benefits and hence fitness
consequences cannot be assessed. In this context, our
approach of employing introgression lines offers additional
insights thatcannot be easily obtained in other ways.Although
the cost of nectar production is debated [10–12], if the cost
is high, it follows that plants may benefit from reducing nectar
production, if at all possible. In this context, two of our results
are of interest. First, we found increased production in the F25
line of MB, a floral volatile compound known to attract hawk-
moths [28–30]. Although there was no evidence in our experi-
ments that increased MB emission had an effect on hawkmoth
preference, our greenhouses were likely scent saturated,
which would have masked any positive effects. Therefore,
the potential benefits of increased scent production should
be tested under field conditions. However, the chemical struc-
ture of MB is so different from sugar  that we currently
consider it unlikely that there is a trade-off in the production
of both substances. More generally, trade-offs with respect
to the quantity of volatile productions are currently poorly
understood (see for example [32, 33]). A second, more likely
explanation for reallocating resources away from nectar
by sugar . We found evidence for increased potential seed
set in the F25 line through hand pollination, while hawkmoth-
pollinated wild-type plants showed an intermediate seed set.
Again, field studies will be informative in this regard because
natural light conditions will have a major influence on sugar
production and, therefore, the potential trade-off between
nectar production and seed set.
Understanding the causes of stable nectar production in flow-
ering plants has been a major challenge. Much remains to be
learned about the costs and benefits of nectar production in
flowering plants. Our novel use of a low-nectar introgression
line allows for comparative cost-benefit analysis of nectar
production in a single species. We have demonstrated that
reduced nectar production in P. axillaris causes reduced
seed set under greenhouse conditions, and that this reduction
is a function of decreased pollinator probing duration. Future
experiments should be conducted under field conditions
within the native range of P. axillaris and include measures of
male reproductive success. From a game theoretical perspec-
tive, our results add to the growing awareness that, though
complex partner control mechanisms may occur in mutual-
isms (for example, reputation-based access to partners ),
simple self-serving mechanisms appear to be most wide-
spread [6–8, 36, 37].
The study was conducted at the Institute of Plant Sciences, University of
Bern, from April 2007 through July 2011 using Petunia axillaris and an
introgression line (F25) with reduced nectar volume but otherwise similar
phenotype. For details on P. axillaris, growth conditions, and the breeding
program resulting in a low-nectar line, see Supplemental Experimental
Figure 1. Hawkmoth Behavior on P. axillaris, Low-Nectar Line F25, and Refilled F25
(A) Left panel, highlighted in gray: box plots of number of flowers visited per plant and M. sexta flight in F25 compared to control P. axillaris plants. Right
panel: number of flowers visited per plant and moth on P. axillaris plants and F25 supplemented with 15 ml nectar. The upper and lower limits of the box
depict the first and third interquartiles; the thicker line is the median. Error bars depict the variance, with maximal value on top and minimal value on the
per flower per hawkmoth on P. axillaris plants and F25 supplemented with 15 ml nectar. The upper and lower limits of the box depict the first and third in-
terquartiles; the middle line is the median. Error bars depict the variance, with maximal value on top and minimal value on the bottom.
Pollinator Probing Time Determines Plant Fitness
Behavioral Experiments with Hawkmoth Manduca sexta
arena and the pollinator species Manduca sexta (L.) (Sphingidae). We used
vegetative propagations of both the low-nectar introgression line (F25) and
P. axillaris to ensure a homogenous genotype. In hawkmoth behavioral
assays, two plants were presented simultaneously, one P. axillaris and
one F25 with either low nectar content or supplemented with 15 ml nectar.
For details on insect care and handling and the experimental setup, see
Supplemental Experimental Procedures.
We measured seed set capacities (number of seeds per capsule) as one
parameterof female fitness of both F25and P.axillaris plants using 15vege-
tative propagations per category, obtained by hawkmoth pollination and by
hand pollination. For details on the seed set experiments, see Supplemental
Correlation of Seed Set to Probing Time
The experimental procedure was similar to seed set experiments, except
that additionally, hawkmoth probing time per flower was measured and
For details on all statistical analyses, see Supplemental Experimental
Supplemental Information includes one figure, one table, and Supplemental
Experimental Procedures and can be found with this article online at http://
The authors would like to thank Martin Heil, Robert Raguso, Nicola Raihani,
and Sara Carlson for critically reading the manuscript and all of our
colleagues for support; Christopher Ball for plant care; Samuel Zeeman
for sugar composition analysis; Oliver Bossdorf for statistical analysis;
Florence Kuhlemeier and Kristin Schu ¨ffler for phenotypic measurement;
Eligio Bossolini for genetic marker analysis; Marc Gre ´millon for photog-
raphy; Danny Kessler and Sylke Dietels for hawkmoth supply; and the
sity of Bern for financial support.
Received: April 25, 2012
Revised: June 4, 2012
Accepted: June 19, 2012
Published online: July 26, 2012
1. Schiestl, F.P. (2005). On the success of a swindle: pollination by decep-
tion in orchids. Naturwissenschaften 92, 255–264.
2. Cozzolino, S., and Widmer, A. (2005). Orchid diversity: an evolutionary
consequence of deception? Trends Ecol. Evol. 20, 487–494.
3. Internicola, A.I., Bernasconi, G., and Gigord, L.D.B. (2008). Should
food-deceptive species flower before or after rewarding species? An
experimental test of pollinator visitation behaviour under contrasting
phenologies. J. Evol. Biol. 21, 1358–1365.
4. Ackerman, J.D., Rodriguez-Robles, J.A., and Melendez, E.J. (1994).
A meager nectar offering by an epiphytic orchid is better than nothing.
Biotropica 26, 44–49.
5. Neiland, M.R.M., and Wilcock, C.C. (1998). Fruit set, nectar reward, and
rarity in the Orchidaceae. Am. J. Bot. 85, 1657–1677.
6. Herre, E.A., Knowlton, N., Mueller, U.G., and Rehner, S.A. (1999). The
evolution of mutualisms: exploring the paths between conflict and
cooperation. Trends Ecol. Evol. 14, 49–53.
7. Bshary, R., and Bronstein, J.L. (2004). Game structures in mutualistic
interactions: what can the evidence tell us about the kind of models
we need? Adv. Stud. Behav. 34, 59–101.
8. Bshary, R., and Bronstein, J.L. (2011). A general scheme to predict
partner control mechanisms in pairwise cooperative interactions
between unrelated individuals. Ethology 117, 271–283.
9. Leimar, O., and Hammerstein, P. (2010). Cooperation for direct fitness
benefits. Philos. Trans. R. Soc. Lond. B Biol. Sci. 365, 2619–2626.
10. Southwick, E.E. (1984). Photosynthate allocation to floral nectar—a
neglected energy investment. Ecology 65, 1775–1779.
11. Pyke, G.H. (1991). What does it cost a plant to produce floral nectar?
Nature 350, 58–59.
12. Leiss, K.A., Vrieling, K., and Klinkhamer, P.G.L. (2004). Heritability of
nectar production in Echium vulgare. Heredity (Edinb) 92, 446–451.
13. Kessler, D., and Baldwin, I.T. (2007). Making sense of nectar scents: the
effects of nectar secondary metabolites on floral visitors of Nicotiana
attenuata. Plant J. 49, 840–854.
14. Fisogni, A., Cristofolini, G., Rossi, M., and Galloni, M. (2011). Pollinator
directionality as a response to nectar gradient: promoting outcrossing
while avoiding geitonogamy. Plant Biol (Stuttg) 13, 848–856.
15. Mitchell, R.J., and Waser, N.M. (1992). Adaptive significance of
Ipomopsis aggregata nectar production: pollination success of single
flowers. Ecology 73, 633–638.
16. Johnson, S.D., and Nilsson, L.A. (1999). Pollen carryover, geitonogamy,
and the evolution of deceptive pollination systems in orchids. Ecology
17. Jersakova, J., Johnson, S.D., Kindlmann, P., and Pupin, A.C. (2008).
Effect of nectar supplementation on male and female components of
pollination success in the deceptive orchid Dactylorhiza sambucina.
Acta Oecol. Int. J. Ecol. 33, 300–306.
18. Meekers, T., and Honnay, O. (2011). Effects of habitat fragmentation on
the reproductive success of the nectar-producing orchid Gymnadenia
conopsea and the nectarless Orchis mascula. Plant Ecol. 212, 1791–
19. Thomson, J.D., and Plowright, R.C. (1980). Pollen carryover, nectar
rewards, and pollinator behavior with special reference to Diervilla
lonicera. Oecologia 46, 68–74.
20. Brandenburg, A., Kuhlemeier, C., and Bshary, R. (2012). Innate adjust-
ment of visitation behavior to rewarding and reward-minimized
mean number of
seeds per capsule
number of seeds/
probing time (s)
Figure 2. Seed Set of P. axillaris and Low-Nectar Line F25
(A) Comparison of mean number of seeds per capsule of low-nectar line
F25 and P. axillaris with standard error, in both hand-pollinated and
M. sexta-pollinated treatments. For statistical output underlying this figure,
see Table S1.
(B) Linear regression model revealing positive relationship between hawk-
moth probing time and number of seeds produced per flower.
Current Biology Vol 22 No 17
Petunia axillaris (Solanacea) plants by hawkmoth Manduca sexta Download full-text
(Sphingidae). Ethology 118, 654–661.
21. Oz, M., Karasu, A., Cakmak, I., Goksoy, A.T., and Turan, Z.M. (2009).
Effects of honeybee (Apis mellifera) pollination on seed set in hybrid
sunflower (Helianthus annuus L.). Afr. J. Biotechnol. 8, 1037–1043.
22. Johnstone, R.A., and Bshary, R. (2002). From parasitism to mutualism:
partner control in asymmetric interactions. Ecol. Lett. 5, 634–639.
23. Brandenburg, A., and Bshary, R. (2011). Variable responses of hawk-
moths to nectar-depleted plants in two native Petunia axillaris
(Solanaceae) populations. Arthropod Plant Interact. 5, 141–148.
24. Luyt, R., and Johnson, S.D. (2001). Hawkmoth pollination of the African
epiphytic orchid Mystacidium venosum, with special reference to flower
and pollen longevity. Plant Syst. Evol. 228, 49–62.
25. Artz, D.R., Villagra, C.A., and Raguso, R.A. (2010). Spatiotemporal vari-
ation in the reproductive ecology of two parapatric subspecies of
Oenothera cespitosa (Onagraceae). Am. J. Bot. 97, 1498–1510.
26. Kiers, E.T., Rousseau, R.A., West, S.A., and Denison, R.F. (2003). Host
sanctions and the legume-rhizobium mutualism. Nature 425, 78–81.
27. Sachs, J.L., Russell, J.E., and Hollowell, A.C. (2011). Evolutionary insta-
bility of symbiotic function in Bradyrhizobium japonicum. PLoS ONE 6,
28. Hoballah, M.E., Stuurman,J.,Turlings, T.C.J.,Guerin,P.M., Conne ´table,
S., and Kuhlemeier, C. (2005). The composition and timing of flower
odour emission by wild Petunia axillaris coincide with the antennal
perception and nocturnal activity of the pollinator Manduca sexta.
Planta 222, 141–150.
29. Klahre, U., Gurba, A., Hermann, K., Saxenhofer, M., Bossolini, E.,
Guerin, P.M., and Kuhlemeier, C. (2011). Pollinator choice in Petunia
depends on two major genetic Loci for floral scent production. Curr.
Biol. 21, 730–739.
30. Schlumpberger, B.O., and Raguso, R.A. (2008). Geographic variation in
floral scent of Echinopsis ancistrophora (Cactaceae); evidence for
constraints on hawkmoth attraction. Oikos 117, 801–814.
31. Clark, D.G., Pichersky, E., Verdonk, J., Dudareva, N., Haring, M., Klahre,
U., and Schuurink, R. (2009). Benzenoids dominate the fragrance of
Petunia flowers. In Petunia, First Edition, T. Gerats and J. Strommer,
eds. (New York: Springer), pp. 51–69.
32. Gouinguene ´, S.P., and Turlings, T.C.J. (2002). The effects of abiotic
factors on induced volatile emissions in corn plants. Plant Physiol.
33. Schmelz, E.A., Alborn, H.T., Engelberth, J., and Tumlinson, J.H. (2003).
Nitrogen deficiency increases volicitin-induced volatile emission,
jasmonic acid accumulation, and ethylene sensitivity in maize. Plant
Physiol. 133, 295–306.
34. Weber, H., Borisjuk, L., and Wobus, U. (1997). Sugar import and metab-
olism during seed development. Trends Plant Sci. 2, 169–174.
35. Pinto, A., Oates, J., Grutter, A., and Bshary, R. (2011). Cleaner wrasses
Labroides dimidiatus are more cooperative in the presence of an
audience. Curr. Biol. 21, 1140–1144.
36. West, S.A., Griffin, A.S., and Gardner, A. (2007). Evolutionary explana-
tions for cooperation. Curr. Biol. 17, R661–R672.
37. Heil, M., Gonza ´lez-Teuber, M., Clement, L.W., Kautz, S., Verhaagh, M.,
and Bueno, J.C. (2009). Divergent investment strategies of Acacia
myrmecophytes and the coexistence of mutualists and exploiters.
Proc. Natl. Acad. Sci. USA 106, 18091–18096.
Pollinator Probing Time Determines Plant Fitness