Nectar robbers deter legitimate pollinators by mutilating flowers

Accepted Ar tic le
Sangeetha Varma1, T. P. Rajesh1, K. Manoj1, G. Asha1, T. Jobiraj2 and Palatty Allesh
Sinu1

1
Dept of Animal Science, Central Univ. of Kerala, Periye 671316, Kerala, India
2
Dept of Zoology, Kodenchery Government Arts and Science College, Kozhikode,
Kerala, India

Corresponding author: Palatty Allesh Sinu, Dept of Animal Science, Central Univ. of
Kerala, Periye 671316, Kerala, India. E-mail: sinu@cukerala.ac.in

Decision date: 04-Feb-2020

This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
[10.1111/oik.06988].

‘This article is protected by copyright. All rights reserved.’

 Abstract
Accepted Ar tic le
Nectar robbing – harvesting nectar illegitimately – can have a variety of outcomes for
plant sexual reproduction and for the pollinator community. Nectar robbers can damage
flowers while robbing nectar, which could affect the behavior of subsequent flower
visitors and, consequently, plant reproduction. However, only nectar manipulation by
nectar robbers has so far received attention. We found a short-tongued bee, Hoplonomia
sp. (Halictidae), mutilating the conspicuous lower petal of the zygomorphic flowers of
Leucas aspera (Lamiaceae) while robbing nectar. We hypothesized that the mutilation
of the conspicuous lower petal deters legitimate pollinators on L. aspera flowers, which,
in turn, might affect plant reproduction. We first assessed the proportion of naturally-
robbed flowers in plant populations for three years to confirm that it was not a purely
local phenomenon due to a few individual bees. We then studied diversity, community
and visitation characteristics of pollinators, nectar dynamics, and fruit set in unrobbed
and robbed open flowers in naturally-robbed populations. The proportion of robbed
flowers varied significantly across sites and years. Robbing did not affect nectar
dynamics in flowers, but it did alter flower morphology, so much so that it reduced
pollinator visitation and altered the pollinator community on robbed flowers. However,
the maternal function of plant reproduction was not affected by nectar robbing. This
study for the first time shows that a nectar robber can have an ecologically significant
impact on floral morphology.

Keywords: floral trait, Leucas aspera, nectar robbing, pollinator behavior, pollination,
pollinator community, pollinator–robber competition, short-tongued bee, Hoplonomia
sp.

‘This article is protected by copyright. All rights reserved.’

 Introduction
Accepted Ar tic le
Animals interact with a variety of plants. Some interactions are mutualistic as in the
case of pollinators and others are antagonistic, as seen in herbivores and nectar robbers
(Irwin 2006). The plant–pollinator mutualistic interaction evolved from a plant–
herbivore interaction through an adaptive radiation by both animals and plants (Herre et
al. 1996). It is, therefore, not rare to see herbivores in different roles – exploiters,
cheaters, and robbers – in modern-day plant–pollinator networks. Among them, nectar
robbers have received much attention from pollination and evolutionary biologists.

Nectar-robbers are flower visitors that, due to a mismatch in the morphological

traits of flowers and their visitors or due to competition, ‘steal’ floral nectar (Traveset et
al. 1998, Irwin and Maloof 2002). A visitor that perforates the flower tissues of the
corolla for accessing nectar illegitimately is the primary nectar robber (Arizmendi 2001,
Irwin and Maloof 2002, Dedej and Delapine 2004, Irwin et al. 2010, Bronstein et al.
2017). Quite often, other nectar-seeking animals use the new nectar access hole for
accessing nectar. They are considered secondary nectar robbers (Irwin and Maloof
2002, Irwin et al. 2010, Richman et al. 2017). The secondary nectar-robbers consist of
either a new set of flower visitors that previously had no access to floral nectar through
the legitimate flower entrance (Varma and Sinu 2019) or some existing legitimate
visitors (Richman et al. 2017, Varma and Sinu 2019). Short-tongued bees are likely to
be primary nectar robbers owing to a mismatch in the floral (long tube) and the animal
trait (short tongue). Surprisingly, long-tongued bees belonging to Apidae – bumblebees,
honey bees, and carpenter bees – are the main nectar robbers (Irwin et al. 2010 and
references therein, Varma and Sinu 2019).

The evolution of floral traits is typically understood with reference to pollinators

(Stebbins 1970, Alexandersson and Johnson 2002). Long tubular flowers have been
interpreted as a result of selection imposed by long-billed or long-tongued specialist
pollinators (Darwin 1877, Alexandersson and Johnson 2002). The matching of floral
and visitor traits is required for effective pollination and plant sexual reproduction
(Garibaldi et al. 2015). It has been hypothesized for quite a long time that robbers,
ubiquitous in plant–pollinator webs, can also have some role in the natural selection of
flower traits and the reproductive mechanism of plants (Galen and Cuba 2001, Irwin et
al. 2001, Lara and Ornelas 2001, Navarro and Medel 2009, Bronstein et al., 2017).

‘This article is protected by copyright. All rights reserved.’

Accepted Ar tic le Nectar-robbed flowers are typically either long tubular flowers with very little
nectar or have nectar spurs (Urcelay et al. 2006, Navarro and Medel 2009, Maruyama et
al. 2015). In the tubular flowers, the robbers make an incision at the base of the corolla.
This does not greatly change the overall appearance/attractiveness of flowers. Normally,
the robber species does not mutilate flower parts that legitimate visitors use to access
flower resources, such as the landing platform (a conspicuous petal) of a zygomorphic
flower. The robbers, therefore, do not leave any specific visual cue on flowers that can
interfere with the visits of legitimate visitors. As a result, legitimate visitors, particularly
pollen foragers, find robbed flowers still attractive, and maintain legitimate visits to
them.

Nectar robbing has variable direct and indirect fitness consequences for plants
(Inouye 1983, Morris 1996, Maloof 2001, McDade and Weeks 2004, Irwin et al. 2010,
Varma and Sinu 2019). Nectar-robbers can have a direct positive effect on the plants’
sexual reproduction if they pollinate flowers and do not damage the reproductive parts
of flowers (Rust 1979, Higashi et al. 1988, Arizmendi et al. 1996, Morris 1996, Maloof
and Inouye 2000, Navarro 2000, Varma and Sinu 2019). Conversely, some robbers
transfer conspecific pollen to stigmas, but may damage the reproductive organs, such as
the ovary (Galen 1999). Such nectar robbers can have a negative effect on plant
reproduction (Galen 1983, Roubik et al. 1985, Traveset et al. 1998, Galen 1999,
Yanwen et al. 2007).

Indirectly, nectar robbers can impair plant reproduction if the robbers affect
visitation rate, diversity, or the behavior of crucial legitimate pollinators and nectar
secretion patterns of plants (Irwin and Brody 1998, Gonzalez-Gomez and Valdivia
2005, Fumero-Caban and Melendez-Ackerman 2013). In such cases, both the plants and
the animals that cooperate in this interaction are likely to be critically affected by nectar
robbing. The impact of nectar robbing is often assessed from the plants’ perspective
(Bronstein et al. 2017, citations in Irwin et al. 2010). Only a few studies assess it from
the perspective of the forager (Zimmerman and Cook 1985, Irwin and Brody 1998,
Maloof 2000, Gonzalez-Gomez and Valdivia 2005, Richman et al. 2017, Varma and
Sinu 2019). Some studies suggested that nectar robbing has an indirect favorable
outcome for plants as it increases both the legitimate and illegitimate visitation rates of
pollinators to robbed flowers (Navarro 2000, Sampson et al. 2004, Varma and Sinu
2019). In none of these studies had the robber considerably damaged the overall

‘This article is protected by copyright. All rights reserved.’

 appearance of flowers or affected the visitation rate of legitimate pollinators due to
Accepted Ar tic le
flower mutilation.

In the present study, we investigated the effect of nectar robbing on the

pollinator community as well as on the maternal function of plant reproduction in
Leucas aspera (Lamiaceae). The primary robber in this study is a short-tongued bee –
Hoplonomia sp. – belonging to the family Halictidae. L. aspera has small zygomorphic
flowers with bilabiate corollas that fuse to form short corolla tubes, which accumulate
some nectar accessible to long-tongued bees (Fig. 1). Visitors use the conspicuous lower
lip of the corolla with translucent nectar guides as landing platform. The anthers and the
stigma are concealed inside an inconspicuous hood-shaped upper lip of the corolla. The
flowers open around 03:00 h and have a life of about 25–29 h. Although numerous
insects visit the flowers, only visitors that open the upper lip of the corolla and touch the
concealed anther filaments and stigma carry out pollination. This suggests a bee-
pollination syndrome in L. aspera (Claben-Bockhoff et al. 2004). Since nectar is
inaccessible for short-tongued bees, they limit their foraging to pollen grains. However,
Hoplonomia sp. makes an incision on the lower lip of the corolla and proceeds to cut it
towards the corolla tube. This leaves the conspicuous lower petal of corolla bifurcated
and gives the robbed flowers a different morphological appearance (Fig. 1). We
hypothesized that robbing in general, and mutilated lower petal in particular, makes
flowers unattractive to legitimate flower visitors, so much so that it affects the nectar
dynamics, visitation rate, diversity, community, and behavior of its pollinators in robbed
flowers. To test this, we asked 1) how does this change in flower morphology affect
pollinators? And 2) how does it impact the reproductive success of the plant? To answer
these questions we 1) compared the diversity, community, and the visitation rate of
legitimate visitors, functional morphologies of visitors as well as the nectar dynamics in
robbed and unrobbed flowers, and 2) studied the maternal reproductive success of
robbed and unrobbed open flowers. Before addressing these questions, we surveyed
several populations of the plant in different parts of India for three years to understand
the extent of damage made by robbers to the flowers of L. aspera and to confirm that it
is not a local phenomenon.

Material and Methods

Study species

‘This article is protected by copyright. All rights reserved.’

 Leucas aspera is a small, erect branched herb attaining a height between 15 and 60 cm
Accepted Ar tic le
(Prajapati et al. 2010). It grows in paddy fields (during fallow periods), open lands,
hedges, and margins of forest thickets, orchards and plantations. The plant is distributed
throughout India from the Himalayan plains to the southern part of Kerala and in Sri
Lanka (Prajapati et al. 2010). It is both a medicinal and a culturally important plant. The
whole plant is used in Ayurveda and the inflorescences are used in Hindu worship
rituals.

Although L. aspera flowers throughout the year both in south and northeast
India, it has two flowering peaks: a minor one in May during the onset of summer
showers, and a major one during August to November towards the end of the southwest
monsoon. The flowers are zygomorphic, white, small (average length of 1 cm) and held
together in auxiliary whorls as verticillaster inflorescences. A plant produces 12.4 (±4.8
SD; n=40 plants; all values are mean±SD) inflorescences, each bearing about 62.0
(±20.6.; n=120 inflorescences) buds. A plant in its whole life produces 744.5 (±389.2;
n=10 plants) flowers at a daily rate of 29.1 (±17; n=820 plants) flowers. The upper lip
of the corolla (hood) hides four didynamous anthers and a bifid stigma. The anthers are
arranged in two pairs as a lower pair and an upper pair. The pointed incurved stigma is
guarded between the lower pair of anthers (Supplementary material Appendix 1 Fig.
A1). The lower petal, about 9 mm long and 7 mm wide, is the most conspicuous part of
the flower. Floral visitors use it as a landing platform to access flower resources. The
corolla tube length is 5.7 mm (±0.3; n=27) with about 100% of its upper side and about
70% of its lower side covered by an oblique toothed green calyx cup (Supplementary
material Appendix 1 Fig. A1). The length of the corolla tube of robbed (5.6±0.2; n=12)
and unrobbed flowers (5.7±0.3; n=15) was not different. The gynoecium is syncarpous
and contains a superior ovary, with four chambers, each with one ovule. The nectar
accumulates to about 3 mm of the tube. The plant secretes nectar throughout the
anthesis period (24-29 h). The flowers are mainly visited by bees, butterflies, moths,
wasps, and ants. The robber species, Hoplonomia sp., is a solitary bee and has a short
tongue measuring about 2 mm. It tears the lower petal of L. aspera flowers (5 mm wide)
into two unequal pieces of 5.39 mm (±1.21; n=15) wide left arm and 2.09 mm (±0.87;
n=15) wide right arm separated by a gap of about 9.8 mm (±1.07; n=15) for the
remaining period of anthesis (Fig. 1). This gives the flowers an entirely different
appearance from the front.

Study sites

‘This article is protected by copyright. All rights reserved.’

 Our casual observations on L. aspera in different parts of south (across the states of
Accepted Ar tic le
Kerala (12˚54’81.48”N; 75˚16’52.02”E) and Karnataka (13˚41’99.31”N; 75˚25’68.43”))
and northeast India (across the Himalayan plains of West Bengal and Assam states
(26˚53’93.4”N; 88˚54’47.5”E)) showed nectar robbing signs on the flowers by
Hoplonomia sp. However, all observational and experimental parts of the study were
carried out in 35 sites in the northern part of the state of Kerala (Kannur district
(12˚02’83.71”N; 75˚26’13.61”E) and Kasaragod district (12˚54’81.48”N;
75˚16’52.02”E)) and the southwestern parts of the state of Karnataka (Sringeri in
Chikmagalur district (13˚41’99.31”N; 75˚25’68.43”)) in south India. The study covered
major flowering peaks over five years (2014-2018).

Extent of nectar robbing in L. aspera flowers

In a pilot study, we watched L. aspera flowers during our official and recreational visits
to different parts of India, and examined the flowers for the signature robbing marks of
Hoplonomia sp. Wherever we saw torn L. aspera flowers, we watched the flowers until
we saw the robber species tearing the petal and subsequently collected and identified the
species. In all cases, Hoplonomia sp. was responsible for cutting the petal of L. aspera
flowers. Later, we studied the proportion of robbed flowers in plant populations during
the major flowering season for three years (2015 – 2017) in two locations, Sringeri
(Chikmagalur district, Karnataka State) and Kasaragod (Kasaragod district, Kerala
state). The aerial distance between Sringeri and Kasaragod was about 150 km. The
average aerial distance between the 3-4 plant populations per year in Sringeri and 6-7
plant populations per year in Kasaragod was about 24 km. Overall, we watched 24 126
flowers on 830 plants in three years.

In each plant population, we selected thirty plants and recorded the number of
robbed flowers at 08:00 h, 11:00 h, 14:00 h and 17:00 h. During each round of
observation, we recorded and removed the robbed flowers from the focal plants, and
kept the remaining flowers on the plants for the next round of observation. In the next
round of observation, we counted the number of flowers available on the plants and
those shed, and recorded the number of robbed flowers both on the plants and on the
ground. We used the latter number to find out the percent of robbed flowers in that
round of observation. The accumulated number of robbed flowers on and off the plants
at the end of the day were used to calculate the proportion of robbed flowers in the sites.

Effect of nectar robbing on nectar dynamics

‘This article is protected by copyright. All rights reserved.’

 We bagged thirty mature flower buds the evening before they opened on ten plants in a
Accepted Ar tic le
population (Aingoth) in Kasaragod to study the effect of nectar robbing on nectar
dynamics in L. aspera flowers. The plant population used for this study had a large
number of robbed L. aspera flowers and the robber species was active in the field from
sunrise onwards. Next day, we restricted visitors on fourteen flowers on five plants by
caging the entire plant using mosquito net bags. On sixteen flowers of another set of
five plants, we allowed the robber species to visit the flower first. After the first visit of
the robber species, we caged the experimental plants using mosquito net bags. We
harvested and recorded the nectar volume on both types of flowers at 07:00 h, 09:00 h,
11:00 h and 13:00 h using 5 µl capillary tubes (Drummond). We inserted the tubes
multiple times until the tubes accumulated no nectar.

Effect of nectar robbing on pollinators

The effect of nectar robbing on the visitation rate and the visitor community was studied
in thirty-five sites of Kasaragod district for five years (2014-2018). In each site, we set
up three types of flowers – naturally-robbed flowers, unrobbed flowers, and manually-
cut flowers, in about 10 m2 area. Since Hoplonomia sp. tear the lower petal of L. aspera
into two pieces, it was required to confirm that the pollinators avoided the robbed
flowers by seeing the damage, and not by the plausible chemical scent that the robber
may use to mark the visited flowers (Goulson et al. 1998, Stout and Goulson 2001,
Williams 1998). We made a slit on the lower petal of the flowers using a pair of
sterilized forceps and fissured it like Hoplonomia sp. did on the flower. We watched
2698 robbed flowers, 11 348 unrobbed flowers, and 1609 manually-cut flowers with a
mean of 19, 68, and 40 flowers respectively per site per year.

We recorded the diversity, visitation rate, and foraging behavior of the floral
visitors on the flowers through scheduled flower watches. We watched all the three
types of flowers for one hour from 07:00 h to 08:00 h and recorded the visitor species,
visitation rate, and the floral resource upon which the visitor foraged. We used these
data to calculate the visitation rate per flower h–1 for overall visitors (total visits of all
visitors/the number of flowers watched), each visitor taxon (total visits of a given
species/the number of flowers watched), and five insect functional groups described
below (total visits of insects of a functional group /the number of flowers watched) in
each category of flowers. The visitation rate per flower h–1 was used as a standard
quantitative measure of visitation rate in all the statistical analyses. To study the
foraging behavior, we recorded the flower resource (pollen or nectar) the visitors
‘This article is protected by copyright. All rights reserved.’
 harvested and the visit type they made on the flowers (legitimate, primary robbing, and
Accepted Ar tic le
secondary robbing). In this part of the study, the naturally-robbed and manually-cut
flowers may or may not have had the visits of legitimate pollinators before they were
robbed. But, in the focal unrobbed flowers, we did not allow the visits of robber species.
We bagged the flower buds until our observation. During the observations, whenever
we encountered a robber species approaching the focal flowers, we chased it away or
netted the insect before it landed on flowers. If we failed to stop the robber visiting the
focal unrobbed flowers, we discarded the robbed flowers from the group of unrobbed
flowers. If a visitor landed on the flower and foraged for nectar and/or pollen, that was
considered as a visit. If a visitor approached the flower, hovered around the lower petal,
but did not land, that was considered as ‘no visit’. Subsequent visits of both pollinators
and Hoplonomia sp. on primary robbed flowers were considered as secondary nectar
robbing visits. Legitimate flower visitors are classified into bees, butterflies and moths,
wasps, and other insects. Bees were further classified into large and small bees based on
their relative size with reference to the lower petal of the flower. Bees smaller than the
length of the lower petal (<5 mm) were classified as small bees and bees equal to or
larger than the length of the lower petal were considered large bees. Occasionally, some
visitors such as ants and small bees, particularly stingless bees, harvested nectar
remnants from the corolla tube or open calyx of shed flowers. Such visits were not
included in the analyses. Representative insects of all flower visitor taxa were collected,
identified to genus or species level, and deposited in the Entomology collection of the
Central University of Kerala.

Effect of nectar robbing on plant reproduction

We studied the effect of nectar robbing on fruit and seed set in two plant populations
(Aingoth (12o38’37.56’’N and 75o08’68.99’E) and Thannoth (12o28’11.72’’N and
75o10’69.87’E)) in Kasaragod in August 2018. We bagged mature flower buds on the
evening before the flowers bloomed. We removed the bags at 06:00 h next day, and one
set of the flowers was assigned to be ‘unrobbed open flowers’ (n=202 flowers on 29
plants). We watched the flowers till 12:00 noon and ensured that these flowers were not
robbed but visited by legitimate visitors, then we covered the flowers using paper bags.
In a second set of flowers (n=192 flowers on 29 plants), we ensured that the first visit
was made by the primary nectar robber species. If a legitimate visitor approached the
focal flowers before the first visit of robber, we chased them away or netted. If we
failed, we discarded those flowers from the data of robbed open flowers. To simulate

‘This article is protected by copyright. All rights reserved.’

 the natural system, we allowed legitimate pollen foragers and secondary nectar robbers
Accepted Ar tic le
(nectar seekers on robbed flowers are considered secondary nectar robbers) on the
robbed flowers until 12:00 noon and covered the flowers using paper bags labeled as
“robbed open flowers”. A third set of flowers (n=166 flowers on 14 plants) was
maintained under the cover of mosquito bags during the entire period of anthesis and
left for studying autogamous selfing. The experimental flowers of three types were
marked using three different paints on the bracts and left for assessing fruit set and seed
set, which we measured six days after the experiment.

Data analyses

First, we examined whether the proportion of robbed flowers varied across the diurnal
periods of the day, sites, and years. We used a generalized linear model with binomial
distribution as an error type to test this. We used the proportion of robbed flowers/site
data as the response variable and year, site, and diurnal period as the fixed categorical
variables in the models. The significance of explanatory variables in the model was
tested using type III ANOVA test available in the R-package ‘car’ (Fox et al. 2016). We
compared the variation in the amount of replenished nectar in robbed and unrobbed
flowers at different time intervals using a repeated measures ANOVA test using the R-
package ez (Easy analysis and visualization of factorial experiments) (Lawrence 2011).

Before studying the effect of nectar robbing on various traits of the pollinator
community, we examined whether the pollinator community of the naturally-robbed and
manually-cut flowers was different or alike using the Analysis of Similarity (ANOSIM)
tool in the R package vegan (Oksanen et al. 2018). We used the visitation rate of
pollinators to naturally-robbed and manually cut flowers to construct the dissimilarity
matrix for different sites. It suggested that the flower visitor community of naturally
robbed and manually-cut flowers were alike (ANOSIM: R=0.06). Therefore, in all the
statistical models, we pooled manually-cut and naturally-robbed flowers, and
considered only two levels of flowers: robbed and unrobbed. This also decreased the
disparity in the numbers of naturally-robbed flowers and unrobbed flowers in plant
populations.

We used generalized linear mixed models (GLMMs) to study the effect of nectar
robbing on the visitation rate and the richness of legitimate visitors. In these models, we
fitted flower type (‘robbed’ and ‘unrobbed’) as a fixed factor, the number of flowers
observed in plant populations as a covariate and year and site ID in nested fashion as

‘This article is protected by copyright. All rights reserved.’

 random factors. We fitted the species richness and visitation rate of overall floral
Accepted Ar tic le
visitors and floral visitors of different functional groups (bees, lepidopterans, wasps,
miscellaneous insects, large bees, and small bees) as response variables in different
models to examine which functional group/s are critically affected by nectar robbing.
For the visitation rate data, we used Gaussian distribution and for species richness, we
used Poisson distribution with log as a link function for error types. Additionally, we
examined the interaction network formed between bee pollinators and the flower type
(robbed and unrobbed) using the quantitative interaction matrix available in the R-
package bipartite (Dorman et al. 2009).

To study the effect of caging (autonomous selfing) on the maternal function of

plant reproduction, we compared the fruit set (binary variable: 0=no fruit; 1=developed
fruit) and seed set (number of seeds developed out of four ovules total) in caged and
unrobbed open flowers. To study the effect of nectar robbing on maternal function of
plant reproduction, we compared the fruit set and the seed set of open unrobbed flowers
and open robbed flowers. We constructed a generalized linear mixed model and a linear
mixed model to analyze the effect of caging and nectar robbing on fruit set and seed set,
respectively. When fruit set was the response variable, we fitted binomial distribution as
the error type in the models. When seed set was the response variable, we used log
x+0.5 number of seeds as the response variable and Gaussian as the distribution type in
the models. In both the models, plants nested in sites were used as the random variable.
The linear and generalized linear mixed models were performed using the R-package
lme4 (Bates et al. 2015) and lmerTest (Kuznetsova et al. 2017). All the statistical
analyses were done using R ver. 3.2.5 (<www.r-project.org>).

Results

Extent of nectar robbing in Leucas aspera populations

Our observations showed that Hoplonomia sp. robbed fresh flowers and revisited the
robbed flowers throughout the day. The proportion of robbed flowers did not differ
significantly across four diurnal periods (F3,78=0.78, p=0.5) (Fig. 2), but varied among
the three years (F2,78=3.52, p=0.03) (Fig. 2) and sixteen sites (F15,78=6.73, p<0.0005).
Nectar robbing did not significantly affect the quantity of resecreted nectar
(robbed flower: 0.49 µl (±0.097) versus unrobbed flower: 0.51 µl (±0.097); F1,101=2.30,
p=0.13). A freshly opened flower (03:00 h) produced about 0.91 µl (±0.097; n=10)
nectar under control conditions. However, the amount of resecreted nectar decreased or

‘This article is protected by copyright. All rights reserved.’

 diminished steadily over the time in both the robbed (F3,56=190.74, p<0.00005) and
Accepted Ar tic le
unrobbed flowers (F3,48=220.77, p<0.00005) (Fig. 3).

Effect of nectar robbing on pollinators

A total of 77 species of insects (including Hoplonomia sp.) visited the focal flowers of
L. aspera in 35 plant populations monitored for this purpose. This includes 35 species
of bees (24 large bees and 11 small bees), 25 species of butterflies and moths, and 16
species of other insects (Supplementary material Appendix 1 Table A1). The unrobbed
flowers were visited by 42 species and robbed flowers were visited by 26 species of
insects.

Visitation rates of all visitors, bees, and large bees (but not visitation rates of
other functional groups) were significantly greater for unrobbed than for robbed flowers
(Table 1; Fig. 4). The species richness of visitors was also significantly higher for
unrobbed flowers and this was true for all functional groups (Table 1; Fig. 4) except for
butterflies/ moths, small bees, and other insects (Table 1; Fig. 4). The interaction
network confirmed that nectar robbing affected the community structure and
composition of bees on flowers (Supplementary material Appendix 1 Fig. A2).

Effect of nectar robbing on maternal reproduction of L. aspera

The caged flowers set fruits and seeds, suggesting that the plant is self-compatible and
has some capacity for autogamous selfing. Fruit set (coefficient±SE, -2.69±0.52, z=-
5.15, p<0.00005) and seed set (-0.28±0.05, z=-5.63, p<0.00005), however, were
significantly lower in caged flowers than in unrobbed open flowers, suggesting that the
floral visitors’ visits improve reproductive fitness (Fig. 5). However, fruit set
(0.06±0.03, z=1.74, p=0.08) and seed set (0.02±0.02, z=1.39, p=0.16) were comparable
in robbed open and unrobbed open flowers (Fig. 5) suggesting that robbing does not
affect maternal function of plant reproduction in L. aspera.

Discussion
Although nectar robbing is ubiquitous in various bee families, genera and species, long-
tongued bees belonging to Apidae predominate in the list (Irwin et al. 2010). Regardless
of taxonomic affiliation, the robbed flowers have a long corolla tube or nectar spur
(Irwin et al. 2010, Maruyama et al. 2015, Richman et al. 2017, Ye et al. 2017). Reports
of short-tongued bees as primary nectar robbers are scarce in the literature despite the

‘This article is protected by copyright. All rights reserved.’

 mismatch of their short tongue lengths with the long corolla tubes of many flowers. The
Accepted Ar tic le
present study will be one among those few studies that demonstrate that a short-tongued
bee can be a primary nectar robber (Inouye 1980, Stout et al. 2000, Newman and
Thomson 2005) and that even a small zygomorphic flower can be robbed for nectar.
Because it has been suggested that nectar robbing is a result of local adaptation
in some individual bees (Bronstein et al. 2017), it was required to confirm that the
robbing that we saw in Leucas aspera by Hoplonomia sp. was not a merely local
phenomenon. It was found across sites for multiple years, but varied spatially and
temporally. The observed rate and pattern of nectar robbing in L. aspera flowers are
comparable to those reported in other plant systems (Zhang et al. 2007, Irwin et al.
2010, citations in Bronstein et al. 2017). Hoplonomia sp. was also a frequent visitor of
at least ten other herbaceous and shrubby plant species co-flowering with L. aspera in
south India, but it behaved as a legitimate visitor in all (Sangeetha Varma and P.A.
Sinu, pers. obs.) but Sesamum radiatum, where it was a secondary nectar robber (Varma
and Sinu, 2019). This suggests that although Hoplonomia sp. exhibits tactic constancy
(Bronstein et al. 2017) on L. aspera flowers, they switch foraging behavior on other
plant species.
Nectar robbing can have direct and indirect consequences for plant reproduction
(Zimmerman and Cook 1985, Irwin and Brody 1998, 1999, Maloof 2001, Irwin 2003,
Irwin et al. 2010). The direct effects could be positive, neutral, or negative, depending,
to some extent, on the reproductive system of plants (Roubik 1982, Arizmendi et al.
1996, Richardson 2004. Burkle et al. 2007, Navarro and Medel 2009; Zhang et al. 2009,
Hazlehurst and Karubian 2016; Varma and Sinu 2019). If robbers pollinate robbed
flowers or if the robbing increases the visits of legitimate pollinators, the net effect of
nectar robbing might be positive. Our results showed that robbing, despite affecting the
visits of large nectaring bees, had no particular effect on nectar replenishment pattern,
richness and visitation rate of small bees, and fruit set and seed set in L. aspera. The
robber in the present study is a “large bee” with reference to the size of the flower, and
touches anthers and stigma while slitting the lower petal of the flower for robbing
nectar. The small pollen-harvesting bees landed directly on the upper petal of the
flowers. So, both the robber and the small bees might have facilitated both self- and
cross-pollination in the flowers. This study therefore agrees with the expectation that
nectar robbing has little effect on maternal function of plant reproduction in self-
compatible plants (Zhang et al. 2009, Irwin et al. 2010, Hazlehurst and Karubian 2016,
Varma and Sinu 2019).

‘This article is protected by copyright. All rights reserved.’

Accepted Ar tic le Indirectly, nectar robbing might affect plant reproduction by affecting diversity,
composition, visitation rate, and behavior of crucial pollinators of plants (Gonzalez-
Gomez and Valdivia 2005, Fumero-Cabán and Meléndez-Ackerman 2013, Hazlehurst
and Karubian 2016, Varma and Sinu 2019). These changes in pollinator traits occur if
robbing alters crucial floral traits that attract pollinators to the flowers, such as nectar
dynamics and flower morphology (Irwin and Brody 1998, Dohzono et al. 2008, Irwin et
al. 2010, Ye et al. 2017). The effect of nectar robbing on nectar dynamics and the
associated changes in pollinator fauna and behaviour have received some attention
(Morris 1996, Irwin and Brody 1998, Dohzono et al. 2008, Irwin et al. 2010, Mayer et
al. 2014, Richman et al. 2017, Ye et al. 2017). In some plants, the altered nectar
resecretion pattern led to desertion of robbed flowers by the legitimate nectar foragers,
which in turn negatively affected plant reproduction (Roubik 1982). Some other plants
increased the frequency of nectar resecretion in robbed flowers over that of the
unrobbed flowers or secreted nectar in robbed flowers normally like the unrobbed
flowers, thereby avoiding deterrence of legitimate visitors (Irwin et al. 2010, Mayer et
al. 2014; Ye et al. 2017). If visitors can’t detect which flowers have been robbed,
legitimate pollinators may continue visiting the robbed flowers and perform cross-
pollination (Sampson et al. 2004), but this scenario is unlikely in L. aspera, given the
damage inflicted by Hoplonomia sp. on the flowers. In L. aspera, nectar replenishment
pattern in both the robbed and unrobbed flowers was alike across the day. However, the
nectar robber revisited the robbed flowers to access the replenished nectar and robbed
new flowers throughout the day, thereby competing with legitimate pollinators through
both exploitation and interference.
How pollinators respond to physically-altered robbed flower has received less
attention, because such cases are reported less. In L. aspera, the robber slits the most
conspicuous part of the flower – the lower petal – altering the overall appearance of the
flower from the front. Some pollinators have behaviorally adapted to robbing and
learned to harvest nectar as secondary nectar robbers (Richman et al. 2017; Varma and
Sinu, 2019). But, this is possible only if the pollinators can at least land on the robbed
flowers. Many large bees, despite approaching the robbed L. aspera flowers, found
them inaccessible and left. Very large bees, such as carpenter bees (Xylocopa spp.)
‘hold’ the two arms of the torn petal and foraged for nectar; but they were infrequent on
L. aspera flowers. The visitation rate of tiny pollen-foraging bees was also not severely
affected by nectar robbing in L. aspera. In L. aspera, the corolla tube is only 5.7 mm
long, but nectar is still accessible only for large-bodied long-tongued bees. It was

‘This article is protected by copyright. All rights reserved.’

 inaccessible for small long-tongued bees, such as Apis florea, Trigona iridipennis,
Accepted Ar tic le
Ceratina hieroglyphica, Braunsapis sp. and C. unimaculata. So, these species were
either pollen foragers or limited to foraging remnant nectar from the open calyx and
corolla tube of shed flowers. However, nectar robbing gave these small long-tongued
bees and three short-tongued bees (Halictus sp3, Halictus sp5 and Seladonia sp.) access
to the nectar. So, nectar robbing benefited a small number of pollinators as we saw in
Sesamum radiatum (Varma and Sinu 2019). Nectar robbing, therefore, altered the
pollinator composition in robbed L. aspera flowers. In summary, nectar robbing in L.
aspera, despite having no impact on plant reproduction, affected the pollinator
community not by altering nectar dynamics, but by altering morphology of the flowers.
The present study also generates additional insights into how plants evolve to
defend against nectar robbers. It has been widely believed that long-tongued pollinators
represent a selective force favoring large and long-tubed flowers (Darwin 1877;
Alexandersson and Johnson 2002). Considering the fact that shorter and smaller flowers
are less represented among robbed flowers, and are visited more by legitimate
pollinators, several authors have suggested that nectar robbers can impose selection
opposite to that imposed by specialist pollinators, e.g. in favour of smaller flowers or
shorter corolla tubes (Galen and Cuba 2001; Lara and Ornelas 2001; Urcelay et al.
2006; Navarro and Medel 2009). Unlike more conventional robbed flowers, L. aspera
has small flowers with a short-tubed corolla, which is protected inside a thick calyx cup,
making them invulnerable to conventional nectar robbing. However, Hoplonomia sp.
opted to tear the entire lower petal and the lower side of corolla tube for stealing nectar,
challenging the generalization that long-tubed flowers are normally preferred for
robbing. Nevertheless, the findings that L. aspera is self-compatible and fruit set and
seed set were not affected by nectar robbing suggest that nectar robbers may not impose
selection on flower form and size in this species. Nectar robbing, after five decades of
serious research, continues to generate several exciting questions.

Acknowledgements – SV thanks Kerala State Council for Science Technology and

Environment PhD. fellowship program for supporting her research. PAS would like to
thank the Western Ghats Cell of the Economic and Planning Affairs of Kerala State
Planning Board and an EMR grant (EMR/007170/2016) from Science, Engineering
Research Board, of Department of Science and Technology of Government of India for
financial support. We thank Jessica Forrest for critical and constructive comments in the
previous versions of the manuscript.

‘This article is protected by copyright. All rights reserved.’

 Funding – The research was funded by the Western Ghats Cell of the Economic and
Accepted Ar tic le
Planning Affairs of Kerala State Planning Board and an EMR grant of Science,
Engineering Research Board, of Department of Science and Technology of Government
of India.
Conflict of interest statement – The authors states no conflict of interest

Author contributions – SV and PAS contributed equally to the work. RTP, MK and AG
assisted in field observations, JT identified the bees, SV and PAS collected data,
analysed data, and wrote the manuscript

References

Alexandersson, R. and Johnson, S.D. 2002. Pollinator-mediated selection on flower-

tube length in a hawkmoth-pollinated Gladiolus (Iridaceae). Proc. R. Soc. London
B. Biol. Sci. 269: 631-636.

Arizmendi, M.C. et al.1996. The role of avian nectar robber and of humming bird
pollinators in the reproduction of two plant species. - Funct.Ecol.10:119-127.

Arizmendi, M.C. 2001. Multiple ecological interactions: nectar robbers and humming
birds in a highland forest in Mexico. Can. J. Zool. 79: 997-1006.

Bates, D.et al.2015. Fitting linear mixed-effects models using lme4. J. Stat. Softw.
67:1–48.

Bronstein, J.L. et al., 2017. The behavioral ecology of nectar robbing: why be tactic
constant? – Curr. Opin. Insect Sci. 21: 14-18.

Burkle, L. A. et al. 2007. Predicting the effect of nectar robbing on plant reproduction:
Implications of pollen limitation and plant mating system. - Am. J. Bot. 94:
1935–1943.

Claben-Bockhoff R. et al. 2004. The staminal lever mechanism in Salvia L. (Lamiaceae)

- a Review. - Plant Biol. 5: 33-41.

Darwin, C. 1877. The different forms of flowers on plants of the same species. London:
J. Murray.

Dedej, S. and Delaplane, K.S. 2004. Nectar-Robbing Carpenter Bees Reduce Seed-
Setting Capability of Honey Bees (Hymenoptera : Apidae) in Rabbiteye Blueberry,
Vaccinium ashei , ‘Climax’. - Environ. Entomol.33: 100–106.

Dohzono, I. et al. 2008. Alien bumble bee affects native plant reproduction through
interactions with native bumble bees. – Ecology 89:3082-3092.

Dorman, C.F.et al.2009. Indices, graphs and null models:analyzing bipartite ecological
networks. Open J. Ecol. 2:7-24.
‘This article is protected by copyright. All rights reserved.’
 Fox, J., Weisberg, S., Adler, D., Bates, D., Baud-Bovy, G., Ellison, S., et al.
Accepted Ar tic le 2016. Package ‘Car.’ An R package to applied regression. Available online
at: http://cran-r.project.org/web/packages/car/car.pdf

Fumero-Caban, J. J. and Melendez-Ackerman, E.J. 2013. Effects of nectar robbing on

pollinator behavior and plant reproductive success of Pitcairnia angustifolia
(Bromeliaceae). - Plant species Biol.28: 224–234.

Galen, C.1983. The effects of nectar thieving ants on seed set in floral scent morphs of
Polemonium viscosum. - Oikos. 41:245-249.

Galen, C.1999.Flowers and enemies: predation by nectar thieving ants in relation to

variation in floral form of an alpine wild flower, Polemonium viscosum. -
Oikos.85:426-436.

Galen, C. Cuba, J. 2001. Down the tube: pollinators, predators, and the evolution of
flower shape in the alpine skypilot, Polemonium viscosum. Evolution 55: 1963-
1971.

Garibaldi, L.A. et al.2015. Trait matching of flower visitors and crops predicts fruit set
better than trait diversity. - J Appl. Ecol.52: 1436-1444.

Gonzalez-Gomez, P. L. and Valdivia, C. E. 2005. Direct and indirect effects of nectar

robbing on the pollinator behavior of Patagona gigas (Trochilidae). - Biotropica
37:693-696.

Goulson, D. et al. 1998. Foraging bumblebees avoid flowers already visited by

conspecifics or by other bumblebee species. – Anim. Behav.55: 199–206.

Hazlehurst, J. A. and Karubian, J. O. 2016. Nectar robbing impacts pollinator behavior

but not plant reproduction. - Oikos. 2: 1–9.

Herre, E.A. 1996. An overview of studies on a community of Panamanian figs.- J.

Biogeogr. 23:593–607.

Higashi, S. et al.1988. Robber-like pollinators: overwintered queen bumblebees

foraging on Corydalis ambigua. - Ecol. Entomol. 13: 411-418.

Inouye, D.W. 1980. The effect of proboscis and corolla tube lengths on patterns and
rates of flower visitation by bumble bees. – Oecologia 45:197-201.

Inouye, D.W.1983. The ecology of nectar robbing. – In: Beattey, B. and Elias, E. (ed.),
The Biology of Nectaries. Columbia Univ. Press, pp. 153–173.

Irwin, R.E. 2003. The impact of nectar robbers on estimates of pollen flow: conceptual
predictions and empirical outcomes. - Ecology 84: 485–95.
Irwin, R.E. 2006. Consequences of direct versus indirect species interactions to
selection on traits: pollination and nectar robbing in Ipomopsis aggregata. - Am.
Nat. 167:315–28.

Irwin, R.E. and Brody, A. K. 1998. Nectar robbing in Ipomopsis aggregata: effect on
pollinator behavior and fitness. - Oecologia116: 519-527.

‘This article is protected by copyright. All rights reserved.’

 Irwin, R.E. and Brody, A. K. 1999. Nectar-Robbing Bumble Bees Reduce the Fitness of
Accepted Ar tic le Ipomopsis aggregata (Polemoniaceae). – Ecology 80: 1703–1712.

Irwin, R.E. and Maloof, J. E. 2002. Ecology in Nectar Robbing over time, space, and
species. - Oecologia133: 525–533.

Irwin, R.E. et al. 2010. Nectar Robbing : Ecological and Evolutionary Perspectives. -
Annu. Rev. Ecol. Evol. Syst. 41: 271–292.

Kuznetsova, A. et al. 2017. lmerTest package: Tests in linear mixed effect models.
J.Stat.Softw. 82: 1-26.

Lara, C. and Ornelas, J.F. 2001. Preferential nectar robbing of flowers with long
corollas: experimental studies of two humming bird species visiting three plant
species. Oecologia 128: 263-273.

Lawrence,M.A. 2011. ez: Easy analysis and visualization of factorial experiments. R

package version 3.0.0.

Maloof, J. E. and Inouye, D. W. 2000. Are Nectar Robbers Cheaters or Mutualists? –

Ecology 81: 2651–2661.

Maloof, J. E. 2001. The effects of a Bumble bee nectar robber on plant reproductive
success and pollinator behavior. - Am. J. Bot. 88: 1960–1965.

Maruyama, P.K. et al. 2015.Nectar robbery by a hermit hummingbird: association to

floral phenotype and its influence on flowers and network structure. - Oecologia178:
783-793.

Mayer, C. et al. 2014. Nectar robbing improves male reproductive success of the
endangered Aconitum napellus ssp. lusitanicum. - Evol. Ecol. 28: 669-685.

McDade, L.A. and Weeks, J.A. 2004. Nectar in humming bird pollinated Neotropical
plants II: Interaction with flower visitors. - Biotropica 36: 216-230.

Morris, W. F. 1996. Mutualism denied? Nectar robbing bumble bees do not reduce
female or male success of Bluebells. - Ecology77: 1451-1462.

Navarro,L. 2000. Pollination ecology of Anthyllis vulneraria subsp.vulgaris (Fabaceae):

nectar robbers as pollinators.-Am.J.Bot.87:980-985.

Navarro, L. and Medel, R. 2009. Relationship between floral tube length and nectar
robbing in Durantaerecta L (Verbenaceae). - Biol. J. Linnean Soc. 96: 392-398.

Newman, D.A. and Thomson, J. D.2005. Effects of nectar robbing on nectar dynamics
and bumble bee foraging strategies in Linaria vulgaris (Scrophulariaceae). -
Oikos110:309-320.

Oksanen, J.et al.2018.vegan: community ecology package. – R package ver 2.5-3

<http://CRAN.R-project.org/package=vegan>.

Prajapati, M. S. et al. 2010. Leucas aspera: A Review. - Pharmacogn Rev.4: 85-87.

‘This article is protected by copyright. All rights reserved.’

 Richman, S. K. et al. 2017. Facilitated exploitation of pollination mutualisms : fitness
Accepted Ar tic le consequences for plants. – J. Ecol. 105: 188–196.

Roubik, D.W.1982. The ecological impact of nectar-robbing bees and pollinating

humming birds on a tropical shrub. - Ecology. 63: 354-360.

Roubik, D.W. et al. 1985. Roles of nectar robbers in reproduction of the tropical treelet
Quasiaamara (Simaroubaceae). – Oecologia 66: 161-167.

Rust, R.W. 1979. Pollination of Impatiens capensis : Pollinators and nectar Robbers. -
J. Kans. Entomol. Soc.52: 297–308.

Sampson, B.J. et al. 2004. Nectar robbery by bees Xylocopa virginica and Apis mellifera
contributes to the pollination of Rabiteye blueberry. - J Econ Entomol. 97: 735-740.

Stebbins, G.L. 1970. Adaptive radiation of reproductive characteristics in angiosperms,

I: pollination mechanisms.- Ann.Rev.Ecol.Sys 1:307-326.

Stout, J. C. et al. 2000. Nectar robbing, forager efficiency and seed set : Bumblebees
foraging on the self-incompatible plant Linaria vulgaris (Scrophulariaceae). - Acta
Oecol.21: 277–283.

Stout, J.C. and Goulson, D. 2001.The use of conspecific and interspecific scent marks
by foraging bumblebees and honeybees. – Anim. Behav.62:183-189.

Traveset, A. et al. 1998. Effect of nectar-robbing birds on fruit set of Fuchsia

magellanica in Tierra Del Fuego : a disrupted mutualism. –Funct. Ecol.12: 459–464.

Urcelay, C. et al. 2006. Relationship between corolla length and floral larceny in the
South American hummingbird-pollinated Campsidium valdivianum (Bignoniaceae).
– Ann. Bot. Fennici. 43: 205-211.

Varma, S. and Sinu, P.A.2019. Nectar robbing in bellflower (Sesamum

radiatum) benefited pollinators but unaffected maternal function of plant
reproduction.-Scientific reports.9:8357.

Williams, C. S. 1998. The identity of the previous visitor influences flower rejection by
nectar-collecting bees. - Anim. Behav. 56: 673–681.

Yanwen, Z. et al. 2007. Effects of nectar-robbing on plant reproduction and evolution. -

Front. Biol (Beijing) 2: 443–449.

Ye, Z.M. et al.2017. Nectar replenishment maintains the neutral effects of nectar
robbing on reproductive success of Salvia przewalskii (Lamiaceae), a plant
pollinated and robbed by bumble bees. – Ann.Bot. 119:1053-1059.

‘This article is protected by copyright. All rights reserved.’

 Zhang, Y.W. et al. 2007. Nectar robbing of a carpenter bee and its effects on the
Accepted Ar tic le reproductive fitness of Glechomalongituba (Lamiaceae). - Plant Ecol.193:1–13.

Zhang, Y.W. et al. 2009. Differential effects of nectar robbing by the same bumble-bee
species on three sympatric Corydalis species with varied mating systems. - Ann.
Bot.104: 33–39.

Zimmerman, M. and Cook, S. 1985. Pollinator foraging, experimental nectar-robbing

and plant fitness in Impatiens capensis.-Am.Midl.Nat.113:84-91.

‘This article is protected by copyright. All rights reserved.’

 Figure Legends
Accepted Ar tic le
Figure 1. (A) a robbed flower and B) two unrobbed flowers of Leucas aspera; (C) a
female Hoplonomia sp is tearing the lower petal of a flower while robbing nectar.

‘This article is protected by copyright. All rights reserved.’

 Figure 2. The percent of robbed flowers did not differ significantly across four diurnal
Accepted Ar tic le
periods (F3,78=0.78, p=0.5), but varied among three years (F2,78=3.52, p=0.03)

‘This article is protected by copyright. All rights reserved.’

 Figure 3. Nectar production in robbed (p<0.00005) and unrobbed flowers (p<0.00005)
Accepted Ar tic le
decreased over time, but in similar fashion.

‘This article is protected by copyright. All rights reserved.’

 Figure 4. Visitation rate (visits/flower h–1) (A-F) and Visitor species richness (G-L) of
Accepted Ar tic le
overall visitors and pollinators of different functional groups per site in robbed and
unrobbed flowers of Leucas aspera. Please see Table 4 for the n for each pollinator
taxon.

‘This article is protected by copyright. All rights reserved.’

 Figure 5. A) Fruit set (proportion of flowers set fruits/plant) and B) seed set (number
Accepted Ar tic le
of seeds in pollinated flowers/plant) in different pollination experiments.

Caged=flowers are caged before opening for studying autonomous selfing (n=166

flowers); robbed.open=robbed flowers kept open for legitimate pollinators’ visits

(n=192 flowers); unrobbed.open=unrobbed flowers kept open for studying open

pollination efficiency (n=202 flowers).

‘This article is protected by copyright. All rights reserved.’

 Table Legend
Accepted Ar tic leTable 1. Model results show the effect of nectar robbing on visitation rate and species

richness of different functional groups of visitors

Response variable Estimate±SE n t-value p-value

Visitation rate
Overall visitors 2.51±0.97 122 2.6 0.01
All bees 1.9±0.86 122 2.2 0.029
Large bees 1.57±0.71 105 2.2 0.03
Small bees 0.01±0.88 94 0.02 0.9
Butterflies and moths -0.32±0.24 35 -1.32 0.3
Other insects 0.61±0.64 26 0.94 0.35
Species richness
Overall visitors 1.24±0.21 122 6.03 <0.00005
All bees 0.88±0.21 122 4.11 <0.0003
Large bees 0.56±0.27 105 2.13 0.03
Small bees 0.46±0.31 94 1.49 0.12
Butterflies and moths 1.09±1.78 35 0.61 0.54
Other insects 0.53±0.89 26 0.59 0.55

‘This article is protected by copyright. All rights reserved.’