Content uploaded by Casper J Van der Kooi
Author content
All content in this area was uploaded by Casper J Van der Kooi on Oct 19, 2017
Content may be subject to copyright.
RESEARCH PAPER
Competition for pollinators and intra-communal spectral
dissimilarity of flowers
C. J. van der Kooi
1,2
, I. Pen
3
, M. Staal
1
, D. G. Stavenga
2
& J. T. M. Elzenga
1
1 Plant Physiology, Groningen Institute for Evolutionary Life Sciences, University of Groningen, Groningen, The Netherlands
2 Computational Physics, Zernike Institute for Advanced Materials, University of Groningen, Groningen, The Netherlands
3 Theoretical Biology, Groningen Institute for Evolutionary Life Sciences, University of Groningen, Groningen, The Netherlands
Keywords
Biotic interactions; flower colouration; plant–
pollinator signalling; principal components
analysis; reproductive character displacement.
Correspondence
J. T. M. Elzenga, Plant Physiology, Centre for
Ecological and Evolutionary Studies, University
of Groningen, Nijenborgh 7, NL-9747 AG
Groningen, The Netherlands.
E-mail: j.t.m.elzenga@rug.nl
Editor
A. Dafni
Received: 21 December 2014; Accepted: 3
March 2015
doi:10.1111/plb.12328
ABSTRACT
Competition for pollinators occurs when, in a community of flowering plants, several
simultaneously flowering plant species depend on the same pollinator. Competition
for pollinators increases interspecific pollen transfer rates, thereby reducing the num-
ber of viable offspring. In order to decrease interspecific pollen transfer, plant species
can distinguish themselves from competitors by having a divergent phenotype. Floral
colour is an important signalling cue to attract potential pollinators and thus a major
aspect of the flower phenotype. In this study, we analysed the amount of spectral dis-
similarity of flowers among pollinator-competing plants in a Dutch nature reserve.
We expected pollinator-competing plants to exhibit more spectral dissimilarity than
non-competing plants. Using flower visitation data of 2 years, we determined the
amount of competition for pollinators by different plant species. Plant species that
were visited by the same pollinator were considered specialist and competing for that
pollinator, whereas plant species visited by a broad array of pollinators were consid-
ered non-competing generalists. We used principal components analysis to quantify
floral reflectance, and found evidence for enhanced spectral dissimilarity among plant
species within specialist pollinator guilds (i.e. groups of plant species competing for
the same pollinator). This is the first study that examined intra-communal dissimilar-
ity in floral reflectance with a focus on the pollination system.
INTRODUCTION
Community structures are shaped by many types of biotic
interactions, e.g. plant–herbivore, plant–pathogen and plant–
pollinator interactions (Gumbert et al. 1999; Sargent & Ackerly
2008; McEwen & Vamosi 2010). Plant–pollinator interactions
are essential for outcrossing in many flowering plant species in
order to set seed. Simultaneously, flowering plant species that
strongly depend on the same shared pollinator are likely to
experience competition for that pollinator. Neighbouring plant
species that compete for pollinators will not only suffer from a
reduction in number of pollinator visits, but will also enhance
interspecific pollen transfer, which often reduces the amount of
viable offspring (reviewed in Ashman & Arceo-G
omez 2013).
Competition for pollinators might therefore influence the com-
munity assembly.
Flowering plants can avoid interspecific pollen transfer by
changing their flowering period, but if this period is relatively
fixed (e.g. by environmental factors, such as light, water and
nutrient availability; see Pleasants 1980), interspecific pollen
transfer can be prevented through phenotypic differentiation
from simultaneously flowering competitors (Heinrich 1975;
Caruso 2001). The latter process is often referred to as charac-
ter displacement (Wilson & Brown 1953; Heithaus 1974; Plea-
sants 1980; Armbruster et al. 1994; Muchhala & Potts 2007).
Character displacement is a change in appearance, effectively
leading to a greater distinction between plants that flower
simultaneously and share pollinators. In order to achieve char-
acter displacement, various floral features can be changed, e.g.
corolla length (Armbruster et al. 1994; Eaton et al. 2012), cor-
olla width (Caruso 2000), inflorescence height (Waddington
1979), nectar production (Eaton et al. 2012) and pollen place-
ment on the pollinator (Brown & Kodric-Brown 1979; Armbr-
uster et al. 1994; Muchhala & Potts 2007). In addition to these
morphological traits, floral colour is an important feature for
pollinator attraction (e.g. Kevan & Baker 1983; Barth 1991;
Schiestl & Johnson 2013; Shrestha et al. 2013; Renoult et al.
2014).
An efficient way of character displacement is to increase flo-
ral colour contrast (i.e. to increase dissimilar spectral proper-
ties) between simultaneously flowering plant species (Levin
1985; Dyer & Chittka 2004; McEwen & Vamosi 2010; de Jager
et al. 2011; Muchhala et al. 2014). In community-wide studies,
reproductive character displacement by means of divergent
floral reflectance has been relatively little studied (but see
McEwen & Vamosi 2010; de Jager et al. 2011; Eaton et al.
2012; Muchhala et al. 2014). An elegant study on a subalpine
plant community from McEwen & Vamosi (2010) reported
significant spectral differences among co-flowering plant spe-
cies. Presumably this represents an example of character dis-
placement, but unfortunately actual flower visitations by
pollinators in this community were not recorded (as is often
Plant Biology ©2015 German Botanical Society and The Royal Botanical Society of the Netherlands 1
Plant Biology ISSN 1435-8603
the case in studies on character displacement; see Waser
1983).
Plant–pollinator interactions can be either specialised, i.e.
when a plant species is pollinated by a single pollinator species,
or generalised, i.e. when a species is serviced by a broad range
of pollinators (reviewed in Johnson & Steiner 2000; Ollerton
et al. 2007). Compared to generalists, specialist plants, which
depend strongly on a specific pollinator, are subjected to
increased selection pressure for strong floral signalling (Heit-
haus 1974). Consequently, spectral dissimilarity is expected to
be higher between specialists than between generalists (Waser
& Ollerton 2006).
In this study we determined the spectral dissimilarity of
flowers among specialists relative to generalists. We used
detailed insect visitation data collected in the Dutch nature
reserve Drentsche Aa (Hoffmann 2005). We grouped plant spe-
cies into either the generalist or a specialist guild based on the
width of the range of pollinator species that visits them. We
expected that in plant species that heavily depended on only a
few pollinator species, competition for pollination would be
stronger than between generalist species, and that therefore
flowers of specialist plant species would exhibit stronger spec-
tral differences. Using principal components analysis (PCA),
we found that the spectral dissimilarities of flowers within each
specialist guild were indeed higher than those within the gener-
alist guild. This is the first study that documents how character
displacement by means of spectral characteristics of flowers
might be related to the specialisation of plant species to pollin-
ators.
MATERIAL AND METHODS
Flower visitation observations
Flower and inflorescence visitation data were obtained from a
comprehensive database collected in 2000 and 2001 in the
Drentsche Aa (Hoffmann 2005). This nature reserve is charac-
terised by homogeneous, open grasslands that are separated by
rows of trees. The Drentsche Aa is rich in flowering plants,
including several rare plant species such as orchids (Grootjans
et al. 2002). Observations of insects on flowering plants were
done from May until October in 2000 and from May until
August in 2001, respectively. In total, the database comprises
observations of over 38,000 individual insects visiting more
than 1 million flowers of 88 different plant species (for details,
see Hoffmann 2005). We pooled insect visitation data per plant
species. To correct for rare flower morphs and species with
reproductive systems that do not rely on pollinators per se (i.e.
self-fertilisation or apomixis), we included only plant species
with a minimum of 20 individual insect visits (following Hoff-
mann 2005). In addition, we combined various Diptera species
to the genus level since pollination is often achieved by groups
of pollinator species (reviewed in Fenster et al. 2004). We
pooled only species with similar ecologies and avoided pooling
of species with different spectral sensitivities (following Vogt
1989).
To determine the importance of pollinators per plant species
rather than per individual plant, we calculated the relative visi-
tor numbers for all plant species. In the database, the vast
majority of insects visiting flowers were identified to the species
level, but for some visiting insects the species or genus could
not be determined (Hoffmann 2005). Therefore, to avoid a bias
in our analysis, we excluded plant species with more than 10%
unknown visitors. The observations on Bombus terrestris
include the less abundant B. lucorum because these bumblebee
species cannot be distinguished in the field (following Hoff-
mann 2005). Due to the large number of observations in the
database it was impossible to determine whether all flower visi-
tors in our study area are indeed pollinators. However, the
effects of e.g. occasional visitation or pollen theft by insects will
be negligible given the large number of visits recorded in the
database. Most importantly, all insect species or genera that we
considered as ‘specialist pollinators’ were documented else-
where to indeed be plant pollinators (Barth 1991; D’Arcy-Burt
& Blackshaw 1991; Goldblatt et al. 2005; Clement et al. 2007;
Ssymank et al. 2008; Garibaldi et al. 2013). We therefore con-
sidered all flower-visiting insects as pollinators.
Specialists and generalists
The degree of specialisation of a plant species to a pollinator
strongly depends on its habitat (reviewed in Richardson et al.
2000; Vazques & Aizen 2006). Preliminary analysis of the rela-
tive visitor numbers in our study area showed a distinct separa-
tion of generalists from specialists at ca. 40% visits (Figure S1).
This visit level has no fundamental justification but can be sub-
stantiated. A lower threshold yielded plant species that were
simultaneously ‘specialist’ for multiple insect guilds. For exam-
ple, some plant species were specialist for two insect guilds,
each guild accounting for 30% of the relative visitation num-
ber. Using the 40% criterion, the remaining visitations (maxi-
mum 60%) always consisted of many different, non-related
insect species. We thus considered plant species visited at least
40% by one insect genus as specialist for that genus (for further
details regarding this criterion, see the Discussion). Two polli-
nator guilds comprised only one plant species (of which one
was specialist for the honeybee Apis mellifera). We excluded
these guilds because it is impossible to compare plant species
within a guild with only one plant species. After applying our
criteria to the plant species in the Drentsche Aa dataset, 39
plant species remained, which were subsequently assigned to
either the generalist guild or to one of the six specialist guilds.
The remaining 39 plants had relatively long flowering periods,
and based on the insect observations, we found that the overlap
of the flowering time for plants within the same guild was at
least 1 month.
Flower species and reflectance measurements
Flower samples were either collected locally from meadows
around Groningen, the Netherlands, or grown from seed
(obtained from Cruydt-Hoeck, Nijeberkoop, the Netherlands).
Reflectance spectra of the flowers were measured with a bifur-
cated fibre-optic probe (Avantes FCR-7UV200; Avantes, Eer-
beek, the Netherlands) using an AvaSpec 2048-2 CCD detector
array spectrometer. The light source was a halogen-deuterium
lamp (AvaLight-D(H)-S); a white diffuse tile (WS-2; Avantes)
was used as a reference. We measured several reflectance spec-
tra from the dominant coloured petal areas (following McEwen
& Vamosi 2010). In these areas, the shape of the spectra
was virtually constant and only the amplitude varied slightly.
We obtained reflectance spectra of additional flowering plant
Plant Biology ©2015 German Botanical Society and The Royal Botanical Society of the Netherlands2
Competition and spectral dissimilarlity of flowers van der Kooi, Pen, Staal, Stavenga & Elzenga
species from the online floral database www.reflectance.co.uk
(Arnold et al. 2010).
To correct for brightness, we subtracted, for each species, the
mean percentage reflectance, and thus specifically compared
spectral quality differences (Cuthill et al. 1999; McEwen & Va-
mosi 2010). We analysed the wavelength range from 300 to
600 nm (which includes the ultraviolet light wavelength range
and excludes the red wavelength range) as insects that are sen-
sitive in the red part of the spectrum (e.g. beetles and butter-
flies; Briscoe & Chittka 2001) only very rarely visited our plant
species (Table S1).
Ideally, floral reflectance is analysed by incorporating polli-
nator spectral sensitivity (Peitsch et al. 1992; Lunau et al. 2011;
Dyer et al. 2012; Shrestha et al. 2013; Burd et al. 2014). How-
ever, for many important flower-visiting insects in our study
area (notably Diptera), we currently have insufficient reliable
information to model colour vision for these species (Lunau
2014). We therefore strictly aim to describe the spectral proper-
ties of co-occurring plants, rather than floral colour as per-
ceived by pollinators.
Statistical analyses
Statistical analyses were conducted using R Statistical Software
(R Core Team 2012). We calculated the spectral differences
using two complementary methods. The spectral differences
were calculated based on the raw spectra and by transforming
the spectra using PCA. PCA allowed us to easily visualise the
spectral differences, whereas calculations based on the raw
spectra provided more statistical power. The PCA was per-
formed using the 39 standardised reflectance spectra with bins
of 1 nm. The principal components 1 (PC1) and 2 (PC2)
together largely explained the variance (see below), in accor-
dance with similar studies (e.g. Cuthill et al. 1999; Grill & Rush
2000; Renoult et al. 2013; Sun et al. 2014). To visualise the dif-
ferences in spectra between guilds, we constructed a PCA scat-
terplot comprising all plant species using the PC1 and PC2
values of each plant species as x- and y-coordinates, respec-
tively. To quantify the spectral contrast, we calculated the
Euclidean pair-wise distances between the plant species within
each guild, resulting in a mean pair-wise distance (MPD) value
per guild. To account for the total variance, and not only the
92% explained by PC1 and PC2, the MPDs were calculated
using all 39 principal components. The MPD was also deter-
mined directly from the raw spectra. For all flowers within a
guild we calculated the absolute difference between the
reflectance values for each nanometre, and then we averaged
the obtained values over the examined wavelength range
(300–600 nm). The guild’s MPD was then calculated by averag-
ing the obtained spectral differences.
The MPDs derived from both the PCA and the raw spectra
were used in two different randomisation tests. We calculated
the difference between the average MPD of the specialist guilds
and the MPD of the generalist guild, denoted by DMPD. Based
on our hypothesis, we expected DMPD to be positive (i.e. more
spectral dissimilarity exists between the flowers that belong to a
group of specialist plants and the flowers of generalist plants)
and therefore we performed a one-tailed test. To test its signifi-
cance, we generated 1,000,000 ‘random’ plant communities by
randomly assigning the 39 plant species to one of the six spe-
cialist guilds or the generalist guild, keeping the sample size for
each guild identical to that in the original plant community
(see Table 1). For each of the random communities we then
calculated and stored DMPD, thus generating a null distribu-
tion of DMPD values. Finally, we tested if the observed spectral
dissimilarity was larger than that expected by chance, by
inspecting the quantile of the observed DMPD value in the null
distribution (V
azques & Aizen 2006).
RESULTS
Specialists and generalists
Throughout our study period, 39 simultaneously flowering
plant species (from 14 angiosperm families) were frequently
visited by insects (Table S1). We excluded 49 plant species that
did not occur frequently and were thus only rarely visited by
insects, or because they were visited mostly by unidentified
insect species (Table S2). A total of 17 frequently occurring
plant species were visited by many different insects, and we
subsequently considered these plant species as generalist
(Table 1; Figure S1). Despite its apomictic mode of reproduc-
tion, Taraxacum officinale was included in the generalist guild
because it affects the community’s colour composition due to
its frequent occurrence and high number of pollinator visita-
tions in our study area (Table S1). A total of 22 frequently
occurring plant species were, at least 40%, visited by one polli-
nator, and we accordingly assigned these plant species to one
of the six specialist pollinator guilds (Table 1; Figure S1). The
degree of specialisation of specialist plant species to an insect
Table 1. Assignment of generalist and specialist plant species to different pollinator guilds and the degree of specialisation (in parentheses).
guild plant species
generalists Achillea millefolia;Aegopodium podagraria;Angelica sylvestris;Chamerion angustifolium;Cirsium arvense;Cirsium palustre;
Epilobium hirsutum;Eupatorium cannabinum;Glechoma hederacea;Jasione montana;Lamium album;Lychnis flos-cuculi;Lycopus
europaeus;Lythrum salicaria;Mentha aquatica;Rorippa amphibia;Taraxacum officinale
Bibio Heracleum sphondylium (47); Hieracium pilosella (45)
Musca Filipendula ulmaria (47); Hieracium aurantiacum (59)
Eristalis Nasturtium officinale (47); Sonchus arvensis (62); Succisa pratensis (69); Valeriana officinalis (46)
Rhingia campestris Ajuga reptans (86); Phyteuma spicatum (60); Silene dioica (59)
Bombus pascuorum Galeopsis tetrahit (52); Linaria vulgaris (43); Lotus corniculatus (59); Stachys palustris (50); Symphytum officinale (62); Trifolium
pratense (55); Trifolium repens (51); Vicia cracca (93); Vicia sativa (82)
Bombus terrestris Lupinus polyphyllus (86); Rhinanthus angustifolius (68)
Plant Biology ©2015 German Botanical Society and The Royal Botanical Society of the Netherlands 3
van der Kooi, Pen, Staal, Stavenga & Elzenga Competition and spectral dissimilarlity of flowers
guild ranged from 43% (moderately specialised) to 93%
(highly specialised).
Floral reflectance analysis
Flowers that appear white to the human eye are always low in
ultraviolet reflectance (Kevan et al. 1996), but spectra with
peaks in the medium or long wavelength range can co-occur
with high reflectance in the ultraviolet wavelength range
(depending on the nature of the pigment; Grotewold 2006; Lee
2007; van der Kooi et al. 2014, 2015). The PCA of flowers in
our study area yielded two principal components, PC1 and
PC2, which accounted for 92.4% of the total variation in the
reflectance spectra (53.7% and 38.7%, respectively). The two
principal components strongly depended on the different
wavelength ranges of the reflectance spectra (Fig. 2a). Flowers
with reflectance peaks in the ultraviolet or short wavelength
range yielded a high PC1 score, and had low, medium or high
PC2 scores. Flowers with high reflectance in the medium and
long wavelength range yielded low PC1 scores and high and
low PC2 scores, respectively. A plot of PC2 versus PC1 thus
formed a reflectance scatterplot, with different floral reflectance
spectra clustering in different sections of the scatterplot
(Fig. 2b).
To quantify the spectral differences, we calculated the MPD
between plant species belonging to the same pollinator guild.
We calculated the MPD using both the PCA-transformed spec-
tra and the raw spectra (see Figure S2). This yielded very simi-
lar results. The MPDs within all six specialist guilds were larger
than MPDs within the generalist guild (Fig. 3a), which is sig-
nificant according to a binomial test (P=0.017). Furthermore,
the average MPD between the specialist guilds and the general-
ist guild was 7.1, which yielded a P-value of 0.050 according to
our randomisation test (Fig. 3b). Similarly, the results based on
the raw spectra were significant, with a P-value of 0.052.
DISCUSSION
The present investigation of the spectral properties of co-flow-
ering plants within pollinator guilds lends support to the
hypothesis that flower communities are, at least in part, struc-
tured by plant–pollinator interactions. The results obtained for
lowland flowers are in fair agreement with earlier findings in a
subalpine community (McEwen & Vamosi 2010) and also with
a recent study on hummingbird-pollinated Solanaceae (Much-
hala et al. 2014).
We could distinguish several pollinator guilds in our study
area. Some plants were pollinated by various insect species,
others mainly by insects of one of six specialist pollinator
guilds (Table 1). Three of six specialist guilds are defined to the
insect genus level, meaning that within these specialist guilds,
plant species were visited by different species from that genus.
Nevertheless, a plant species visited by various insect species of
one insect genus with similar ecologies is more specialist than a
plant species visited by only a few insect species from distinct
lineages with different ecologies (G
omez & Zamora 2006). The
pollinator guilds observed in our study can thus be regarded as
functional pollinator groups.
Fig. 1. Flower reflectance spectra of four exemplar plant species: Lythrum
salicaria (1), Jasione montana (2), Heracleum sphondylium (3) and Sonchus
arvensis (4).
(a)
(b)
Fig. 2. (a) Coefficients of the first two principal components of the PCA per-
formed on the spectra of the species listed in Table 1. PC1 represents the rel-
ative amount of short to long wavelength reflectance. PC2 represents the
relative amount of medium wavelength reflectance to both short (<400 nm)
and long (>500 nm) wavelength reflectance. (b) Flower reflectance scatter-
plot based on the PC1 and PC2 values of our 39 plant species. Similarly as
perceived by the human eye, different floral reflectance spectra cluster in
different sectors of the scatterplot, as indicated by ellipses. The positions of
the four exemplar plant species with different spectra (see Fig. 1) are indi-
cated with numbers (1–4).
Plant Biology ©2015 German Botanical Society and The Royal Botanical Society of the Netherlands4
Competition and spectral dissimilarlity of flowers van der Kooi, Pen, Staal, Stavenga & Elzenga
The generalist flowers in our study are visited by a large
array of insect species (Table S1). These generalists are not
restricted to a specific region in the reflectance scatterplot,
indicating that the generalist guild comprises plant species
with different spectral properties (Fig. 2b). Spectral dissimilar-
ity among generalists seems to contradict our hypothesis that
generalist plants are more similar in spectral reflectance than
specialists. However, the MPD value of the generalists is smal-
ler than in any of the specialist groups (Fig. 3), meaning that
within the generalist guild the spectral properties are more
similar than among specialists. Furthermore, the derived spec-
tral dissimilarities are larger than expected for a random
assembly. The same spectral dissimilarities were obtained
based on distance between the raw spectra, and thus the trans-
formation of spectra into principal components did not sig-
nificantly alter our results. Interestingly, for generalists,
spectral dissimilarity might not be a prerequisite for survival
in the community –in fact, the opposite might be true –
because similar reflectance spectra might increase visitation
from generalist pollinators (e.g. Schiestl & Johnson 2013).
Furthermore, generalist plants may have converged in floral
colouration, yet can be specialist in e.g. floral morphology
(Eaton et al. 2012; Low et al. 2015).
The number of potential pollinators in a community is
reduced if the morphology of the plant requires a specific mor-
phological fit to the pollinator (Cresswell 1998; Schemske &
Bradshaw 1999; G
omez & Zamora 2006; Mur
ua & Esp
ındola
2015). In other words, a ‘complex’ morphology might render a
plant specialist. Complex morphologies are often associated
with specific angiosperm families, e.g. in Fabaceae and Lamia-
ceae, where anthers and stigmas are not as easily accessible as
in members of the Apiaceae and Asteraceae. In addition, occur-
rence of floral pigments, and hence floral reflectance, might be
conserved within families (Levin 1985; Grotewold 2006). Floral
reflectance and morphology might thus potentially be linked
through ancestry. However, generalist plant species are located
throughout the reflectance scatterplot, and both the generalist
and specialist guilds include multiple plant families, indicating
that phylogenetic constraints do not exist (Fig. 2b). This is cor-
roborated by many studies that report phylogenetic effects on
floral reflectance are negligible in large communities (e.g.
Schemske & Bradshaw 1999; Sargent & Otto 2006; McEwen &
Vamosi 2010; Eaton et al. 2012; Muchhala et al. 2014).
We note that a theoretical framework for the effects of
plant–pollinator relations on plant community composition is
not fully developed. A theoretical framework might provide a
more solid basis for the criteria by which plant species can be
grouped into generalist and specialist guilds. In the present
study, the grouping of plant species proved to be rather
straightforward, since plants were either visited by numerous
different insect species or predominantly by a single group
together with a limited number of other species. However, the
chosen minimal percentage of at least 40% visits covered by the
principal pollinator is somewhat arbitrary and has no funda-
mental justification.
We conclude that spectral dissimilarity is most likely a prere-
quisite for pollinator-competing plants. When flowers of dif-
ferent species within a community have dissimilar reflectance
spectra, pollinators are less likely to switch between species
(Dyer & Chittka 2004), and thereby interspecific pollen transfer
is reduced. Even though our results are not highly significant,
pollinators are documented as capable of detecting small spec-
tral differences (e.g. Papiorek et al. 2013; Renoult et al. 2013).
Clearly, for specialists, spectral dissimilarity is an efficient sig-
nalling cue, as it can be perceived by insects from far longer
distances than is the case with morphological traits such as cor-
olla length (Schemske 1976). Interestingly, a study performed
on a South African flower community showed that flowers
with similar spectral properties carry a fitness cost, yet co-flow-
ering plant species had rather similar spectral characteristics
(de Jager et al. 2011). The latter might be explained by a high
prevalence of a particular type of pollinator or by the species
composition in that specific community, as it largely consisted
of plant species that belong to the same genus. This might indi-
cate that the specific set of conditions present in the Drentsche
Aa nature reserve (e.g. plant and pollinator species composi-
tion) create a level of competition for pollinators in which
character displacement can be observed.
The dissimilarity in spectral properties between co-flowering
plants in specialised pollinator guilds that we encountered
in our study area might also occur in other communities.
(a)
(b)
Fig. 3. Observed mean pair-wise distance (MPD) between the PCA-trans-
formed floral reflectance spectra within each pollinator guild. (a) All plant
species within specialist guilds have more spectral contrast than generalist
plants. The MPDs between plant species within each guild were calculated
using the Euclidean distance based on the 39 principal components describ-
ing the spectral properties of the flowers. (b) Distribution of specialist–gener-
alist spectral differences, calculated over the 1,000,000 communities with
randomly assigned guilds. The observed spectral contrast (vertical line) is
higher than the spectral contrast expected by chance.
Plant Biology ©2015 German Botanical Society and The Royal Botanical Society of the Netherlands 5
van der Kooi, Pen, Staal, Stavenga & Elzenga Competition and spectral dissimilarlity of flowers
Comparisons with other habitats are particularly relevant, as
plant reproductive strategies often differ between habitats (Lin-
hart & Feinsinger 1980; Waser & Ollerton 2006). In addition,
since both plant and insect diversity often depend on the envi-
ronment, the selective pressures on floral traits, such as spectral
reflectance, might be different between environments (Gum-
bert et al. 1999; L
azaro et al. 2015). Future studies on plant–
pollinator interactions will help explain the differences in plant
reproductive strategies between populations.
ACKNOWLEDGEMENTS
The authors thank Frank Hoffman for kindly providing the
Drentsche Aa dataset and Adrian G. Dyer, Aidan Vey and three
referees for their suggestions for improvements. We also thank
Ger Telkamp, Hein Leertouwer and Jacob Hogendorf for assis-
tance in acquiring samples of the plant species and their practi-
cal support.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online
version of this article:
Figure S1. Relative pollinator visitation of plants used in this
study.
Figure S2. Raw reflectance spectra for the plants within each
pollinator guild.
Table S1. Total insect visitation counts (obtained from
Hoffmann 2005) of the 39 plant species investigated in this
study.
Table S2. Overview of plant species that were excluded from
the analysis and the reason for exclusion.
REFERENCES
Armbruster W., Edwards M., Debevec E. (1994) Floral
character displacement generates assemblage struc-
ture of Western Australian triggerplants (Stylidium).
Ecology,75,315–329.
Arnold S.E., Faruq S., Savolainen V., McOwan P.W.,
Chittka L. (2010) FReD: the floral reflectance data-
base –a web portal for analyses of flower colour.
PLoS One,5, e14287.
Ashman T., Arceo-G
omez G. (2013) Toward a predic-
tive understanding of the fitness costs of heterospeci-
fic pollen receipt and its importance in co-flowering
communities. American Journal of Botany,100,
1061–1070.
Barth F.G. (1991) Insects and flowers. The biology of a
partnership. Princeton University Press, Princeton,
NJ, USA.
Briscoe A.D., Chittka L. (2001) The evolution of color
vision in insects. Annual Review of Entomology,46,
471–510.
Brown J.H., Kodric-Brown A. (1979) Convergence,
competition, and mimicry in a temperate commu-
nity of hummingbird-pollinated flowers. Ecology,60,
1022–1035.
Burd M., Stayton C.T., Shrestha M., Dyer A.G. (2014)
Distinctive convergence in Australian floral colours
seen through the eyes of Australian birds. Proceed-
ings of the Royal Society Series B: Biological Sciences,
281, 20132862.
Caruso C.M. (2000) Competition for pollination influ-
ences selection on floral traits of Ipomopsis aggregata.
Evolution,54, 1546–1557.
Caruso C.M. (2001) Differential selection on floral
traits of Ipomopsis aggregata growing in contrasting
environments. Oikos,94, 295–302.
Clement S.L., Hellier B.C., Elberson L.R., Staska R.T.,
Evans M.A. (2007) Flies (Diptera: Muscidae: Calli-
phoridae) are efficient pollinators of Allium ampelo-
prasum L. (Alliaceae) in field cages. Journal of
Economic Entomology,100, 131–135.
Cresswell J. (1998) Stabilizing selection and the struc-
tural variability of flowers within species. Annals of
Botany,81,463–473.
Cuthill I.C., Bennett A.T.D., Partridge J.H.C., Maier
E.J. (1999) Plumage reflectance and the objective
assessment of avian sexual dichromatism. The Amer-
ican Naturalist,153, 183–200.
D’Arcy-Burt S., Blackshaw R. (1991) Bibionids (Dip-
tera, Bibionidae) in agricultural land –a review of
damage, benefits, natural enemies and control.
Annals of Applied Biology,118, 695–708.
Dyer A.G., Chittka L. (2004) Biological significance of
distinguishing between similar colours in spectrally
variable illumination: bumblebees (Bombus terres-
tris) as a case study. Journal of Comparative Physiol-
ogy A,190, 105–114.
Dyer A.G., Boyd-Gerny S., McLoughlin S., Rosa M.G.,
Simonov V., Wong B.B. (2012) Parallel evolution of
angiosperm colour signals: common evolutionary
pressures linked to hymenopteran vision. Proceed-
ings of the Royal Society Series B: Biological Sciences,
279, 3606–3615.
Eaton D.A., Fenster C.B., Hereford J., Huang S., Ree
R.H. (2012) Floral diversity and community struc-
ture in Pedicularis (Orobanchaceae). Ecology,93,
S182–S194.
Fenster C.B., Armbruster W.S., Wilson P., Dudash
M.R., Thomson J.D. (2004) Pollination syndromes
and floral specialization. Annual Review of Ecology,
Evolution and Systematics,35, 375–403.
Garibaldi L.A., Steffan-Dewenter I., Winfree R., Aizen
M.A., Bommarco R., Cunningham S.A., Kremen C.,
Carvalheiro L.G., Harder L.D., Afik O., Bartomeus
I., Benjamin F., Boreux V., Cariveau D., Chacoff
N.P., Dudenhoeffer J.H., Freitas B.M., Ghazoul J.,
Greenleaf S., Hipolito J., Holzschuh A., Howlett B.,
Isaacs R., Javorek S.K., Kennedy C.M., Krewenka
K.M., Krishnan S., Mandelik Y., Mayfield M.M.,
Motzke I., Munyuli T., Nault B.A., Otieno M., Pete-
rsen J., Pisanty G., Potts S.G., Rader R., Ricketts
T.H., Rundlof M., Seymour C.L., Schueepp C.,
Szentgyoergyi H., Taki H., Tscharntke T., Vergara
C.H., Viana B.F., Wanger T.C., Westphal C., Wil-
liams N., Klein A.M. (2013) Wild pollinators
enhance fruit set of crops regardless of honeybee
abundance. Science,339, 1608–1611.
Goldblatt P., Manning J., Bernhardt P. (2005) The flo-
ral biology of Melasphaerula (Iridaceae: Crocoideae):
is this monotypic genus pollinated by march flies
(Diptera: Bibionidae)? Annals of the Missouri Botani-
cal Garden,92, 268–274.
G
omez J., Zamora R. (2006) Ecological factors that
promote the evolution of generalization in pollina-
tion systems. In: Waser N., Ollerton J. (Eds), Plant-
pollinator interactions. University of Chicago Press,
Chicago, IL, USA, pp 145–166.
Grill C.P., Rush V.N. (2000) Analysing spectral data:
comparison and application of two techniques. Bio-
logical Journal of the Linnean Society,69, 121–138.
Grootjans A., Bakker J., Jansen A., Kemmers R. (2002)
Restoration of brook valley meadows in the Nether-
lands. Hydrobiologia,478, 149–170.
Grotewold E. (2006) The genetics and biochemistry of
floral pigments. Annual Review of Plant Biology,57,
761–780.
Gumbert A., Kunze J., Chittka L. (1999) Floral colour
diversity in plant communities, bee colour space and
a null model. Proceedings of the Royal Society Series
B: Biological Sciences,266, 1711–1716.
Heinrich B. (1975) Bee flowers: a hypothesis on flower
variety and blooming times. Evolution,29, 325–334.
Heithaus E. (1974) Role of plant–pollinator interac-
tions in determining community structure. Annals of
the Missouri Botanical Garden,61, 675–691.
Hoffmann F. (2005) Biodiversity and Pollination.
Flowering plants and flower-visiting insects in agri-
cultural and semi-natural landscapes. Ph.D. thesis,
University of Groningen, the Netherlands. Online
access: http://irs.ub.rug.nl/ppn/289762693.
de Jager M.L., Dreyer L.L., Ellis A.G. (2011) Do pollin-
ators influence the assembly of flower colours within
plant communities? Oecologia,166, 543–553.
Johnson S.D., Steiner K.E. (2000) Generalization versus
specialization in plant pollination systems. Trends in
Ecology & Evolution,15, 140–143.
Kevan P., Baker H. (1983) Insects as flower visitors and
pollinators. Annual Review of Entomology,28, 407–
453.
Kevan P., Giurfa M., Chittka L. (1996) Why are there
so many and so few white flowers? Trends in Plant
Science,1, 252–527.
van der Kooi C.J., Wilts B.D., Leertouwer H.L., Staal
M., Elzenga J.T.M., Stavenga D.G. (2014) Iridescent
flowers? Contribution of surface structures to optical
signaling. New Phytologist,203, 667–673.
van der Kooi C.J., Dyer A.G., Stavenga D.G. (2015) Is
floral iridescence a biologically relevant cue in
plant–pollinator signaling? New Phytologist,205,18–
20.
Lazaro A., Lundgren R., Totland Ø. (2015) Pollen limi-
tation, species’ floral traits and pollinator visitation:
different relationships in contrasting communities.
Oikos,124, 174–186.
Lee D. (2007) Nature’s palette. The science of plant
color. University of Chicago Press, Chicago, IL, USA.
Levin D.A. (1985) Reproductive character displace-
ment in Phlox.Evolution,39, 1275–1281.
Linhart Y.B., Feinsinger P. (1980) Plant–hummingbird
interactions: effects of island size and degree of spe-
Plant Biology ©2015 German Botanical Society and The Royal Botanical Society of the Netherlands6
Competition and spectral dissimilarlity of flowers van der Kooi, Pen, Staal, Stavenga & Elzenga
cialization on pollination. Journal of Ecology,68,
745–760.
Low S.L., Wong S.Y., Ooi I.H., Hesse M., St€
adtler Y.,
Sch€
onenberger J., Boyce P.C. (2015) Floral diversity
and pollination strategies of three rheophytic Schis-
matoglottidae (Araceae). Plant Biology, doi:10.1111/
plb.12320.
Lunau K. (2014) Visual ecology of flies with particular
reference to colour vision and colour preferences.
Journal of Comparative Physiology A,200, 497–512.
Lunau K., Papiorek S., Eltz T., Sazima M. (2011)
Avoidance of achromatic colours by bees provides a
private niche for hummingbirds. Journal of Experi-
mental Biology,214, 1607–1612.
McEwen J.R., Vamosi J.C. (2010) Floral colour versus
phylogeny in structuring subalpine flowering com-
munities. Proceedings of the Royal Society Series B:
Biological Sciences,277, 2957–2965.
Muchhala N., Potts M.D. (2007) Character displace-
ment among bat-pollinated flowers of the genus
Burmeistera: analysis of mechanism, process and pat-
tern. Proceedings of the Royal Society Series B: Biologi-
cal Sciences,274, 2731–2737.
Muchhala N., Johnsen S., Smith S.D. (2014) Competi-
tion for hummingbird pollination shapes flower
color variation in Andean Solanaceae. Evolution,68,
2275–2286.
Mur
ua M., Esp
ındola A. (2015) Pollination syndromes
in a specialised plant–pollinator interaction: does
floral morphology predict pollinators in Calceolaria?
Plant Biology,17, 551–557.
Ollerton J., Killick A., Lamborn E., Watts S., Whiston
M. (2007) Multiple meanings and modes: on the
many ways to be a generalist flower. Taxon,56, 717–
728.
Papiorek S., Rohde K., Lunau K. (2013) Bees’ subtle
colour preferences: how bees respond to small
changes in pigment concentration. Naturwissens-
chaften,100, 633–643.
Peitsch D., Fietz A., Hertel H., Desouza J., Ventura D.,
Menzel R. (1992) The spectral input systems of hyme-
nopteran insects and their receptor-based color vision.
Journal of Comparative Physiology A,170,23–40.
Pleasants J. (1980) Competition for bumblebee pollina-
tors in Rocky Mountain plant communities. Ecology,
61, 1446–1459.
R Core Team (2012) R: a language and environment for
statistical computing. R Foundation for Statistical
Computing, Vienna, Austria.
Renoult J.P., Thomann M., Schaefer H.M., Cheptou
P.O. (2013) Selection on quantitative colour varia-
tion in Centaurea cyanus: the role of the pollinator’s
visual system. Journal of Evolutionary Biology,26,
2415–2427.
Renoult J.P., Valido A., Jordano P., Schaefer H.M.
(2014) Adaptation of flower and fruit colours to
multiple, distinct mutualists. New Phytologist,201,
678–686.
Richardson D.M., Allsopp N., D’Antonio C.M., Milton
S.J., Rejmanek M. (2000) Plant invasions –the role
of mutualisms. Biological Reviews,75,65–93.
Sargent R.D., Ackerly D.D. (2008) Plant–pollinator
interactions and the assembly of plant communities.
Trends in Ecology & Evolution,23, 123–130.
Sargent R.D., Otto S.P. (2006) The role of local species
abundance in the evolution of pollinator attraction
in flowering plants. The American Naturalist,167,
67–80.
Schemske D.W. (1976) Pollinator specificity in Lan-
tana camara and L. trifolia (Verbenaceae). Biotropic-
a,8, 260–264.
Schemske D.W., Bradshaw H. (1999) Pollinator prefer-
ence and the evolution of floral traits in monkey
flowers (Mimulus). Proceedings of the National Acad-
emy of Sciences USA,96, 11910–11915.
Schiestl F.P., Johnson S.D. (2013) Pollinator-mediated
evolution of floral signals. Trends in Ecology & Evolu-
tion,28, 307–315.
Shrestha M., Dyer A.G., Boyd-Gerny S., Wong
B.B.M., Burd M. (2013) Shades of red: bird-polli-
nated flowers target the specific colour discrimina-
tion abilities of avian vision. New Phytologist,198,
301–310.
Ssymank A., Kearns C.A., Pape T., Thompson F.C.
(2008) Pollinating flies (Diptera): a major contribu-
tion to plant diversity and agricultural production.
Biodiversity,9,86–89.
Sun M., Gross K., Schiestl F.P. (2014) Floral adaptation
to local pollinator guilds in a terrestrial orchid.
Annals of Botany,113, 289–300.
Vazques D.P., Aizen M.A. (2006) Community-wide
patterns of specialization in plant–pollinator interac-
tions revealed by null models. In: Waser N.M., Oller-
ton J. (Eds), Plant-pollinator interactions. From
specialization to generalization. University of Chicago
Press, Chicago, IL, USA, pp 200–219.
Vogt K. (1989) Distribution of insect visual chro-
mophores: functional and phylogenetic aspects.
In: Stavenga D.G., Hardie R.C. (Eds), Facets of
vision. Springer, Heidelberg, Germany, pp 134–
151.
Waddington K.D. (1979) Divergence in inflorescence
height: an evolutionary response to pollinator fidel-
ity. Oecologia,40,43–50.
Waser N. (1983) Competition for pollination and flo-
ral character differences among sympatric plant spe-
cies: a review of evidence. In: Jones C., Little R.
(Eds), Handbook of experimental pollination biology.
Van Nostrand Reinhold, New York, NY, USA, pp
277–293.
Waser N., Ollerton J. (2006) Plant-pollinator interac-
tions. From specialization to generalization. University
of Chicago Press, Chicago, IL, USA.
Wilson E., Brown W. (1953) The subspecies concept
and its taxonomic application. Systematic Zoology,2,
97–111.
Plant Biology ©2015 German Botanical Society and The Royal Botanical Society of the Netherlands 7
van der Kooi, Pen, Staal, Stavenga & Elzenga Competition and spectral dissimilarlity of flowers
