Content uploaded by Felipe Siverio
Author content
All content in this area was uploaded by Felipe Siverio on Jun 03, 2024
Content may be subject to copyright.
88
FLOWER-VISITING LIZARDS AS KEY ECOLOGICAL ACTORS FOR AN ENDEMIC
AND CRITICALLY ENDANGERED PLANT IN THE CANARY ISLANDS
Aarón González-Castro1,2* & Felipe Siverio2
1Department of Animal Biology, Edaphology and Geology (University of La Laguna - Faculty of Sciences - Section of
Biology), Avda. Astrofísico Francisco Sánchez, s/n, E-38206 La Laguna, Tenerife, Canary Islands, Spain.
2Canary Islands’ Ornithology and Natural History Group (GOHNIC), La Malecita, s/n, E-38480 Buenavista del Norte,
Tenerife, Canary Islands, Spain.
Abstract—Oceanic islands are places where biological assemblages are relatively
simple, as compared to the mainland. On islands, however, pollinator assemblages
may to be composed of a taxonomically disparate group of organisms (e.g. insects,
lizards, and birds), some of them with opportunistic nectar-feeding behaviour.
Here we investigated some components of pollination effectiveness of Lotus
maculatus (Fabaceae), an endangered Canary Islands endemic. In a flower
exclusion experiment, we bagged flowers and compared their subsequent fruit and
seed set to that of control flowers. Number of interactions with vertebrate and
invertebrate flower visitors was counted and it was recorded whether interactions
were legitimate (potentially pollinating) or non-legitimate (nectar robbing).
Additionally, we estimated pollen loads on lizards and looked for any relationship
between reproductive success of individual plants and number of visits made by
the top three flower-visiting species (in terms of both frequency of occurrence at
censuses and number of floral visits). Bagged flowers fruited less and with fewer
seeds than control flowers. The only observed flower-visiting vertebrate was the
Tenerife lizard Gallotia galloti, whose interactions were always legitimate and with
around a half of captured individuals carrying pollen grains. The most frequent
flower-visiting insect was the honeybee Apis mellifera followed by the solitary bee
Lasioglossum arctifrons. The honeybee, however, was only a nectar robber, and the
solitary bee was not an effective pollinator, but rather a pollen gatherer. Fruit set
by individual plants was positively related only to frequency of visits by the lizard.
Thus, the lizard seems to play a key role in the conservation management of L.
maculatus.
Keywords—endemic mutualism, Gallotia galloti, Lotus maculatus, Macaronesia,
oceanic island, pollination effectiveness
INTRODUCTION
Mutualisms have been highlighted as a key
process in species coevolution and biodiversity
maintenance (Thompson 2005). For instance,
plant-pollinator interactions are crucial to the
reproductive success of 87.5% of all flowering
plant species (Ollerton et al. 2011). In the case of
islands, their isolation is a barrier to species
colonization, leading to a depauperate and
disharmonic biota compared to nearby mainland
(Carlquist 1974), which in turn leads insular plant-
pollinator networks to include relatively few
species and to have different species groups being
under- or over-represented (Olesen & Jordano
2002). Consequently, island plant species include
new interactions (that are more rarely seen on
mainland) with opportunistic vertebrates like
insect-eating passerine birds and lizards, which
include nectar or pollen into their diet (e.g. Elvers
1977; Pérez-Mellado & Casas 1997; Traveset & Sáez
1997; Rodríguez-Rodríguez & Valido 2008; Hansen
& Müller 2009; Siverio & Rodríguez-Rodríguez
2012; García & Vasconcelos 2017; Abrahamczyk
2019; Fuster et al. 2019).
In the Macaronesian Islands, flowers of several
species of the genus Canarina L. (Campanulaceae),
Echium L. (Boraginaceae), Isoplexis Lindl. Ex Benth.
Journal of Pollination Ecology,
36(8), 2024, pp 88-103
DOI: 10.26786/1920-
7603(2024)777
Received 28 November 2023,
accepted 8 April 2024
*Corresponding author:
agonzalc@ull.edu.es
Article
April 2024 Generalist lizards as effective pollinators of a threatened plant 89
(Scrophulariaceae) and Lotus L. (Fabaceae), among
others, are visited and pollinated by opportunistic
birds (Vogel et al. 1984; Olesen 1985; Valido et al.
2004; Rodríguez-Rodríguez & Valido 2008;
Ollerton et al. 2009; Ortega-Olivencia et al. 2012).
These plants have been included in the so-called
‘Macaronesian bird-flower element’ (Vogel et al.
1984; Olesen 1985; Ojeda Alayón 2013); a group of
around 16 plant species that seem to have
converged to ornithophily because of a set of floral
traits like corolla colour, large volume of dilute and
hexose-rich nectar, absence of scent, and a trend
towards losing papillate cells in the epidermis of
petals (e.g. Dupont et al. 2004; Valido et al. 2004;
Ojeda et al. 2016). Besides birds, lizards have also
been reported as flower visitors for some of these
plant species (Elvers 1977; Olesen & Valido 2003;
Rodríguez-Rodríguez & Valido 2008; Ortega-
Olivencia et al. 2012; Siverio & Rodríguez-
Rodríguez 2012; Esposito et al. 2021).
To our knowledge, no bird species has yet been
recorded in the wild as flower visitor of any of the
five Canary endemic and threatened Lotus species
(section Rhyncholotus (Monod) D. D. Sokoloff): L.
pyranthus P. Pérez and L. eremiticus A. Santos (La
Palma), L. berthelotii Masf. and L. maculatus Breitf.
(Tenerife), and L. gomerythus A. Portero, J. Martín-
Carbajal & R. Mesa (La Gomera). The only
exception is a record of visits by two passerine bird
species to L. berthelotii in an urban garden (Ollerton
et al. 2009). On the other hand, the Tenerife lizard
Gallotia galloti (Oudart, 1839), an endemic to
Tenerife and La Palma, has been recorded as a
frequent visitor of L. maculatus (Siverio &
Rodríguez-Rodríguez 2012) and sporadic visitor of
L. berthelotii (Ollerton et al. 2009), which suggests
this reptile might be a pollinator candidate of these
plants. However, pollination effectiveness of this
lizard to these Lotus species remains to be assessed.
Here, we describe the pollination of L.
maculatus, a threatened Tenerife endemic.
Although G. galloti visits flowers of L. maculatus
frequently (Siverio & Rodríguez-Rodríguez 2012),
its importance as effective pollinator has yet to be
corroborated. Our aims are to 1) evaluate the
ability of L. maculatus flowers to fruit without
flower visitation; 2) compare contribution to the
quantity component of pollination effectiveness
provided by different flower-visiting animals to L.
maculatus; 3) assess if G. galloti could contribute to
pollination of L. maculatus by carrying pollen
grains; and 4) compare the contribution of
different flower-visiting animals to the fruit set of
L. maculatus. This information is vital to the
conservation management of this extremely
threatened plant species.
MATERIALS AND METHODS
NATURAL HISTORY OF LOTUS MACULATUS
Lotus maculatus is a prostrate or pendant
legume that produce a greatly variable number of
zygomorphic flowers (mean: 48.9 flowers per
plant, range: 7-360; unpublished). Flowers are
yellow tending towards orange with a dark brown
stripe in the banner and have a nectar reservoir at
the base. These flowers have a lifespan of seven
days (range: 6-8 d; unpublished). Pollen
presentation to flower visitors follows the piston
mechanism typical in legumes, which depends on
animal-exerted pressure for anthers and stigma
exposure and pollen release. Fruits have a highly
variable number of ovules (mean: 6.8 ovules per
fruit, range: 1-23; unpublished), viable seeds (mean:
4.9 seeds per fruit, range: 1-17; unpublished), as well
as seed to ovule ratio (mean: 0.85, range: 0.36-0.88;
unpublished). Further details about the plant and a
description of the flower can be found in Breitfeld
(1973) and Hind (2008). It is endemic to Tenerife,
listed as ‘Critically Endangered’ (IUCN 2023), and
included in both national and regional catalogues
of threatened species as well as in the Annex I of
the Bern Convention. Two natural populations are
or were known from Tenerife: the first one (i.e. its
locus classicus) is at the coast of the northern
municipality of El Sauzal (Breitfeld 1973), and the
second one is on a rocky islet off the north-east
coast of the island (Hernández 1993). This latter
one may have disappeared (but see Rodríguez
Navarro & Fariña Trujillo 2011). At present, we
regard the population from El Sauzal as the only
natural one. In addition, about ten plantings have
been established by the conservation authorities
on Tenerife (Cabildo Insular de Tenerife) to ensure
its survival.
Both L. maculatus and its close relative L.
berthelotii seem to need pollinators to ensure fruit
set as they rarely self-pollinate (Owens 1985;
Calero & Santos 1988). The former is visited by
insects and lizards, but also a single visit by the
90 González-Castro & Siverio J Poll Ecol 36(8)
introduced house mouse Mus musculus has been
observed (Siverio & Rodríguez-Rodríguez 2012).
STUDY SITES
The study was carried out during the plant’s
flowering season, between March and July 2016 at
Punta Puertito de El Sauzal (the plant’s locus
classicus), and between March and May 2017 at five
different localities from northern Tenerife: 1, the
already mentioned Punta Puertito de El Sauzal (13
m a.s.l., area c. 0.2 ha, and with 22 and 18 flowering
individuals within a population of 40 and 48
individuals in 2016 and 2017, respectively); 2,
Punta de El Clavo (36 m a.s.l., 0.6 ha, and 6
individuals); 3, Punta del Sol (28 m a.s.l., 0.06 ha,
and 15 individuals); 4, La Sabinilla (30 m a.s.l., 0.02
ha, and 12 individuals); and 5, La Fajana (7 m a.s.l.,
0.03 ha, and 3 individuals) (Fig. 1). The four latter
localities were among the plantings established by
the Cabildo Insular de Tenerife. At three of them,
all individuals flowered, except for La Fajana,
where two out of three individuals produced
flowers.
At all study sites, the habitat is characterized by
xerophytic and shrubby plants, like Euphorbia
lamarckii Sweet, Artemisia thuscula Cav., and Kleinia
neriifolia Haw., and especially by halophilous
species like Astydamia latifolia (L. f.) Baill., Salsola
divaricata Masson ex Link in Buch, and Schizogyne
sericea (L. f.) DC. Mean annual rainfall and
temperature were similar at all the study sites and
ranged between 200-300 mm and 16-20 ºC (Marzol
2000).
VISITOR-EXCLUSION EXPERIMENT
To test if L. maculatus can produce fruits
without animal visitation, we made an exclusion
experiment in the natural population (Punta
Puertito de El Sauzal). In 2016 we selected 10
flowering plants. Although 12 more individuals
flowered in 2016, they did so at the end of the
study period (late June) and with a very low
number of flowers, thus it was not possible to
increase sample size without altering the balanced
experimental design (i.e. same number of flowers
per treatment and individual). In 2017 the
experiment was repeated with 10 new individuals
that did not flower in 2016. However, flowering
was poor, which hampered us to use more plants
without affecting the hand-pollination experiment
(see below) and censuses. On each of the total 20
plants from the two years, we randomly selected
six flowers (N = 120 flowers). Three of these (N = 60
flowers for 20 plants) were enclosed in muslin bags
to exclude animals (exclusion treatment) and the
Figure 1. Map showing the
location of the study sites on the
island of Tenerife. The study site
codes refer to: Punta de El Clavo
(1), Punta Puertito de El Sauzal
(2), Punta del Sol (3), La Sabinilla
(4), and La Fajana (5).
April 2024 Generalist lizards as effective pollinators of a threatened plant 91
remaining three (N = 60) were left exposed to
flower visitors (control treatment). Treatment
assignment to each flower was also made
randomly. Later, we estimated fruit set (i.e.
number of flowers that became fruit). From the
sample of fruits, we also estimated seed set (i.e.
number of seeds per fruit).
In 2017 we also initiated a hand-pollination
experiment to describe the breeding system. We
made three experimental treatments in the natural
population: (1) ‘A’, autogamous crossing, where
flowers were bagged and left unmanipulated to
test for autonomous self-pollination; (2) ‘G’,
geitonogamous crossing with pollen from another
flower but on the same plant; and (3) ‘X’,
xenogamous crossing with pollen from other
individuals located at distances ranging between
5.98 m and 16.18 m from recipient plants, but in the
same population. Lastly, we used a group of
untreated flowers left exposed to flower visitors as
a control ‘C’. Unfortunately, flowering was scarce
and reduced the sample size to six individuals. In
addition, the flowers are fragile, and many were
aborted due to manipulation. Thus, our results of
this experiment were dropped out of the study.
CENSUSES OF FLOWER VISITORS
In 2016, preliminary censuses of flower visitors
were made at El Sauzal. However, the diversity of
potential flower visitors was too low. Therefore,
we re-arranged censuses in 2017, expanding our
observations to more localities (i.e. to planting
sites). From March to May, we regularly visited
each study site and made censuses from 08:00 to
18:15 to quantify animal flower visitation. Working
within this range of times allowed us to match with
the activity of potential flower visitor groups (i.e.
birds, lizards, and insects). We made 15-minute
censuses of floral visitors at a total of 53
individually marked plants from all study sites.
However, number of individuals surveyed had to
vary across sites from two at La Fajana to 18 at El
Sauzal. In total, we made 233 censuses for insects
and 226 for vertebrates, encompassing 114.7 hours
(ranging from 15.0 to 34.7 hours at La Fajana and
Punta del Sol, respectively). Across-site variation
in observation time was due to differences in
flowering phenology length. After each 15-
minutes census, we randomly moved to other
plants and alternated between censuses for
vertebrate and insect visitors so as not to introduce
any bias between visitor groups.
For vertebrates, censuses were carried out with
binoculars and/or a field scope (20–60) at a
prudential distance so as not to interfere with their
behaviour near the focal plant. Censuses for insects
were made c. 1.5 m from the focal plant. Small
invertebrates may spend much time (i.e. longer
than 15 minutes) inside flowers. Thus, after each
census we inspected the interior of 10-20 flowers of
the focal plant for small insects, e.g. ants and small
beetles, and the number of flowers with animals
was recorded. At each census, we recorded the
total number of open flowers of the focal plant and
the number of flowers visited by animals. To
standardize the number of flower visits across
censuses, we divided it by the number of open
flowers and the time in minutes each census lasted.
Additionally, during 2016 we accumulated 130
hours of observation by nightly census and camera
traps. However, they were performed just in the
locus classicus and we observed no flower visits.
Each flower visit by an animal was categorized
as either legitimate or illegitimate. In the case of
Lotus spp., visit legitimacy is influenced by flower
size and petal arrangement. Species taxonomically
grouped within the section Rhyncholotus, like L.
maculatus, have relatively large flowers and the
keel (where anthers and stigma are hidden) is
displayed in upper position, with a space of c. 30
mm between tip of the keel and nectar reservoir
(Fig. 2A). On the other hand, other Lotus species,
like those within section Pedrosia (Lowe) Valdés
(Fig. 2B), have relatively small flowers, the keel is
displayed in lower position, and space between
keel and nectar reservoir is shorter than in
Rhyncholotus flowers. When animals visit flowers
like those within section Pedrosia, they alight on the
keel and their body weight is enough to release the
piston mechanism for anthers and stigma exposure
and pollen release, putting them in contact with
the animal’s body. Meanwhile, in the case of
Rhyncholotus, flower visitors accessing nectar must
be large and heavy enough to reach the keel and to
provoke the piston mechanism. Given this context,
flower visits were categorized as legitimate if the
animal contacted the keel, hence potentially
touching anthers and stigma, acting therefore as a
potential pollinator. When insects alighted on the
flower keel to gather pollen, visits were also
92 González-Castro & Siverio J Poll Ecol 36(8)
Figure 2. The differential flower size and arrangement of petals within the sections Rhyncholotus (A) and Pedrosia (B) (Lotus
spp.). Flowers are shown in vertical (A) or horizontal (B) position as they naturally appear displayed to animals. Letters refer to
the banner (B), the keel (K), and the wings (W). Rhyncholotus flowers are larger than Pedrosia ones. The keel is displayed in
upper position in Rhyncholotus and in a lower position in Pedrosia. Notice that given the size of flowers, Rhyncholotus needs to
be visited by relatively larger animals than Pedrosia for the anthers and stigma to be contacted by animals when they access the
nectar through the space between keel and banner.
considered as legitimate because they might
potentially transfer pollen between sexual parts of
flowers. Visits were considered as illegitimate
when the animal accessed the nectar at the base of
flowers via the space between sepals (or by
piercing them), thus without any contact with
anthers and stigma, acting as nectar robber.
FLOWER VISITORS AS POLLINATORS
Gallotia galloti was the only vertebrate
performing flower visits, and we assessed if pollen
grains got attached to its scales. We did not analyse
pollen loads of insects, because we assumed that
these animals are capable of carrying pollen on
their body surface or in their corbiculae and scopae
when they perform legitimate visits.
To estimate the lizards’ capacity to carry pollen
grains, we visited each site once, except for
Puertito de El Sauzal (visited twice), and placed
two pitfall traps baited with tomatoes to capture
lizards. Pitfall traps were operative from 11:30 to
17:00, when activity of lizards was the highest
(unpublished). To avoid direct contact between
lizards and tomatoes in the trap, we placed a false
bottom made of metal mesh upon which lizards
stayed, whereas the tomato bait was below the
mesh. In order to avoid pollen transfer between
lizards or pollen loss, traps were continuously
monitored with binoculars, and whenever a lizard
was caught the trap was immediately removed,
and the lizard inspected for pollen. In accordance
with the protocol by Pérez-Mellado et al. (2000),
we pressed transparent adhesive tape to the neck,
head, and throat of the lizard. Then, the strip of
tape was placed on a microscope slide for later
study under an optical microscope (20 and 40).
We also measured the lizard snout-vent length
(hereafter, SVL), noticed its sex (i.e. male, female,
or undetermined), age (i.e. juvenile or adult), and
marked the lizard with xylene-free paint to avoid
repeated samples. In general, the whole sampling
process took less than ten minutes and lizards
were released at the same site with no damage.
To assess the relative contribution of G. galloti
and insects to the fruit set of L. maculatus, we
correlated the reproductive success of each
individual plant monitored during censuses with
the total number of flower visits it received by
lizards and the two most frequent flower-visiting
insect species, i.e. Lasioglossum arctifrons (Saunders,
1903) and the honeybee Apis mellifera Linnaeus,
April 2024 Generalist lizards as effective pollinators of a threatened plant 93
1758. Reproductive success of individual plants
was estimated by following a similar procedure to
that shown by Gómez (2003). At each individual
used for flower-visitor censuses, we counted the
number of flowers and fruits once per week until
the plant stopped flower and fruit production. As
flowers and fruits are very fragile and tend to be
dropped if manipulated, we could not mark them
to control for aborted flowers and fruits. In
addition, it is difficult to monitor all abortion on
the ground, because of the dense prostrate growth
of the species. Therefore, we divided the maximum
number of fruits (i.e. corresponding to the week
with the highest number of fruits on its branches)
by the maximum number of flowers produced
(corresponding to the week with the highest
number of flowers) to estimate the fruit set as a
measure of reproductive success of each
individual plant.
STATISTICAL ANALYSES
Regarding the exclusion experiment, to test for
any differences in fruit set and seed set from the
exclusion and control treatments, two Generalized
Linear Mixed Models (GLMM) were performed by
using the ‘lme4’ package (Bates et al. 2015) for R (R
Core Team 2023). The first GLMM was performed
with a binomial error distribution, where the
response variable was the success (1) or failure (0)
of each target flower to fruit. The second GLMM
was made with a Poisson error distribution for
count data, where the number of seeds per fruit
was the response variable. In both GLMMs the
exclusion treatment (i.e., exclusion vs. control) was
the explanatory variable, and the identity of each
individual plant was the random factor. To test for
the effect of exclusion treatment, we used the
function ‘Anova’ from the R package ‘car’ (Fox &
Weisberg 2019).
We are aware that our sample size (three
flowers per treatment and individual plants) might
be considered too small to make reliable statistical
inference. Therefore, we used a null model based
on permutations to make statistical inference
(Anderson 2001). This null model was made under
the assumption that experiment outputs
(proportion of flowers setting fruits and number of
seeds per fruit) occur randomly across different
levels of the exclusion treatment. Permutations of
data were run 9999 times and the GLMM was run
on each simulated dataset, so with the ‘Anova’
function we obtained 9999 simulated Chi-squared
values (χ2s) for exclusion treatment as explanatory
variable. Then, to calculate the Monte Carlo p-
value for exclusion treatment, the χ2 calculated on
the original dataset was compared to the
distribution of χ2s values obtained from
permutations as
(1) 𝑃 = 𝐾+1
𝑅+1
There, K is the number of permutations leading
to a χ2s value equal or higher to the χ2 calculated
on the original data, and R is the total number of
permutations.
To assess if different flower visitors differed
regarding the legitimacy of their visits (i.e.
potentially pollinating visits), we performed a G-
test for count data in a contingency table by using
the ‘DescTools’ package (Signorell 2023) for R.
For captured lizards that carried pollen grains,
we tested for the relationship between lizard traits
(i.e. SVL, sex and age) and amount of pollen
sampled from their body by using a Generalized
Linear Model (GLM) with normal error
distribution. For some individuals, we could not to
determine sex or age. Therefore, to have a sample
size as large as possible to assess the effect of each
explanatory variable, the GLMs to test for each of
them were run separately. For this analysis we
excluded individuals from La Fajana. We did so
because we observed that lizards in the planting at
La Fajana were prevented from interacting with L.
maculatus flowers due to protection fences, and
lack of pollen grains could be due to these fences
rather than any lizard traits. Significance of sex and
age was estimated with the function ‘Anova’ from
the R package ‘car’ (Fox & Weisberg 2019). As
number of pollen grains in samples was very
variable, we log-transformed the variable to
ensure its normality.
Lastly, to assess the relationship between
number of floral visits by animals to each
individual plant and its reproductive success (i.e.
fruit set) we performed a GLM with binomial error
distribution for proportion data, with plant
reproductive success as a response variable and
number of visits made by lizards, L. arctifrons, and
A. mellifera as explanatory variables. Due to great
variability in the number of visits by each flower
visitor, explanatory variables were log-
transformed. When individual plants grew very
94 González-Castro & Siverio J Poll Ecol 36(8)
close to each other, it was difficult to assign floral
visits to a given individual. This may affect our
estimation of the relationship between number of
floral visits received by a given plant and its
reproductive success. Therefore, for this particular
analysis we excluded observations to groups of
individuals growing close together. It resulted in a
sample size of 23 individual plants.
RESULTS
VISITOR-EXCLUSION EXPERIMENT
Exclusion of flower visitors resulted in
significantly less fruit and seed set than the control
(χ2 = 10.59; D.F. = 1; P = 0.001; Monte Carlo-P =
0.0012; Fig. 3A, and χ2 = 26.51; D.F. = 1; P < 0.001;
Monte Carlo-P = 0.02; Fig. 3B, respectively).
FLOWER VISITORS
During the entire study, vertebrates made 183
flower visits and invertebrates 433 visits. Gallotia
galloti was the only flower-visiting vertebrate; no
birds visited the flowers although they were active
around the plants. Among the invertebrates, A.
mellifera and L. arctifrons made 79.5% and 16.6% of
all flower visits, respectively, whereas the
remaining 3.2% of visits were mainly by flies
(Diptera), ants (Formicidae), and beetles
(Coleoptera). Each flower received several visits by
the same flower visitor species, both within each
census period and between different censuses.
Figure 3. Results of an exclusion
experiment to assess the effect
of enclosing flowers on fruit set
(A; expressed as the percentage
of flowers that produced fruits)
and seed set (B; as number of
seeds per fruits). Horizontal
thick lines in each box indicates
the median; lower and upper
box limits refer to first and third
quartile, respectively; vertical
lines are the 1.5 interquartile
range; and dots are outliers.
April 2024 Generalist lizards as effective pollinators of a threatened plant 95
Figure 4. Flower visits were considered as legitimate when the animal accessed the nectar through the space between banner
and keel or touched the keel, where anthers and stigma are hidden. The endemic Tenerife lizard Gallotia galloti (A) always access
the nectar legitimately. Lasioglossum arctifrons (B) tended to alight on the keel seeking for pollen, so it potentially touches the
reproductive parts of the flower and could be a pollinator. Honeybees Apis mellifera (C) mostly accessed the nectar at the base
of the flower, acting as nectar robber. After applying Bonferroni’s correction factor, the proportion of legitimate visits (D) was
significantly different among animals, as shown by different letters at the top of bars. Photo credits: Beneharo Rodríguez (A
and C) and Yurena Gavilán (B).
Gallotia galloti always made legitimate visits (Fig.
4A). Almost always, corolla parts returned to their
normal position after visits by lizards, but
sometimes the corolla was left disarticulated after
a visit by lizards. Most of the visits by L. arctifrons
were legitimate, as it tended to alight on the keel
seeking for pollen (Fig. 4B). Apis mellifera always
made illegitimate visits, thus acting as a nectar
robber (Fig. 4C). This difference across flower
visitors in proportion of legitimate visits was
highly significant (G3 = 716.1; P < 0.001; Fig. 4D).
QUALITY OF FLOWER VISITORS AS POLLINATORS OF LOTUS
MACULATUS
Considering data from all five study sites, 34
lizards were captured, and we obtained pollen
grains from 18 of them (Table 1; Fig. 5A). It is
noteworthy that in the natural population at Punta
Puertito de El Sauzal, we sampled the highest
number of pollen grains from 63% of captured
individuals (Table 1). The amount of pollen varied
widely among individuals (range 1-1004 pollen
grains), but there was no relationship with lizard
body size (parameter estimate = -0.014; T-value = -
1.05; P = 0.3). Although samples from males tended
to have less pollen grains than females, we found
no difference between sexes (χ2 = 2.73; D.F. = 1; P =
0.1) or between juvenile and adult lizards (χ2 =
0.03; D.F. = 1; P = 0.87).
Reproductive success (i.e. fruit set) of each
individual plant was positively related to number
of visits by lizards (parameter estimate = 1.46; Z =
26.74; P < 0.001; Fig. 5B), not related to number of
visits by L. arctifrons (parameter estimate = 0.04; Z
= 1.26; P = 0.21; Fig. 5C), and negatively
96 González-Castro & Siverio J Poll Ecol 36(8)
Table 1. Sampling of pollen grains of Lotus maculatus on lizards. Number of individuals of Tenerife lizard Gallotia galloti captured
(N) at each study site, number of individuals carrying pollen grains (NIp), and number of pollen grains (Np) counted per individual
(mean ± standard deviation). *The natural population and locus classicus of L. maculatus.
Study site
Study site code
N
NIp
Np
Punta de El Clavo
1
8
4 (50%)
7.13 11.58
Punta Puertito de El Sauzal*
2
19
12 (63.2%)
167.47 304.73
Punta del Sol
3
1
1 (100%)
1
La Sabinilla
4
1
1 (100%)
4
La Fajana
5
5
0 (0%)
-
Figure 5. Pollen grains of Lotus maculatus (A) under an optical microscope (40×). The panels show the relative reproductive
success of each individual plant (dots) against the number of visits by Tenerife lizard Gallotia galloti (B), Lasioglossum arctifrons
(C) and honeybee Apis mellifera (D). Solid line in panels B-D represents the reproductive success predicted by the model for each
visitor. Photo credits: Aarón González-Castro (A).
April 2024 Generalist lizards as effective pollinators of a threatened plant 97
related to number of visits by A. mellifera
(parameter estimate = -0.62; Z = -25.32; P < 0.001;
Fig. 5D).
DISCUSSION
In this study we show that the Tenerife endemic
and threatened L. maculatus flowers rarely set fruit
by spontaneous self-pollination, that G. galloti
might be an effective (and perhaps the most
important) pollinator, and that visitation by
honeybees has a negative impact upon fruit set.
Our results of the visitor-exclusion experiment
agree with previous studies on lizard-pollinated
species, showing that flowers prevented from
visits by animals have a lower reproductive
success than control flowers (e.g. Pérez-Mellado &
Casas 1997; Traveset & Sáez 1997). It was
surprising that one out of 60 excluded flowers
fruited. The result might be a consequence of an
error while manoeuvring flowers to be bagged or
the fact that self-pollination in Rhyncholotus is quite
difficult but not impossible (Owens 1985). This
very low spontaneous self-pollination is partly in
agreement with Calero & Santos (1988) and might
be explained by the late-acting self-incompatibility
found in other Lotus species (Lundqvist 1993;
Ollerton & Lack 1998). Like other Fabaceae, L.
maculatus has a stigmatic cuticle, preventing the
entrance of the pollen tube into the stigma (Owens
1985; Rodríguez-Riaño et al. 2004; Valtueña et al.
2010; Ojeda & Santos-Guerra 2011). Effective
pollination observed in the closely related L.
berthelotii, as well as in other Fabaceae, depends on
the rupture of the stigmatic cuticle (Heslop-
Harrison & Heslop-Harrison 1983; Owens 1985;
Heenan 1998). Contact of control flowers with
animals might have facilitated cuticle rupture, thus
breaking down self-incompatibility and enhancing
fruiting success. Consequently, foreign pollen may
perhaps not be a prerequisite for fruiting, but
spontaneous self-pollination needs animals to
break the stigmatic cuticle.
VERTEBRATES AS POLLINATORS OF LOTUS MACULATUS
The only vertebrate recorded as a flower visitor
of L. maculatus was G. galloti. Lizard pollination of
an island plant is not surprising as it has been
observed worldwide (Olesen & Valido 2003).
Besides G. galloti, the nocturnal Delalande's gecko
Tarentola delalandii (Duméril & Bibron, 1836), an
endemic to Tenerife and La Palma, is also a flower
visitor for different plants in the Canary Islands
(Hernández-Teixidor et al. 2020; Fariña & Mangani
2020; Koppetsch et al. 2020), like other Tarentola
species in Macaronesia (Pinho et al. 2018) and
other geckos worldwide (Olesen & Valido 2003).
However, after 130 hours of night observation and
use of camera traps in 2016, we did not record any
flower visits made by geckos (unpublished).
Therefore, G. galloti seems to be the most probable
reptile pollinator of L. maculatus, although the role
of Delalande’s gecko as a pollinator should not be
completely disregarded.
Although some G. galloti lizards consume
flowers of L. maculatus, most recorded interactions
during our study were legitimate pollinations.
Lizards accessed the nectar through the front of the
flower, between keel and banner petals, pressing
the flower down and provoking the piston
mechanism to release the pollen and cause the
stigma to emerge, and thus touching anthers and
stigma with head and neck, causing pollination.
Therefore, the high frequency of legitimate
interactions, together with the fact that they can
carry pollen grains and the positive relationships
between number of flower visits and reproductive
success of plants, strongly suggest that G. galloti is
currently the most important and perhaps sole
pollinator of L. maculatus. Numerous studies
demonstrate that lizards might be effective
pollinators, as they visit flowers and are able to
carry pollen (e.g. Elvers 1977; Pérez-Mellado &
Casas 1997; Traveset & Sáez 1997; Rodríguez-
Rodríguez & Valido 2008; Hansen & Müller 2009;
Ortega-Olivencia et al. 2012; García & Vasconcelos
2017; Jaca et al. 2018; Pinho et al. 2018; Hernández-
Teixidor et al. 2020; Koppetsch et al. 2020; Esposito
et al. 2021). However, data about the contribution
of Macaronesian reptiles to plant reproductive
success or their ability to carry pollen grains are
still scarce (but see Rodríguez & Valido 2008; Jaca
et al. 2018; Hernández-Teixidor et al. 2020).
The lack of any relationship between lizard
traits and pollen load might be explained by the
wide body size span of male, female, and juvenile
lizards. Adult males, for example, may exert too
much pressure on flowers, forcing them backward,
and thus reducing the possibility for pollen to pop
out or for the lizard to touch the keel. Nonetheless,
the low number of pollen grains found in some
samples is noteworthy (Table 1). Although low
98 González-Castro & Siverio J Poll Ecol 36(8)
number of pollen grains may be considered as
contamination (Romero-Egea et al. 2023), in the
picture shown in Fig. 4A it is possible to see a
relatively large pollen spot on the lizard’s head.
Therefore, the low number of pollen grains might
be due to low adherence of pollen to animal scales
rather than contamination. Besides, samples
containing between one and six pollen grains
could be considered as contaminated (Romero-
Egea et al. 2023). Although contamination of
sampling tools (e.g. adhesive tape or microscope
slide) might have occurred, it is necessary to
highlight that five out of our 18 pollen-occurring
samples contained less than ten pollen grains, and
four of them were from localities where L.
maculatus has been planted, where lizard activity
can be lower, or plant protection fences tend to
prevent flower visit by lizards (see below). If
manipulation-related contamination were a
pervasive issue in the study, samples with less
than six pollen grains (i.e. potentially
contaminated) would occur at similar proportions
across all study sites.
It was surprising that no flower visits by birds
were recorded, although this plant traditionally
has been proposed to be putatively ornithophilous
(Olesen 1985; Dupont et al. 2004). The explanation
may be that the habitat is not suitable for passerine
birds (Siverio & Rodríguez-Rodríguez 2012). This
may be true for the only natural population, at El
Sauzal, but in the present work, interactions with
birds were not observed at any study site, even in
some localities (i.e. the plantings at La Fajana, La
Sabinilla, Punta del Sol, and Punta de El Clavo),
where abundance of opportunistic passerine birds
is high.
Ollerton et al. (2009) recorded flower visits by
two passerine species –mainly the Canary Islands
chiffchaff Phylloscopus canariensis– to L. berthelotii in
a gardened area. As both Lotus species produce
large volume of dilute and hexose-rich nectar
(Dupont et al. 2004), nectar composition does not
seem to be the reason for the absence of visits by
birds to L. maculatus. A plausible explanation
might be the different colours of flowers, because
L. berthelotii produces intensely red flowers, a
common (though not exclusive) colour among
ornithophilous plants (Proctor et al. 1996), whereas
L. maculatus flowers are yellow towards orange.
Despite birds are attracted to yellow flowers, this
trend towards orange-yellowish colours and a
changing contrast to the green foliage background
could make L. maculatus less prone than L.
berthelotii to receive visits by birds (Ollerton et al.
2009). More studies on flower colorimetry of Lotus
species within the section Rhyncholotus, as well as
focusing on other traits that do not fit the
ornithophilous syndrome, may improve our
understanding of this interspecific difference in
attractiveness of Lotus flowers to birds (e.g.
Ollerton 2024; Rodríguez-Sambruno et al. 2024).
Another plausible explanation for the lack of
visits by birds might be that localities where birds
were abundant correspond to non-natural
populations of L. maculatus. At these planting sites,
the number of L. maculatus individuals is lower
than in the natural population, thus birds may be
habituated to exploit other food resources than
nectar of L. maculatus, especially if ecological
interactions may depend on individual bird
behaviour (Aplin et al. 2013, 2014). In this sense, it
is possible that interactions with a “new” species
in non-natural plantings might depend on the
presence of “bold” individual birds willing to
explore this new resource, which thereby would
trigger the use of L. maculatus nectar by “shy”
individuals in the population. The natural
population at El Sauzal, is perhaps the last
stronghold, at the margin of the original
distribution of L. maculatus, where the plant might
have been relegated due to effect of alien
herbivores (i.e. rabbits and goats). Previous
populations at higher altitudes –where birds are
abundant– might have interacted with pollinating
birds. Therefore, we cannot disregard the
possibility that interactions of L. maculatus with
flower-visiting birds have been lost before any of
mutualists have gone extinct, as reported for seed
dispersal mutualisms (McConkey & O’Farrill
2016).
INSECTS AS POLLINATORS OF LOTUS MACULATUS?
The probably introduced A. mellifera was the
most frequently recorded flower visitor, but it only
made illegitimate visits. It accesses the nectar at the
base of the flower, thus acting as a nectar robber.
Indeed, the relationship between number of
honeybee visits and plant reproductive success
was negative. If a bee made any attempt to access
the nectar through the front of the flower, between
keel and banner petals (in a similar way to
April 2024 Generalist lizards as effective pollinators of a threatened plant 99
legitimate visits made by lizards), it was too light
to make enough pressure to trigger pollen release.
Furthermore, A. mellifera is too small to reach the
anthers and stigma in case they were already
exposed outside the keel after a visit by lizards.
Therefore, apparently legitimate visits by bees
would hardly imply an effective pollination of L.
maculatus. This contrasts with legitimate honeybee
visits to flowers of Lotus tenellus (Lowe) Sandral, A.
Santos & D. D. Sokoloff (unpublished), a species
sympatric with L. maculatus in some of our study
sites but taxonomically grouped in section
Pedrosia. Lotus tenellus has smaller flowers than L.
maculatus and its keel, as in other Fabaceae, is
arranged downward (Fig. 2). When a honeybee
visits a flower of L. tenellus it alights on the keel and
its body weight is large enough to release the
pollen and force the anthers and stigma out of the
keel. Thus, here the honeybee might act as a
pollinator of L. tenellus, as also in other Lotus
species (e.g. Benachour et al. 2007; Siqueira et al.
2018). Therefore, the legitimacy of A. mellifera as a
flower visitor of Lotus spp. seems to be more
dependent on flower size and petal arrangement,
rather than on an intrinsic stereotyped feeding
behaviour of bees. A similar conclusion may be
reached for other insects (i.e. ants, beetles, and
flies) which weigh too little to trigger the pollen
release, despite these animals sometimes accessing
the nectar in a way that could be considered as
apparently legitimate.
The native, solitary bee L. arctifrons made >80%
of visits that could be considered as apparently
legitimate. It alights on the keel –where the stigma
and anthers are hidden– to collect, and probably
eat, pollen (Fig. 3C). However, we found no effect
of number of visits by this insect on plant
reproductive success. Therefore, although
Lasioglossum spp. and other halictid bees have been
reported as pollinators of other species (Singer &
Cocucci 1999; Howard et al. 2021), including some
Fabaceae (Gros 2001), it seems unlikely that L.
arctifrons was a quantitatively important pollinator
of L. maculatus, though it may happen for flowers
previously visited by lizards. Sometimes, after
visits by lizards, flowers may be left with anthers
and stigma exposed, then it may be possible that
pollen grains carried by L. arctifrons to be
transferred to a stigma. Nonetheless, such a
sequential double interaction has not been
corroborated and remains hypothetical.
RECOMMENDATION TO ENSURE THE REPRODUCTIVE SUCCESS
OF LOTUS MACULATUS
Although other vertebrates cannot be
disregarded as pollinator candidates of L.
maculatus, our results suggest that the endemic G.
galloti might be currently the most important and
perhaps the sole pollinator of this threatened
species. Nonetheless, to compare the effectiveness
and definitive importance of G. galloti as pollinator
of L. maculatus it would be desirable to perform
some additional experiments that we could not
carry out due to the relatively large number of
flowers needed for exclusions, which was
incompatible with the threatened status of L.
maculatus. Beyond of visitor-exclusion
experiments, it would be required additional
experiments to selectively exclude lizards and
invertebrate visitors, as well as single-visit
experiments, where recently open flowers are
observed until a visit is recorded, and
subsequently bagged to exclude further visitors.
Whereas all these experiments are quite difficult to
be addressed in the wild, they could be easily
performed in greenhouses.
In any case, conservation of L. maculatus is
urgent and any information gathered from studies
like ours should be useful for conservation of this
endangered plant. Despite local administration
conducting several plantings of L. maculatus
throughout the northern coast of Tenerife, it is
difficult to witness natural regeneration by
seedling recruitment. Study sites where lizards
performed flower visits more often were also the
sites where plants showed a higher fruit
production. Therefore, to enhance the success of
conservation actions made by the local
administration, it is necessary to ensure this
mutualistic interaction.
Since the flowering of L. maculatus mostly
occurs in winter and spring, plantings should be
made in zones exposed to sunlight –e.g. coastal
tips of land and rocky islets–, where lizard activity
can be high; but without disregarding other zones
inland at slightly higher elevations, where
interactions with potential flower-visiting birds
could be enhanced. Also, the design of protection
fences for plants against introduced herbivores
should allow lizards to get into the fenced areas
and be planned with space enough for the plants
to grow prostrate to the ground, facilitating lizard
100 González-Castro & Siverio J Poll Ecol 36(8)
access to flowers, instead of allowing plants to
climb the fences as occurred at several planting
sites. Indeed, it has been demonstrated that
protection fences may have negative consequences
for threatened species (Lorite et al. 2021), so that it
is necessary to keep monitoring fenced plants.
Also, it is important to ensure diverse and
abundant populations of potential flower visitors,
because nectar of L. maculatus is a novel resource
for lizards and birds in human-made plantings,
and the occurrence of these plant-vertebrate
interactions might rely on the presence of bold
individuals willing to explore this new resource
and transmit their feeding behaviour through the
population (see Aplin et al. 2013, 2014; Pérez-
Cembranos & Pérez-Mellado 2014).
Acting mostly as nectar robbers for L.
maculatus, the role of A. mellifera –apart from
having a “sterile” interaction with the plant–
appears to be unfavourable to this threatened
species, as honeybees deplete nectar resource level
in flowers and may exclude other pollinators from
the community (e.g. Carbonari et al. 2009; Valido
et al. 2019). Therefore, it would be desirable to
avoid honeybee hives in places close to
populations of L. maculatus. Lastly, it is known that
G. galloti and other lizard species from Canary
Islands are exposed to introduced predators like
feral cats and snakes (Medina & Nogales 2009;
Piquet & López-Darias 2021) and it has been
demonstrated that key mutualistic interactions
disappear even before the interacting species
become extinct (McConkey & O’Farrill 2016).
Therefore, it is necessary to prevent both
competition and predation risk to lizards to ensure
the further maintenance of pollination and hence
the survival of L. maculatus.
ACKNOWLEDGEMENTS
This research was framed within a technical assistance
commissioned by the Cabildo Insular de Tenerife,
institution that also provided all the necessary
permissions to work in natural protected areas and with
threatened species (register number 10446). We are
especially grateful to Manuel Siverio and Yurena
Gavilán, Salvador de la Cruz, and Helena Morales for
helping during the fieldwork, and Antonio Pérez
Delgado for insect identification. We are also indebted
to Beneharo Rodríguez, who kindly dedicated several
days to take some of the photographs that complement
the text. Alfredo Valido made useful comments on an
early version of the manuscript and helped to obtain
some data for flowers lifespan. The associate editor (Jeff
Ollerton), Jens M. Olesen and two anonymous reviewers
made useful comments and suggestions to improve the
manuscript.
AUTHOR CONTRIBUTION
AGC and FS conceived the idea, designed the study, and
performed fieldwork. AGC led statistical analyses and
manuscript writing with substantial input of FS. Both
authors read and give approval to the final version of the
manuscript.
DISCLOSURE STATEMENT
Authors declare no conflict of interest.
DATA AVAILABILITY STATEMENT
The datasets analysed for this study can be found in the
figshare repository at doi: 10.6084/m9.figshare.25559724.
REFERENCES
Abrahamczyk S (2019) Comparison of the ecology and
evolution of plants with a generalist bird pollination
system between continents and islands worldwide.
Biological Reviews 94:1658-1671. https://doi.org/10.
1111/brv.12520
Anderson M.J. (2001) Permutation tests for univariate or
multivariate analysis of variance and regression.
Canadian Journal of Fisheries and Aquatic Sciences, 58,
626–639. https://doi.org/10.1139/f01-004
Aplin LM, Farine DR, Morand-Ferron J, Cole EF,
Cockburn A, Sheldon BC (2013) Individual
personalities predict social behaviour in wild networks
of great tits (Parus major). Ecology Letters 16:1365–1372.
https://doi.org/10.1111/ele.12181
Aplin LM, Farine DR, Mann RP, Sheldon BC (2014)
Individual-level personality influences social foraging
behaviour in wild birds. Proceedings of the Royal
Society B 281:20141016. https://doi.org/10.1098/rspb.
2014.1016
Bates D, Maechler M, Bolker B, Wallker S (2015) Fitting
Linear Mixed-Effects Models Using lme4. Journal of
Statistical Software 67:1–48 https://doi.org/10.18637/
jss.v067.i01
Benachour K, Louadi K, Terzo M (2007) Rôle des abeilles
sauvages et domestiques (Hymenoptera: Apoidea)
dans la pollinisation de la fève (Vicia faba L. var. major)
(Fabaceae) en région de Constantine (Algérie). Annales
de la Société Entomologique de France 43:213–219.
https://doi.org/10.1080/00379271.2007.10697513
Breitfeld C (1973) Lotus maculatus, eine bisher
unbeschriebne Art von Tenerife. Cuadernos de
Botánica Canaria 17:27–31.
April 2024 Generalist lizards as effective pollinators of a threatened plant 101
Calero A, Santos A (1988) Biología reproductiva de
especies amenazadas en la flora canaria. Lagascalia 15
(supplement):661–664.
Carbonari V, Polatto LP, Aives VV (2009) Evaluation of
the impact on Pyrostegia venusta (Bignoniaceae) flowers
due to nectar robbery by Apis mellifera (Hymenoptera,
Apidae). Sociobiology 54:373–382.
Carlquist SJ (1974) Island biology. Columbia University
Press, New York. https://doi.org/10.5962/bhl.title.63768
Dupont YL, Hansen DM, Rasmussen JT, Olesen JM
(2004) Evolutionary changes in nectar sugar
composition associated with switches between bird
and insect pollination: The Canarian bird–flower
element revisited. Functional Ecology 18:670–676.
https://doi.org/10.1111/j.0269-8463.2004.00891.x
Elvers I (1977) Flower-visiting lizards on Madeira.
Botaniska Notiser 130:231–234.
Esposito F, Costa R, Boieiro M (2021) Foraging behavior
and pollen transport by flower visitors of the Madeira
Island endemic Echium candicans. Insects 12:488.
https://doi.org/10.3390/insects12060488
Fariña B, Mangani R (2020) Nectarivoría de Tarentola
delalandii sobre la planta exótica Hoya kerrii en Tenerife,
Islas Canarias. Boletín de la Asociación Herpetológica
Española 31:53-54. http://www.herpet-ologica.org/
BAHE/31_2/BAHE31-2_240_04_HNat14.pdf
Fox J, Weisberg S (2019) An R Companion to Applied
Regression. Third Edition. Thousand Oaks CA.
Fuster F, Kaiser-Bunbury C, Olesen JM, Traveset A
(2019) Global patterns of the double mutualism
phenomenon. Ecography 42:826-835. https://doi.org/
10.1111/ecog.04008
García C, Vasconcelos R (2017) The beauty and the beast:
Endemic mutualistic interactions promote community-
based conservation on Socotra Island (Yemen). Journal
for Nature Conservation 35:20–23. https://doi.org/
10.1016/j.jnc.2016.11.005
Gómez JM (2003) Herbivory reduces the strength of
pollinator-mediated selection in the Mediterranean
herb Erysimum mediohispanicum: consequences for
plant specialization. The American Naturalist 162:242–
256. https://doi.org/10.1086/376574
Gros CL (2001) The effect of introduced honeybees on
native bee visitation and fruit set in Dillwynia juniperina
(Fabaceae) in a fragmented ecosystem. Biological
Conservation 102:89–95. https://doi.org/10.1016/S0006-
3207(01)00088-X
Hansen DM, Müller CB (2009) Reproductive ecology of
the endangered enigmatic Mauritian endemic Roussea
simplex (Rousseaceae). International Journal of Plant
Sciences 170:42–52. https://doi.org/10.1086/593050
Heenan, PB (1998) The pollination system and stigmatic
cuticle of Clianthus puniceus (Fabaceae). New Zealand
Journal of Botany 36:311–314. https://doi.org/10.
1080/0028825X.1998.9512571
Hernández E (1993) La flora vascular de los roques de
Anaga (Tenerife, Islas Canarias). Vieraea 22:1–6.
Hernández-Teixidor D, Díaz-Luis N, Medina FM,
Nogales M (2020) First record of geckos visiting
flowers in the Paleartic Ecozone. Current Zoology
66:447–448. https://doi.org/10.1093/cz/zoz051
Heslop-Harrison J, Heslop-Harrison Y (1983) Pollen-
stigma interaction in the Leguminosae: the
organization of the stigma in Trifolium pratense. Annals
of Botany 51:571–583. https://doi.org/10.1093/oxford
journals.aob.a086503
Hind N (2008) 619 Lotus maculatus. Leguminosae-
Papilionoideae. Plant in Peril 30. Curtis’s Botanical
Magazine 25:146–157. https://doi.org/10.1111/j.1467-
8748.2008.00613.x
Howard SR, Prendergast K, Symonds MRE, Shrestha M,
Dyer AG (2021) Spontaneous choices for insect-
pollinated flower shapes by wild non-eusocial halictid
bees. Journal of Experimental Biology 224:jeb242457.
https://doi.org/10.1242/jeb.242457
IUCN (2023) The IUCN Red List of Threatened Species.
Version 2023-1. https://www.iucnredlist.org. Accessed
on 27/02/2024.
Jaca J, Nogales M, Traveset A (2018) Reproductive
success of the Canarian Echium simplex (Boraginaceae)
mediated by vertebrates and insects. Plant Biology
21:216–226. https://doi.org/10.1111/plb.12926
Koppetsch T, Sánchez Romero G, Fischer E, Wolfgang B
(2020) A new record of nectarivory for Tarentola
delalandii (Duméril and Bibron, 1836) pollinating the
introduced palm Dypsis lutescens (H. Wendl.) Beentje
and J. Dransf. (Arecaceae) on Tenerife, Canary Islands.
Herpetology Notes 13:415–419.
Lorite J, Salazar-Mendías C, Pawlak R, Cañadas EM
(2021) Assessing effectiveness of exclusion fences in
protecting threatened plants. Scientific Reports
11:16124. https://doi.org/10.1038/s41598-021-95739-4
Lundqvist A (1993) The self-incompatibility system in
Lotus tenuis (Fabaceae). Heredity 119:59–66.
https://doi.org/10.1111/j.1601-5223.1993.00059.x
Marzol V (2000) El Clima. In: Morales G, Pérez R (eds),
Gran Atlas Temático de Canarias. Editorial Interinsular
Canaria, Santa Cruz de Tenerife.
McConkey KR, O’Farrill G (2016) Loss of seed dispersal
before the loss of seed dispersers. Biological
Conservation 201:38–49. https://doi.org/10.1016/j.
biocon.2016.06.024
Medina FM, Nogales M (2009) A review on the impacts
of feral cats (Felis silvestris catus) in the Canary Islands:
implications for the conservation of its endangered
fauna. Biodiversity and Conservation 18:829–846.
https://doi.org/10.1007/s10531-008-9503-4
102 González-Castro & Siverio J Poll Ecol 36(8)
Ojeda I, Santos-Guerra A (2011) The intersection of
conservation and horticulture: bird-pollinated Lotus
species from the Canary Islands (Leguminosae).
Biodiversity and Conservation 20:3501–3516.
https://doi.org/10.1007/s10531-011-0138-5
Ojeda DI, Valido A, Fernández de Castro AG, Ortega-
Olivencia A, Fuertes-Aguilar A, Carvalho JA, Santos-
Guerra A (2016) Pollinator shifts drive petal epidermal
evolution on the Macaronesian Islands bird-flowered
species. Biology Letters 12: 20160022.
https://doi.org/10.1098/rsbl.2016.0022
Ojeda Alayón, DI (2013) The Macaronesian bird-
flowered element as a model system to study the
evolution of ornithophilous floral traits. Vieraea 41:73-
89. https://doi.org/10.31939/vieraea.2013.41.06
Olesen JM (1985) The Macaronesian bird-flower element
and its relation to bird and bee opportunists. Botanical
Journal of the Linnean Society 91:395–414.
https://doi.org/10.1111/j.1095-8339.1985.tb01010.x
Olesen JM, Jordano P (2002) Geographic patterns in
plant–pollinator mutualistic networks. Ecology
83:2416–2424. https://doi.org/10.1890/0012-9658(2002)
083[2416:GPIPPM]2.0.CO;2
Olesen JM, Valido A (2003) Lizards as pollinators and
seed dispersers: an island phenomenon. Trends in
Ecology and Evolution 18:177–181. https://doi.org/10.
1016/S0169-5347(03)00004-1
Ollerton J (2024) Birds & Flowers: An Intimate 50 million
year relationship. Pelagic Publishing, Exeter. https://
doi.org/10.53061/FFQO3130
Ollerton J, Lack AJ (1998) Relationships between
flowering phenology, plant size and reproductive
success in Lotus corniculatus (Fabaceae). Plant
Ecology 139:35–47. https://doi.org/10.1023/A:10097983
20049
Ollerton J, Cranmer L, Stelzer RJ, Sullivan S, Chittka L
(2009) Bird pollination of Canary Island endemic
plants. Naturwissenschaften 96:221–232. https://doi.
org/10.1007/s00114-008-0467-8
Ollerton J, Winfree R, Tarrant S (2011) How many
flowering plants are pollinated by animals? Oikos
120:321–326. https://doi.org/10.1111/j.1600-0706.2010.
18644.x
Ortega-Olivencia A, Rodríguez-Riaño T, Pérez-Bote JL,
López J, Mayo C, Valtueña FJ, Navarro-Pérez M (2012)
Insects, birds and lizards as pollinators of the largest-
flowered Scrophularia of Europe and Macaronesia.
Annals of Botany 109:153–57. https://doi.org/10.
1093/aob/mcr255
Owens SJ (1985) Seed set in Lotus berthelotii Masferrer.
Annals of Botany 55:811–814. https://doi.org/10.1093/
oxfordjournals.aob.a086960
Pérez-Cembranos A, Pérez-Mellado V (2014) Local
enhancement and social foraging in a non-social
insular lizard. Animal Cognition 18:629–637.
https://doi.org/10.1007/s10071-014-0831-3
Pérez-Mellado V, Casas JL (1997) Pollination by a lizard
in a Mediterranean island. Copeia 1997:593–595.
https://doi.org/10.2307/1447565
Pérez-Mellado V, Ortega F, Martín-García S, Perera A,
Cortázar G (2000) Pollen load and transport by the
insular lizard, Podarcis lilfordi (Squamata, Lacertidae) in
coastal islets of Menorca (Balearic Islands, Spain).
Israel Journal of Zoology 46:193–200. https://doi.
org/10.1092/QMY9-PXWF-AG43-RP6F
Pinho CJ, Santos B, Mata VA, Seguro M, Romeiras MM,
Lopes RJ, Vasconcelos R (2018) What is the giant wall
gecko having for dinner? Conservation genetics for
guiding reserve management in Cabo Verde. Genes
9:599. https://doi.org/10.3390/genes9120599
Piquet JC, López-Darias M (2021) Invasive snake causes
massive reduction of all endemic herpetofauna on
Gran Canaria. Proceedings of the Royal Society B
288:20211939. https://doi.org/10.1098/rspb.2021.1939
Proctor M, Yeo P, Lack A (1996) The natural history of
pollination. Timber Press, Portland, Oregon.
R Core Team (2023) R: A language and environment for
statistical computing. R Foundation for Statistical
Computing, Vienna, Austria. URL https://www.R-
project.org/
Rodríguez Navarro ML, Fariña Trujillo B (2011) On the
current presence of Lotus maculatus Breitf. (Fabaceae) in
the “roque de Tierra” of Anaga (Tenerife, Canary
Islands). Vieraea 39:229–232. https://doi.org/10.
31939/vieraea.2011.39.22
Rodríguez-Riaño T, Ortega-Olivencia A, Devesa Alcaraz
JA (2004) Reproductive biology in Cytisus multiflorus
(Fabaceae). Annales Botanici Fennici 41:179–188.
https://www.annbot.net/PDF/anbf41/anbf41-179.pdf
Rodríguez-Rodríguez MC, Valido A (2008) Opportun-
istic nectar-feeding birds are effective pollinators of
bird-flowers from Canary Islands: experimental
evidence from Isoplexis canariensis (Scrophulariaceae).
American Journal of Botany 95:1408–1415.
https://doi.org/10.3732/ajb.0800055
Rodríguez-Sambruno C, Narbona E, del Valle JC, Valido
A (2024) Bird-flower colour on islands supports the
bee-avoidance hypothesis. Functional Ecology 38:600–
611. https://doi.org/10.1111/1365-2435.14493
Romero-Egea V, Robles C, Traveset A, Del Río L,
Hervías-Parejo S (2023) Assessing the role of lizards as
potential pollinators of an insular plant community
and its intraespecific variation. Animals 13:1122.
https://doi.org/10.3390/ani13061122
Signorell A (2023) DescTools: Tools for descriptive
statistics. https://CRAN.R-project.org/package=Desc
Tools
April 2024 Generalist lizards as effective pollinators of a threatened plant 103
Singer RB, Cocucci AA (1999) Pollination mechanism in
southern Brazilian orchids which are exclusively or
mainly pollinated by halictid bees. Plant Systematics
and Evolution 217:101–117. https://doi.org/10.1007/
BF00984924
Siqueira E, Oliveira R, Dotterl S, Cordeiro GD, Alves-
dos-Santos I, Mota T, Schlindwein C (2018) Pollination
of Machaerium opacum (Fabaceae) by nocturnal and
diurnal bees. Arthropod-Plant Interactions 12:633–645.
https://doi.org/10.1007/s11829-018-9623-z
Siverio F, Rodríguez-Rodríguez MC (2012) Gallotia
galloti (Canary Lizard). Nectarivory. Herpetological
Review 43:333–334.
Thompson JN (2005) The geographic mosaic of
coevolution. University of Chicago Press, Chicago
USA. https://doi.org/10.7208/chicago/9780226118697.
001.0001
Traveset A, Sáez E (1997) Pollination of Euphorbia
dendroides by lizards and insects: Spatio-temporal
variation in patterns of flower visitation. Oecologia
111:241–248. https://doi.org/10.1007/PL00008816
Valido A, Dupont YL, Olesen JM (2004) Bird–flower
interactions in the Macaronesian islands. Journal of
Biogeography 31:1945–1953. https://doi.org/10.1111/j.
1365-2699.2004.01116.x
Valido A, Rodríguez-Rodríguez MC, Jordano P (2019)
Honeybees disrupt the structure and functionality of
plant-pollinator networks. Scientific Reports 9:4711.
https://doi.org/10.1038/s41598-019-41271-5
Valtueña FJ, Rodríguez-Riaño T, Espinosa F, Ortega-
Olivencia A (2010) Self- sterility in two Cytisus species
(Leguminosae, Papilionidae) due to early-acting
inbreeding depression. American Journal of Botany
97:123–135. https://doi.org/10.3732/ajb.0800332
Vogel S, Westerkamp C, Thiel B, Gessner K (1984)
Ornithophilie auf den Canarischen Inseln. Plant
Systematics and Evolution 146:225–248. https://doi.
org/10.1007/BF00989548
This work is licensed under a Creative Commons Attribution 4.0 License. ISSN 1920-7603
