Introduction

Floral nectar is a chemically complex aqueous solution in which the main components comprise sugars, followed by amino acids (Nicolson and Thornburg 2007). In recent decades considerable progress has been made in providing evidence that points to the involvement of nectar chemistry in the interactions between plants and a variety of organisms (Nepi 2014; Stevenson et al. 2017). Although there is wide variability in nectar traits (Pacini et al. 2003; Nocentini et al. 2013; Irwin et al. 2014), a general paradigm shared by plants is balancing nectar chemical composition in order to not deter specific pollinators exceeding their tolerance thresholds (Baker and Baker 1975; Adler 2000; Nicolson 2007; Wright et al. 2013; Stevenson et al. 2017). For example, a small increase in nectar sugar concentration can increase its viscosity (Harder 1986; Nicolson and Thornburg 2007), which is strongly related to the energy required by nectar consumers to visit flowers (Corbet 1978; Josens and Farina 2001; Borrell and Krenn 2006; Nepi and Stpiczyńska 2006; Kim et al. 2011).

After sugars the most abundant nectar solutes are the amino acids (Baker and Baker 1982; Nepi et al. 2012; Bogo et al. 2019). A study conducted by Inouye and Waller (1984) showed a general decline in nectar consumption in honeybees as amino acid concentrations increased, despite evidence supporting the preference for amino acid enriched sugar solutions in insects (Alm et al. 1990; Bertazzini et al. 2010; Bogo et al. 2019). Amino acids also contribute to the taste of nectar, stimulating specific insects’ labellar chemoreceptors (Gardener and Gillman 2002). Among protein amino acids, Inouye and Waller (1984) found that phenylalanine and leucine were phagostimulant for honeybees at all concentrations tested, even at those that in the case of other amino acids resulted in deterrence. In the same way, a preference in honeybees for proline enriched artificial nectar was reported (Carter et al. 2006; Bertazzini et al. 2010), as well as a strong phagostimulatory activity (Nicolson and Thornburg 2007; Petanidou 2007).

Beside primary metabolites (such as sugars and amino acids) an array of secondary metabolites with different chemical natures have been identified in nectar and all of them positively or negatively affect attractiveness to pollinators, showing effects which depend on metabolite concentration and pollinators’ sensitivity (Baker and Baker 1977, 1982; Faegri and van der Pijl 1979; Adler 2000; Stevenson et al. 2017). Among them non-protein amino acids (NPAAs) have been detected in nectar (Nicolson and Thornburg 2007; Petanidou 2007; Nepi et al. 2012). Despite that they can constitute a large portion of the amino acidic content of floral nectar, little is known about their role in determining pollinators’ preferences and feeding behaviour. For some of those, such as γ-aminobutyric acid, a phagostimulant function has been reported in some caterpillars and adult beetles (Mitchell and Harrison 1984; Schoonhoven et al. 2005), whilst Bogo et al. (2019) found that both bumblebees and honeybees showed higher consumption of sucrose solution enriched with ß-alanine, but exhibited the effect at different concentrations.

Understanding preferences and aversions to nectar traits is crucial since they likely influence the patterns of floral visitation by nectar consumers and thus the plant inbreeding and outbreeding rate within a population. Minimal inbreeding is predicted when pollinators visit a small fraction of the open flowers on a plant (Iwasa et al. 1995; Ohashi and Yahara 2001): this behaviour may be enhanced by within-plant variation in nectar, as occurs in plants showing gender-biased nectar production (Feinsinger 1978; Pyke 1978; Rathcke 1992). Despite many studies having already addressed the subject of gender-biased nectar composition, most of them investigated the existence of bias in relation to nectar volume or sugar content only (Langenberger and Davis 2002; Canto et al. 2011; Fisogni et al. 2011; Stpiczyńska et al. 2015; Antoń et al. 2017; Jacquemart et al. 2019; Konarska and Masierowska 2020) and few reported the observation of insect visit bias (Carlson and Harms 2006 and references therein).

In this study we focused on the many-flowered hermaphrodite species Echium vulgare L., a self-compatible plant which shows both herkogamy and incomplete protandry, that avoids self-pollination within the same flower, but within which geitonogamy can still occur (Rademaker et al. 1999). Melser et al. (1999) reported evidences of inbreeding depression in E. vulgare, finding a significant decline in siring success when selfing occurs. A study on geitonogamy conducted by Rademaker et al. (1999), though, found a consistently lower percentage of selfing rate than expected. Also, they reported that bumblebees visited only a small fraction of the flowers on E. vulgare as a result of the presence of different flower stages simultaneously occurring on a single individual plant.

E. vulgare represents an important food resource for many insect visitors, despite containing toxic pyrrolizidine alkaloids in both nectar and pollen (Lucchetti 2017). The pollen contains high concentrations of pyrrolizidines, whilst more than 500 times lower concentrations are found in nectar (Lucchetti et al. 2016). For this reason, only a few taxa show oligolecty or floral constancy on E. vulgare by actively collecting pollen for larval nourishment (Cane and Sipes 2006; Burger et al. 2010; Filella et al. 2011), even if its flowers are visited by a wide spectrum of insect taxa among which bumblebees have often been reported as main pollinators (Corbet 1978; Klinkhamer and de Jong 1990; Pappers et al. 1999; Rademaker et al. 1999).

Here, we examined if floral visitation pattern may be influenced by variations in the chemical composition of nectar through different floral stages, and thus we investigated (i) whether E. vulgare produces a gender-biased nectar for volume, sugar and amino acid composition and (ii) if flower visitation rates of insects looking for nectar varied among different floral stages.

Material and methods

Study site

The activity in the field was carried out in June 2018 and took place in the Parco Belpoggio, a public park managed since 2010 by the WWF, in San Lazzaro di Savena (Bologna, Italy). The area is situated close to the protected area Parco dei Gessi Bolognesi e Calanchi dell’Abbadessa (44° 27′ 14.5″ N 11° 22′ 58.3″ E). The studied population was located on an open prairie along the public pathway.

Study species

Echium vulgare L. is a perennial hemicryptophyte belonging to the family Boraginaceae. It is distributed in Europe, Asia and North America and it shows a long flowering period, ranging between June and October. Flower anthesis lasts 3–4 days and flowers show an incomplete protandry (Melser et al. 1997): the anthers are often dehiscent already at the bud stage, while the stigma becomes receptive only hours after the flower opening.

In this study we considered three phases of floral development: closed flower (Bud), functionally male (M) and functionally female (F) flowers. The male phase was represented by an open flower presenting pollen with non receptive stigma, whilst the female phase was recognised as soon as the stigma became bifid and receptive.

Plant phenology

On the first day of the study we counted all plants and inflorescences per plant constituting the population (an area of approximately 600 m2) and we observed all open flowers to assess whether the phenomenon of gynodioecy, firstly described in E. vulgare populations by Darwin (1877), occurred in our study population. Each day, prior to visitor observations, on the same patch we recorded the number of flowers per developmental stage. Two fixed patches were alternatively considered: the first one was a single plant carrying 6 inflorescences while the second one was made up of 6 plants carrying one or two inflorescences each.

Nectar quality

Sampling

We collected nectar samples by means of Drummond Microcaps (3–5 μL; Drummond Scientific Co., Broomall, PA), we transferred samples to Eppendorf tubes filled with 100 μL of pure ethanol, and then we took them to the laboratory in thermal bags where they were kept at 5 °C until analyses. We collected each sample from multiple flowers at the same floral stage in order to reach a minimum volume of 2 μL needed for the sugar and amino acid analyses. In order to let the nectar accumulate, flowers were bagged in the morning for 2 h prior to sampling; all nectar present in the selected flowers was collected.

We collected a total of 8 nectar samples, each one from 3 to 13 male flowers belonging to 1–7 plants, and a total of 8 samples from 2 to 9 female flowers belonging to 1–3 plants. Both sugar and amino acid compositions were investigated on these samples. We then collected 14 additional samples from 1 to 22 buds belonging to 1–10 plants. Since the amount of nectar presents in the buds was very low, the minimum volume of 2 µL needed for amino acid analysis could not be reached and thus these samples were tested for sugar composition only.

Sugar analysis

Sugar content was analysed by HPLC technique through a Waters LC1 with refractive index detector (Waters 2410) connected to the output of a REZEX RCM Monosaccharide column (Phenomenex, 300 mm × 7.8 mm, grain 8 µm) maintained at 85 °C. Water (MilliQ, pH 7) was used as mobile phase at a flow rate of 0.6 mL min−1; 20 µL of sample and standard solutions of sucrose, glucose and fructose were also injected (Nocentini et al. 2012).

Amino acids analysis

Amino acid analysis was performed by gradient HPLC with an ion exchange Novapack C18 (15 mm × 4.6 mm) cartridge with guard column maintained at 37 °C and a Waters 470 scanning fluorescence detector (excitation at 295 nm, detection at 350 nm). A solvent composed of TEA-phosphate buffer (pH 5.0) mixed with a 6:4 acetonitrile–water solution was used as mobile phase at a flow rate of 1.0 mL min−1. According to AccQtag protocol (Waters Corp.), the selected volume of each reconstituted sample was amino acid derivatized (Cohen and Micheaud 1993) with AQC fluorescent reagent and 0.02 M borate buffer (pH 8.6). In addition to all the protein amino acids, standard solutions of β-alanine, citrulline, l-homoserine, α-aminobutyric acid (AABA), γ-aminobutyric acid (GABA), hydroxyproline, ornithine and taurine were also used (Nocentini et al. 2012).

Flower visitors’ observations

We carried out observations on flower visitors on the two fixed patches described previously, on 7 non-sequential days. Every survey consisted of two 15-min periods separated by 10 min of rest, adapting the protocol of Fisogni et al. (2016). Every day we performed 1 to 3 surveys, between 10:30 am and 3:00 pm and under favourable weather conditions, for a total of 200 min of observation. Once a visitor left the patch, we counted the following approaching insect belonging to the same taxon as a different individual. Recorded data concerned the food resource collected (nectar or pollen, observing if the insect inserted its mouth-parts deeply inside the corolla or if it manipulated the anthers) and the number of male and female flowers approached per visit. We also recorded the visitor’s taxon, indicating the taxonomic level in as much detailed as possible, and its sex.

After each observation period, we performed a 15-min period of net sampling throughout the area, collecting insects that alighted on flowers of E. vulgare. Captured individuals were put in separate vials with ethyl acetate and brought to the laboratory where they were pinned in entomological boxes and inspected under a dissecting microscope for taxonomic identification.

Data analysis

Sugar and amino acid quantities and the mean nectar volume were calculated per single flower. Total sugar concentration was calculated as the sum of sucrose, fructose and glucose concentrations.

Data on nectar composition were grouped by floral stage and tested to assess homogeneity of variances and normality of distribution (Bartlett test and Shapiro Wilk test).

Data on sugars per flower, total sugar concentration and sucrose per flower were square root transformed to achieve normality. When the transformed data failed to match normality, we applied the corresponding non-parametric analyses.

To investigate whether the floral stage affected sugar content and volume a one-way ANOVA followed by Tukey’s HSD post hoc test with Benjamini–Hochberg correction for ‘false discovery rate’ (Verhoeven et al. 2005) were performed. When distribution was not normal a Kruskal Wallis H-test followed by a Mann Whitney pairwise comparison with Benjamini–Hochberg correction were carried out instead.

Data on single amino acid concentrations were ln transformed to achieve normality when needed and a Student t-test was applied in all analyses.

For both phenological stages (functionally male and functionally female flowers), three diversity indices were calculated on the nectar amino acid composition. The first index was the reciprocal Simpson’s diversity index 1-D of the nectar amino acidic spectrum. D was calculated as \(D=\sum_{i=1}^{n}{\left(\frac{ni}{n}\right)}^{2}\), where ni is the abundance of the ith amino acid and n is the total mean concentration (Ranjbar et al. 2017). This index ranges from 0 (one amino acid dominates the spectrum) to 1 (all amino acids equally represented) (Harper 1999).

The second was the Shannon’s H- index, by taking into account mean amino acid concentrations as well as the total mean concentration of amino acids. The index is calculated as \(H=-\sum_{i}\frac{ni}{N}\mathrm{ln}\frac{ni}{N}\), where ni is the mean concentration for the ith amino acid and N is the total number of amino acids (Magurran 2004). This index varies from 0 for a spectrum with only a single amino acid to high values for a spectrum with many amino acids, each represented by relatively low concentrations (Harper 1999; Hubálek 2000; Fattorini et al. 2016).

The third one was the Buzas and Gibson’s evenness index, a measure of the relative abundance of the different amino acids within the floral stage. The index is calculated as the proportion of equally dominant amino acid in the phenological stage \(E={e}^{H}/S\), where H is Shannon’s H index and S is the number of amino acids within the floral stage. This index ranges from 0 (highest dominance by a single amino acidic species) to 1 (all amino acids have the same abundance) (Buzas and Hayek 2010; Fattorini et al. 2016).

Insect visit data were first analysed by comparing the observed number of male and female flowers visited to the expected ones by χ2 test. The expected number of visits was calculated on the basis of the ratio between the functionally male and the functionally female flowers occurring in the population.

Frequencies of male flowers visited by each taxon were compared by a Kruskal Wallis H-test followed by a Mann–Whitney pairwise comparison with Benjamini–Hochberg correction.

All data are presented as mean ± SE and all statistics were performed using R software (version 3.6.1) with the significance level set at 0.05.

Results

Plant phenology

In June 2018, the studied population contained 47 flowering individuals, all hermaphrodites. The mean number of inflorescences per plant was 3.17 ± 0.44, while the mean number of cymes per inflorescence was 14.30 ± 0.81. Moreover, the mean number of male flowers per inflorescence was 2.69 ± 0.171, while the mean number of female flowers per inflorescence was 21.07 ± 0.858. On the basis of the data collected on the population structure the ratio of male and female floral stages in the observation patches was determined at 1:9.

Nectar analyses

Sugars and volume

Mean nectar volume per flower showed a clear trend of increasing in relation to floral age, with volume in buds statistically lower than in both male- and female-phase flowers (U = 15, p = 0.009 and U = 2, p = 0.001, respectively). A significant difference for mean sugar quantity per flower was also reported between buds and female-phase flowers (Tukey’s HDS: p = 0.028), whilst sugar concentration did not differ significantly among floral stages (Table 1).

Table 1 Comparison of nectar volume, sugar and amino acid (AA amino acids, PAA protein amino acids, NPAA non-protein amino acids) compositions among the three phenological stages (bud, male and female flowers)
Full size table

A more in depth analysis on sugars reported that hexose sugar quantity per flower in the bud stage differed significantly from both male- and female-phase flowers (U = 12, p = 0.008 and U = 19, p = 0.018, respectively), whilst sucrose quantity per flower found in bud differed statistically only from the average amount found in the female stage (Tukey’s HDS: p = 0.021; Table 1). Mean percentage of sucrose per flower did not appear to be significantly different among floral stages (Table 1).

Amino acids

There was no significant difference for total, protein, and non-protein amino acid quantity per flower between male and female flowers, while the ratio between protein and non protein amino acid concentrations was significantly higher for male-phase flowers (Table 1).

The only amino acid with a statistically significant difference was phenylalanine (t15 = 2.94, p = 0.011), showing a higher concentration in male floral phase (M = 352.7 ± 63.2 nmol mL−1 and F = 143.6 ± 32.6 nmol mL−1; Fig. 1).

Fig. 1
figure 1

Amino acid concentrations (nmol mL−1) detected in functionally male (dark bars) and in functionally female (light bars) flowers (mean ± SE). Amino acids hydroxyproline, homoserine, citrulline, cysteine, histidine, glutamine, asparagine and l-thyronine were not detected in either floral stages and thus not shown in the graph. The asterisk denotes a statistically significant difference according to Student t-test. NPAA non-protein amino acids, PAA protein amino acids

Full size image

Among all protein amino acids, proline and phenylalanine showed the highest concentrations: the former appeared to reach higher concentrations in the functionally female stage (674.8 ± 243.5 nmol mL−1), whilst the latter in the functionally male stage (352.7 ± 63.2 nmol mL−1). Among non protein amino acids, in both male and female stages GABA showed the highest concentration (51.4 ± 12.2 nmol mL−1 and 202.0 ± 73.4 nmol mL−1, respectively).

The number of different amino acids (richness) detectable in the male stage was significantly lower than number of amino acids in the female stage (t15 = 3.54, p = 0.003; 16.5 ± 0.6 and 19.0 ± 0.3, respectively), while no differences were found in Simpson, Shannon and Evenness indices between male and female stages (Table 2).

Table 2 Comparison of diversity indices calculated on nectar amino acid concentration between male and female phases (8 samples for both floral phases)
Full size table

Insect visit analyses

Flower visitors’ abundance

A total of 215 insect visits were recorded on Echium vulgare during 200 min of field surveys (Table 3).

Table 3 Echium vulgare visitors recorded in June 2018 (215 visits in total), their abundance and the percentage of them looking for nectar as reward
Full size table

Visitors belonged to three orders: Hymenoptera (87.4%), Lepidoptera (9.8%) and Diptera (2.8%). The order Hymenoptera was mainly represented by individuals belonging to the family Megachilidae (59%), followed by the family Halictidae (26.5%) and Apidae (14%). The order Lepidoptera was represented mainly by individuals belonging to the species Macroglossum stellatarum (43%) and the family Pieridae (43%). The order Diptera was represented only by 6 individuals belonging to the families Bombyliidae and Syrphidae. The most frequent visitors were solitary bees of the species Hoplitis adunca (42%).

Flower visitor observations

Among the 215 insects visiting the plant, we fully recorded data for 189 individuals. Statistical analyses were carried out only on the 112 individuals which were looking for nectar and for which the number of total visits exceeded 5 (Macroglossum stellatarum, Pieridae, Anthidium florentinum, Apis mellifera and Hoplitis adunca). The family Pieridae was analysed as a single taxon in order to reach a total number of visits above 5. Since Hoplitis adunca was the most abundant taxon and the only species strongly oligolectic on Echium, we therefore decided to analyse the sexes separately.

Although nectar is produced before flower opening and insects can force the bud searching for nectar (personal observation), this event occurred very rarely. Consequently, we did not consider the phenological stage bud in these analyses.

For each insect taxon, we compared the number of visits to male and female flowers with the expected ones, calculated according to the ratio 1:9 between male and female flowers registered in the studied population.

Regarding the number of male flowers visited, no significant difference was reported for lepidopterans (Pieridae spp., Macroglossum stellatarum) and for females Hoplitis adunca, while Anthidium florentinum, Apis mellifera and Hoplitis adunca males visited more male flowers than expected (Table 4). The number of female flowers visited was never statistically different from that expected.

Table 4 Male (a) and female (b) flowers visited by each taxon (mean ± SE)
Full size table

The frequency of male flowers visited in relation to the total number of flowers visited among taxa was statistically different (H4 = 14.01, p = 0.016). Statistical analyses confirmed that the female Hoplitis adunca visited fewer male flowers than did Anthidium florentinum (U = 65, p = 0.002), Apis mellifera (U = 48, p = 0.002) and Macroglossum stellatarum (U = 28.5, p = 0.043; Fig. 2).

Fig. 2
figure 2

Frequency of male flowers visited by each taxon. Different letters denote statistical differences according to Kruskal Wallis H-test followed by Mann-Withney pairwise comparison with Benjamini–Hochberg correction (p < 0.05)

Full size image

Discussion

Our studied population did not show the phenomenon of gynodioecism, as all flowers were hermaphrodite, and our data confirmed the ratio of 1:9 found by Rademaker et al. (1999) between functionally male and functionally female flowers.

Our analyses confirmed that nectar is secreted in the bud, as reported by Chwil and Weryszko-Chmielewska (2011). Contrary to Klinkhamer and de Jong (1990), we found that nectar volume, as well as sugar quantity per flower, increased with the age of the flower (from bud to female phase), although the positive trend between male and female phases was not statistically significant. Both quantity of hexose sugars and sucrose per flower increased with the age of the flower, the latter reaching a mean almost sevenfold higher in functionally female flowers than the mean amount found in the bud stage and almost twice the amount found in functionally male flowers. At the same time, the mean percentage of sucrose per flower appeared to be lower in male-phase flowers, even though not significantly, meaning that the total sugar increase in relation to floral age is due to the rise of nectar volume, since total sugar concentration and composition remained constant during the entire flower phenology. The existence of nectar homeostasis mechanisms which actively maintain a constant nectar sugar concentration to ensure pollinator visits has been previously reported in other species (Nepi and Stpiczyńska 2008; Nepi et al. 2011).

When we compared the number of insect visits on male and female flowers observed to the expected ones, all bee taxa except female Hoplitis adunca showed a higher number of visits to male flowers than expected. This result could be explained by the higher proportion of protein amino acids found in the male stage: preferences have often been reported in bees for protein amino acid enriched solutions (Inouye and Waller 1984; Bertazzini et al. 2010; Hendriksma et al. 2014), suggesting that flower visitors may actively choose to visit functionally male flowers. Comparable results have been reported by Klinkhamer and de Jong (1990) and by Rademaker et al. (1999) on bumblebees: when calculating the probabilities of visits on different floral stages, the oldest female stage was less likely to be visited than a male-phase flower. Females of Hoplitis adunca are the only bees collecting both pollen and nectar on E. vulgare: this different foraging behaviour might explain the difference from the other bee species.

Individuals of Lasioglossum sp. were observed visiting the flower and collecting pollen only. A tendency for afternoon trips for nectar only have been reported for the subfamily Halictinae by Michener (2003) so we cannot conclude that Lasioglossum sp. does not exploit E. vulgare nectar since the species may simply collect the resource at different time of the day.

Despite Lepidoptera having been reported to prefer nectar rich in PAAs (Baker and Baker 1986; Erhardt and Rusterholz 1998), our study reports that Pieridae butterflies visited as many male flowers as expected, indicating that these insects did not actively look for functionally male flowers (containing a higher proportion of protein amino acids). A study conducted by Alm et al. (1990) showed that male individuals of the species Pieris rapae do not discriminate between artificial nectars containing sugar only or sugar solution enriched with protein amino acids, and Romeis and Wackers (2000) reported that feeding and source-selection in Pieris brassicae is elicited by sucrose more than protein amino acids. We report a similar result for the species Macroglossum stellatarum, but to date no study has been done in order to assess amino acid preferences in the species and whether taste receptors on the proboscis can sense their presence in nectar remains unsubstantiated (Stöckl and Kelber 2019).

Nectar of male-phase flowers in E. vulgare presented, among all the amino acids, the highest concentration of phenylalanine, representing an average of 35% of total amino acid content. Phenylalanine is an essential protein amino acid (de Groot 1953) and several studies proved that it exerts a phagostimulatory effect on several insects, especially on honey bees, and it is strongly correlated with pollinator preferences (Inouye and Waller 1984; Hendriksma et al. 2014; Tiedge and Lohaus 2017; Seo et al. 2019). Consequently, this could explain the higher frequency of visit on male flowers than expected. A correlation between phenylalanine concentration and nectar feeding by Megachilids, that were the more numerous pollinators in our study, was demonstrated in a phriganic community, a plant association typical of the East Mediterranean (Petanidou et al. 2006).

Proline, instead, represented the most concentrated amino acid in functionally female flowers, and the second in the early-stage functionally male flowers (representing more than 30% and almost 20% of the total amino acid content, respectively). This non-essential amino acid, commonly found in nectar (Nicolson and Thornburg 2007), can stimulate the insect salt cell increasing intensity of feeding behaviour (Hansen et al. 1998; Wacht et al. 2000). Proline also represents an energy substrate to fuel the earliest or most expensive stages of insect flight (Micheu et al. 2000; Gade and Auerswald 2002), resulting in short-term bursts of energy production (Teulier et al. 2016).

Finally, in both male- and female-phase flower nectar GABA showed the highest concentration among the non-protein amino acids representing more than 5% and 9% of total amino acid content, respectively. Recent studies indicated that GABA could affect both insects’ physiology and behaviour, feeding rate and flight muscles performances (Shelp et al. 2017; Felicioli et al. 2018; Bogo et al. 2019). Besides GABA, or possibly the combination of GABA and NaCl, can constitute an important nectar phagostimulant and its presence correlates with visits by an array of pollinators such as long tongued bees, ex-anthophorid and andrenid bees, as well as anthomyiid and syrphid flies (Petanidou 2007 and reference therein).

The spectrum of visitors recorded through our observations confirm that reported by previous studies stating that flowers of E. vulgare are visited by hummingbird hawkmoths (Aguado Martìn et al. 2017), bees, bee flies (Proctor et al. 1996) and syrphids (Willmer and Finlayson 2014). Also, even though the species has often been reported as mainly pollinated by bumblebees (Corbet 1978; Klinkhamer and de Jong 1990; Pappers et al. 1999; Rademaker et al. 1999), we observed only one individual of Bombus pascuorum visiting the flowers. Pollinators of wide spread plant species can vary in relation to their geographical distribution (Armbruster 1985; Thompson 2006; Pérez-Barrales et al. 2007) and, moreover, as reported by Lázaro et al. (2010), the plant and pollinators assemblages of an entire community may also influence the composition of visitors of a particular species by determining, for instance, the strength of competition or the intensity of attraction to that species rather than another. Thus, the scarcity of bumblebees observed on Echium vulgare in 2018 may either depend on several factors and/or reflect a temporal fluctuation in the species composition of the pollinator community, as previously reported by many studies (Cane et al. 2005; Petanidou et al. 2008; Dupont et al. 2009).

Conclusions

The inbreeding avoidance hypothesis states that some mechanisms develop within a species in order to prevent breeding among related individuals and its damaging effects on fitness (Darwin 1876, 1877; Charlesworth and Charlesworth 1987). In dichogamous species, gender-biased nectar often occurs (Carlson and Harms 2006; Stpiczyńska et al. 2015; Konarska and Masierowska 2020), and this, according to the mentioned above hypothesis, may contribute to decrease geitonogamous selfing through its effects on a pollinator’s behaviour (Carlson and Harms 2006). Our results suggest that the quality of nectar offered by the two sexually distinct floral phases may target different insect needs, thus affecting simultaneously different behavioural traits and ensuring an optimal pattern of visit among functionally different floral stages, unequally present in the population throughout the anthesic period. The more nutritional nectar found in the less frequent sexual phase occurring in the population (male flowers) may enhance movements among plants by encouraging “better-resource hunt”, whilst the flight efforts accomplished for doing so may be sustained by a rapidly oxidable fuel such proline offered in female-phase flowers. In the light of this hypothesis, it appears clear that gender-biased nectar studies in dichogamous, many-flowered species should be undertaken in relation to the occurrence of floral sexual phases in the population (when a bias in the frequency of sex occurrence exists).

Despite no study yet providing strong scientific evidence that gender-biased nectar in fact reduces inbreeding (Carlson and Harms 2006), it is reasonable to assume that by offering variable quality nectar through sexually different floral phases the plant may produce a mosaic of food targeting different pollinator behavioural traits aiming to promote cross-pollination.