Influence of sexual dimorphism and dichromatism on reproductive success in a rare native cactus

Identifying plant sexual dimorphic traits is critical in advancing our knowledge on plant–pollinator interactions. For example, dimorphism in floral colors, or sexual dichromatism, is a crucial mediator of pollinator choice on foraging decisions. We studied Cylindropuntia wolfii, a model system, with diverse flower colors and a functionally dioecious sexual system. However, evidence suggests that sexual reproduction is limited in this species as it has a low seed set especially in naturally pollinated fruits. Thus, it is critical to this native species’ conservation to investigate its relationship with pollinators. Our goals were to: (a) investigate the sexual dimorphism including the sexual dichromatism in the flowers of the cactus, and (b) determine whether sexually dimorphic traits affect the pollinator attraction of both the sexes. We measured several quantitative and qualitative traits and compared them between male and female flowers. Then we recorded the pollinator visitation rate in nature for both sexes and tracked pollinator color preference using fluorescent dyes as pollen analogues. Our study showed that male flowers of C. wolfii are bigger and brighter, and they attract more potential pollinators than females, supporting the hypothesis that sexual dimorphism influences pollinator visitation preference. Fluorescence dichromatism, in which female flowers’ anthers fluoresce more than male flower anthers suggest this could be female flowers’ strategy to compensate for their dark colors and small size. The results from this study showed that C. wolfii exhibits sexual dichromatism and fluorescence dichromatism, which is a novel finding in plant research.


Introduction
Dioecy, a sexual system where female and male flowers are carried in different individuals, is present in ~ 6% of angiosperms and its occurrence is observed in ~ 40% of the angiosperm families (Renner 2014).Dioecy is reported to have evolved independently in many lineages at least 871 times (Renner 2014).One common factor that sexual separation is accompanied by is sexual dimorphism (Correns 1928;Lloyd and Webb 1977;Barrett and Hough 2013).Sexual dimorphism is defined as the morphological differences between males and females in their secondary characters, which includes all traits other than reproductive organs.
In dioecious plants, sexual dimorphic traits include floral size (Delph et al. 1996), flower longevity (Primack 1985), number of flowers per individual (Delph et al. 2005), scent (Ashman 2009), nectar production (Bawa and Opler 1975).However, sexual dimorphism in terms of color, also known as sexual dichromatism, has not been reported in plants to the best of our knowledge.In contrast, it has been widely reported in animals such as birds (Badyaev and Hill 2003), fishes (Kodric-Brown 1998) and reptiles (Olsson et al. 2013).Studies on sexual dimorphism in plants have identified few species that exhibit intersexual mimicry to overcome pollinator visit discrimination.This mimicry is especially prevalent in plants where the females do not offer any rewards such as pollen or nectar (Delph et al. 1996;Ashman 2009).While this mimicry is not always successful, theory predicts that it should not effectively affect reproduction as long as there is no pollinator limitation (Yang et al. 2022).
The evolution of sexual dimorphism has been explained by sexual selection, especially in animals (Darwin 1851;Bateman 1948).Darwin (1851) suggested that two factors contribute to sexual selection of traits: male-male competition and female choice.In the past decades, this principle has also been increasingly applied to plants (Delph and Ashman 2006).Male-male competition in plants results in male flowers becoming more attractive to pollinators by evolving a brighter display, more potent scent, and other relatable traits (Stanton 1994).However, there can be other architectural traits that are sexually selected but do not have an effect on biotic pollinator attraction, in the wind pollinated species (Tonnabel et al. 2019).Even within the insect pollinated species, forces other than pollinator-mediated selection like fertility selection in females, can select for traits that improve the reproductive success (Barbot et al. 2022).From the female's perspective, studies have observed selection for fertility traits such as ovule number and flower number (Delph and Herlihy 2012;Barbot et al 2023), although flower number can also be a result of pollinator mediated selection.Studies have shown that the mating success (finding mates) increases with reproductive success (siring offspring) in males while they are not correlated in females and hence the difference in selection acting on different floral traits (Tonnabel et al. 2021;Kwok and Dorken 2022).
On the other hand, cryptic female choice is predicted to occur in plants, but there is little evidence as it usually occurs a long time after fertilization through pre-selected abortion of ovules during the embryogenesis (Moore and Panell 2011).Recently, Tonnabel et al. 2022 reported female choice in wind pollinated species which involved traits such as pollen tube growth, size of stigma and expression of pollen proteins that facilitate competition for ovules.This suggests that sexual selection continues to occur even after the pollen dispersal phase of plant reproduction.
Studies on plant sexual dimorphism will advance our knowledge on plant-pollinator interactions and the influence of pollinators' preference on plant sexual separation.The key factors that contribute to the study of plant-insect pollination and their conservation in the face of climate change and loss of biodiversity are floral strategies and floral resource availability (Vanbergen and Insect Pollinators Initiative 2013).One of the crucial mediators of plant and pollinator interaction is flower color polymorphism (Weiss 1991).While many factors contribute towards pollinator choice for a particular flower, a large number of studies have suggested that flower color widely influences their choice on foraging decisions (Dafni et al. 1990;Ômura and Honda 2005;Dötterl et al. 2014;Reverte et al. 2016).Diurnal insect pollinators have trichromatic vision which covers a broader range of spectrum than humans.Their vision includes visible light in the UV, fluorescent, blue, and green spectra (Chittka 1992;Mori et al. 2018).Based on phylogenetic analysis, it has been shown that pollinators were pre-adapted to trichromatic vision long before the widespread radiation of angiosperms.This evidence leads to the hypothesis that flowers have evolved their beautiful colors to visually attract the pollinators (Chittka 1997;Reverte et al. 2016).Among floral color traits, UV patterns play a major role in pollinator preference since UV-absorbance and reflection guide the pollinators to foraging parts (Lunau et al. 2017).Several studies have demonstrated the importance of UV patterns in bees through artificial manipulation of UV patterns in flowers (Welsford and Johnson 2012;Rae and Vamosi 2013;Brock et al. 2016;Klomberg et al. 2019).Recently, a study by Rao and Ostroverkhova (2015) demonstrated the strong attraction of a wide variety of wild bees to fluorescence under UV excitation.Therefore, studies in plants displaying dimorphism in floral traits and the effect of floral color polymorphism, including UV and fluorescence patterns, on pollinator attraction is needed.Sexual dichromatism in fluorescence has only been recently reported in blue winged parrotlets (Barreira et al. 2012), highlighting the need to further study the role of fluorescence patterns on sexual dimorphism.Moreover, the relationship between sexual sterility in plants and the specific autofluorescent properties of anthers remains an area of scientific investigation that has not yet been explored fully (Araki et al. 2020;Nakashima et al. 1984).
Cylindropuntia wolfii (Cactaceae) is an exceptional model to study sexual dimorphism and dichromatism, and its influence on pollinator attraction owing to its six distinct color morphs within the same geographical location (Fig. 1) and its functionally dioecious sexual system (Ramadoss et al. 2022).The sexual system is called functionally dioecious as their flowers carry vestigial organs that are non-functional.Both the male and female flowers have anthers but males have pollen on them and females lack pollen (Ramadoss et al. 2022).This species is rare and native to California and Baja California.Cylindropuntia wolfii is ranked 4.3 on the Inventory of Rare and Endangered Plants by the California Native Plant Society (1968 onwards) implying it has limited distribution but the populations are not threatened.Moreover, our previous study on C. wolfii showed that it had a lower seed set compared to its close relatives (Ramadoss et al. 2022).This seed set was significantly lower in naturally pollinated fruits compared to manually pollinated ones.Thus it would be beneficial to this native species' conservation to explore its relationship with pollinators.Our goal here is to investigate the sexual dimorphism of the cactus C. wolfii and determine its effect on pollinator attraction.
In particular, we had two particular goals and predictions: (1) Because C. wolfii has different flower color morphs and it is functionally dioecious, we determined whether the flowers show sexual dimorphism including dichromatism.As observed in other dioecious species, we predict that there will be sexual dimorphism traits such as display diameter, tepal width, tepal length, number of flowers, and because they exhibit color polymorphism we tested for sexual dichromatism in tepal and filament color.If we observe sexual dichromatism in the flowers, that could be reflected in the emission of UV and/or fluorescence patterns as we have different flower color morphs.The emission intensity patterns of UV and green fluorescence holds significance due to its visibility to bees and its potential to impact pollinator attraction (Rao and Ostroverkhova 2015;Mori et al. 2018).( 2) We evaluated whether the presence of sexual dimorphism influenced the pollinator visitation rate.We predict that if there is sexual dimorphism between the male and female flowers, then that will influence flowers' pollinator visitation rate with more pollinators visiting the sex that has more attractive traits (e.g.larger and brighter flowers).

Study site and species
The present field work was carried out in Mountain Springs (Imperial County, CA) from April 2021-July 2022.Cylindropuntia wolfii is a rare endemic shrub which is narrowly distributed in the Sonoran Desert of southern California and Baja California.It is densely branched with stout cylindrical stems and sheaths on their spines.The flowers have six tepal colors ranging from green to red with green to dark pink filaments on their stamens (Fig. 1).

Flower dimorphism measurements
To determine sexual dimorphism, we measured and compared several quantitative and qualitative characters between the male and female flowers.The qualitative floral traits refer to flower color and filament color, and quantitative factors refer to flower display diameter, tepal length, tepal width, and number of flowers per plant.For the qualitative factors, we recorded the flower color and sex of only 53 plants along a 600 m transect in Mountain Springs (Imperial County, California).The flower colors were attributed based on a self-built color map with the help of herbarium color reference chart (McCamy et al. 1976).The six flower color morphs were further distinguished into two groups as brighter (green, greenish yellow and yellow) and darker (yellowish orange, orange and red) colors.During our pollinator survey, we visually and statistically observed a pattern (supplementary Fig. 2) where brighter morphs attracted more than the darker morphs so we split the six color morphs into two groups based on their color gradient, grouping 3 morphs per group.The reduced sample size is due to asynchronous blooming and reduced number of plants.The filament color was recorded from 94 plants we collected from the field.A Fisher's independence test was performed to check if the qualitative factors are related to sex.
Then, quantitative factors which are display diameter, tepal length and tepal width were measured and compared between the two sexes.The survey was done in 2022 with a sample size of 94 plants with at least seven plants per sex and color combination.A general linear model of ANOVA was used to assess the significant differences in quantitative traits between sexes after testing for normality using Shapiro's test and homoscedasticity using Bartlett's test.We then explored the correlation between color morph and flower size to assess if brighter colored flowers, regardless of sex, are bigger using a two-way ANOVA.

Ultraviolet (UV) and green fluorescence (GF) measurement
To determine sexual dimorphism in terms of UV and GF, we measured their emission intensity in male and female flowers.At least seven plant fragments with flower buds were collected for each color and sex combination and tracked until blooming to assess dimorphic differences.The fragments were potted in 70% soil, 15% perlite, and 15% sand mixture and grown in a growth chamber.The photoperiod was set to 12 h with day temperature of 25 °C and night 15 °C.The bloomed flowers were collected and mounted under a NIKON stereomicroscope (SMZ-25/18) at 0.5X magnification.Photographs were taken using RBG (Red Green Blue) filter for human eye view and Blue, green, and UV filter for bee's eye view (Fig. 2) based on the false color photography protocol by Verhoeven et al. (2018).In addition, GFP (Green Fluorescent Protein) filter cube photographs were also taken to capture the green autofluorescence of anthers in male and female flowers.Photographs were analyzed using Nikon BRanalysis software to measure the emission intensity of UV and GF per unit area, from the flowers.The emission intensities were compared by ANOVA and post hoc tests in R and R studio.
To further investigate the autofluorescence in female flower anthers, buds from C. wolfii females were collected from the pre-anthesis stage as that is when the premature degradation of tapetum occurs (Ramadoss et al. 2022).The anthers were placed in glass vials containing 70% ethanol for preservation.The anthers were frozen in the freeze medium and then cut into sections with a thickness of 10 µm using a Leica CM1950 cryostat.The cryo-sectioned samples were placed on SuperFrost Plus microscope slides (Fisher Scientific, USA) and observed under Nikon stereo microscope under regular vs GFP light for observing the location of autofluorescence.

Pollinator visitation rate
To determine the pollinator visitation rate between the sexes, we followed three steps.First, we determined whether C. wolfii flowers attract potential pollinators by manually surveying the diurnal floral visitors during the peak flowering season, April-May of 2021.We surveyed 24 plants including both sexes and all colors except for male red as they are hard to find, for 10 min each during the daytime (between 8 am and 2 pm) in the Mountain Springs field site on a clear sunny day.The number of visits and types of visitors (i.e., bees, moths, beetles) were noted down.The percentage of visits from each type of visitor was calculated.Two samples from each pollinator species were collected and identified.
In the second step, we quantified the pollinator visitation rate per flower for the two sexes and tested whether one sex or color attracts pollinators more than the other, we manually surveyed the plants and noted the number of pollinators' visits for each flower.We surveyed 75 plants for all six colors with at least 11 plants per color and 32 males and 43 females that had one branch with one flower in the field and observed them for 10 min each.This survey was conducted for diurnal floral pollinators during the peak flowering season, April-May of 2021.We used a two-way ANOVA to test if pollinator visits were influenced by sex, color or interaction of both.We log transformed the visits data to achieve a normal distribution.The visits data then satisfied the assumptions of normality and homogeneity of variance.We controlled for the time of day by using time as a random factor in ANOVA.We also assessed the influence of flower color on pollinator visitation, categorizing it into brighter hues encompassing green, greenish yellow, and yellow, and darker shades, which included yellowish orange, orange, and red flower colors.
Finally, in order to corroborate that the potential pollinators exhibiting biased visits to these flowers were indeed responsible for pollen transportation, we employed the fluorescent dye powder.These fluorescent dyes act as pollen analogues, and this method is a good indicator of pollen dispersal (Fenster et al. 1996).We used three colors of dye: green, violet, and orange due to restricted availability.Subsequently, we organized the six color morphs into pairs based on their proximity along the color gradient, resulting in the creation of three primary flower color categories, grouping them as follows: green = green/greenish yellow; yellow = yellow/yellowish-orange; red = orange/red.It was ensured that significant differences in pollinator visits were not observed within these pairs.We selected male plants that are of these three colors in four separate locations along a 188 m transect.Since females do not have any pollen we did not include them in this experiment.The selected plants were marked with flags and GPS points (Supplementary Fig. 1).We then applied the color dye using a sterile brush on the anthers.The following colors were green, violet, and orange corresponding to green, yellow, and red flowers respectively.To avoid cross contamination, three individuals performed the dye deposition separately.The deposition was carried out on a clear sunny day early morning to avoid dye carryover by abiotic factors.We visited the study site that same day after sunset to observe the flowers in the area under UV light.We surveyed only the plants with flowers present within a 950 m 2 quadrant around the dyed flowers (Supplementary Fig. 1).We manually noted the presence or absence of fluorescent dye in 55 flowers of the neighboring plants around the four locations.Then we calculated the percentage of flowers that had the different colored dyes on them from the flower counts data.

Males flowers are bigger and brighter
Our study on sexual dimorphism in floral traits revealed three quantitative traits were significantly different between sexes with males always having a higher value for each: tepal length (P < 0.01), tepal width (P < 0.01), and display diameter (P < 0.01) (Fig. 3).In addition the color of the flowers also influenced the size traits such as diameter (P < 0.01), tepal length (P = 0.01), tepal width (P < 0.01) but we did not see a consistent pattern that explains brighter color flowers regardless of sex being bigger.For display diameter, yellow flowers regardless of sex are bigger than green (P < 0.01).For tepal length, yellowish orange flowers are bigger than green (P = 0.01).For tepal width, yellowish orange flowers are bigger than green (P = 0.01) and orange flowers (P < 0.01).
We found that flower color and sex are related (X 2 = 18.63;P = 0.03).For example, male flowers are usually associated with green and yellow shades while the females are mainly associated with orange and red shades (Table 1).In addition, the filament color (Fig. 4) differed between the six color morphs and had a correlation with brighter colors generally displaying brighter filament colors (green and pink) Fig. 3 Box plots showing Cylindropuntia wolfii males have a significantly larger display diameter (A), tepal width (B) and tepal length (C) than females ( P < 0.01) (Fig. 4).There was no significant difference between the number of flowers in males and females.Both sexes had a mean of about 26 flowers per plant.

Female flowers fluoresce more than males
A total of 113 flowers were imaged under RGB, UV, and GFP filter cubes.We sampled at least seven flowers per sex and color.The mean UV emission intensity per-unit area and mean GF emission intensity per-unit area were measured using the Nikon BR elements application.The intensity values were compared between the two sexes and among the six colors using a linear model in ANOVA and Tukey's post hoc test.The data fit the normality and homoscedasticity assumptions of ANOVA.Sex and color were fixed factors and individuals were random factors.We checked if the emission intensity of the individuals changed as a factor of sex or color or interaction between sex and color.The ANOVA analysis showed that there were no significant differences in the UV emission intensity between the sexes (P = 0.09) or color (P = 0.11).However, the GF emission intensity in the anthers differed significantly between the sexes.Anthers of females were shown to have a higher emission of GF compared to that of males (Fig. 5; P < 0.01).This could be due to the obstruction of fluorescence by pollen in male anthers.Further, the GF emission intensity of tepals did not vary significantly as a factor of sex or flower color.When the cryo-sectioned female anther sections were viewed under the GFP filter light, autofluorescence was observed in the areas of degraded tapetum (Fig. 6).

Males attract more potential pollinators
We observed that C. wolfii flowers attracted diverse insects such as beetles (5%), moths (5%), and potential pollinators like bees (90%).Two species that we collected were identified as Apis mellifera (European honey bee) and Diadasia spp.(native cactus bee).We identified the beetles as belonging to the Nitops genus and the moth as a member of the hummingbird moth taxon (Hemaris spp.).
Our ANOVA results on the potential pollinator visitation rate showed that sex and tepal color influenced the pollinator visitation rate.Males attract more potential pollinators than females (P = 0.02) (Fig. 7A).In terms of colors, we observed that brighter flower colors attracted more pollinators than darker flower colors (P = 0.02) (Fig. 7B) but there was no significant interaction between flower color and sex in influencing pollinator visits (P = 0.12).The time of the day did not have a significant effect (P = 1).
Our experiment using dye pigments to track pollen movement and validate pollinator preference showed that out of the total 55 flowers, the majority of them (39/55) did not have any dye.About 20% (11/55) had green dye on them, which corresponds to the green/greenish yellow male flowers.None of the flowers (0/55) had the violet dye which was applied to yellow/yellowish orange flowers.Five flowers (5/55) had the orange dye that corresponds to orange/ red flowers.

Discussion
Our study focused on measuring the dimorphic floral traits of a functionally dioecious native cactus, C. wolfii, which is a valuable model system to study because of the sexual dimorphism/dichromatism relation and the effects of varying floral traits on pollinator floral preferences within a single population.The flowers of C. wolfii carry vestigial organs i.e., non-functional anthers in female flowers and non-functional pistil in the male flowers (Ramadoss et al. 2022).This is considered as a 'deceit strategy' where the plants try to deceit the pollinators into visiting their non-rewarding flowers equally as rewarding flowers to ensure successful sexual reproduction (Dafni 1984).This strategy of developing both sexes whorls in unisexual flowers and later arresting the development of the opposite sex called as cryptic dioecy is observed in at least 78 species from 18 families (Mayer and Charlesworth 1991), commonly in Anacardiaceae (Zohary 1952) and Rosaceae (Cronk and Muller 2020).Despite this strategy, this rare species is limited in its sexual reproduction (Ramadoss et al. 2022) and studying its floral ecology might provide us with some valuable insights.
Our results on sexual dimorphism showed that male flowers were larger and brighter than female flowers.Large flower size is a trait that is often selected by pollinator-mediated selection as they attract more pollinators and facilitate  et al. 2008et al. , Delph et al. 1996)).In C. wolfii, we observed that the quantitative traits such as tepal length, tepal width, and tepal display diameter were significantly higher in male flowers than female flowers.Intersexual mimicry theory suggests that in rewardless females (i.e.females that lack pollen or nectar) qualitative traits such as scent composition do not differ significantly between the sexes to ensure unbiased pollination and successful reproduction (Ashman 2009).However, C. wolfii did not completely support this theory.With respect to the qualitative trait-color, females are usually found to be in orange-red shades while males tend to be mostly in brighter colors of green-yellow shades.However, there is no strict dimorphism cutoff in the flower color trait.For instance, males are generally more brightly colored, but they can also exhibit darker flower colors.We predict that males are biased towards brighter colors because it is potentially advantageous as pollinators, especially bees, tend to be attracted to brighter colors, such as yellow flowers, and the male reproductive success depends on pollen dispersal (Acharya et al. 2022).On the other hand, females are biased towards darker shades as brighter colors are more apparent and not only attract pollinators but also other florivores and herbivores which might affect the fruit production (McCall and Irwin 2006).Additionally, the color of the flower also seems to have an effect on the size of the flower.In general, the green flowers appear to exhibit reduced size dimensions in comparison to their yellow/yellowish orange counterparts.
We found an interesting relation between flower color and filament color in which lighter colored flowers usually have lighter colored filaments.A similar association was observed in Bixa orellana flowers where deeper colors had darker filaments and it was suggested that this color combination could act as visual cues to pollinators because with the background colors blending the anthers could stand out with better contrast (Joseph and Siril 2013).In fact, a study on bumble bees has shown that the contrast of flowers against their background is more critical for identifying their targets than color patterns within the flower (Whitney et al. 2013).Alternatively, within-flower contrast helps in guiding pollinators towards the nectar (Goldblatt et al. 1998;Hempel de Ibarra and Vorobyev 2009;Sletvold et al. 2016).Additionally, factors such as scent and nectar composition also play a major role in female attractivity (Bawa and Opler 1975;Ashman 2009).Previous studies in Mammillaria magnimamma (Callejas-Chavero et al. 2021) have shown that females have smaller nectar chambers and that in Opuntia quimilo (Diaz and Cocucci 2003) have lower nectar supply than hermaphrodites.Future studies on dioecious Cylindropuntia could focus on measuring these differences between the sexes and how they may influence pollinator attraction.
Moreover, it is critical to consider patterns that are visible by insects only, such as UV and fluorescence.Floral fluorescence is proposed to be a visual signal for insect pollinators (Gandia-Herrero et al. 2005).Several plant species have been reported to have auto blue or green fluorescence in their petals, stigma, ovaries and nectaries (Iriel and Lagorio 2010;Lagorio et al. 2015;Mori et al. 2018).Therefore, we examined the insect perspective by measuring emission intensity in the UV and GF spectra.Our results suggested that there was no significant difference in UV emission intensity between sexes and/or colors.However, we observed that GF emission intensity of anthers differed as a factor of sex.The anthers of females had a higher emission intensity than the Fig. 7 Box plots showing pollinator visitation rate with respect to sex (A) and flower color (B).In (A), males attract significantly more potential pollinators than females.In (B) Bright green-yellow flower colors (related to male sex) attracted significantly more potential pollinators than dark (related to female sex) flowers male anthers.Though fluorescence signals are proposed to be involved in communication of plant-pollinator systems, there are few studies that performed behavioral experiments to understand the perceiving capabilities of bees for fluorescence (Gandia-Herrero et al. 2005;Garcia-Plazaola et al. 2015).A recent study by Mori et al. (2018) showed that honeybees in the field were attracted to blue fluorescence of anthers and pollen.Thus, brighter anther fluorescence patterns found in females could be attractive to bees that possess green light sensitivity and attractivity to brighter targets.This could be one way the female flowers are compensating for their dull flower colors.
We propose that the mechanism by which anthers of female flowers have brighter fluorescence patterns is associated with the sterility cellular mechanism.Autofluorescence in pollen and anthers has been reported in some plants (Castro et al. 2010).Persistent autofluorescence in anthers of female flowers only has been reported in mutant soybeans Glycine max.It was suggested that due to early degeneration of tapetum in male sterile flowers, the autofluorescence pigment is not transferred to the microspores and ends up piling in the tapetum, which leads to persistent autofluorescence in male-sterile flowers only (Nakashima et al. 1984).Premature disintegration of the tapetum is a hallmark of male sterility in C. wolfii (Ramadoss et al. 2022).In the pre-anthesis stage of female anthers of C. wolfii, autofluorescence was detected inside the anther where the microspores deteriorated suggesting that C. wolfii follows a similar pattern as that of the mutant female soybeans which highlights the potential for fluorescence sexual dichromatism in other functionally dioecious species.Additionally, premature tapetal disintegration is not uncommon in dioecious Cactaceae Opuntia (Orozco-Arroyo et al. 2012;Flores-Rentería et al. 2013), Consolea (Strittmatter et al. 2002(Strittmatter et al. , 2006(Strittmatter et al. , 2008) ) and Echinocereus (Hernández-Cruz et al. 2018).Particularly, in O. stenopetala autofluorescence has been reported in male sterile flowers at later stages of development due to abnormal callose retention in the anthers (Flores-Renteria et al. 2013).These observations underscore the potential for fluorescent dichromatism in those species as well.Sexual dichromatism in fluorescent signals has only been recently observed in birds (Barreira et al. 2012) and studies about its role on animal evolution are still limited (Ancillotto et al. 2022).We report for the first time fluorescence sexual dichromatism in a plant species which has important implications for future research on the evolution of sex in plants.
Our results on pollinator visitation showed that male flowers attract more pollinators than females.However, discerning the relative influence of color or size on pollinator visitation rates is not possible as we collected these data independently.Our results showed that males had larger flowers than females.This could be evolved by growthsexual investment tradeoff where females generally allocate more resources towards reproduction (forming fruits and offspring) thereby affecting their vegetative growth, while males directly allocate their resources towards growth (Stearns 1992;Rabska et al. 2022).Several studies have found flower size to positively influence pollinator visitation (Bell 1985;Eckhart 1991;Dudash et al. 2011;Lazaro et al. 2013;Barbot et al. 2023) as larger flowers are expected to be more attractive.Alternatively, the pollinator visitation bias observed could be due to sexual dichromatism in our species.Our investigation revealed a statistically significant preference for brighter tepal colors in attracting a higher number of pollinators compared to darker hues.This observation aligns with the notion that bees have limited capacity to perceive the color red, but exhibit heightened sensitivity to the color green, as substantiated by prior studies (Kevan et al. 1996;Chittka and Waser 1997;Acharya et al. 2022).This was in agreement with our observations that about 67% of the flowers that received dye moved by pollinators was green suggesting that the green/greenish yellow flowers were visited more frequently than the other colors.Based on our empirical data, the enhanced pollen dispersal from green/ greenish yellow male flowers explains the higher prevalence of brighter coloration in male flowers as opposed to darker variants.This suggests that male flowers evolved to become more attractive through intrasexual competition that occurs as a consequence of sexual selection (Stearns 1992, Paterno et. al 2020;Barbot et al 2023).
Our experiment with dyes to track pollen movement found low pollinator visitation.About 71% of the flowers we surveyed did not have any dye on them which might be an indication of low pollinator density or low pollinator visits or inefficient movement of pollen to the C. wolfii flowers in general.Concordantly, we have previously observed that natural pollinations of C. wolfii resulted in a reduced average seed set (0.05) compared to that of hand pollination (0.25) (Ramadoss et al. 2022).Studies have theorized and empirically shown that while male attractiveness is optimal during high pollinator visits, this strategy could lead to inefficient sexual reproduction when the pollinator density is low due to the biased visit to the most attractive sex, which is usually the males (Vamosi and Otto 2002;Moquet et al. 2022).Experimental studies have shown that plants can abort fruits that have reduced seed number due to low pollen deposition to save their resources (Stephenson 1981).We acknowledge the presence of multiple variables beyond pollination that may exert influence on reduced seed production such as resource availability, predation, and environmental conditions, which have not been subjected to empirical investigation in our study.
Studying sexually dimorphic traits is essential for understanding plant-pollinator interactions, especially on species with restricted distributions like C. wolfii, which is limited with low seed production.Our results suggest that C. wolfii exhibits sexual dimorphism with male flowers being bigger than females.Additionally, they exhibited sexual dichromatism with males usually having brighter tepals and females having darker tepals.We also observed an unprecedented phenomenon of fluorescent sexual dichromatism, which had not been previously observed in plants.The mechanism for autofluorescence of sterile anthers might be prevalent in other functionally dioecious species and more studies need to be undertaken.Our pollinator survey showed that males attracted more pollinators than females which could be a consequence of the observed sexual dimorphism.This observation could potentially account for the reduced seed production observed in this endemic species, as when pollinator density is low female plants have less opportunity for ovules to be fertilized and therefore produce mature fruits.
Our research breaks new ground by unveiling sexual dichromatism in plants, including fluorescent dichromatism, previously observed solely in birds.Our findings reshape plant sexual system research.

Fig. 1
Fig. 1 Six color morphs are found in C. wolfii flowers-Green (A), Greenish yellow (B), Yellow (C), Yellowish orange (D), Orange (E) and Red (F).In addition, three different filament color morphs are present, for example (A) green filaments (C) pink filaments, and (D) red filaments

Fig. 2
Fig. 2 Human eye view (A) vs Bee's eye view (B) of a red C. wolfii flower

Fig. 5
Fig. 5 Boxplot showing GF emission intensity of anthers in females being higher (i.e. more fluorescent) than males

Table 1
Counts of sexes found in different tepal colors.Males are gradually decreasing and females are gradually increasing with darker shades of colors