Morph bias in inflorescences and individual plants reduces opportunities for geitonogamy in a monomorphic enantiostylous species

Bee pollination, Eastern Himalayas, Gesneriaceae, Reproductive strategies, Self-pollination, Stylar polymorphism, Quantum dots

In hermaphroditic flowers, the spatial arrangement of male and female reproductive structures can affect their reproductive success (Webb & Lloyd, 1986; Richards, 1997; Barrett et al., 2000; Barrett, 2002; Li et al., 2009; Barrett, 2010). Hermaphroditic plants with inter-mating .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 3 floral morphs, such as plants with stylar polymorphisms, are predicted to have evolved to counter selfing and to promote outcrossing (Barrett et al., 2000; Barrett & Harder, 2005; Pauw, 2005; Barrett, 2010; De Almeida & de Castro, 2019). In species with stylar polymorphisms, on a population scale, pollination and its outcome depend on the number of flowers of each floral morph that are displayed simultaneously within and between plants (Waites & Ågren, 2004; Brys & Jacquemyn, 2010). For instance, in stylar polymorphism like heterostyly, individual plants display only one type of floral morph. This facilitates pollen movement among individuals carrying dissimilar morphs, that is, inter-morph cross-pollination, thus ensuring higher outcrossing rates within a population (Keller et al., 2014). However, in enantiostyly, which represents a distinct type of stylar polymorphism characterised by left (L-morph) or right-handed (R-morph) flowers, the morphs can be present on the same individual or segregated between individuals (Table 1). Enantiostylous flowers are identified by left or right-handed orientation of the style (non-reciprocal enantiostyly) or both style and stamen (reciprocal enantiostyly). The two morphs may be present in two separate individuals, resulting in dimorphic enantiostyly (DE), or within the same individual, as observed in monomorphic enantiostyly (ME). A majority of the ME species are reported to be self-compatible (Table 1; Wang et al., 1995; Gao et al., 2006; Tang & Huang, 2005; Ren et al., 2013; Richman & Venable, 2018; Mora-Carrera et al., 2019; Paudel et al., 2024; Johnson et al., 2025) and thus, the presence of both the morphs within an individual can pose a risk of geitonogamous pollen transfer. Floral handedness in enantiostylous species has been shown to enhance the movement of pollen between flowers of different morphs, resulting in a disassortative pollen movement (i.e., L to R or R to L), thus reducing self-pollen transfer (geitonogamy) when compared with plants that lack enantiostyly (Table 1; Richards, 1997; Barrett et al., 2000; Jesson et al., 2003; Jesson & Barrett, 2005). Between DE and ME types of enantiostyly, geitonogamy is lowest in DE species, whereas ME species may exhibit intermediate levels of geitonogamy (Jesson & Barrett, 2005; Saltini et al., 2025). Thus, rates of geitonogamy may vary with the type of enantiostyly, where the main differentiating factor is the distribution of L and R morphs within an individual. Among ME species, a population may consist of individual plants with an equal distribution of L and R flowers (Wang et al., 1995; Gao et al., 2006; Martins, 2008; Ren et al., 2013; Richman & Venable, 2018; Mora-Carrera et al., 2019; Braga et al., 2022) or a skewed .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 4 proportion of morphs, where individual plants exhibit partial or complete bias towards either one of the morphs (Dulberger & Ornduff, 1980; Tang & Huang, 2005; De Almeida et al., 2018; Robertson et al., 2025). Since ME species bear both L and R morphs on the same individual plant, both the visitation sequence of pollinators within an individual as well as the relative abundance of L and R morphs, are predicted to contribute towards geitonogamous pollen transfer (Table 1; Jesson & Barrett, 2005; Tang & Huang, 2005; Mora-Carrera et al., 2019; Barrett & Fairnie, 2024). However, in ME individuals with an equal number of L and R morphs, that is, an isoplethic morph ratio, geitonogamous pollen transfer may be inevitable since both active morph-switches and random switches by pollinators will result in high geitonogamous pollen transfer. Thus, a reduced opportunity for geitonogamous pollen transfer may be achieved in an ME species if morph ratios are highly skewed within an individual. A highly skewed morph ratio within a plant can ensure that even under random pollinator movement within an individual, the disassortative pollen movement (L to R and R to L) is mostly between individuals. Enantiostylous forms and their morph ratios Reproductive compatibility Expected rate of successful self- and cross-pollination events given the type of enantiostyly and reproductive compatibility. Autogamy Geitonogamy Allogamy Dimorphic enantiostyly (Plants have only one type of morph per individual) Self-compatible (Johnson et al., 2023; Jesson & Barrett, 2002) Exiguous Exiguous High Self-incompatible (Minnaar & Anderson, 2021) Absent Absent High Monomorphic enantiostyly (Plants have isoplethic morph ratio) Self-compatible (Wang et al., 1995; Gao et al., 2006; Ren et al., 2013; Richman & Venable., 2018; Mora-Carrera et al., 2019) Exiguous Intermediatea to Higha, b Intermediatea Self-incompatible (Morais et al., 2020) Absent Absent Intermediatea .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 5 Monomorphic enantiostyly (Plants have a biased proportion of morphs) Self-compatible (Tang & Huang, 2005) Exiguous Exiguousb to lowb Intermediatea to High Self-incompatible (Dulberger & Ornduff, 1980; Robertson et al., 2025) Absent Absent Intermediatea to High Non-polymorphic hermaphroditic species (Included for comparison with enantiostylous floral forms) Self-compatible (Brys et al., 2013) Intermediate to High Intermediate to High Lowa Self-incompatible (Li et al., 2013) Absent Absent Intermediatea to High Table 1 The potential for successful self-pollination (autogamy and geitonogamy) and cross-pollination (allogamy) events in different types of enantiostylylous plants. Floral arrangement of L morph (L), R morph (R), and hermaphroditic flowers (F) within two inflorescences (1 and 2) and within a plant are illustrated. The pollination evaluation is based on the morph ratios characteristically known for the specific type of enantiostyly and their known reproductive compatibility (references included). The superscripts refer to pollination outcomes under two specific conditions: a–when within-plant pollen loss is expected, b–when floral arrangement can affect pollen movement within an inflorescence on a plant. To the best of our knowledge, the effect of the relative abundance of morphs within the plant on opportunities for geitonogamy in natural populations of ME plants remains to be tested. Our present research aims to fill this void by investigating the ME species Didymocarpus podocarpus (Gesneriaceae), an understorey, perennial, mostly lithophytic herb restricted to the eastern Himalayas. We carried out in situ studies on two populations of D. podocarpus and tested the hypothesis that geitonogamous events will be reduced when plants have a biased morph ratio, especially due to disassortative pollen movement between morphs. Though ME has been reported from multiple genera within Gesneriaceae (Harrison et al., 1999; Gao et al., 2006; Ling et al., 2020; Prasanna, 2023), enantiostyly remains an understudied topic. The study was conducted across two populations of D. podocarpus to ensure that the morph ratio is not a population-specific character. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 6 The aims of our investigation were as follows: (1) to quantify monomorphic enantiostyly by documenting morph ratios in natural populations across multiple years; (2) to perform hand-pollination experiments to establish autonomous self-pollination, and to check intra- and inter-morph compatibility; (3) to evaluate pollen transfer patterns by the pollinators in the two morphs, and (4) to finally integrate morph ratios, reproductive compatibilities of the two morphs, and pollinator visitation patterns to infer the potential of geitonogamy within a plant. Based on these objectives, we address the following key questions: (i) Do the morph ratios within a plant and a population deviate from an isoplethic ratio? (ii) Do self and cross treatments yield similar reproductive success (fruit set and seed set) in the two morphs? (iii) Do morph ratio and pollinator visitation patterns within an individual affect opportunities for geitonogamous pollen transfer?

Study species and study sites Didymocarpus podocarpus (Gesneriaceae) is a perennial herb distributed in the Himalayan foothills of Northeast India, Nepal, Bhutan, and China. It is mostly lithophytic (occasionally terrestrial), found on moss-covered rocks. The species flowers between late July and early September, and the flowering within an individual plant lasts for a maximum of 15–20 days. The plants produce 1–9 cymose inflorescences, each bearing ≤ 10 open flowers (mean ± SE number of open flowers is 6.28 ± 0.41). The species exhibits monomorphic, reciprocal enantiostyly with distinct L and R morphs (Fig. 1; Fig. S1). The flowers are tubular, last for five days, and have a nectary at the base of the tube. Stigma receptivity did not differ between the two morphs across their floral longevity, based on the hydrogen peroxide test (Kearns & Inouye, 1993). All experiments were conducted in two natural populations of D. podocarpus within India, the Takdah Reserve Forest, West Bengal (27°01’58” N, 88°20’23” E, elevation 1812m), and Gangtok, Sikkim (27°20’30” N, 88°37’16” E, elevation 2001m), during August in the years 2022, 2024, and 2025. We studied a total of 153 plants for the estimation of morph ratio (section 2.2), out of which 72 were used in quantifying pollinator visitation sequence within individual plants (section 2.6). Out of 153, only 109 plants were available to measure fruit set. Since experimentation on this species in shade house conditions is challenging due to both collection .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 7 restrictions from the wild and survivability rates in the shade house, all studies were carried out in the wild populations. The primary pollinator (98.8% visits) of D. podocarpus was identified as Bombus breviceps (Apidae; Williams, 2022), and throughout this study, the term ‘Bombus bee’ refers only to this species. Fig. 1 Floral and fruit morphology of enantiostylous D. podocarpus. (a) From left to right of the image - L morph and R morph. Arrows point to the reciprocal placement of anther (A) and stigma (S) on the L and R morphs. (b) Floral dissection (from left to right) - adaxial surface of calyx, abaxial surface of calyx, abaxial surface of corolla tube, abaxial surface of corolla tube showing fusion of the anther to the corolla tube, and pistil. (c) mature fruit, and microscopic image of seeds. Estimation of morph ratios (MR) to quantify L bias and R bias In ME, since both morphs can be present on the same inflorescence, an inflorescence can show L bias, R bias, or an isoplethic ratio. To estimate morph ratio within an inflorescence, within an individual, and within a population, we sampled 30–33 individual plants at random, across three years in Takdah, in 2022 (T22), 2024 (T24), and 2025 (T25), and two years in Gangtok in 2024 (G24) and 2025 (G25). For each individual, we counted the number of open L and R morphs on all the inflorescences. Next, to quantify the relative abundance of the morphs within inflorescences and individual plants, we calculated the ‘morph ratio’ (MR) of inflorescences and individual plants as: MR = the total number of flowers of L morph ÷ the total number of flowers. Thus, MR represents the relative bias of inflorescences or plants with reference to only the L morphs and is a metric that retains morph identity (L bias or R bias). That is, when all flowers in an inflorescence or plant are L morphs, the MR value = 1, and can be identified as L bias. Later, we introduce another metric, morph bias (MB), which designates the overall morph bias of an .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 8 inflorescence or plant, without retaining morph identity. To investigate yearly variation in the MR of plants, we fitted a Generalised Linear Mixed Model (GLMM) with a binomial distribution using the lme4 package (Bates et al., 2015) in R (version 4.4.0; R Core Team, 2024). The year of sampling was treated as a fixed effect, and the population as a random factor. A chi-square goodness-of-fit test was used to evaluate whether the total number of L and R morph flowers within populations across years deviated from a 1:1 expectation. To investigate whether individual plants exhibit L or R biased inflorescences, based on the MR values, we categorised each inflorescence within a plant into three groups: inflorescences with (i) complete bias (MR values of 0 for R bias and 1 for L bias); (ii) partial bias (0.01–0.44 for partial R bias and 0.56–0.99 for partial L bias), and (iii) isoplethic ratio (0.45–0.55). The number of inflorescences in each category of morph ratio within an individual plant was compared using a GLMM with a negative binomial distribution, where morph ratio was treated as a fixed effect, and plant identity and population were treated as random effects. The significance of the individual fixed effects was tested using Wald’s χ2 test (type III) using the car package (Fox & Weisberg, 2018), and multiple contrast analysis was carried out using the emmeans package (Lenth, 2025) with Sequential Bonferroni Correction. All statistical analyses were carried out in R (version 4.4.0; R core team, 2024). Reproductive compatibility between morphs To quantify the self and cross-compatibility between the two morphs, we carried out the following hand-pollination treatments: (i) inter-morph outcrossing (n = 40 flowers from 29 plants), (ii) intra-morph outcrossing (n = 41 flowers from 30 plants), (iii) inter-morph geitonogamy (n = 42 flowers from 25 plants), (iv) intra-morph geitonogamy (n = 40 flowers from 25 plants), and (v) unmanipulated, bagged flowers to check for autonomous self-pollination (n = 52 flowers from 34 plants). Flowers were bagged pre- and post-treatment to avoid pollinator visitation, emasculated before anther dehiscence, and hand-pollination was carried out within the first 18 hours post-anthesis. Each treatment was carried out on flowers of different plants, with no plant receiving more than one treatment. A maximum of two flowers per plant was used for any given treatment. For outcrossing, mixed pollen from multiple plants at least 3 m apart was used. Mature fruits, except the ones herbivored, were collected after three weeks. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 9 The seeds of D. podocarpus are tiny - c. 0.3 × 0.1 mm and are reported to be ‘numerous’. Therefore, seed counting for mature fruits was carried out using ImageJ software (Rueden et al., 2017). All seeds were carefully removed from the capsule and placed in a white weighing boat. High-resolution images of these weighing boats were taken using a digital camera (Nikon D3500), and the total number of seeds per capsule was quantified using the ‘analyse particle’ function in the ImageJ software following the method mentioned in Ochogavía (2022). This was carried out for a total of 84 fruits: (i) inter-morph outcrossing (n = 25), (ii) intra-morph outcrossing (n = 21), (iii) inter-morph geitonogamy (n = 23), (iv) intra-morph geitonogamy (n = 15). We fitted separate GLMM models to check whether the fruiting success rate and seed count differed between (i) different hand pollination treatments, (ii) between L and R morphs, (iii) between outcrossing and selfing treatments, and (iv) between inter-morph and intra-morph treatments. Fruiting success was analysed using a binomial distribution, whereas seed count was analysed using a Poisson distribution. Seed count between L and R morphs was compared using a negative binomial distribution due to overdispersion and poor fit of the Poisson distribution. Plant identity was used as a random effect in all models. The significance of the individual fixed effect and multiple contrast analysis was carried out using the method mentioned under the section ‘estimation of morph ratio, and to quantify L and R bias’. Furthermore, to compare the pollen production between the two morphs, we counted the pollen grains from mature buds of both morphs. Paired mature buds (one L and one R morph) were collected from ten individual plants (five each from Takdah and Gangtok). Undehisced anthers were stored in microcentrifuge tubes containing 1 ml of 70% ethanol. The anthers were macerated, and pollen grains were counted using a Neubauer chamber following the method mentioned in Dafni (1992) and Kearns & Inouye (1993). The difference in pollen count was compared using a paired t-test. Pollen placement on the pollinator To examine if the pollen from the two morphs is differentially placed on the Bombus bees, we carried out controlled manipulative experiments using stained pollen grains and euthanised bees. The pollen grains of the L and R flowers were labelled using heavy-metal-free CuInSexS2-x/ZnS .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 10 (core/shell) quantum dots with Zinc Oleate ligands dissolved in hexane (5 mg/mL), following the

mentioned in Minnaar and Anderson (2019). To distinguish the pollen grains of L and R morphs, we labelled them using 2 μL of yellow (580 nm) and red (650 nm) quantum dots, respectively. Experiments were carried out on 10 freshly opened flowers of each morph, and the individual euthanised Bombus bees (n = 10) were manually inserted into a labelled L morph flower, followed by an R morph using sterile forceps, mimicking their natural probing behaviour (Fig. 5a). To quantify the number of labelled pollen grains deposited on the pollinator, we categorised the body of the bees into four parts (Fig. 5a): (i) left proximal region (left side of head and thorax), (ii) left distal region (left side of the abdomen), (iii) right proximal region (right side of the head and thorax), and (iv) right distal region (right side of the abdomen). The number of fluorescent pollen grains was counted in these four regions using a 60X hand microscope with an inbuilt UV light source. We tested for statistical differences in pollen placement on the pollinator using GLMM with a negative binomial distribution, where the interaction of morph type and body parts of the bee was the fixed effect, and bee identity was the random effect. The significance of the fixed effects and multiple contrast analysis was carried out as mentioned in the section ‘Estimation of morph ratios (MR) to quantify L bias and R bias’. Pollen movement between morphs using quantum dots To confirm if pollen transfer to stigmas is highest between the two morphs (L to R and R to L) in natural conditions, we tracked the pollen using quantum dots. We selected 12 plots (eight from Takdah and four from Gangtok) of 3 × 2 m with a minimum distance of 1.5 km between plots. To mimic the natural ratio of morphs within a population in our experimental plots, we maintained the morph ratio at an isoplethic ratio by manually removing flowers when necessary. Since D. podocarpus has a restricted distribution in its native habitat and the available flowering density allowed only a single experimental setup, in the interest of the species and its conservation, we did not set up alternate experiments, such as with variable morph ratios. All the inflorescences within the plot were bagged a day before the quantum dot application to avoid pollen loss. To ensure recording of pollen movement in natural conditions, pollen grains of approximately one-fourth of the total number of L and R flowers were labelled between 06:00 and 07:00 h using yellow and red colored quantum dots. After 8 hours of natural pollination, all flowers within the plot were collected between 15:00 and 16:00 h and immediately screened for the .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 11 number of fluorescent pollen grains deposited on the stigmas using a 60X hand microscope with an inbuilt UV light source. We tested if inter-morph pollen transfers were higher than intra-morph pollen transfer by fitting a negative binomial GLMM, with the interaction between morph type of the donor and recipient as fixed effects, and plant and plot identity as random effects. The significance of the individual fixed effect was tested, and pairwise comparisons were carried out using the method mentioned in the section ‘Estimation of morph ratios (MR) to quantify L bias and R bias’. Quantifying pollinator constancy towards a floral morph Pollinator preference for a morph and its subsequent foraging behaviour can influence pollen movement between morphs. To explore if individual pollinators show a preference for one morph while foraging across multiple plants, which may result in biased morph switches, we quantified the floral constancy of Bombus bees towards a specific morph in a single foraging bout. We tracked individual Bombus bees (n = 50) for up to 10–30 consecutive floral visits, and recorded their total intra-morph switches (indicating high pollinator constancy towards a morph) and inter-morph switches (indicating low pollinator constancy towards a morph) during a single foraging bout. We used a paired-sample t-test to assess whether the frequency of intra-morph switches significantly differed from that of inter-morph switches. Effect of morph bias on pollinator movement within the plant Complementary to the above experiment, we next tested whether pollinator switches within a plant may be governed by the plant’s morph bias. For this, we first measured pollinator switches within a plant in 72 of the 153 experimental plants (40 in Takdah and 32 in Gangtok). Each plant was observed for two 25-minute periods from 07:00 to 15:00 h during the peak pollinator activity time, resulting in a total of 60 hours of observation (2 slots × 25 min × 72 plants). We recorded 76 Bombus bees and their within-plant inter-morph switches (R to L and L to R) and intra-morph switches (R to R and L to L) during one visitation bout. We next assigned a morph bias value for each experimental plant, where ‘morph bias’ (MB) was computed as: MB = 1- (number of flowers of the rare morph ÷ number of flowers of the abundant morph). Here, MB = 0 would represent individuals with equal proportions of both morphs, while MB =1 would indicate a complete skewness towards one morph. This metric identifies only if a plant deviates .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 12 from an isoplethic morph ratio and does not identify the type of bias, that is, L bias or R bias. In ME, since both morphs are present on a plant, and all inter-morph switches (L to R and R to L) will result in self-pollen transfer (geitonogamy), the direction of the switch is not critical. To examine the effect of morph bias on the frequency of inter-morph switches, we fitted a GLMM with a binomial distribution, with MB as a fixed effect and plant identity and population as random variables. Finally, we quantified the effect of MB on the pollination success of plants by counting the total number of fruits, normalised by the total number of flowers, per plant, in a total of 109 (out of 153) naturally pollinated plants. To test the effect of morph bias on the fruit set, we used a binomial GLMM with morph bias as a fixed effect and population as a random variable.

Morph ratio (MR) within populations We calculated the MR of a total of 226 inflorescences from 153 plants, from two populations (Table S1). Of the total number of inflorescences, 85.39% and 79.46% exhibited a biased MR value (MR value outside the isoplethic range of 0.45–0.55) in the Takdah and Gangtok population, respectively (Fig. 2a). Similarly, of the 153 individuals, 74.19% and 75% of individual plants exhibited a biased MR value in the Takdah and Gangtok populations, respectively (Fig. 2b). GLMM analysis revealed that the MR of plants across the years and for the two populations was not significantly different (χ² = 5.39, df = 4, p = 0.24; Fig. S2), and therefore, in all subsequent analyses, population identity has not been retained. Among all the inflorescences examined from the two populations (226 inflorescences), 49.11% exhibited a complete bias towards a single morph, whereas 34.32% showed a partial morph bias, and only 16.57% of inflorescences were noted to display isoplethic morph ratio. The number of inflorescences across the three categories of morph ratios within individual plants differed significantly (χ² = 46.6, df = 2, p < 0.001). Individual plants exhibited a significantly higher number of inflorescences with complete bias compared to inflorescences with partial bias (Estimate = 0.36, z = 2.81, p < 0.005) and isoplethic ratio (Estimate = 1.09, z = 6.8, p < 0.001). However, within a population, the total number of L and R morph flowers across the years did .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 13 not significantly deviate from an isoplethic ratio (T22: χ² = 3.47, p = 0.06; T24: χ² = 1.38, p = 0.24; T25: χ² = 0.04, p = 0.85; G24: χ² = 0, p = 1; G25: χ² = 0.59, p = 0.44; Table. 2). Fig. 2 Observed morph ratio distributions in (a) inflorescences and (b) individual plants within two populations of D. podocarpus. Zero value on the y-axis represents R-bias, and one represents L-bias, while 0.5 represents anabsolute isoplethic ratio. The isoplethic region is shown in grey shade (0.45–0.55). The red open circle represents the average of all the morph ratios distributed within the population. When the average value (red open circle) is at 0.5, it suggests that the population displays an equal number of L and R morphs (Gangtok population), while a slight shift towards the lower part of the isoplethic region suggests a slightly higher number of R morphs in the population (Takdah population). Reproductive compatibilities of morphs We carried out a total of 215 hand pollinations, of which 103 sired fruits. The per cent fruiting success of each pollination treatment was recorded as follows: (i) inter-morph outcrossing: 80%, (ii) intra-morph outcrossing: 60.9%, (iii) inter-morph geitonogamy: 66.6%, (iv) intra-morph geitonogamy: 45%, and (v) autonomous self-pollination with zero success rate (Fig. 3a). We observed a significant effect of the type of pollination treatment (excluding the autonomous self-pollination) on the fruiting success (χ² = 8.33, df = 3, p < 0.05), but the pairwise comparisons with sequential Bonferroni correction showed significant differences in fruit set only between inter-morph outcrossing and intra-morph geitonogamy (Table S2). The two morphs .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 14 did not differ in their fruiting success (Estimate = -0.29, z = -0.8, p = 0.42). A marginally significant difference in fruit set was observed when all outcrossing treatments were compared with all selfing treatments, which included both intra and inter-morph crosses (Estimate = -0.75, z = -1.8, p = 0.07). However, the fruit set of all inter-morph pollination treatments was significantly higher than all intra-morph pollination treatments (Estimate = -0.99, z = -2.44, p = 0.01; Fig. 3a), irrespective of whether these crosses were carried out within a plant (selfing) or between plants (outcrossing). We did not observe a significant effect of the type of pollination treatment on seed count (χ² = 3.15, df = 3, p = 0.37; Fig. 3b), and it did not differ between the two morphs (Estimate = -0.03, z = -0.19, p = 0.85). No significant difference was observed between all types of outcrossing and selfing treatments (Estimate = 0.09, z = 0.37, p = 0.71). However, consistent with the fruiting success, all inter-morph treatments produced significantly higher seeds compared to intra-morph treatments (Estimate = -0.5, z = -2.11, p < 0.05; Fig. 3b). Intra-morph treatments (both within and between plants) exhibited a 28.33% reduction in fruiting success and 23.3% reduction in mean seed count per fruit compared to that of inter-morph treatments (both within and between plants). Finally, pollen count was also noted to be different between the morphs, where flowers of the R morph show significantly higher pollen count when compared to the L morph (t = -3.18, df = 9, p-value = 0.su01; Fig. 4a). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 15 Fig. 3 Results of fruiting success and seed count from four pollination treatments: inter-morph outcrossing (inter-out), inter-morph geitonogamy (inter-geit), intra-morph outcrossing (intra-out), and intra-morph geitonogamy (intra-geit). (a) Fruiting success rate showing significant effect between inter-morph treatments and intra-morph treatments, and (b) mean number of seeds per fruit (± SE) showing significant effect between inter-morph treatments and intra-morph treatments. * indicates a statistical significance at p < 0.05. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 16 Fig. 4 Comparative pollen count in the two morphs and on the pollinator. (a) Comparison of pollen count between L and R morph flowers (mean ± SE). * indicates a statistical significance at p < 0.05. (b) Asymmetrical pollen deposition was noted on a B. breviceps after natural visits to approximately 30 flowers on different D. podocarpus plants. Arrows point to the larger pollen patch on the bee’s right side of the thorax, the contact area of the R morph (R), while the smaller pollen patch on the bee’s left side represents the contact area of the L morph (L). Pollen placement on the pollinator All B. breviceps individuals (n=50) observed during their natural visitation to D. podocarpus probed the flowers using their proboscis by positioning their proximal part (head and thorax) at the rim of the flower. Since the anther and stigma are situated at the rim of the floral opening, the anthers of the L and R morphs were observed to come in contact with the left and right proximal regions of the Bombus bees, respectively (Fig. 4b). This observation was further corroborated by

from the pollen placement experiment using euthanized bees and pollen stained using quantum dots which show that the left and right proximal regions of the bees received pollen from the two morphs differentially (χ² = 75.65, df = 3, p < 0.001; Fig. 5b). The following pair-wise comparison shows that the left proximal region of the Bombus bee received a significantly higher number of L morph pollen grains than the R morph (Estimate = 2.6, z = 4.89, p < 0.001), and the right proximal region received more R morph pollen grains than the L morph .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 17 (Estimate = -3.54, z = -6.64, p < 0.0001; Fig. 5b). We further found that the R morph deposited significantly higher labelled pollen grains on the pollinator compared to the L morph (χ² = 6.68, df = 1, p < 0.01). Fig. 5 Pollen deposition across different parts of B. breviceps. (a) The natural probing position exhibited by B breviceps and the four regions in which pollen deposition was quantified: left proximal (L-Pro), left distal (L-Dist), right proximal (R-Pro), and right distal (R-Dist). (b) Mean (± SE) number of pollen grains labelled with quantum dots recovered from the euthanised B. breviceps after manual probing. Asterisks indicate statistical significance: ***, p ≤ 0·001; *, p ≤ 0·05; ns, non-significant. Pollen movement using quantum dots All 12 plots in which pollen tracking experiments were carried out consisted of a total of 582 flowers, in which we found movement of labelled pollen grains in stigmas of 217 flowers (37.28%; Table S3). Inter-morph pollen transfer identified on the stigmas was significantly higher than the intra-morph pollen transfer (χ² = 126.3, df = 1, p L × L, donor × recipient; Estimate = -2.35, z = -10.86, p < 0.0001) as well as .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 18 when the R morph was the recipient flower (that is, L × R > R × R; Estimate = 1.01, z = 4.98, p < 0.0001). However, irrespective of intra- or inter-morph comparisons, stigmas of the L morph received significantly higher labelled pollen than the R morph (χ² = 59.12, df = 1, p < 0.0001; Fig. 6). Fig. 6 Pollen deposition on the stigmas of the two morphs in Takdah (T) and Gangtok (G) population. (a) Mean (± SE) number of inter-morph (black) and intra-morph (grey) quantum dot-labelled pollen grains deposited on the stigma of L and R morph recipients. *** indicate statistical significance at p ≤ 0·001. (b) A schematic representation of the percentage of pollen movement recorded within morphs and between morphs of D. podocarpus. Illustrations by Abhishek Thakur. Pollinator constancy and pollinator movement within a plant Paired t-test results did not show any significant difference between the number of intra-morph switches (pollinator constancy on a morph) and inter-morph switches (absence of pollinator constancy) by Bombus bees (n = 50) during their single foraging bout (t = 0.53, df = 49, p = 0.6). We noted a significant negative correlation between the frequency of inter-morph switches and MB (Estimate = -2.54, z = -3.69, p < 0.001; Fig. 7a). Specifically, pollinators exhibited higher inter-morph switches when the morphs were in equal proportion (isoplethic), and the frequency of these switches decreased as the morph bias approached one (complete bias, Fig. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 19 7a). Finally, GLMM analysis also identified a significant positive association between the MB and the fruit set (Estimate = 0.35, z = 3.44, p < 0.001; Fig. 7b). That is, individual plants with a biased morph ratio exhibited higher fruit sets. Fig. 7 Relationship between morph bias (MB) and pollinator switches, and MB and fruiting success. (a) Correlation between the morph bias of individual plants and the frequency of inter-morph switches by the pollinator. (b) Correlation between the morph bias of individual plants and the natural fruit set. The values range from 0 (isoplethic), which represents an equal number of both morphs, to 1, which represents complete skewness towards either one of the morphs. The grey shaded area along the x-axis represents plants with isoplethic morph ratio. Statistical significance is indicated in the top right corner.

The relative abundance of morphs within an individual plant and its effect on opportunities for selfing (geitonogamy) remain one of the least explored aspects of monomorphic enantiostylous plants. The key findings from our study are that while individuals of the monomorphic enantiostylous D. podocarpus produced inflorescences with a biased morph ratio, the overall number of morphs within the population met the isoplethic ratio. Further, we noted inter-morph pollen transfer to be higher than intra-morph pollen transfer and hand-pollination treatments between morphs yielded higher fruit set and seed set than treatments carried out within morphs. Therefore, we propose that the predominant disassortative pollen movement and reduced inter-morph switches by the pollinators within individuals with higher morph bias may play a critical role in reducing self-pollination (geitonogamy) and increasing cross-pollination. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 20 Individuals with biased morph ratios show reduced geitonogamous events The frequent morph bias observed within inflorescences and the majority of individuals (Table S1) in the two populations of D. podocarpus across multiple years suggests ecological factors that favour a biased morph ratio and a strong selective advantage for the underlying developmental process. Similar morph bias has also been reported from other ME species (Dulberger & Ornduff, 1980; Tang & Huang, 2005; De Almeida et al., 2018; Robertson et al., 2025). Additionally, in D. podocarpus, we noted that the morph ratio of individual plants mirrored the most common morph ratio noted among all of its inflorescences, suggesting that individual plants and their inflorescences tend to exhibit a bias towards the same morph (Fig. S3). In ME species, a morph bias within inflorescences and within individual plants can reduce the frequency of geitonogamous selfing events. This is because when there are multiple flowers on an individual plant, interactions specifically between flowers within an inflorescence can affect their reproductive outcome since pollinators tend to visit flowers within the same inflorescence, causing geitonogamy (Harder & Barrett, 1995; Harder et al., 2000; Harder et al., 2004). Therefore, any inflorescence trait that might influence mating outcomes, such as a biased morph ratio, is likely under a strong selective pressure (Harder et al., 2000), as noted in the majority of individuals of the two D. podocarpus populations. Enantiostyly in D. podocarpus is unique because it differs from expected floral traits in enantiostylous flowers, such as the presence of heteranthery, the lack of nectary, and an open floral design (outward facing; Jesson & Barrett, 2003). Instead, D. podocarpus lacks heteranthery, has nectaries, and exhibits tubular flowers where the stigma and anther are exerted at the rim of the floral tube (Fig. 1). As expected from the floral morphology, and consistent with other studies, we found that inter-morph pollen transfers (L to R and R to L) were higher than intra-morph (L to L and R to R) pollen transfers, confirming disassortative pollen movement within D. podocarpus (Fig. 6; Jesson & Barrett, 2005; De Almeida et al., 2013; Minnaar & Anderson, 2021). Although the inter-morph pollen transfer in ME species can result in geitonogamous pollination, the extent of geitonogamy will be low in ME species compared to non-enantiostylous species, where all flowers within a plant are potential mates (Jesson & Barrett, 2005; Saltini et al., 2025). Studies on heterostylous species have shown that, apart from the arrangement of sex organs within flowers, the floral display of individual plants and the .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 21 morph composition of neighbouring plants affect the degree of disassortative pollination and quality of pollen deposited on the stigma (Waites & Ågren, 2004; Brys & Jacquemyn, 2010). While these factors are also true for ME species, the morph ratio within individual plants adds an additional layer of complexity.

from our study provide empirical evidence of the theoretical prediction that the extent of geitonogamy that may occur in ME species depends on a) the relative abundance of L and R morph flowers within an individual plant, and b) the visitation sequence by the pollinator (Jesson & Barrett, 2005; Saltini et al., 2025). We found that a biased morph ratio within individual plants limited the occurrence of geitonogamous events, and thus may influence both the male and female fitnesses by limiting the effect of both pollen and ovule discounting. The potential outcome of reduced selfing was also reflected in the fruit set, where the fruit set increased with an increase in the morph bias. We expected that the predominant pollen movement between flowers of different morphs might mandate an equal proportion of both morphs in a population. Our results validate this because, despite the biased morph ratio in the inflorescence and in the individual plants, the total number of L and R morph flowers within the two populations across multiple years did not significantly vary from the isoplethic ratio. This result is also consistent with the majority of previous studies on ME species (Fenster, 1995; Ren et al., 2013; Morais et al., 2020; Contreras-Varela et al., 2023). This suggests that, while individuals exhibit a biased morph ratio, the isoplethic ratio at a population level is likely driven by negative frequency-dependent selection since any deviation from this ratio can lead to lower pollination success (Jesson & Barrett, 2002). Flowers of the same morph show reduced compatibility Studies examining reproductive compatibilities among morphs have reported that stylar polymorphisms, like distyly and tristyly, are linked to heteromorphic incompatibility associated with variations in pollen size and number, size of the stigmatic papillae, and stigma shape between morphs, potentially leading to intra-morph pollen recognition at the stigma, style, or ovary, thus reducing fruiting success and seed count (Richards, 1977; Barrett & Cruzan, 1994; Massinga et al., 2005; Costa et al., 2017). We noted moderate intra-morph incompatibility in the .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 22 form of lower fruit set and lower seed count per fruit in D. podocarpus, irrespective of whether the intra-morph treatment was within or between plants. These observations strongly suggest that ‘morph type’ is being recognised regardless of whether the pollen is from within the individual or a different individual. Intra-morph incompatibility has been previously reported in the dimorphic enantiostylous species Wachendorfia paniculata (Ornduff & Dulberger, 1978), although a subsequent study failed to corroborate this (Jesson & Barrett, 2002). We propose that the moderate incompatibility between flowers of the same morphs in D. podocarpus can act as an additional mechanism to ensure disassortative pollen movement (Costa et al., 2017), where the fruiting success of pollen transfer between flowers of the same morph is significantly reduced. Finally, in our experiments to track pollen movement, despite an equal number of L and R morphs in our experimental plots, the stigmas of L morphs received significantly higher numbers of labelled inter-morph pollen compared to stigmas of R morphs (Fig. 6). Similarly, R morphs deposited significantly higher pollen on the pollinator compared to the L morphs (Fig. 4b; Fig. 5b). In concordance with this, the comparison of pollen count between the two morphs show that the R morph indeed produces significantly higher pollen compared to the L morph in D. podocarpus (Fig. 4a). This difference was observed only when the morph was a pollen donor, and the fruiting success of the two morphs as a female parent was not different in our hand pollination experiments. These results suggest that the two morphs may have dissimilar sexual roles due to their differential contribution to the population’s overall pollen pool. The R morph in D. podocarpus appears to display higher male function through higher pollen donation, whereas the L morph may display higher female function by receiving a higher number of pollen grains. These patterns loosely share similarities with a few andromonoecious systems, where male flowers are reported to produce more pollen compared to that of hermaphrodite flowers (Narbona et al., 2005; Dai & Galloway, 2012). While male flowers of andromonoecious plants are solely males, in the monomorphic enantiostylous D. podocarpus R morphs may emphasise more on male function by producing and exporting more pollen without being exclusively male.

.CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 23 We demonstrate that the morph ratio of inflorescences and in individual plants has a significant influence on the pollination success in monomorphic enantiostylous species. A biased morph ratio within inflorescences and individuals, in combination with disassortative pollen movement and intra-morph incompatibility, acts as a multi-pronged strategy to reduce opportunities for selfing (geitonogamy). Further, the presence of intra-morph incompatibility and morph-specific pollen count differences together suggests that there may be differences in sexual roles between the two morphs that remain to be explored. Monomorphic enantiostylous plants represent one of the unique innovations in labile reproductive strategies among all organisms, where they do not represent true and fixed genetic polymorphisms as known from other stylar polymorphisms. Instead, they represent morphs whose ratios are probably maintained within a population, although developmentally regulated within a plant. This opens unique insights into how morphs are distributed within a plant and population in order to optimise high outcrossing rates, reproductive success, and genetic admixtures.

We express our gratitude to the forest departments of Sikkim and West Bengal for the necessary permits. We would also like to thank the local community, especially Mr Provesh Gurung and his family, for their hospitality and support. We would like to thank Sanika Goray, Nevil S, Manila Chingtham, Rohan Dandavate, and Jyotil Dave for their assistance during the fieldwork and Kirti and Arti for helping with the seed count. We acknowledge Dr Paul Williams, Natural History Museum, London, and Dr A Rameshkumar, Zoological Survey of India, for helping us identify the Bombus bee specimens. We thank Dr N. S. Prasanna for giving us the details of the study sites. We thank Abhishek Thakur for the illustrations. We would like to acknowledge the Science and Engineering Research Board (SERB ECR/2017/001073; POWER Grant SPG/2021/000793) and DBT-NER (BT/PR24525/NER/95/754/2017) and Ministry of Human Resource Development (Ministry of Education), for the research fund awarded to VG, IISER Bhopal for infrastructural support, the University Grant Commission (UGC; 191620076099) for a fellowship to RSB and the American Society of Plant Taxonomists (ASPT) Graduate Student Research Grant, International Association for Plant Taxonomist (IAPT) Graduate Student Research Grant and Society for Tropical Ecology Student Research Grant awarded to RSB for fieldwork. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 24 Competing interests The authors have no conflict of interest to declare. Author contributions RSB and VG conceived and designed the study. RSB carried out the field experiments and collected the data. RSB carried out the statistical analyses with inputs from VG. RSB drafted the manuscript, and VG provided conceptual advice and edited the manuscript. Both authors gave final approval for publication. Data availability The dataset and R-scripts used for this study will be made available through a public repository once the study is published. Literature cited Barrett, S. C., & Cruzan, M. B. (1994). Incompatibility in heterostylous plants. In Genetic control of self-incompatibility and reproductive development in flowering plants (pp. 189–219). Dordrecht: Springer Netherlands. Barrett, S. C., & Fairnie, A. L. (2024). The neglected floral polymorphism: Mirror-image flowers emerge from the shadow of heterostyly. Evolutionary Journal of the Linnean Society, 3(1), kzae004. https://doi.org/10.1093/evolinnean/kzae004 Barrett, S. C., & Harder, L. D. (2005). The evolution of polymorphic sexual systems in daffodils (Narcissus). New Phytologist, 165(1), 45–53. https://doi.org/10.1111/j.1469-8137.2004.01183.x Barrett, S. C., Jesson, L. K., & Baker, A. M. (2000). The evolution and function of stylar polymorphisms in flowering plants. Annals of Botany, 85(suppl 1), 253–265. https://doi.org/10.1006/anbo.1999.1067 Barrett, S. C. (2002). The evolution of plant sexual diversity. Nature Reviews Genetics, 3(4), 274–284. https://doi.org/10.1038/nrg776 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 25 Barrett, S. C. (2010). Darwin’s legacy: The forms, function, and sexual diversity of flowers. Philosophical Transactions of the Royal Society B: Biological Sciences, 365(1539), 351–368. https://doi.org/10.1098/rstb.2009.0212 Bates, D., Maechler, M., Bolker, B., Walker, S., Christensen, R. H. B., Singmann, H., … & Bolker, M. B. (2015). Package lme4 (Version 12 [1], 2). Braga, B. L. P., Matias, R., Consolaro, H., Souza, J. T., & Almeida, N. M. (2022). Inaccuracy of sexual organ position and spatial variation of styles in monomorphic enantiostylous flowers of Senna rugosa (Fabaceae-Caesalpinioideae). Flora, 293, 152112. https://doi.org/10.1016/j.flora.2022.152112 Brys, R., & Jacquemyn, H. (2010). Floral display size and spatial distribution of potential mates affect pollen deposition and female reproductive success in distylous Pulmonaria officinalis (Boraginaceae). Plant Biology, 12(4), 597–603. https://doi.org/10.1111/j.1438-8677.2009.00259.x Brys, R., Geens, B., Beeckman, T., & Jacquemyn, H. (2013). Differences in dichogamy and herkogamy contribute to higher selfing in contrasting environments in the annual Blackstonia perfoliata (Gentianaceae). Annals of Botany, 111(4), 651–661. https://doi.org/10.1093/aob/mct031 Contreras-Varela, X., García-Franco, J. G., & Aguirre-Jaimes, A. (2023). Floral and reproductive biology of Commelina erecta L. (Commelinaceae) in a coastal dune system on the Gulf of Mexico. Flora, 302, 152259. https://doi.org/10.1016/j.flora.2023.152259 Costa, J., Castro, S., Loureiro, J., & Barrett, S. C. (2017). Experimental insights on the function of ancillary pollen and stigma polymorphisms in plants with heteromorphic incompatibility. Evolution, 71(1), 121–134. https://doi.org/10.1111/evo.13082 Dafni, A. (1992). Pollination ecology: A practical approach. Oxford University Press. Dai, C., & Galloway, L. F. (2012). Male flowers are better fathers than hermaphroditic flowers in andromonoecious Passiflora incarnata. New Phytologist, 193(3), 787-796. https://doi.org/10.1111/j.1469-8137.2011.03966.x .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 26 De Almeida, N. M., & de Castro, C. C. (2019). Enantiostyly in angiosperms. In Asymmetry in plants (pp. 319–328). CRC Press. De Almeida, N. M., De Castro, C. C., de Lima Leite, A. V ., Novo, R. R., & Machado, I. C. (2013). Enantiostyly in Chamaecrista ramosa (Fabaceae-Caesalpinioideae): Floral morphology, pollen transfer dynamics, and breeding system. Plant Biology, 15(2), 369–375. https://doi.org/10.1111/j.1438-8677.2012.00651.x De Almeida, N. M., Souza, J. T., Oliveira, C. R. S., Bezerra, T. T., Novo, R. R., Siqueira Filho, J. A., … & Castro, C. C. (2018). Functional dimorphic enantiostyly in monomorphic enantiostylous species of the subtribe Cassiinae (Fabaceae-Caesalpinioideae). Plant Biology, 20(4), 797–801. https://doi.org/10.1111/plb.12718 Dulberger, R., & Ornduff, R. (1980). Floral morphology and reproductive biology of four species of Cyanella (Tecophilaeaceae). New Phytologist, 86(1), 45–56. https://doi.org/10.1111/j.1469-8137.1980.tb00778.x Fenster, C. B. (1995). Mirror-image flowers and their effect on outcrossing rate in Chamaecrista fasciculata (Leguminosae). American Journal of Botany, 82(1), 46–50. https://doi.org/10.1002/j.1537-2197.1995.tb15647.x Fox, J., & Weisberg, S. (2018). An R companion to applied regression. Sage Publications. Gao, J. Y ., Ren, P. Y ., Yang, Z. H., & Li, Q. J. (2006). The pollination ecology of Paraboea rufescens (Gesneriaceae): A buzz-pollinated tropical herb with mirror-image flowers. Annals of Botany, 97(3), 371–376. https://doi.org/10.1093/aob/mcj044 Harder, L. D., & Barrett, S. C. (1995). Mating cost of large floral displays in hermaphrodite plants. Nature, 373(6514), 512–515. https://doi.org/10.1038/373512a0 Harder, L. D., Barrett, S. C., & Cole, W. W. (2000). The mating consequences of sexual segregation within inflorescences of flowering plants. Proceedings of the Royal Society B: Biological Sciences, 267(1441), 315–320. https://doi.org/10.1098/rspb.2000.1002 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 27 Harder, L. D., Jordan, C. Y ., Gross, W. E., & Routley, M. B. (2004). Beyond floricentrism: The pollination function of inflorescences. Plant Species Biology, 19(3), 137–148. https://doi.org/10.1111/j.1442-1984.2004.00110.x Harrison, C. J., Möller, M., & Cronk, Q. C. B. (1999). Evolution and development of floral diversity in Streptocarpus and Saintpaulia. Annals of Botany, 84(1), 49–60. https://doi.org/10.1006/anbo.1999.0887 Jesson, L. K., & Barrett, S. C. (2002). Enantiostyly in Wachendorfia (Haemodoraceae): The influence of reproductive systems on the maintenance of the polymorphism. American Journal of Botany, 89(2), 253–262. https://doi.org/10.3732/ajb.89.2.253 Jesson, L. K., & Barrett, S. C. (2003). The comparative biology of mirror-image flowers. International Journal of Plant Sciences, 164(S5), S237-S249. Jesson, L. K., & Barrett, S. C. (2005). Experimental tests of the function of mirror-image flowers. Biological Journal of the Linnean Society, 85(2), 167–179. https://doi.org/10.1111/j.1095-8312.2005.00480.x Jesson, L. K., Barrett, S. C., & Day, T. (2003). A theoretical investigation of the evolution and maintenance of mirror-image flowers. The American Naturalist, 161(6), 916–930. https://doi.org/10.1086/375176 Johnson, S. D., Midgley, J. J., & Illing, N. (2023). The enantiostylous floral polymorphism of Barberetta aurea (Haemodoraceae) facilitates wing pollination by syrphid flies. Annals of Botany, 132(6), 1107-1118. https://doi.org/10.1093/aob/mcad118 Johnson, S. D., Midgley, J. J., & Illing, N. (2025). Functions of heteranthery and enantiostyly for wing pollination by pollen-collecting bees in Dilatris ixioides (Haemodoraceae). Annals of Botany, mcaf189. https://doi.org/10.1093/aob/mcaf189 Kearns, C. A., & Inouye, D. W. (1993). Techniques for pollination biologists. University Press of Colorado. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 28 Keller, B., Thomson, J. D., & Conti, E. (2014). Heterostyly promotes disassortative pollination and reduces sexual interference in Darwin’s primroses: Evidence from experimental studies. Functional Ecology, 28(6), 1413–1425. https://doi.org/10.1111/1365-2435.12274 Lenth, R. (2025). emmeans: Estimated marginal means, aka least-squares means (Version 1.11.1-00001) [R package]. https://rvlenth.github.io/emmeans/ Li, X. X., Zou, Y ., Xiao, C. L., Gituru, R. W., Guo, Y . H., & Yang, C. F. (2013). The differential contributions of herkogamy and dichogamy as mechanisms of avoiding self-interference in four self-incompatible Epimedium species. Journal of Evolutionary Biology, 26(9), 1949–1958. https://doi.org/10.1111/jeb.12193 Li, Y . X., Quan, Q. M., & Sun, G. L. (2009). Effect of floral morphology on fruit set in Epimedium sagittatum (Berberidaceae). Plant Systematics and Evolution, 279, 51–58. https://doi.org/10.1007/s00606-009-0145-0 Ling, S. J., Lu, T., & Ren, M. X. (2020). A novel type of mirror-image flowers caused by lateral bending of the floral tube in a bumblebee-pollinated plant. Flora, 272, 151694. https://doi.org/10.1016/j.flora.2020.151694 Martins, D. J. (2008). Pollination observations of the African violet in the Taita Hills, Kenya. Journal of East African Natural History, 97(1), 33–42. https://doi.org/10.2982/0012-8317(2008)97[33:POOTA V]2.0.CO;2 Massinga, P. H., Johnson, S. D., & Harder, L. D. (2005). Heteromorphic incompatibility and efficiency of pollination in two distylous Pentanisia species (Rubiaceae). Annals of Botany, 95(3), 389–399. https://doi.org/10.1093/aob/mci040 Minnaar, C., & Anderson, B. (2019). Using quantum dots as pollen labels to track the fates of individual pollen grains. Methods in Ecology and Evolution, 10(5), 604–614. https://doi.org/10.1111/2041-210X.13155 Minnaar, C., & Anderson, B. (2021). A combination of pollen mosaics on pollinators and floral handedness facilitates the increase of outcross pollen movement. Current Biology, 31(14), 3180–3184. https://doi.org/10.1016/j.cub.2021.04.074 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 29 Mora-Carrera, E., Castañeda-Zárate, M., Fornoni, J., Boege, K., & Domínguez, C. A. (2019). On the adaptive value of monomorphic versus dimorphic enantiostyly in Solanum rostratum. Annals of Botany, 123(1), 205–212. https://doi.org/10.1093/aob/mcy162 Morais, J. M., Consolaro, H. N., Bergamini, L. L., & Ferrero, V . (2020). Reproductive biology and pollinators in two enantiostylous Qualea species (V ochysiaceae) in the Brazilian Cerrado. Plant Biology, 22(5), 872–880. https://doi.org/10.1111/plb.13091 Narbona, E., Ortiz, P. L., & Arista, M. (2005). Dichogamy and sexual dimorphism in floral traits in the andromonoecious Euphorbia boetica. Annals of Botany, 95(5), 779-787. https://doi.org/10.1093/aob/mci077 Ochogavía, A. C. (2022). Quantifying the reproductive progression of sunflower using FIJI (Image J). MethodsX, 9, 101879. https://doi.org/10.1016/j.mex.2022.101879 Ornduff, R., & Dulberger, R. (1978). Floral enantiomorphy and the reproductive system of Wachendorfia paniculata (Haemodoraceae). New Phytologist, 80(2), 427-434. https://doi.org/10.1111/j.1469-8137.1978.tb01577.x Paudel, B. R., Fairnie, A. L., McCarren, S., Anderson, B., Illing, N., & Barrett, S. C. (2024). Loss of buzz pollination results in chronic pollen limitation in an enantiostylous plant. South African Journal of Botany, 171, 592-601. https://doi.org/10.1016/j.sajb.2024.06.019 Pauw, A. (2005). Inversostyly: A new stylar polymorphism in an oil-secreting plant, Hemimeris racemosa (Scrophulariaceae). American Journal of Botany, 92(11), 1878–1886. https://doi.org/10.3732/ajb.92.11.1878 Prasanna, N. S. (2023). Phylogenetic and biogeographic approaches to understand the diversification of the Asian genus Didymocarpus Wall. (Gesneriaceae) (Doctoral dissertation). Indian Institute of Science Education and Research, Bhopal. http://hdl.handle.net/10603/525010 R Core Team. (2025). R: A language and environment for statistical computing [Software]. R Foundation for Statistical Computing. https://www.R-project.org/ .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 30 Ren, M. X., Zhong, Y . F., & Song, X. Q. (2013). Mirror-image flowers without buzz pollination in the Asian endemic Hiptage benghalensis (Malpighiaceae). Botanical Journal of the Linnean Society, 173(4), 764–774. https://doi.org/10.1111/boj.12101 Richards, A. J. (1997). Plant breeding systems. Garland Science. Richman, S. K., & Venable, D. L. (2018). Aggregate enantiostyly: Floral visitor interactions with a previously unreported form of floral display. Journal of Pollination Ecology, 22, 49–54. https://doi.org/10.26786/1920-7603(2018)seven Robertson, C., Xue, H., Saltini, M., Fairnie, A. L., Lang, D., Kerstens, M. H., … & Lenhard, M. (2025). Spiral phyllotaxis predicts left–right asymmetric growth and style deflection in mirror-image flowers of Cyanella alba. Nature Communications, 16(1), 3695. https://doi.org/10.1038/s41467-025-58803-5 Rueden, C. T., Schindelin, J., Hiner, M. C., DeZonia, B. E., Walter, A. E., Arena, E. T., & Eliceiri, K. W. (2017). ImageJ2: ImageJ for the next generation of scientific image data. BMC bioinformatics, 18(1), 529. https://doi.org/10.1186/s12859-017-1934-z Saltini, M., McCarren, S., Illing, N., Barrett, S. C., & Anderson, B. (2025). Selective contexts favouring asymmetry: A theoretical and empirical study of pollen transfer in mirror-image flowers. bioRxiv, 2025-09. https://doi.org/10.1101/2025.09.23.677969 Tang, L. L., & Huang, S. Q. (2005). Variation in daily floral display and the potential for geitonogamous pollination in two monomorphic enantiostylous Monochoria species. Plant Systematics and Evolution, 253, 201–207. https://doi.org/10.1007/s00606-005-0309-5 Waites, A. R., & Ågren, J. O. N. (2004). Pollinator visitation, stigmatic pollen loads and among‐population variation in seed set in Lythrum salicaria. Journal of Ecology, 92(3), 512-526. https://doi.org/10.1111/j.0022-0477.2004.00893.x Wang, G., Miura, R., & Kusanagi, T. (1995). The enantiostyly and the pollination biology of Monochoria korsakowii (Pontederiaceae). Acta Phytotaxonomica et Geobotanica, 46(1), 55-65. https://doi.org/10.18942/bunruichiri.KJ00001079078 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint 31 Williams, P. H. (2022). The Bumblebees of the Himalaya: An Identification Guide. Royal Belgian Institute of Natural Sciences. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted January 30, 2026. ; https://doi.org/10.64898/2026.01.29.702576doi: bioRxiv preprint