Dimorphic Pollen of Heterostylous Waltheria indica Identified in Honey Bee Pollen Loads: Implications for Pollination Ecology

Abstract

During an analysis of the pollen load composition in the Apis cerana Fabr. honey bee colonies maintained in the Pondicherry region of southern India, an unidentified and taxonomically unresolved pollen morphotype was repeatedly recorded. Subsequent field observations, corroborated by light microscopic pollen studies, revealed that this morphotype corresponded to Waltheria indica L. (Malvaceae), a heterostylous shrub or subshrub that produces distinct long- and short-styled floral morphs associated with pollen dimorphism. Archived bee pollen slides were re-examined and compared with reference pollen collected from both floral morphs of W. indica using light microscopy (LM). The two pollen morphotypes differed considerably in size, shape and exine ornamentation: long-styled (pin) flowers produced smaller, reticulate pollen grains, whereas the short-styled (thrum) flowers yielded larger, microechinate grains. Pollen recovered from bee loads corresponded primarily to the larger morphotype from the thrum flowers. Field observations confirmed that A. cerana foraged on both floral morphs, supporting the potential for intermorph pollen transfer. This study demonstrates that heterostyly-associated pollen dimorphism can influence pollen spectra obtained from bee foraging studies, underlining the need to consider floral polymorphism for accurate palynological identification and pollination interpretations.

INTRODUCTION

Heterostyly is a floral polymorphism characterised by the reciprocal positioning of stamens and stigmas in distinct floral morphs, often accompanied by dimorphism in pollen size and stigma structure. Such adaptations promote outcrossing and occur widely among tropical plant taxa (Silveira et al., 2017). However, pollen dimorphism resulting from heterostyly can complicate palynological identification, particularly in studies that analyse pollen collected by insects or from environmental samples (Mercuri et al., 2021). Palynology provides valuable taxonomic and ecological insights through the detailed examination of pollen and spores, whose morphological features, such as exine ornamentation, size, shape and aperture type, serve as reliable indicators of plant identity and evolutionary relationships (Erdtman, 1952; Faegri & Iversen, 1975; Moore et al., 1991). These characters often reflect adaptations linked to pollination mechanisms and have contributed significantly to resolving taxonomic ambiguities among closely related species.

The genus Waltheria L. (Malvaceae) comprises about 60 species of shrubs and subshrubs, distributed mainly in tropical and subtropical regions, with most species confined to the Americas (Saunders, 1993). Waltheria is recognised as an entomophilous plant in the neotropics, though it is also found in fossil pollen records there (Bush, 1995; Fonseca et al., 2021), whereas in South India, it has been reported only in melisso- and entomo-palynological studies (Lazar et al., 2023, 2024; Ponnuchamy et al., 2014). This could be attributed to the fact that most Waltheria species, including Waltheria indica, are indigenous to the neotropical areas and are believed to have been introduced and naturalised in the Indian subcontinent (Nirmala & Sridevi, 2021; Sreejith et al., 2020). W. indica is recognised for its diverse medicinal applications, with studies documenting its ethnobotanical uses, phytochemical constituents and pharmacological activities (Nirmala & Sridevi, 2021; Yabré et al., 2024). Several members of the genus exhibit a heterostylous or distylous nature, producing distinct long- and short-styled floral morphs that facilitate cross-pollination through reciprocal positioning of anthers and stigmas between the floral morphs, which reduces self-pollination and promotes intermorph pollen transfer by pollinators. This floral polymorphism is commonly accompanied by pollen dimorphism, wherein long-styled morphs bear smaller, reticulate pollen grains, while short-styled morphs produce larger, microechinate grains (Silveira et al., 2017). W. indica shows notable variation in floral morphology across its geographic range. Most regional botanical sources describe the species as uniformly homostylous, including reports from Paraguay, Central America, China and India (Saunders, 2007; Sharma & Sanjappa, 1993). However, some studies from India have reported the coexistence of both heterostylous and homostylous forms (Bahadur & Srikanth, 1983). These authors documented the presence of heterostylous, self-incompatible morphs occurring alongside homostylous, self-compatible morphs, indicating clear intraspecific floral polymorphism. Heterostylous populations appear to be restricted to isolated pockets in Andhra Pradesh and Tamil Nadu, where bees such as Apis cerana and various butterfly species act as primary floral visitors (Raju & Rao, 2022). Despite the species’ wide distribution, detailed investigations into its pollen morphology and the broader implications for its pollination ecology remain limited. During doctoral research on the pollen load composition of A. cerana colonies in the Pondicherry region of southern India, an unidentified pollen morphotype was consistently recorded but could not be assigned to any known local flora despite extensive comparisons with reference slides. Subsequent fieldwork and microscopy revealed that this morphotype originated from W. indica L., a widespread, common ruderal shrub bearing heterostylous flowers. The present study documents the morphological characteristics of the dimorphic pollen grains of W. indica, confirms their occurrence in honey bee pollen loads, and discusses their implications for both palynological identification and pollination-ecological studies.

MATERIALS AND METHODS

Study Area and Sampling

The study area is characterised by a tropical coastal climate, with mixed dry deciduous scrublands, disturbed habitats and agricultural farms. Foraging activity of A. cerana was regularly monitored two to three times a week, and pollen loads from the bees’ corbiculae were collected during routine sampling from managed A. cerana colonies maintained in the Puducherry region, southeastern India, in the framework of the PhD thesis of Lazar (2024). A total of 46 pollen load slides from four of these sites, Merveille (MER) (11°57′ 10.84″N, 79°45′ 57.98″E), Gratitude Farm (GF) (11°57′ 2.45″N, 79°46′ 5.8″E), Arul Ashram (AA) (12°01′ 34.02″N, 79°51′ 20.63″E) and Institut Français de Pondichéry (IFP) Garden (11°56′ 12.97″N, 79°50′ 8.87″E), that contained at least one of the pollen morphotypes of W. indica were used in the present study (Figure 1). Of these slides, 38 were from the site at MER in Project Ecolake, SAICE (Sri Aurobindo International Centre of Education). For more details on this primary site and sampling methodology, please refer to Lazar et al. (2023, 2024).

Figure 1.

Source: Google Earth Pro.

Re-examination of Unidentified Pollen

Archived pollen load slides were re-examined under a light microscope (Olympus BX53, under 400× and 1,000× magnifications). The unidentified pollen type was photographed and measured using an ocular micrometre. This was used to characterise the unknown pollen morphotype and to compare it with reference material. Independently, fresh pollen from plants matched with authenticated herbarium specimens (HIFP_028210, HIFP_028211) was mounted as reference slides of the distinct morphotypes of W. indica after acetolysis.

Reference Pollen Collection and Floral Morphological Examination

Fresh flowers of W. indica were collected from several natural populations in the Pondicherry region. The species exhibits heterostyly, with distinct long-styled (pin) and short-styled (thrum) floral morphs. Pollen from each morph was collected separately, acetolysed following Erdtman (1960), and mounted in glycerine jelly for light microscopic examination.

For floral morphological examination, flowers at different developmental stages were carefully dissected and observed under a light microscope. The heterostylous condition of the species was clearly identified. Photomicrographs were taken to document the inflorescence, a single flower, stamens and stigma, and a close-up view of the anther (Figure 2). Voucher specimens were registered and deposited in the IFP Herbarium, Puducherry, India, and in the Auroville Virtual Herbarium, Auroville, Tamil Nadu.

Figure 2.

Comparative Photos of Heterostylous Flowers in Waltheria indica, Illustrating Short-styled (Left) and Long-styled (Right) Morphs. (A) Inflorescence, (B) Single Flower, (C) Stamen and Stigma, (D) Close Up View of Anther.

Pollen Comparison and Data Analysis

For each morph, 30 pollen grains were measured to determine the average size. Diagnostic characters such as size, shape, aperture configuration and exine sculpturing were compared with the unidentified pollen from bee loads to assess/determine taxonomic correspondence. Multivariate statistics (principal component analysis (PCA) and canonical correspondence analysis (CCA)) were carried out on the complete pollen data from the corbicular loads of the 46 samples used in this study using the PAlaeontological STatistics (PAST) software (Hammer & Harper, 2001). This was done in order to understand the seasonal and site-wise variation in the bee foraging preferences vis-à-vis environmental variables such as rainfall and humidity, and to understand if there was any pattern in the way the bees foraged on the two pollen morphotypes of W. indica.

Field Observations of Bee Visitation

Observations were conducted at W. indica populations in the vicinity of the study areas during morning hours. Visits by A. cerana and other pollinators were recorded for both floral morphs (Figure 3). The mode of foraging (pollen or nectar collection) and the floral morph visited were noted to understand potential differential pollen transfer between morphs.

Figure 3.

Pollinators Observed on Waltheria indica During the Field Visits. (A) W. indica (Long-styled) Habit, (B) W. indica (Short-styled) Habit, (C) Apis dorsata Collecting Nectar, (D) Apis cerana Collecting Nectar, (E) Carpenter Bee, (F) Metallic Bee, (G) Junonia lemonias, (H) Danaus chrysippus, (I) Acraea terpsicore, (J) Jamides celeno.

RESULTS

Light microscopic examination of W. indica pollen revealed two distinct morphotypes corresponding to the long-styled (pin) and short-styled (thrum) floral forms. Pollen trait measurements from plants at the study site are summarised in Table 1. Differences in floral characters between the distinct long-styled (pin) and short-styled (thrum) floral morphs are presented in Table 2.

Table 1.

Diagnostic Pollen Characters of the Two Distinct Morphotypes Observed in Waltheria indica.

Floral Morph	Shape	Aperture	P (µm) Min–Max (Mean)	E (µm) Min–Max (Mean)	Exine Ornamentation
Type I: Long-styled (pin)	Spheroidal-oblate spheroidal—Prolate spheroidal	3–4 colporate	34.5–52.15 (45.51)	37.03–53.32 (46.84)	Reticulate
Type II: Short-styled (thrum)	Spheroidal- oblate spheroidal	4–5 colporate	39.7–64.5 (52.6)	44.6–66.0 (57.1)	Microechinate

Table 2.

Diagnostic Floral Characters of Waltheria indica.

Sl No.	Plant Part	Waltheria indica (Short Style) Thrum	W. indica (Long Style) Pin
1	Stamens	6.0–6.5 mm	2 mm
2	Stamen filament tube	Fused	Basal fused, upper free
3	Anther colour	White	Yellow
4	Style–length	1.5 mm	5.0–5.2 mm

Both pollen morphotypes are radially symmetrical, isopolar and colporate. Yet light microscopy (LM) clearly shows that they differ in size, number of apertures and exine ornamentation. Pollen from the long-styled morph is smaller with a more prominently reticulate exine, whereas pollen from the short-styled morph is larger and exhibits an echinate sculpturing pattern. These morphological distinctions were consistently observed under LM and confirmed through samples collected from separate plants (Figure 4). Pollen of both morphotypes was identified in the corbicular loads of A. cerana, indicating that this forager visits both the floral morphs.

Figure 4.

Light Micrographs of Pollen Grains of Waltheria indica Type 2 (Left on the Green Background - (A&B) Polar View; (C&D) Equatorial View -Echinate) and W. indica Type 1 (Right on the Blue Background - (A&B) Polar View; (C&D) Equatorial View -Reticulate).

The differences in the floral characters between the two floral morphs are presented in Table 2 and Figure 2. One floral morph exhibited long stamens measuring 6.0–6.5 mm with a comparatively short stigma of approximately 1.5 mm. In contrast, the opposite floral morph showed shorter stamens measuring 2 mm and a longer stigma measuring 5.0–5.2 mm. In addition to differences in the relative lengths of stamens and stigmas, clear variation was also observed in anther colour and the morphology of the stamen filament.

Re-examination of archived A. cerana pollen-load slides also revealed two distinct pollen morphotypes matching those of W. indica. Type 1 pollen was generally smaller with a thicker exine, while type 2 pollen was comparatively larger with finer sculpturing. The unidentified pollen morphotype detected in archived bee pollen load samples was larger (~57 µm, spheroidal to oblate spheroidal and echinate), corresponding exactly with the type 2 pollen of the short-styled morph (Thrum). This confirms that the previously unassigned pollen originated from the thrum morph. Both pollen morphs were also detected in the bee pollen loads in different proportions. Field observations further confirmed that A. cerana and other pollinators visited both floral morphs frequently, collecting both nectar and pollen (Figure 3).

The CCA integrating pollen taxa, environmental variables, sampling months and number of taxa occurrence shows clear seasonal structuring of pollen assemblages (Figure 5(A)). Associations between taxa and environmental vectors indicate that rainfall and humidity are the main drivers influencing honeybee foraging patterns, while the clustering of months along shared gradients reflects seasonal shifts in resource availability. Rainfall emerges as the strongest determinant of pollen composition, with high-rainfall months shifting toward the right side of Axis 1 and aligning with moisture-responsive tree taxa, especially those from the IFP site, such as Peltophorum pterocarpum, Piper and Calophyllum inophyllum (Figure 5(A)). Most other months cluster near the centre, dominated by consistently occurring herbaceous taxa (Asteraceae, Poaceae, Sida, Commelina, Coldenia/Cleome, Lamiaceae, Amaranthaceae, Alternanthera). Humidity shows a moderate influence, while the number of taxa aligns with late-monsoon richness patterns along Axis 2 (Figure 5(A)).

Figure 5.

(A) Canonical Correspondence Analysis (CCA) of the Pollen Data from the 46 Corbicular Loads Considered in this Study Depicting the Association/Relationship Between Pollen Assemblages and Environmental Explanatory Variables (Rainfall, Humidity and Floral Availability (Total Number of Pollen Taxa/Sample)). (B) Principal Component Analysis (PCA) of the Pollen Data from the 46 Corbicular Loads Considered in This Study, Illustrating the Separation of Samples Containing the Two Morphs Along the Two Axes.

Further, PCA supported the distinction between the two pollen types. PC1 accounted for most of the variation and clearly separated the morphotypes, with type 1 clustering along the positive side of PC1 and type 2 forming a distinct cluster/group towards the upper-right quadrant (Figure 5(B)). PC2 reflected temporal variation, with February samples across multiple years clustering lower on the axis and June–July samples showing higher PC2 values. Site vectors indicated minor but noticeable influences, with GF samples more closely associated with type 2 pollen, whereas AA and MER exhibited a broader spatial distribution.

Field observations confirmed that A. cerana visited both long-styled and short-styled floral morphs of W. indica, collecting nectar and pollen and frequently moving between morphs during peak morning hours, suggesting effective intermorph pollen transfer. These consistent visitation patterns indicate that A. cerana likely facilitates intermorph pollination, supporting the occurrence of the two pollen morphotypes observed in bee-load samples that originate from natural heterostyly in the species. In addition to the primary pollinators, bees (Bush, 1995), several other floral visitors were also observed on W. indica (Figure 3). Additionally, a few butterfly species were observed foraging actively on the flowers, along with frequent visits by bees. These included the Lemon pansy (Junonia lemonias), Plain Tiger (Danaus chrysippus), Tawny Coster (Acraea terpsicore) and Common Cerulean (Jamides celeno). As the butterflies feed on nectar and visit multiple inflorescences repeatedly, they contribute to pollen transfer. The presence of a diverse assemblage of pollinators and floral visitors indicates W. indica’s ecological importance in promoting local insect biodiversity and maintaining pollination networks.

DISCUSSION

The recognition of dimorphic pollen in W. indica illustrates how floral polymorphism can influence palynological interpretation. Intraspecific variation in pollen size and exine structure can complicate taxonomic assignments when analysing pollen loads, honey or soil samples. Such patterns are well documented in heterostylous taxa; for example, Primula, Turnera and Hedyotis exhibit clear morph-linked differences in stigma and anther positions that correlate with pollen traits (Bahadur & Ramaswamy, 1984; Rama Swamy & Bahadur, 1984; Barrett, 2013). Similar morph-specific palynological differentiation is also observed in Pentas lanceolata, where pin and thrum flowers differ significantly in pollen grain size and stigmatic papilla structure (Bahadur, 1970, 1984). In W. indica, this dimorphism is sufficiently pronounced to be detected under routine LM, enabling the clarification of an ‘unknown morphotype 2’ previously recorded in A. cerana pollen-load data sets, and confirming that it corresponds to the second floral morph of W. indica (Figure 4). From an ecological standpoint, frequent visitation of both floral morphs by A. cerana promotes legitimate intermorph pollen transfer and maintains population-level genetic diversity. For palynologists, recognising such intraspecific variation is essential for building accurate pollen reference collections, particularly in tropical ecosystems where floral polymorphism is widespread. Pollen from corbicular loads is generally better preserved than sediment and other environmental samples. The approach used here may be extended to frequently occurring unidentified pollen in those contexts, too. Though the use of scanning electron microscope (SEM) and confocal laser scanning microscopy (CLSM) can enhance the morphological characterisation of the two pollen morphotypes and ascertain if the larger pollen from the short-styled morph is indeed microechinate, it is beyond the purview of the present study.

Challenges in determining botanical affinities often arise from limited regional floristic knowledge and limited pollen reference data (Erdtman, 1969; Moore et al., 1991), as well as from morphological similarity among common tricolporate pollen types, particularly within families such as Asteraceae, Rhamnaceae and Melastomataceae/Combretaceae. Pollen occurring at low frequencies (<1%) may reflect plant traits such as low pollen production or aspects of bee foraging behaviour, including incidental or indirect collection (Carvalho et al., 1999). The absence of certain taxa in pollen loads, despite high local floral diversity, may reflect limited floral attractiveness, restricted access to rewards, or interspecific competition among foraging insects (Camillo & Garófalo, 1989; Lazar et al., 2023). Eusocial bees typically exploit dense floral patches offering abundant nectar or pollen resources (Imperatriz-Fonseca et al., 1984), and although they are broadly generalist foragers (Aguiar, 2003), they still display distinct floral preferences. In our samples, Cocos, Borassus, Syzygium, Turnera, Peltophorum and Portulaca were the most frequently represented taxa. Even low-frequency pollen taxa can provide meaningful ecological insights. Cultivated species such as Cucumis and Sesamum indicate agricultural landscape use, while resource-rich forest and scrub–jungle habitats (e.g., the MER site) may have attracted more foraging bees, influencing hive placement.

The arrangement of the corbicular samples along the rainfall vector highlights the strong dependence of certain pollen sources on monsoon activity (Figure 5(A)). Tree taxa positioned on the right side of Axis 1 reflect their peak flowering during wetter months, suggesting they provide important, but seasonally restricted resources to bees. Herbaceous and anemophilous taxa clustered near the origin indicate consistent availability across years, supporting continuous foraging even during dry phases. The mild separation along the humidity and richness vectors suggests that transitional months, when moderate moisture and high floral diversity coincide, contribute to increases in pollen taxa collected by bees. The CCA illustrates a seasonally dynamic foraging landscape, with monsoon-driven pulses of tree pollen superimposed on a stable herbaceous baseline (Figure 5(A)). Overall, the CCA indicates that seasonal moisture conditions, especially rainfall, structure pollen assemblages more strongly than interannual variation.

PCA results further confirm that the two pollen morphotypes represent stable biological categories (Figure 5(B)). Clear separation along PC1 may reflect intrinsic morphological differences between pollen from long-styled and short-styled flowers, while the persistence of this pattern across years, sampling months and study sites indicates that these traits are not environmentally induced. Seasonal patterns detected along PC2, such as the consistent clustering of February samples and higher PC2 values in monsoon season samples, may suggest minor plasticity, potentially linked to environmental conditions. Although site-level effects were generally weak, the closer association of type 2 pollen with GF samples indicates that local micro-environmental conditions may influence pollen occurrence or preservation. Overall, this study provides a robust modern reference for identifying pollen dimorphism in W. indica and demonstrates its consistent representation in the foraging spectrum of A. cerana in both ecological and melissopalynological contexts. In the Pondicherry region, W. indica has been identified as a bee-friendly plant (Lazar et al., 2024).

CONCLUSION

Re-examination of archived A. cerana pollen-load samples from Pondicherry confirmed that the pollen morphotype designated as type 2 corresponds to the larger pollen form of the heterostylous species W. indica. Recognising pollen dimorphism linked to heterostyly contributes not only to improving taxonomic accuracy but also enhances understanding of honey bee foraging patterns in diverse tropical landscapes. The study demonstrates that W. indica produces two distinct and stable pollen morphotypes, both regularly collected by A. cerana and also visited by other bees. Our pilot study supports a heterostylous basis of pollen dimorphism in W. indica and provides a reliable reference framework for identifying this species in pollen-load analyses and broader palynological studies. Overall, our results highlight a seasonally dynamic foraging landscape where bees integrate ephemeral resources with stable herbaceous taxa, emphasising the need to conserve diverse floral sources across seasons.