Search for a command to run...
Seed dispersal is a fundamental process in plant life cycles and is vital for maintaining ecosystem function (Hansen & Müller, 2009). Balanophora species lack chlorophyll and are holoparasitic members of Balanophoraceae. They produce numerous minute, one-seeded dry fruits (approximately 300 μm long) anchored at the base of specialized bracts called claviform bodies (Eberwein et al., 2009; Hansen, 1972). Although these fruits were initially assumed to be wind-dispersed, their spherical form, the low-airflow conditions of their understory habitat, and the ground-level position of infructescences likely limit effective dispersal (Suetsugu, 2020). Traits such as bright red infructescences and indehiscent fruits suggest that endozoochory is widespread in Balanophora. Although the fruits are dry, the fleshy claviform bodies likely provide visual cues and nutritional rewards similar to the pulp of bird-dispersed fruits (Suetsugu, 2020). Recent studies show that species with red infructescences are at least partly dispersed by diurnal birds (Suetsugu, 2020; Suetsugu & Hashiwaki, 2023). However, some species exhibit different traits. For example, Balanophora subcupularis produces dull-colored infructescences with a yeasty odor and is dispersed by ants and camel crickets (Suetsugu & Hashiwaki, 2025). Here, we examined Balanophora tobiracola, which emits a fermented odor detectable to humans. Initial observations indicated frequent nocturnal visits by cockroaches. Because fermented or yeasty scents have evolved in diverse plant lineages to attract pollinators (Goodrich & Jürgens, 2018; Stökl et al., 2010), similar volatiles, whether produced by the plant or through microbial fermentation, may serve as cues for seed dispersers. However, the scent composition of Balanophora has not been characterized. We therefore investigated whether (i) cockroaches are the primary visitors to B. tobiracola, (ii) seeds excreted by cockroaches and other consumers remain viable, and (iii) the emitted volatiles resemble those associated with fermentation. The study was conducted on Yakushima Island, Kagoshima Prefecture, Japan, and detailed methods are provided in Appendix S1: Section S1. Infrared-triggered cameras detected no birds or mammals feeding on B. tobiracola, and no vertebrate feeding marks were observed. Ant species such as Crematogaster osakensis and Nylanderia flavipes, known dispersers of B. subcupularis, were not observed transporting B. tobiracola fruits. Unlike B. subcupularis, in which the fruits and claviform bodies detach readily (Suetsugu & Hashiwaki, 2025), the fruits of B. tobiracola remain firmly attached, likely preventing ant-mediated dispersal. In contrast, direct observations and time-lapse photography recorded frequent visits by cockroaches and crickets (Figure 1, Video S1; Table 1). Strong fruit attachment may represent an adaptation favoring dispersal by insects capable of mechanically detaching the fruits (Figure 1). Although Margattea satsumana and Megamareta pallidiola were the most frequent visitors, the cricket Duolandrevus ivani consumed more fruit material per visit, indicating that both cockroaches and D. ivani act as primary consumers of B. tobiracola infructescences. Feeding trials showed that M. satsumana, M. pallidiola, and D. ivani excreted intact seeds at rates of 5.9% ± 2.1%, 5.4% ± 1.6%, and 11.4% ± 1.7%, respectively (mean ± SE; Figure 2). Seed viability tests using TTC staining revealed no significant differences between excreted seeds and seeds from intact infructescences (19.5% ± 5.4% for M. satsumana, 28.6% ± 7.9% for M. pallidiola, 18.7% ± 4.3% for D. ivani, and 22.1% ± 2.5% for intact fruits; p > 0.5 for all). As no other dispersers were detected, these insects likely function as legitimate seed dispersers. Cockroaches and crickets routinely move and defecate within leaf litter and root crevices where B. tobiracola grows, suggesting that their movements may deposit seeds near host roots suitable for establishment. Although insects are rarely recognized as endozoochorous dispersers, the extremely small seeds of many non-photosynthetic plants with limited nutrient reserves may be particularly suited to insect-mediated dispersal (de Vega et al., 2011; Suetsugu, 2018). Notably, M. satsumana also acts as the primary pollinator of B. tobiracola (Suetsugu, 2025), illustrating a case of double mutualism, in which a single insect species provides both pollination and seed dispersal. This is the first reported case of cockroaches performing both roles, adding to the few known examples of invertebrate-mediated double mutualism (García et al., 2012). Similarly, B. subcupularis relies on ants and camel crickets for both pollination and seed dispersal (Suetsugu & Hashiwaki, 2025). The shaded understory habitats and small stature of Balanophora species may limit partner availability, potentially favoring such multifunctional interactions. Scent analyses showed that whole infructescences of B. tobiracola emitted numerous fermentation-associated volatiles, including acetic acid (13.0% ± 3.0%; mean ± SE), ethyl acetate (3.9% ± 1.1%), methyl acetate (3.5% ± 0.9%), acetoin (2.6% ± 0.9%), diacetyl (1.0% ± 0.0%), 4-ethylphenol (0.8% ± 0.5%), 1-propionylethyl acetate (0.2% ± 0.0%), 2-methylbutyl acetate (0.2% ± 0.0%), isoamyl acetate (0.2% ± 0.1%), and isobutyl acetate (0.1% ± 0.0%) (Appendix S1: Section S2; Stökl et al., 2010). Acetoin mediates chemical communication in several cockroach species (Fukushima et al., 1987; Sreng, 1993) and is also emitted by the cockroach-pollinated Clusia blattophila (Vlasáková et al., 2008). B. tobiracola also emits high levels of styrene (60.9% ± 0.1%), a rare plant volatile with a town-gas-like odor (Steenhuisen et al., 2012), which may reinforce the perception of decay or fermentation and enhance the attraction of detritivorous cockroaches and crickets. Future research should clarify how volatile cues mediate interactions between B. tobiracola and its insect dispersers. Behavioral assays such as Y-tube olfactometer tests and field choice experiments can identify which compounds or blends attract cockroaches and crickets. It will also be important to determine whether these volatiles are synthesized by the plant, consistent with chemical mimicry (Midgley et al., 2015), or originate from microbial communities on infructescence surfaces, indicating a third-party mutualism (Peris et al., 2017). Approaches such as surface sterilization, headspace sampling of aseptic tissues, microbial culturing, and comparative scent profiling can address this issue. In conclusion, this study demonstrates insect-mediated seed dispersal in B. tobiracola, where fermentation-associated scents and ground-level position of infructescences facilitate encounters with cockroaches and crickets. Although insect-mediated seed dispersal has recently been documented in several non-photosynthetic plants, most reported cases involve a single insect group, such as camel crickets or tenebrionid beetles (de Vega et al., 2011; Suetsugu, 2018; Suetsugu & Hashiwaki, 2025). By contrast, our results show that B. tobiracola engages multiple insect groups. Because seed dispersal often represents a diffuse mutualism, similar systems may be more widespread than currently recognized. Future work should assess the broader ecological relevance of insect-mediated seed dispersal and clarify the chemical and behavioral mechanisms underlying these interactions. We thank Hiroaki Yamashita, Toshihiro Saito, and Hiromu Hashiwaki for invaluable field support. We also thank Akira Yamawo and Naoyuki Nakahama for assistance with insect identification, and Kazuma Takizawa, Hidehito Okada, and Michiko Ishida for laboratory work. This study was supported by PRESTO (JPMJPR21D6, KS and JPMJPR21D3, YO) from the Japan Science and Technology Agency. The authors declare no conflicts of interest. Data (Suetsugu et al., 2025) are available in Figshare at https://doi.org/10.6084/m9.figshare.28720124. Appendix S1. Video S1_Metadata. Video S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.