Hunt and pollinate: Hornet pollination of the putative generalist genus Angelica

Most plants of the family Apiaceae produce large, white umbrella-like inflorescences composed of many small open flowers that secrete nectar and produce pollen (Knuth, 1908). This system allows nearly any type of insect to access floral resources and gives rise to the family's reputation for generalist pollination (Bell, 1971; Olesen et al., 2007; Ollerton et al., 2007). Indeed, two members of the Apiaceae are listed among the top 10 generalist plants in Olesen et al. (2007), as they attract more than 100 insect species. Insects that visit apiaceous flowers include diverse taxa including Diptera, Coleoptera, Lepidoptera, and Hymenoptera (Knuth, 1908; Lindsey, 1984). Furthermore, the flower visitor composition of apiaceous plants often fluctuates among years and populations (Davila & Wardle, 2008; Lamborn & Ollerton, 2000; Zych, 2007; Zych et al., 2019). Only in a handful of cases are specific insects known to contribute disproportionately to pollination (Lindsey, 1984); for example, pollen transfer in Ostericum palustre is predominantly achieved by flies despite seemingly generalist visitation (Zych et al., 2014). On 6 September 2015, I observed a vespid wasp, the Asian giant hornet (Vespa mandarinia: Vespidae), visiting dark-red inflorescences of Angelica decursiva (Apiaceae) and feeding on nectar in the Madarao highlands of Nagano Prefecture, Japan (Figure 1a,b). Although the observations were brief, I noted many hornet visits and that their bodies were dusted with pollen. On 9 September 2018, I again observed extensive hornet visitation in A. decursiva and green-flowered A. hakonensis (Figure 1c,d; Video S1) at the Nikko Botanical Garden in Tochigi Prefecture. Predatory wasps including hornets are common visitors of Apiaceae and are not major pollinators (Knuth, 1908; Lukas et al., 2020), but they were predominant in a certain year in fly-pollinated O. palustre (Zych et al., 2014). Given the intensive visitation and the fact that green and dark-red floral displays are common in plants pollinated by wasps (Brodmann et al., 2012), I hypothesized that these two Angelica species possessed pollination systems centered on hornets. To address this hypothesis, I first identified and quantified the flower visitors of A. decursiva and A. hakonensis in the Nikko Botanical Garden located in the Nikko National Park, where A. decursiva is naturally distributed and A. hakonensis, which originates in the central Kanto area of Japan, is semi-naturalized (Figure 1c,d). Flower visitors were counted on sunny days with no wind in September 2018, 2019, and 2020 (see Appendix S1: Table S1 for details), and visitors were captured from A. hakonensis to examine pollen transfer. The hornets were found to be dominant across the years and some visitations by colletid and apid bees, tachinid and muscid flies, butterflies, and ants were observed (Table 1; Appendix S1: Table S2). Hornets regularly carried more than 100 pollen grains per individual, whereas other visitors generally carried fewer grains (Table 1; Figure 1e). The pollen was attached to the ventral side of bodies in hornets, flies, and bumblebees, and to the entire body in small bee species (Table 1). The high visitation frequency and abundant pollen suggest that hornets account for a large proportion of the pollen transfer. The fine mesh treatment yielded almost no seed production. Seed set in hornet-excluded inflorescences was significantly lower than in the natural pollination treatment in A. decursiva and A. hakonensis, but greater than in the fine mesh treatment in A. decursiva (p < 0.001; generalized linear mixed model, followed by Tukey's honestly significant difference tests; Figure 2; Appendix S1: Section S2). These results suggested that hornets played a significant role in the pollination of both Angelica species tested. It should be noted that hornets were not the sole pollinators of the two tested species, as the hornet exclusion treatment also resulted in some seed sets (Figure 2). This finding appears to contradict the evidence that hornets are the predominant pollen vector, far exceeding other insects in visitation frequency and body pollen (Table 1). This apparent discrepancy may be attributed to the hornets' harassment of other visitors. For example, I observed a hornet that was foraging on an inflorescence hunt flower visitor upon encountering it (Figure 1f). This observation aligns with a previous report that hornets discouraged the visitation of other insects in Mentha suaveolens (Rojas-Nossa & Calviño-Cancela, 2020). Therefore, excluding hornets would probably result in increased visits by other insect species, leading to moderate compensation in seed set. The visitor fauna is generally variable in time and space in Apiaceae (Lamborn & Ollerton, 2000; Zych et al., 2019). Thus it is unlikely that this study revealed the whole diversity of floral visitors of studied Angelica species. However, compared with previous studies (Lukas et al., 2020; Zych et al., 2014), flower visitors have been consistently biased toward hornets over the years (Appendix S1: Table S2). Other than frequent visitation, the hornets may be qualified as functional pollen vectors for two reasons. First, the hornets do not harvest the pollen adhered to their bodies, resulting in abundant body pollen (Zych et al., 2014, 2019). This contrasts with bees that actively harvest pollen to feed their larvae. Second, contribution to seed set suggests that the adhesion of pollen to a hornet's body may not reduce pollen fertility, as reported in ants (Galen & Butchart, 2003). The finding of this study may corroborate the observation that the plants pollinated by hornets often possess dark-red and green floral displays (Brodmann et al., 2012). Whereas the dark florets in the umbel center are considered to increase attractiveness to beetle pollinators in Apiaceae (Claßen-Bockhoff et al., 2023), the functional significance of unicolored inflorescence is unclear. Still, it is expected that plants with a similar floral display, such as the dark-red-flowered A. gigas, green-flowered A. pseudoshikokiana, and genus Sanicula, which includes both green- and dark-red-flowered species, may be pollinated by hornets. In contrast, some plants in the genus Bupleurum have bright yellow flowers, which are hallmarks of classical myophilous pollination (pollination by flies; Zych et al., 2014). Given the diversity of floral characteristics within the family, Apiaceae may include pollination systems that are not as highly generalized as commonly thought. I deeply appreciate the Subject-matter Editor Professor John Pastor and the two anonymous reviewers for their helpful comments and suggestions during the reviewing process. I thank the staff of Nikko Botanical Gardens for the care of the plants. This study was supported by a JSPS Grant-in-Aid for Young Scientists to K.M. (grant no. 20K15859). The author declares no conflicts of interest. Data and code (Mochizuki, 2023) are available in Figshare at https://doi.org/10.6084/m9.figshare.24013431.v3. Appendix S1. Video S1. Video S1 Legend. Video S2. Video S2 Legend. 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.