Long-term camera trapping needed to identify sunbird species that pollinate the endangered South African orchid Satyrium rhodanthum

Information about the pollination system of plant species is relevant for both basic and applied purposes. However, characterisation of pollination systems may be challenging, especially if pollinator visits are inherently infrequent, as is often the case for plant species with high levels of specialisation (Benadi & Pauw, 2018). Several studies have shown that increasing observation time may result in more accurate characterisation of the pollinator spectrum (Ollerton et al., 2003; Petanidou & Ellis, 1993), but there are human limits to observation times and the proximity of human observers may also disrupt interaction behaviour. Recent technological advances that enable remote surveillance using motion-triggered cameras have facilitated research into specialised interactions by increasing the time over which pollinators can be recorded without disrupting natural activity. These methods have led to the discovery of several novel interactions between plants and both vertebrate and insect pollinators (Amorim et al., 2020; Balducci et al., 2020; Cozien et al., 2019; Krauss et al., 2018; Steenhuisen et al., 2015). Cameras also allow for detailed analysis of pollinator behaviour, which may otherwise be challenging if interactions are observed from afar (e.g. Balducci et al., 2020). Satyrium rhodanthum Schltr. is one of five bird-pollinated members of the genus (Johnson, 1996; Johnson et al., 2011; Johnson & Van der Niet, 2019; Van der Niet et al., 2015). It is characterised by red, unscented flowers and it is known from only four populations that occur in mistbelt grassland of the summer rainfall region of South Africa (Johnson & Bytebier, 2015; Rushworth, 2015; Van der Niet et al., 2015). An earlier study, based on 5 years of observations and camera trapping, reported pollination of this species by a single sunbird species, Chalcomitra amethystina (Shaw, 1812), at a site in KwaZulu-Natal, South Africa (Van der Niet et al., 2015). Since the range of S. rhodanthum overlaps with that of several sunbird species (Southern African Bird Atlas Project 2; Brooks & Ryan, 2022), any conclusions that this species is pollinated by a single sunbird species may be premature. Inadequate sampling often leads to underestimation of diversity in ecological studies and this also applies to studies of pollination systems (Jordano, 1987). We hypothesised that functionally similar sunbird species would also visit S. rhodanthum and consequently extended our long-term camera trapping study to better characterise the pollinator assemblage for this species and to obtain additional information on pollinator behaviour, such as pollen grooming, which may impact plant fitness. Pollinator observations were carried out by direct observations and the use of Bushnell NatureView Trophy Trail Cameras (see Table 1 for duration of observations). In the interest of protecting the endangered S. rhodanthum, the coordinates of the four known sites where S. rhodanthum occurs are withheld. Voucher specimens are deposited in the Bews Herbarium (NU) (MCZ1470, MCZ1483, MCZ1485). Details of pollinator observations between 2007 and 2013 are provided in Van der Niet et al., (2015). In 2017, 2018, 2020 and 2021, up to six cameras were deployed for 5–28 days per year during peak flowering in October and November. Cameras were placed 25–60 cm from S. rhodanthum inflorescences, which results in a c. 80% chance that birds will be detected by these cameras (Ortmann & Johnson, 2021). Cameras were initially set to record both day and night (24 h). However, in the absence of any observations of nocturnal pollinator activity from either direct observations or camera footage (Van der Niet et al., 2015), cameras were set to record during daylight only to prevent false triggering during night hours and thereby optimise chances of recording rare and irregular feeding bouts over long periods of several weeks. Video length was set to 30–60 s and intervals between trigger events were 5 s. Pollinator identity and behaviour were assessed by reviewing footage. A total of 22 interactions between sunbirds and S. rhodanthum were recorded between 2007 and 2021 (Table 1), of which five consisted of direct observations and the remainder of recordings from motion-triggered cameras. Visits were observed during 4 of 9 years, and at only two of the four sites at which S. rhodanthum occurs (Table 1). The recorded flower visits were between 04:47 AM and 06:24 PM, with more than half in the early morning hours before 08:30 AM (Table 2). At Highflats, all interactions involved a single sunbird species, C. amethystina. At Jolivet, the majority of interactions (N = 9) involved the sunbird species Cinnyris afer (Linnaeus, 1766), although these were never directly observed, whereas C. amethystina was recorded visiting three times, and one sunbird visitor could not be identified. Visitation by C. afer had not previously been recorded for S. rhodanthum (Van der Niet et al., 2015), but is unsurprising since the range of C. afer overlaps with that of S. rhodanthum and it is functionally similar to C. amethystina (Skead, 1967). Furthermore, both of these sunbird species are known pollinators of Satyrium neglectum ssp. woodii (Schltr.) A.V. Hall, which occurs in the same region as S. rhodanthum (Johnson & Van der Niet, 2019). Visits by C. afer were only recorded during 2 years in 2020 and 2021, despite 9 years of cumulative observation effort, spanning 37 observers and 257 camera days (Table 1). This study therefore emphasises the importance of performing pollinator observations over multiple years to characterise the full spectrum of flower visitors (Petanidou & Ellis, 1993) and the value of motion-triggered cameras for recording rare interactions. Pollinaria were clearly visible at the tip of the culmen in 12 of the 18 recorded interactions that involved flower foraging (Table 2, Figure 1), confirming that both sunbird species act as pollinators. Pollinaria that were attached to the tip of the culmen of a C. afer male clearly contacted a stigma of S. rhodanthum. Pulsing of the bird's throat indicates consumption of nectar, illustrating how sunbird feeding behaviour affects pollen transfer (Video S1). Characteristic grooming behaviour by sunbirds, attempting to dislodge pollinaria by swiping the bill against orchid stems (Johnson, 1996; Johnson & Van der Niet, 2019; Van der Niet et al., 2015), is also clearly evident in 5 of the 18 foraging events (Table 2), but not in four recorded interactions in which birds did not carry pollinaria, suggesting that pollinarium removal is the main purpose of grooming. Persistence of pollinaria despite grooming behaviour is consistent with suggestions that the large viscidia that characterise bird-pollinated Satyrium species, including S. rhodanthum (Van der Niet et al., 2015), function as an adaptation to avoid pollen loss as a result of bill-swiping behaviour (Johnson, 1996). Results from this study show that characterisation of the pollination system of a plant species may require extensive sampling over many years, encompassing several sites. Our results, in which we recorded 22 interactions over a combined 294 observer and camera days (an average of one interaction per 13 days), also suggest that standardised protocols for observing pollinators, such as the timed observations that are frequently implemented in pollination network studies, would almost certainly fail to document such infrequent interactions (Gibson et al., 2011). Instead, the use of motion-triggered cameras proved invaluable for documenting pollinator visits. We conclude that a combination of direct observations and use of motion-triggered cameras, which can be deployed repeatedly over entire days, is likely to be most effective for documenting highly specialised pollination systems. Kate and Graham Grieve discovered the Satyrium rhodanthum population in which most of the work for this study was conducted. Barry Cole allowed access and facilitated protection of a site where this endangered orchid occurs. James Harvey and Anina Coetzee assisted with sunbird identification. Ezemvelo KZN wildlife provided permits for this work (OP 4521/2016, OP 118/2018). The authors declare no conflict of interest. The data presented in this study are available from the corresponding author on request. 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.