This study explores the relationship observed between ‘guests’ – foreign inhabitants of social colonies – and the density and fecundity of eusocial-like staghorn ferns (Platycerium bifurcatum, Polypodiaceae). Our observations suggest that guests in staghorn colonies have a range of commensal and negative relationships, paralleling those seen in eusocial animal species.

The genus Platycerium comprises 18 epiphytic species with a pantropical distribution (Kreier & Schneider, 2006). The ‘common staghorn fern’ (P. bifurcatum) occurs on the Sunda Islands, Eastern Australia, Lord Howe Island and New Guinea (Hoshizaki, 1972; Kreier & Schneider, 2006). Our research focused on staghorn ferns occurring in Lord Howe Island (LHI), a volcanic island of <15 km² in area (31°33'S, 159°05'E), located app. 500 km east of Australia in the Tasman Sea. Although the island supports a diverse range of habitats (Harris et al., 2005), lowland areas are dominated by stunted tropical dry forest, where P. bifurcatum colonies are located within arm’s reach (Dawes et al., 2020).

To investigate the frequency and diversity of ‘guests’ in staghorn colonies, we inspected all colonies located within a 5-meter distance on either side of three narrow tracks winding through the island’s tropical dry forest (total track length approximately 5 km; Figure S1). Morphological differences between guests and staghorn colonies, as well as among guests themselves, facilitated their identification. We recorded the relationship between guest presence and density and fecundity of staghorn colonies by measuring three variables for each colony: 1) colony density, 2) per capita fecundity and 3) colony-level fecundity. These traits reflect information about colony health and spatial dynamics. Colony density was determined by dividing the total number of individuals by the colony’s volume (cm³) calculated via the product of width, height and depth measurements. To evaluate the fecundity-related parameters, we subdivided colonies into lower, middle and upper sections, reflecting differential reproductive patterns (Burns et al., 2021). From each section, three mature individuals were randomly selected, and the number of reproductive fronds were counted per individual following Burns et al. (2021). Per capita fecundity was then calculated by dividing the total number reproductive fronds by the number of assessed individuals (mostly nine, unless insufficient mature individuals were present). Lastly, colony-level fecundity was calculated by multiplying per capita fecundity by colony density. To expand our sample size of guest-inhabited colonies for analysing the relationship with their host, we also searched for colonies with guests off-track, in all accessible areas of dry forest. This approach allowed us to compile a dataset that likely encompasses most inhabited colonies on the island.

To assess differences in colony density, per capita fecundity, and colony-level fecundity among colonies inhabited by different guest species or not inhabited at all, we conducted three one-way ANOVAs using R v.4.0.4 with the ‘car’ package. Subsequently, we performed post hoc Tukey HSD tests to identify significant mean differences between inhabited and uninhabited colonies. To investigate colony density variations between on- and off-track locations, we employed a Mann-Whitney U test from the R base package. Similar tests were applied to fecundity metrics. As the outcomes of these tests were not significant for most colonies (Figure S2), both on-track and off-track colonies were analysed together to answer the second research question. All data were tested for normality and transformed where necessary.