Discussion
We investigated here for the first time components determining the chafer assemblage composition, comparing the impact of ecoclimatic influences with macrohabitat and locality stochastics on the similarity of investigated entities. Locality stochastics represent a not further investigated multi-factor ensemble that includes all biotic and abiotic environmental conditions at local scale such as macrohabitat, biogeography, edaphic conditions, land use, predation, local climate, rainfall, radiation, and others. We also explored the patterns of lineage membership and body size resulting from assemblage composition across larger scale entities (forest type, elevation) versus smaller scale entities (localities, macrohabitats).
The comparison of chafer assemblage composition at different eco-spatial scales revealed that assemblages were shaped mainly by locality stochastics, to minor extent to the ecoclimatic conditions, and not by macrohabitat. This was true for the entire chafer assemblage, as well as single lineages or different body size classes. NMDS plots of faunal similarity showed the largest overlap among macrohabitat entities. In contrast to that, overlap for clusters of forest types, elevation zones, and localities were limited. However, contrasts between localities were less pronounced in medium-sized and large specimens.
Investigated macrohabitats were quite different (e.g., forest, grassland, abandoned plantations). They are known to provide multiple niches (Bosc et al. 2019) for chafer species, however, only a few species were recorded that were specific to these habitats. Most species and resulting assemblages sorted rather by locality rather than by macrohabitat. This could be partly explained by the trapping method (light traps) used, as fully winged chafer beetles may be attracted from other habitats over certain distances within the same locality. However, the fact that we found no correlation between species composition (for total assemblage) and geographic distance (Figure 5), even for adjacent localities situated in the same forest type also in the same mountain range (e.g., L2, L4), may indicate either that species generally might tend to disperse also over moderate to longer distances or that species disperse very little. Limited dispersal is supported further by molecular evidence (Ranasinghe et al. in review), since different, the same here investigated localities shared almost no haplotypes. This latter conclusion would be not surprising as previous studies have also shown high turnover rates of assemblage composition at higher elevations independently from geographical distance (García Lopez et al. 2010). However, the resulting significant correlation for the assemblage of small-bodied specimens, which is definitively linked to their limited dispersal capacity and mirrored by their higher endemism (Fabrizi and Ahrens 2014), might indicate that lacking significance on our study might be a result of an insufficient number of samples and species. Larger species were generally less common and are also less represented in higher altitudes. Influence from palaeogeographical and biogeographical factors should also be considered in this context (Kemp et al. 2017) as several sampling localities are situated in the central highlands within complex mountain systems (escarpments, ridges, or peaks) which can act as geographical barriers. The latter can particularly triggered geography-driven speciation, as shown by diversification of Sericini in Asian mountains (Ahrens 2007; Eberle et al. 2016).
Some of the divergent composition patterns retrieved for the full assemblage (Figure 3A, F), which are in turn not encountered for any of the single lineages, reveal that occurrences of entire lineage members may also impact on the apparent differentiation (e.g., wet lowland forest vs. submontane forest, EZ 1 vs EZ 2). The latter case is caused by the more poorly sampling/ absence of larger-bodied species (e.g., Dynastinae), in higher elevations, since low temperatures obviously might not favour larger species with long larval development (Danks 1992). In fact, even in mountain ranges with larger amplitudes of elevations, the altitudinal differentiation of the fauna in phytophagous chafers is rather poor (Ahrens 2004) compared to other insects (Mani 1968).
The strong turnover for localities is in line with the rather high degree of endemism in many phytophagous scarabs (Ahrens and Fabrizi 2016), despite their considerable size. Their assemblage patterns across local spatial scales can be explained not only by poor dispersal capacities, but also by short emergence times compared to the length of their life span: their root feeding, endogenous larvae do not disperse. Their emergence during early night-time often falls together with heavy monsoon rains which narrows down the time window for potential dispersal flights.
Other lineages composed of larger species, such as Dynastinae, have greater dispersal ability compared to smaller Rutelinae and Melolonthinae (García Lopez et al. 2013), and this has an impact on the faunal divergence pattern of assemblages as revealed by pronounced larger cluster overlaps across different spatial scales.
Seasonality and weather fluctuation may strongly impact the expressed patterns of assemblage composition in ecofaunistic analyses (De Oliveira et al. 2021). In tropical climate, rainy seasons and dry seasons are alternating in shorter intervals with quite constant temperature and humidity throughout the year and food resources being continuously available. Thus, minor fluctuations to species’ presence and numbers may occur even in the tropical ecosystems. Many of our localities (except L1-L3, L9) did not show a significant seasonal species turnover, while those which did experience generally stronger dry-wet fluctuations than other localities according to their position in the island.
In final conclusion, we need to remember that at local level all ecological, climatical, and spatial components sum up in their effect increasing the complexity of influences on the assemblages. This points the way for future, more detailed studies, in which localities of similar eco-spatial situations shall be addressed. Yet, since phytophagous chafers are for many tropical crops common pests, and damage can often also be caused by a multispecies autochthonous community with endemic species, (Ahrens et al. (2009), we need more knowledge here, as this might positively affect simultaneously pest and biodiversity management.