Hypothesis 3: Species interactions are stronger in mature islands
Our third goal was to investigate the relative importance of species
interactions in islands of different ages. Our overall hypothesis was
that species interactions, and not the mere accumulation of species,
reshape the community structure in mature islands leading to an
optimised use of the available niche space—explaining why the
contribution of individual species might be reduced in mature islands.
In this scenario, we would expect that functional distance across pairs
of species will exert a stronger effect on both ecological and
geographical niches in mature islands. We do not expect to find such a
strong effect in young islands whose diversity dynamics are dominated by
speciation and where species interactions play a weaker role shaping
their diversity; nor in senescent islands, where species interactions
might play a marginal role, depending on how extinction affects
different communities.
To analyse the extent of functional differentiation between coexisting
species, we modelled the functional distance of each pair of species
co-occurring in the same island as a function of their geographical and
ecological range overlap. We calculated the geographical range overlap
as the distance between the centroids of the distribution area of each
pair of species. For environmental niche overlap, we extracted seven
bioclimatic variables from WorldClim 2 database (Flick et al. 2017)
using the values at the coordinates of the cave entrance or MSS traps
(see Supplementary methods). All extracted variables were highly
correlated with altitude (all Pearson r correlation
> ±0.7), and therefore we used the overlap of the
altitudinal range of each species as a proxy for niche overlap. We
modelled the correlation between the ecological and geographical
distances and the functional distance between pairs of species within
each of the islands using partial Mantel tests using the functionmantel.partial in the R package vegan version 2.5 (Oksanen et al.
2020),