The peninsula effect predicts that the number of species should decline from the base of a
peninsula to the tip. However, evidence for the peninsula effect is ambiguous, as different
analytical methods, study taxa, and variations in local habitat or regional climatic conditions
influence conclusions on its presence. We address this uncertainty by using two analytical
methods to investigate the peninsula effect in three taxa that occupy different trophic levels:
trees, millipedes, and birds. We surveyed 81 tree quadrants, 102 millipede transects, and
152 bird points within 150 km of coastal dune forest that resemble a habitat peninsula along
the northeast coast of South Africa. We then used spatial (trend surface analyses) and nonspatial
regressions (generalized linear mixed models) to test for the presence of the peninsula
effect in each of the three taxa. We also used linear mixed models to test if climate (temperature
and precipitation) and/or local habitat conditions (water availability associated with
topography and landscape structural variables) could explain gradients in species richness.
Non-spatial models suggest that the peninsula effect was present in all three taxa. However,
spatial models indicated that only bird species richness declined from the peninsula base to
the peninsula tip. Millipede species richness increased near the centre of the peninsula,
while tree species richness increased near the tip. Local habitat conditions explained species
richness patterns of birds and trees, but not of millipedes, regardless of model type. Our
study highlights the idiosyncrasies associated with the peninsula effectÐconclusions on the
presence of the peninsula effect depend on the analytical methods used and the taxon studied.
The peninsula effect might therefore be better suited to describe a species richness pattern
where the number of species decline from a broader habitat base to a narrow tip, rather
than a process that drives species richness.