Table 1.
Definitions of local garden and landscape characteristics and transformation of data in the analyses and transformations used in correlation and regression analyses.
Table 2.
Species richness per garden (n = 35).
Fig 1.
Effects of urbanisation (distance to city centre; three classes), garden size (area with vegetation; three classes) and local garden characteristics on the species richness of six groups of invertebrates.
Plots show significant responses (P < 0.05) from GLMs (see Model 1 in S7 Table for more details). P-values for the response of the GLMs are shown. Displayed are deviance residuals for species richness from full models after stepwise reduction omitting the respective factor. This procedure corrected for other factors in the GLM. For native plant species richness, habitat richness and structural diversity, residuals from regressions of these factors on total garden area were used, because all three variables were correlated with garden size. Positive values in the bar plots indicate a higher than expected species richness. “–”indicates factors that were omitted from the models in the stepwise procedure. “ns” indicates factors that were retained in the model, but were not significant. For each family only represented by juvenile spiders, which were not identified to species level, we added an extra species to the count for the category spiders supplemented.
Fig 2.
Results of constrained analyses of principle coordinates visualizing similarities in species compositions of gastropods (a), spiders (b), millipedes (c), woodlice (d), ants (e), and rove beetles (f) in gardens located at different distances to the city centre (three classes; dark blue refers to gardens in the centre of the city, blue to gardens at intermediate distance and light blue to gardens at long distance from the city centre).