To what extent does Tobler's 1st law of geography apply to macroecology? A case study using American palms (Arecaceae)

Table 3 Multiple regression analyses of species richness (r) and species composition (c)

	β _(r)	R ² _(r)	β _(c)	R ² _(c)
*The Americas*		0.16^D		0.56^D
Environmental distance	0.365		0.259
Geographic distance	0.092		0.619
*Amazon*		0.73^A		0.81^C
Environmental distance	-0.074		0.311
Geographic distance	0.896		0.644
*Andes*		0.30^C		0.44^D
Environmental distance	0.576		0.578
Geographic distance	-0.083		0.159
*Caribbean*		0.23^C		0.74^D
Environmental distance	0.526		0.088
Geographic distance	-0.132		0.818
*C. America*		0.31^D		0.53^D
Environmental distance	0.427		0.344
Geographic distance	0.225		0.522

The standardized regression coefficients (β) for the best models are given. Significance levels were tested using 999 permutations. (p-values are not indicated as all results were significant (p < 0.001) due to the large sample size). The distance matrix on species richness has been calculated using Euclidean distance and the distance matrix on species composition has been calculated using D = 1- Sørensen Index. Four combinations of environmental and geographical matrices have been used and the combination for each dataset giving the best model is shown here. The letters refer to:
^A) All environmental variables including precipitation (mm yr^-1), number of wetdays (yr^-1), mean annual temperature (°C), number of vegetation types, topographic range, pH, sand (%), Ca²⁺, and CEC; linear geographic distance measured in kilometres.
^C) Climatic related variables including precipitation (mm yr^-1), number of wetdays (yr^-1), and mean annual temperature (°C); linear geographic distance.
^D) Climatic related variables; ln-transformed geographic distance.

ISSN: 1472-6785