[R-sig-Geo] Using spatial autocoregressive models

[Tim Whitfeld]

I'm new to the list and am hoping for some assistance with analysis for 
a paper that investigates patterns of tree diversity in the tropics. For 
this project I surveyed all trees in a 19 quarter hectare plots in a 
tropical lowland rainforest. I'm interested in testing whether the 
observed correlation between patterns of diversity (measured by mean 
evolutionary distance between all trees in each plot) and total tree 
area (measured by total cross sectional area of tree trunks at 1.3 m 
above ground level, i.e., the basal area) is confounded by plot location 
(measured by UTM coordinates). In other word, I want to test whether 
spatial auto correlation is driving the observed relationship. I have 
performed a partial Mantel test that indicates no spatial 
autocorrelation but I'd also like to test this with a more sophisticated 
approach using simultaneous autogressive models. I understand SAR models 
can be implemented in R using the spdep package. I have used R with 
other packages but am no expert. Does anyone have an example or a walk 
through of how to run the type of analysis I'm interested in and also 
how to interpret the results?
With thanks, Tim