hello list, I have N agricultural fields (some of them share borders, some don't) with ph as the variable of interest. Since I want to test the spatial autocorrelation, I need to come up with the w matrix. The first and easy way that came to my brain was the binary relationship, i.e., the w matrix is all 1s and 0s with 1 meaning two fields sharing border and 0 not. So I checked the distribution of the N fields, and I manually defined a NxN matrix with 1s and 0s. I list the variable of interest (ph) as a vector with N elements, and plan to use moran or moran.test to test the spatial autocorrelation. However, I had difficulty to run the code. Anyone has suggestions about my case where the areal/lattice data are irregular? Please also let me know if moran or moran.test is the correct command to use. Thanks. XY