I would like to thank everyone who responded to my posting - all of the
responses were helpful. I was interested in calculating spatial
autocorrelation coefficients for town-level data. In this case the data
represents the relative abundances of tree species in the state of
Massachusetts, USA. I have decided to convert the town polygons into
centroids for the analysis since lag distances are more meaningful than
adjacencies in this case.

The recommendations I received include:

Michael Batcher suggested a free program called Rook's Case
http://www.uottawa.ca/academic/arts/geographie/lpcweb/sections1/software/sawsoft/usrmanul.htm
this is a Microsoft Excel Add-in; it is very easy to use, and is what I
will probably use when I do the analyses.

Max Lu suggested S+ SpatialStatistics (an ArcView extension is available).
http://www.insightful.com/products/arcview/arcviewfl.html He also suggested
SpaceStat
Once my office computer guru returns from vacation, I will look into S+
SpatialStatistics vs SpaceStat.

Lionel Davoust uses R http://www.r-project.org/, which is a public domain
version of S+. My office has S+ so I might as well go with the original
version.

Craig Hanson and Huizhen Niu use Luc Anselins SpaceStat; it looks like a
powerful and easy to use program since it can link with ArcView.
http://www.spacestat.com/ I am going to compare SpaceStat and S+
SpatialStatistics for later analyses... both look great.

Ben Wheeler suggested a website
http://ftp.ukbf.fu-berlin.de/sozmed/DismapWin.html I looked at it but found
that the commercial packages are much simpler to use

Again, Thanks to all who responded,

Brian

___________________
Brian Hall
PO Box 68
Harvard Forest
Petersham, MA 01366

(978) 724-3302 ext 254
brhall@fas.harvard.edu
____________________