who has noted some minor errors in one of the examples in my paper 'A simple test for spatial pattern in regional health data', Statistics in Medicine, 13, 1037-1044 (1994). Specifically, in the adjacency matrix for the counties of Ireland (p. 1042), the following cell entries (row, column) should be changed the entries for cells (E, L), (G, P), (T, S) and (U, L), which are shown as zeros (0), should be changed to ones (1); and cells (E, U), (G, 0), (T, E), and (U, E), which are shown as ones (l), should be changed to zeros (0).
These errors came about partly as a result of several errors and inconsistencies present in the original data table shown in Cliff and Ord (reference 6), and partly because of my misreading of their map; I had interpreted two small, narrow lakes to represent county boundaries, when in fact they do not.
When these changes are made to the adjacency matrix, minor adjustments occur in the calculation of the D statistic. We now have A = 55 (no change) and D = 394155 = 7-164. When combined with the unchanged mean E(D) = 8.667 and variance var(D) = 0.495, we obtain a new test statistic z = -2.14 (replacing the previous -2.03).
The other examples are correct as printed.