there will never be *along the diagonal of such a matrix, the diagonal is each sample paired with itself,so this obviously can never produce a significant difference
there will also never be a *in the top triangle of this matrix, since it is the mirror image of the bottom triangle That is ,if the intersection of the row for Australia and the column for Canada has a *then the intersection of the column for Australia and the row for canada need not also have a *since it is the same combination of samples