A comparison of summer goose counts for lakes in Flanders compared to the lake area, either with islands (dashed line) or without islands (solid line). The lines are the results of linear models of the log of the average individual count on a lake and the log of the area of the lake. The models assume a constant relationship between average individual count of geese and the lake’s area (A) Branta canadensis (R2 = .16, F(2, 119) = 10.98, p < .001) (B) Anser anser (R2 = .12, F(2, 118) = 8.16, p < .001) (C) Branta leucopsis (R2 = .09, F(2, 118) = 6.16, p < .01) (D) Alopochen aegyptiaca (R2 = .10, F(2, 118) = 6.77, p < .01). There is a significantly larger number of Canada (t = 3.79, p < .001), greylag (t = 2.22, p < .05) and barnacle geese (t = 3.42, p < .001) on lakes with islands. There is a significant positive relationship between the lake area and counts of Canada (t = 2.58, p < .05), greylag (t = 3.30, p < .001) and Egyptian geese (t = 3.58, p < .001).