The kriging method and other various linear estimators are compared when applied to the problems typical of environmental studies. It is shown that in a number of the rather typical situations the essential assumptions of the kriging method are violated and that other approaches can successfully compete with it.