“…However, it should be noted that experience is still very limited, both in number and in iength of experiments. Problems with such a design might be expected to increase as an experiment continues since the divergence of replicate systems may increase with time Smith et al 1982!. This tendency, however, may be largely neutralized if, as has thus far occurred, the tendency for nutrient loading to control ecosystem behavior continues to reinforce itself by radically constraining potential behavior.…”