tinue to work together to develop suitable engineering plans. Engineers and ecologists certainly can not solve these problems alone, but they do have important responsibilities in efforts to keep human environmental impacts within acceptable bounds.
With the exception of selected subdisciplines (Mitsch, in this volume), there is so little dialog between engineers and ecologists that it is hard to know where to start. What would it mean to engineer within ecological constraints? What are the ecological constraints? Which constraints are most important? Can some be ignored? How would one distinguish satisfactory and unsatisfactory designs? What does it mean to keep human environmental impacts within "acceptable" limits? Four related topics arise frequently during discussion of these broad questions: problem definition, uncertainty, lessons from environmental control efforts, and the difficulty of finding short-term solutions that do not aggravate long-term problems. We summarize key themes regarding these four issues below.
"Everything should be made as simple as possible, but no simpler."
—attributed to Albert Einstein
A plan is not likely to be successful if the problem that it is intended to solve is not accurately defined. Whether ecological constraints are met will depend upon whether the definition of the problem included those constraints. Moreover, when substantial consequences of a project, product, or management plan are not appreciated at the design stage, costly problems may arise later. For example, almost as soon as the Kissimmee River channelization project was completed, the severe ecological consequences became apparent and much more expensive work was begun to reverse the damage (Shen, in this volume; Wodraska and yon Haam, in this volume). Other examples could be cited, including the impacts of dams and forestry practices upon salmon and other fish, the tendency of pesticides to kill a pest's natural predators and lead to pesticide-resistant pests, transportation of exotic species by ships and airplanes, impacts of chlorofluorocarbons (CFC) on stratospheric ozone, and even the impact of carbon dioxide emissions on the global atmosphere. This is not to suggest that the activities that led to these problems should not have occurred, merely that had these consequences been appreciated and taken into consideration from the outset, it might have been possible to reduce or eliminate impacts through design or management modifications.
As the problem-definition stage of engineering efforts continues to expand in response to environmental concerns, and as our dependence on ecosystem "services" becomes better understood (Cairns, in this volume), these sorts of problems should become less frequent. To this end, several general criteria and