guidelines have been offered. Schaeffer (in this volume) suggests that engineers adopt the medical oath to "do no harm," in this case to the environment. Norton (in this volume) suggests that engineers adopt the Pareto welfare criterion of economics, in which case an engineering design would be considered satisfactory if its effects were positive for some individuals and neutral for all others. Holling (1992) argues that efforts to identify ecological constraints may benefit from a focus in which the environment and its many components (including humans) are systematically viewed as a hierarchy of systems within systems.
Daly and his colleagues (Costanza et al., 1991; Daly, 1990) have suggested acceptable boundaries for human environmental impacts. Rates of extraction of renewable resources should not exceed regeneration rates. Rates of waste emission should not exceed the assimilative capacity of the environment.2 Rates of extraction of nonrenewable resources should not exceed the rates at which substitutes are found and developed. These guidelines imply a variety of long-term performance standards. For example: pumping from aquifers should not exceed recharge rates; pollutant concentrations should not increase; soil depth should not decline; harvesting should not cause reductions in population sizes. These guidelines can be useful for identifying unsatisfactory circumstances, but it remains to be seen whether they will be elaborated in ways that will make them directly useful to engineers struggling to satisfy particular ecological constraints in the context of particular engineering problems.
Available evidence suggests that when ecological constraints are clearly defined, engineers can develop designs or management plans with the potential for meeting them. Shen (in this volume) describes two such examples. On the Niobrara River the challenge was to release water from a dam such that the downstream reaches of the river would maintain the broad shallow morphology that is required by migrating whooping cranes. In south Florida the challenge was to restore the Kissimmee River such that flood waters would inundate the floodplain frequently and return to the river channel slowly. In both cases seemingly satisfactory water management plans were developed. Lindstedt-Siva et al. (in this volume) describe another example from ARCO's experience of exploring for oil in tropical rain forests. Roads built into remote rain forest regions facilitate human immigration and subsequent forest destruction. Lindstedt-Siva and her colleagues were determined to explore for oil without building roads. They took their lead from offshore exploration operations, and used helicopters rather than trucks to move equipment.
These examples suggest potential for fruitful collaboration between ecologists and engineers. In all three examples, recognition of important ecological processes led to the identification of key ecological constraints, and engineers used those constraints to develop designs that appeared capable of satisfying them. Plans for the proposed dam on the Niobrara River were canceled because of budgetary constraints, and it is too soon to assess the success of the other designs. In addition, satisfaction of particular ecological constraints is not neces-