been virtually no research that compares the relative similarity of undisturbed reference sites, it is hard to make firm conclusions on this point. In their analysis of reference site selection, White and Walker (1997) found that there is so much variation on a regional scale that there may never be a perfect match for a site to be restored. They propose that a conceptual restoration model should be one of interpolation among multiple sites and sources of information, including temporally. Finally, Neiring (in Kusler and Kentula 1990) posits that the end point of the successional process cannot be predicted because it is not an orderly process; that is, we cannot predict what exact species will be at the “end” of the process. In summary, it appears that there is no general trajectory for the development of wetland ecosystems or individual components (structural or functional attributes) of a single watershed.
The significance of these results is not that equivalency among reference and newly managed environments is not reached or that mitigation efforts should not be done. These results demonstrate that (1) ecological equivalency may not be reached within a few months or for several years or even decades, depending on the attribute that is of interest; (2) the ecosystem does not move smoothly to an equilibrium or at the same rate for all components; and (3) some components, including ones identified as important in permits currently being issued, may never reach equivalency with the natural reference wetland. An obvious conclusion from these results, besides the general paucity of scientific analyses, is that the generally observed 5-year limit on monitoring is insufficient when evaluating whether a site has achieved parity with a reference system. Further, the amount of mitigation required should be based on the amount needed to fully offset the permitted wetland losses. To accomplish that, mitigation ratios (area of mitigation to area lost) will often need to be increased to achieve functional equivalency, rather than simply matching wetland area.
The above review of wetland restoration and creation outcomes has not differentiated projects on the basis of their starting conditions. Some of the projects discussed occurred at sites where damages were relatively minor (e.g., Cooper et al. 1998); others were at highly altered sites with artificial (dredge spoil) substrate (e.g., Zedler and Callaway 1999). As discussed earlier, it is very likely that the ability to achieve desired outcomes is a function of the degree of site degradation. Degradation, in itself, is a complex concept, involving not only the local site but also its watershed. Zedler (1999) suggests that generalizations are not easily drawn among restoration efforts involving sites with different degrees or types of degradation, suggesting instead that ecologists seek predictability of outcomes in a matrix of situations, where degree of degradation is one axis and the type of restoration effort is the other. The general pattern