than those in natural wetlands. The total nitrogen content of soil is usually positively related to organic matter content; hence, created wetland soils may contain less total nitrogen than do reference wetlands (Langis et al. 1991; Bishel-Machung et al. 1996). Wilson and Mitsch (1996) reported lower concentrations of phosphorus, potassium, calcium, and magnesium in constructed versus natural wetlands. Craft and colleagues (1991) reported lower concentrations of plant-available nutrients in a created salt marsh than in a reference salt marsh, which may have implications for plant primary productivity (Zedler and Langis 1991). Differences in the activity of soil microbes that control nutrient availability may also occur in created versus restored wetlands (Groffman et al. 1996).
Seeds and rhizomes retained in soil can remain viable for many years, germinating when conditions are right. Germination rates are a function of both environmental conditions (e.g., inundation, exposed mineral soils, and temperature) and seed conditions. For example, certain seeds have higher germination rates after a period of cold storage (stratification) or after abrasion of the seed covering (scarification; Thullen and Eberts 1995).
The presence of a seed bank can be either beneficial or detrimental. It is beneficial when desirable plant species naturally revegetate restored wetlands, reducing the cost and effort of planting programs (Mitsch and Wilson 1996; Mitsch et al. 1998). This approach is used most often in the restoration of preexisting wetlands (Galatowitsch and van der Valk 1996). Similarly, soil material with an intact seed bank can be transplanted from a wetland slated for destruction to a wetland creation site to promote revegetation (Brown and Bedford 1997). The presence of a seed bank can be detrimental, however; if it is dominated by weedy or invasive species, it would be viewed as undesirable. On several of the committee's site visits, the removal of soil was recommended for this reason. Planting of wetland vegetation in restored wetlands may not significantly influence the seed bank. Of 136 seed bank taxa found in soils from the shoreline of a reservoir constructed 6 years prior, only 10% were from species that had been planted during the restoration process (Collins and Wein 1995).
The experience at Hole-in-the-Donut in Everglades National Park (see Appendix B) illustrates the importance of restoring the substrate to achieve restoration of native plant communities. The soil that had been artificially created by rock plows during agricultural usage of the site provided substantially different conditions for plant growth than the oolitic limestone that supported the native sawgrass plant communities. Only by scraping off the artificially created soil was it possible to deter invasion by the invasive Brazilian pepper. When that was done, sawgrass regained its competitive advantage and was able to recolonize naturally without human intervention in the form of planting or weed control.