collections are addressed by chapter according to approach (table 4-2). since every variable of potential interest cannot be measured, a set of conditions and organisms, indicators of ecosystem health or impacts, must be selected. When selecting from several competing indicators, error relative to effect size provides a useful consideration. Alternatives to harvesting, including non-destructive techniques to estimate important indicators such as aboveground biomass, should be considered, several estimation methods are discussed in chapter 7. Most often, investigations will benefit from more than one type of information. For example, a survey to detect changes in an entire seagrass community could include this combination of approaches, plant collection for herbaria, a photographic record of the beds, and a quantitative underwater survey of the relative abundance of each seagrass species present. Monitoring programs also must consider different sampling scales in selecting methodologies, an excellent review is given by Virnstein.