the average daily water temperatures to which organisms are exposed. A species in a warmer downstream reach typically matures earlier than does an upstream population of the same species and may actually have multiple generations in a year compared to only one generation a year at the colder upstream site (Lehmkuhl, 1973; Newell and Minshall, 1978; Mackay, 1979). In temperate climates, the period of maximum community maturity and richness typically ranges from late fall to early spring, depending upon local and regional factors that influence temperature, such as elevation and latitude. In addition to this fall-winter community, most temperate streams have a spring-summer community that is generally less rich taxonomically but whose period of maximum maturity closely approximates summer low-flow periods. Characterization of both fall-winter and spring-summer communities is desirable. However, the choices of when and how often to sample must be made in conjunction with the other elements of the NAWQA Program.
Because of the importance of water temperature in insect development, the timing of sample collection must take water temperature into consideration to account for developmental differences among sites. Water temperatures, at least daily maximum and minimum, need to be monitored at ecological survey sites associated with basic fixed sites. These data are useful in scheduling sampling activities, making comparisons within and among basins, and interpreting study-unit, regional, and national differences.
Current and antecedent discharge conditions must also be considered in determining the appropriate time to collect samples. Access to the sampling site can be limited during seasonal high-flow conditions. Unusually high flows can wash out or bury substantial parts of the benthic invertebrate community and redistribute, create, or eliminate instream and riparian habitats within the sampling reach (Irvine, 1986; Resh and others, 1988; Poff and Ward, 1989). Invertebrate communities require a period of time to adjust following such unusually high discharges. Therefore, sampling should be delayed approximately 4 weeks following a discharge event with a recurrence interval greater than 5 years. This delay reduces the likelihood of misinterpreting channel habitat characteristics and making errors such as sampling substrates that were dry prior to the increase in flow. Extreme flows with recurrence intervals greater than 10 years can so alter the physical conditions in the sample reach that many years or de cades can be required for communities to return to "normal" conditions. In such cases, recovery of the community probably would not occur within the current NAWQA Program cycle, so additional delays in sampling would make little difference in assessing the community. Subsequent interpretation of community structure in these circumstances must take into consideration the unusual antecedent discharge and altered habitat conditions. Selecting a season of minimum-flow variability for ecological sampling minimizes the probability of encountering an extremely high discharge before or during sampling.
The logistical challenges of organizing field crews, especially when arrangements have been made for assistance from other agencies or researchers, can cause major difficulties in implementing the strategy of maximizing community development and minimizing the influence of antecedent high flows. Postponement of a field effort can be complicated by an even greater flow in the intervening period, especially when local climatic conditions are highly variable. Also, flow conditions are seldom uniform across basins; thus, part of a basin