Summary Detritus processing by a small woodland stream is analysed by following the loss of weight of 10 g, single species accumulations of riparian leaves. The daily loss rates are expressed as exponential coefficients after the data are fitted by least squares. Comparisons are made between two sites on a small hardwater trout stream during two seasons. Leaf processing rates form a continuum from a low of 0.5%/day to a high of 2.0%/day. Differences between species of leaf are observed, but significant differences between fall and winter processing and between the two sites are not. The response of the invertebrate community to differences in leaf species is investigated using controlled, artificial streams where significant differences in the effect of the invertebrates are related to the ability ofthe leaf to be processed. Evidence suggests that differential invertebrate colonization of leaf packs is a function of microbial colonization and conditioning. The data are used to develop a general scheme of leaf pack processing.

Studies were conducted in four distinct geographic areas (biomes/sites) in northern United States to examine changes in key ecosystem parameter: benthic organic matter (BOM), transported organic matter (TOM), community production and respiration, leaf pack decomposition, and functional feeding—group composition along gradients of increasing stream size. Four stations ranging from headwaters (1st or 2nd order) to midsized rivers (5th to 7th order) were examined at each site using comparable methods. The results for each parameter are presented and discussed in light of the River Continuum Concept of Vannote et al. (1980). The postulated gradual change in a stream ecosystem's structure and function is supported by this study. However, regional and local deviations occur as a result of variations in the influence of: (1) watershed climate and geology, (2) riparian conditions, (3) tributaries, and (4) location—specific lithology and geomorphology. In particular, the continuum framework must be visualized as a sliding scale which is shifted upstream or downstream depending on macroenvironmental forces (1 and 2) or reset following the application of more localized "micro"—environmental influences (3 and 4). Analysis of interactions between BOM and TOM permitted evaluation of stream retentiveness for organic matter. Headwaters generally were most retentive and downstream reaches the least. Estimates of organic matter turnover times ranged between 0.2 and 14 yr, and commonly were 1—4 yr. Both turnover times and distances were determined primarily by the interaction between current velocity and stream retention. Biological processes played a secondary role. However, the streams varied considerably in their spiraling of organic matter due to differences in the interplay between retentiveness and biological activity. Differences in the relative importance of retention mechanisms along the continuum suggest that headwater stream ecosystems may be functionally more stable, at least to physical disturbances, than are the r intermediate river counterparts.

Four significant areas of thought, (1) the holistic approach, (2) the linkage between streams and their terrestrial setting, (3) material cycling in open systems, and (4) biotic interactions and integration of community ecology principles, have provided a basis for the further development of stream ecosystem theory. The River Continuum Concept (RCC) represents a synthesis of these ideas. Suggestions are made for clarifying, expanding, and refining the RCC to encompass broader spatial and temporal scales. Factors important in this regard include climate and geology, tributaries, location-specific lithology and geomorphology, and long-term changes imposed by man. It appears that most riverine ecosystems can be accommodated within this expanded conceptual framework and that the RCC continues to represent a useful paradigm for understanding and comparing the ecology of streams and rivers.

Benthic community metabolism was studied on four stream systems located in different biomes in the United States: the eastern deciduous forest (Pennsylvania, PA, and Michigan, MI), the high desert (Idaho, ID), and the coniferous forest (Oregon, OR). Studies were designed to test the hypothesis advanced within the River Continuum Concept that a transition in community metabolism will occur from a predominance of heterotrophy in headwaters to a predominance of autotrophy in mid-sized reaches, with a return to heterotrophy further downstream. Both gross primary productivity (GPP) and community respiration (CR 24 ) increased with downstream direction on all systems. Net daily metabolism (NDM, or GPP -CR 2 4 ) shifted from heterotrophy (-NDM, GPP < CR 24 ) to autotrophy (+NDM, GPP > CR 2 4 ) with downstream direction at all sites, supporting the hypothesis. Annual metabolism in the most upstream reach of all sites was dominated by respiration; however, the farthest downstream reach was not necessarily the most autotrophic. Site-specific factors affected manifestation of the trend. Photosynthesis predominated annual metabolism in reaches (designated 1-4 in order of increasing size) 2-4 in ID, 3 and 4 in OR, and 4 in MI. In PA annual photosynthesis was slightly greater than respiration only at Station 3. Photosynthesis was predominant most consistently in ID and respiration most often in PA. About half the reaches that were heterotrophic annually were autotrophic at one or more seasons. Annual means of benthic GPP, CR 24 and NDM ranged from 0.16 to 3.37, 0.36 to 2.88 and -0.73 to 0.50 g 02 m 2 d 1, respectively. Metabolic rates were usually high in PA and Ml (and sometimes ID) and almost always lowest in OR. Parameters accounting for most variance in multiple linear regression analyses of the combined metabolism data from all sites were indicators of stream size, photosynthetically active radiation, temperature, and chlorophyll a concentration.