Depth-integrated, discharge-weighted water samples were collected over 1,800km ofthe Amazon River on eight cruises at different stages of the hydrograph, 1982-1984 and coarse (CPOC, > 63 rm) particulate organic carbon as weight percentage of suspended sediment varied between 0.9-1.5% for FPOC and 0.5-3.49/o for CPOC. Concentrations of FPOC ranged from 5 mg liter-' upriver to 2 mg liter-' downriver in the mainstem and from 6 mg liter-' inthe Rio Madeira to i 1 in the Rio Negro. CPOC had similar distribution patterns. but with concentrations 15.-30% those of FPOC. Dissolved organic carbon (DOC) averaged 4-6 mg liter-r in the mainstem and up to 12 ml; liter -I in the Rio Negro. Upriver dissolved inorganic carbon (DIC) concentrations of about 1,200 @LM were diluted by tributaries and floodplain drainage to 600 PM at the most downriver site. Evasion ofCO,, invasion of O,, and in situ oxidation were of comparable magnitude, 3-8 pmol m-2 s-r.The average export of total organic carbon (TOC) was 36.1 Tg yr-i (8.5 g m-2 yr-I), of which 62% was DOC, 34% was FPOC, and 4% was CPOC. TOC inputs were insufficient to support in situ oxidation by a factor of at least two. A relatively small, rapidly cycling pool of labile organic matter may coexist with a much larger pool of more refractory material.A central problem in riverine ecology is to determine how the dynamics of carbon vary as flowing waters increase in size from first-order springs and seeps to the world's great rivers. With the advent of the energy r To whom correspondence should be addressed.

Effective management of the Amazon's commercial fish populations requires an understanding of the factors controlling their production. A fundamental step in the investigation of fish production is to identify the plant groups that contribute energy to fish foodwebs.Stable isotope data for plants and 35 fish species were used to identify autotrophic carbon sources for the central Amazon fish community. Adult fish, aquatic macrophytes, tree parts, periphyton, and phytoplankton were collected in lakes and other flooded environments along the central Amazon floodplain and analyzed for carbon stable isotope composition (o 13 C) by mass spectroscopy. o 13 C values for plants ranged from -39.4 to -11.9o/oo with averages of-33.3, -28.8, -27 .6, -26.2, and -12.8o/oo for phytoplankton, flooded forest trees, C 3 aquatic macrophytes, periphyton, and C macrophytes, respectively. The average for all C 3 plants (phytoplankton, trees, C3 macrophytes, and periphyton) was -29.1o/oo, while the average for C 4 plants (mainly C 4 macrophytes) was -12.8o/oo. Mean o 13 C values for adult fish ranged from -37.0 to -19.8o/oo with an average of-28.8o/oo. Fish and plant data were used in an isotope mixing model to estimate the relative contribution of different plant groups to fish carbon. C 4 macrophytes, which contributed over half of the primary production on the floodplain, accounted on average for only 2.5-17.6% (minimum to maximum) of the carbon in fish. The C 3 plants, as a group, were the primary carbon source for 34 fish species, and accounted for an average of 82.4-97.5% of the carbon in all species. Phytoplankton, a minor C 3 producer, accounted for a minimum of 36.6% of fish carbon on average, and was the principal carbon source for the commercially important characiform detritivores. Several alternative hypotheses are proposed to explain the apparent selective transfer of C 3 carbon through Amazon fish foodchains.

Concentrationsof COZ, 02, CH4, and N,O in the Amazon River system reflect an oxidationreduction sequence in combination with physical mixing between the floodplain and the mainstem. Concentrations of CO, ranged from 150 PM in the Amazon mainstem to 200-300 PM in aerobic environments and up to 1,000 PM in oxygen-depleted environments of the floodplain. Apparent oxygen utilization (AOU) ranged from 80 to 250 PM. Methane was highly supersaturated with respect to atmospheric equilibrium. Concentrations ranged from 0.06 PM in the mainstem to 100 PM on the floodplain. Concentrations of N,O were slightly supersaturated in the mainstem (-13 nM) but were undcrsaturated on the floodplain (averaging 9 nM). Fluxes calculated from these concentrations indicated decomposition of -1,600 g C m-2 yr-' of organic carbon in Amazon floodplain waters. Analysis of relationships between CH,, 02, and CO, concentrations indicated that about 50% of carbon mineralization on the floodplain is anaerobic, with 20% lost to the atmosphere as CH,. The predominance of anaerobic metabolism leads to consumption of N20 on the floodplain. Elevated concentrations of CH, in the mainstem probably reflect input from the floodplain, while high levels of CO, in the mainstem are derived from a combination of floodplain drainage and in situ respiration.-The rates of oxidation and reduction of carbon in rivers change dramatically with increasing river size (Vannote et al. 1980). Small, temperate streams are relatively well described; the sequence and extent of metabolic processes operating in very large rivers is essentially unknown. Many large river l To whom correspondence should be addressed.

Detritivorous fishes form an important part of the ichthyomass in the Amazon basin. Most of these fishes are contained in the orders Characiformes and Siluriformes (catfishes). The Characiformes constitute more than 30% of the total fish yield in the Amazon basin, whereas the catfishes are of minor importance. Stable isotope data indicate that Characiformes species receive most of their carbon through food chains originating with phytoplankton, while the Siluriformes receive a significant part of their energy from other plant sources.

Using a simple isotope mixing model, we evaluated the relative proportion of water vapour generated by plant transpiration and by soil evaporation at two sites in the Amazon basin. Sampling was carried out at two different soil covers (forest and pasture), in a seasonal tropical rainforest at eastern Amazon where major deforestation is the result of land‐use change, and compared to a less seasonal central Amazon forest. In both forests, vapour from transpiration was responsible for most, if not all, of the water vapour generated in the forest, while it could not be detected above the grassy pastures. Thus the canopy transpiration may be a major source of water vapour to the forest and perhaps to the atmosphere during the dry season. The results are discussed in relation to predictive models based on net radiation that usually are not able to distinguish between transpiration and evaporation.

Summary Several studies have shown that land use has a strong influence on river chemistry and its biotic components. Most of these studies focused on nitrogen in temperate American and European catchments. Much less is known about the relationship between stream conditions and land use in tropical areas of developing countries. Besides climate, there are three important differences between attributes of temperate and tropical catchments: non‐point sources are the dominant contributor of pollution in USA, whereas point source pollution is the most important in our study; use of fertilizer is much smaller in developing countries, and the type of agriculture and management practices are distinct. We test whether the chemical composition of streams and their macroinvertebrate communities can be related to land use. Accordingly, we compared the variability of chemical composition and macroinvertebrate communities in the streams of two catchments (Pisca and Cabras) belonging to the same ecoregion, but having different types of land use. The main land use in the Pisca catchment in 1993 was sugar cane (62%), followed by pasture (22%) and urban centres (10%). In contrast, the main land use in the Cabras catchment was pasture (60%), followed by annual crops (13%) and forest (10%); urban centres occupied only 2% of the catchment. In the Cabras catchment, most of the parameters correlated with a land use index (LUI) ( Fig. 2). However, only conductivity, major cations and major anions (with exception of sulfate) had a statistically significant correlation coefficient. More than 90% of the variance was explained for these parameters. DIC, NO3 and richness of invertebrates (RI) also strongly correlated with LUI (R2 = 0.75), although these correlation coefficients were not significant. Total suspended solids (TSS) had a significant correlation with LUI (R2 = 0.98), but, the correlation was inverse. In the Pisca catchment, conductivity, major cations (with exception of potassium), major anions, and DIC, DO, and DOC had a strong and statistically significant correlation with LUI. Correlation coefficients were also high for respiration rate, although the correlation was not statistically significant. 2 Relationships between variables and LUI (land use index) for the Cabras (closed circle) and Pisca (open circle) catchments. Both catchment were pooled together in this figure, however, statistical tests were performed separately for each catchment.

We describe the sources and routing of the Amazon River flood wave through a 2000‐km reach of the main channel, between São Paulo de Olivença and Obidos, Brazil. The damped hydrograph of the main stem reflects the large drainage basin area, the 3‐month phase lag in peak flows between the north and south draining tributaries due to seasonal differences in precipitation, and the large volume of water stored on the floodplain. We examined several aspects of the valley floor hydrology that are important for biogeochemistry. These include volumes of water storage in the channel and the floodplain and the rates of transfer between these two storage elements at various seasons and in each segment of the valley. We estimate that up to 30% of the water in the main stem is derived from water that has passed through the floodplain. To predict the discharge at any cross section within the study reach, we used the Muskingum formula to predict the hydrograph at downriver cross sections from a known hydrograph at upstream cross‐sections and inputs and outputs along each reach. The model was calibrated using three years of data and was successfully tested against an additional six years of data. With this model it is possible to interpolate discharges for unsampled times and sites.

-The 13C : l*C of suspended particulate organic C (POC), dissolved organic C (DOC), and dissolved inorganic C (DIC) were measured during 1982-l 984 at 11 main-channel and 7 tributary stations over an 1,800 km reach of the Amazon River between Vargem Grande and Obidos, Brazil. The measured 613C range vs. marine carbonate (PDB) was -32 to -269~ for suspended POC, -30 to -289m for DOC, and -26 to -129~ for DIC. The 613C of the fine particulate organic C (FPOC) decreased downriver from Vargem Grande, with values lowest during the fallingwater portion of the runoff cycle; these trends were the result primarily of input of IC-depleted FPOC from tributaries draining the lowland regions of the Amazon basin and floodplain soils. The 613C of the FPOC at Obidos implies that at least 35% of the POC exported by the Amazon River is derived from the lowland portion of the Amazon basin. The 613C of DIC decreased downriver with the lowest values measured during falling water; these trends were due primarily to within-river respiration and tributary input. The 613C of the DIC suggests that -40% of the organic matter being respired in the river is C, plant material derived from floodplain grasses.Within its basin, the discharge of chemically diverse tributaries is blended by the Amazon River, resulting in main-channel concentrations of chemical species and sediment that are close to the world average for rivers (Stallard and Edmond 1983). In the main channel of the river, we found that -20% of the total C was particulate organic