We examined primary production, respiration, and nutrient dynamics in littoral areas of the mainstem and lagoons of the Cinaruco, a nutrient-poor river in the Venezuelan llanos. Gross primary productivity (GPP) was relatively high, given the poor nutrient conditions in this river. Seasonal variability in net ecosystem production (NEP) was also high, with highest values occurring in the dry season (March-April) when fish biomass and chlorophyll a (both water-column and benthic) levels were greatest (dryseason NEP ¼ 542 mg C m À2 d À1 , wet-season NEP ¼ 303 mg C m À2 d À1 ). NEP and algal biomass (measured as chlorophyll a) were higher in lagoons than at river sites, with more pronounced differences between these 2 habitats during the dry season. Strong N limitation was evident. Dissolved inorganic N (DIN) concentrations always were ,2 lM and typically were ,0.5 lM. Molar ratios of DIN:SRP (soluble reactive P) varied little seasonally and were always ,8. As in other Neotropical aquatic systems, water-column productivity was an important source of organic matter and was .23 benthic productivity, even in shallow regions of the riverine ecosystem. Low nutrient levels combined with high rates of autochthonous productivity in the littoral zone of this river suggest extremely rapid nutrient cycling rates and support the view that the littoral regions may be important in providing the organic matter that maintains secondary production, consistent with the Riverine Productivity Model and other observations in the Orinoco Basin.

1. Through analyses of a 34-year record of phytoplankton, zooplankton and physicochemical parameters from Lake Kinneret, Israel, we show that distinct and persistent phytoplankton assemblage states occurred from winter to summer. 2. The most obvious characteristic of these states was the presence or absence of a spring bloom of the dinoflagellate, Peridinium gatunense. 3. Analyses of the data within the framework of the alternative states model revealed a possible complex triggering mechanism, and system hysteresis. 4. A change in zooplankton biomass and body size coincident with changes in predation pressure associated with the collapse of the Kinneret Bleak, Acanthobrama terraesanctae, fishery appeared to be the 'slow changing' variable in the context of the alternative states model. Alternative phytoplankton states were only possible after this variable crossed a threshold in 1993-94, following the collapse of the fishery. 5. When alternative states were possible, some physicochemical parameters and the structure of the zooplankton assemblage appeared to control which phytoplankton state emerged in a given year. In years without a P. gatunense bloom, important physicochemical parameters in winter included low NO 3 loading, high water temperature, high water level, a deeper thermocline, low transparency, high concentrations of NO 3 and Cl in the epilimnion, and low concentration of epilimnetic total phosphorus. In addition, the cladoceran Chydorus sphaericus and adults of the copepod Mesocyclops ogunnus were observed in winter in years without a bloom. 6. Zooplankton biomass and body size of some taxa have recovered since the 1993-94 collapse of the fishery, yet incidence of both phytoplankton states in Lake Kinneret was still possible. Within the framework of the alternative states model, this suggests that the slow changing variable threshold where alternative states became possible is different from the threshold where alternative states will no longer be possible. In other words, the system is characterised by a hysteresis.

Our study evaluated spatiotemporal variation of physicochemical parameters and benthic and sestonic algal biomass in littoral areas of the main channel and floodplain lakes of the Cinaruco River, a lowland tributary of the Orinoco River in Venezuela. The Cinaruco is characterized by a highly predictable annual flood pulse, high transparency, and extremely low conductivity and suspended sediment load. During 2002 and 2003, 10 sites in the main channel and floodplain lagoons were sampled on 12 occasions. Correspondence analysis based on water physicochemical variables and algal biomass showed that ;80% of the variation in the multivariate space among sites and sampling dates was explained by 2 axes. During the high-water periods, river and lagoon sites showed high similarity in water variables and algal biomass. Conductivity, SiO 3 , and chlorophyll a concentrations (both benthic and sestonic) were low during the highwater phase. On the other hand, water variables and algal biomass differed between river and lagoon sites during the low-water period. The absence of flow in lagoons and consistently low algal biomass (sestonic and benthic) in river sites were the most important features of the spatial variability between main channel and lagoon sites during low-water phases. Benthic chlorophyll a was highly uniform at small spatial scales and significantly heterogeneous at large spatial scales. The annual flood regime of the Cinaruco, which drives the concentrations of dissolved materials and affects material interchanges between aquatic and terrestrial systems, also appears to be responsible for creating strong patterns of seasonal and spatial variation in benthic algal crops.

Variable hydrology of rivers strongly affects biophysical factors that influence primary production and population densities, thereby affecting the relative influence of bottom‐up and top‐down processes in trophic networks. Many tropical floodplain rivers have sustained seasonal flood pulses driven by precipitation patterns of the Intertropical Convergence Zone. These changes in flow alter concentrations of dissolved nutrients, aquatic primary productivity, and per‐unit‐area densities of aquatic organisms. Therefore, one would predict that the strength of top‐down effects of animals on basal resources should shift as the annual flood pulse progresses. We conducted a series of field experiments in a Neotropical lowland river to test for effects of hydrologic phase, habitat (in‐channel vs. floodplain aquatic habitat), and benthic‐feeding fish and meiofauna on particulate organic matter, chlorophyll, and benthic microalgae. Net ecosystem productivity of this oligotrophic river is higher during the low phase of the annual flood cycle, which is also when resident fishes are at highest densities and there is a seasonal influx of migratory benthic‐feeding fish. We therefore hypothesized that top‐down effects of benthic‐feeding fish would fluctuate temporally, with strongest effects during low water levels. We found that fish controlled the abundance of particulate organic matter and algae on solid substrates, but not on sand, during falling‐ and low‐water phases within both channel and floodplain habitats. Except for diatom assemblages, which responded to fish exclusion, the taxonomic structure of algal and meiofauna assemblages was not significantly influenced by fish‐exclusion treatments, but varied in relation to habitat type and hydrologic phase. Meiofauna densities were highest during the low‐water period; experimental exclusion of meiofauna during this period had a significant effect on accumulation of particulate organic matter in sand. By controlling abundance of important basal resources, fishes and meiofauna have a large potential to influence other components of this tropical ecosystem. Our findings emphasize the predictable, gradual, changes in consumer–resource interactions associated with the seasonal flood pulse in tropical river systems.