1. The taxonomic composition and biomass of the phytoplankton and the taxonomic composition of the phytobenthos of the San Joaquin River and its major tributaries were examined in relation to water chemistry, habitat and flow regime. Agricultural drainage and subsurface flow contribute to a complex gradient of salinity and nutrients in this eutrophic, 'lowland type' river.

2. Because of light-limiting conditions for growth, maintenance demands of the algae exceed production during summer and autumn in the San Joaquin River where there is no inflow from tributaries. In contrast to substantial gains in concentration of inorganic nitrogen and soluble reactive phosphorus during the summer of normal-flow years, net losses of algal biomass (2-4 ??g L^-1 day^-1 chlorophyll a) occurred in a mid-river segment with no significant tributary inflow. However, downstream of a large tributary draining the Sierra Nevada, a substantial net gain in algal biomass (6-11 μg L^-1 day^-1) occurred in the summer, but not in the spring (loss of 1-6 μg L^-1 day^-1) or autumn (loss of 2-5 ??g L^-1 day^-1).

3. The phytoplankton was dominated in summer by 'r-selected' centric diatoms (Thalassiosirales), species both tolerant of variable salinity and widely distributed in the San Joaquin River. Pennate diatoms were proportionally more abundant (in biomass) in the winter, spring and autumn. Abundant taxa included the diatoms Cyclotella meneghiniana, Skeletonema cf. potamos, Cyclostephanos invisitatus, Thalassiosira weissflogii, Nitzschia acicularis, N. palea and N. reversa, and the chlorophytes Chlamydomonas sp. and Scenesdesmus quadricauda. Patterns in the abundance of species indicated that assembly of the phytoplankton is limited more by light and flow regime than by nutrient supply.

4. The phytobenthos was dominated by larger, more slowly reproducing pennate diatoms. Few of the abundant species are euryhaline. The diatoms Navicula recens and Nitzschia inconspicua and cyanophytes, Oscillatoria spp., were the principal late-summer benthic species upstream in the mainstem and in drainages of the San Joaquin Valley. Many of the other abundant diatoms (Amphora veneta, Bacillaria paxillifer, Navicula symmetrica, Nitzschia amphibia, N. fonticola, N. palea, Pleurosigma salinarum) of late-summer assemblages in these segments also are motile species. While many of these species also were abundant in segments downstream of confluences with rivers draining the Sierra Nevada, the relative abundance of prostrate (Cocconeis placentula var. euglypta, Navicula minima) and erect or stalked (Achnanthidium deflexum, Achnanthes lanceolata, Gomphonema kobayasii, G. parvulum var. lagenula) diatoms and Stigeoclonium sp. was greater in these lower San Joaquin River segments.

5. A weighted-averaging regression model, based on salinity and benthic-algal abundance in the San Joaquin River and segments of its major tributaries within the San Joaquin Valley, yielded a highly significant coefficient-of-determination (r² = 0.84) and low prediction error between salinity inferred from the species and that observed, indicating that salinity tolerance is a primary constraint on growth and assembly of the phytobenthos. The same measures of predictability indicated poor performance of a model based on inorganic nitrogen. However, with a greater representation of tributaries (including segments within the Sierra Nevada foothills) in the sample set, an inorganic nitrogen model also yielded a highly significant coefficient-of-determination (r² = 0.87) and low prediction error between the species-inferred and the observed concentration. As with the salinity model (r² = 0.94) for the enlarged data set, a systematic difference (increased deviation of residuals) existed at high inorganic nitrogen concentrations. These results indicate substantial interaction between salinity and inorganic nitrogen as constraints on the structure of benthic-algal communities of the San Joaquin River basin.