An artificial stream has been used to examine the population and production dynamics of benthic algae that grow in the hard-water, nutrient-rich streams of southern England. The channel, made of glass-reinforced plastic is 60 m in circumference and filled with flintsca. 40 mm in diameter. Water 0.3 m deep above the gravel was recirculated at 0.4 m s-1by an archimedean screw pump. Water was supplied from the aquifer at a rate of 4.2 m3h-1and left the channel from an overflow notch. The chemical composition of the inflow water was relatively constant so that differences between inflow and outflow could be directly related to biological and non-biological changes in the channel. Experiments in the dark showed that calcium carbonate precipitated, reducing calcium concentration and alkalinity to 80% of inflow values. Reactive phosphate precipitated to 28% of inflow concentrations. Preliminary experiments were done in the light between April and December 1976 and detailed experiments between April 1977 and July 1978. The initial colonization phase, in both 1976 and 1977, was dominated by diatoms. Chlorophyll densities increased by over 300-fold in 14 days in 1977. The change in cell numbers of four dominant species, over a 7 day period, confirmed doubling times of less than 2 days. Maximum densities of 500-600 mg chlam-2were reached 5 weeks after the dark covers had been removed. The biomass remained constant for the next 14 days and during this phase the rate of silicon uptake was 4.5 g m-2day-1. Substantial uptake was also recorded, while the biomass was declining, between the seventh and ninth weeks. Over a period of 60 days, 107 gSi m-2were taken up and photosynthetic studies indicated that 137 gC m-2has also been taken up. In terms of biomass this would correspond to 2.3-4.6 g chlam-2, whereas the maximum observed biomass was 0.5-0.6 g m-2. Loss of fine particulate material over 60 days was 0.0558 g of pigment per square metre and sedimentation into the lower layers of the gravel was 0.735 g of pigment per square metre. During weeks 10, 11 and 12 substantial quantities of silicon and phosphorus were released into the water and for a brief period calcium bicarbonate concentrations in the channel approached inflow concentrations. The diatoms of the initial colonization phase were succeeded by a lime-encrusted growth of blue-green and green algae (Lyngbya kützingii,Chamaesiphon polymorphusandGongrosira incrustans) with densities through the winter of 1977 of 200 mg chi a m -2.Cladophora glomeratadeveloped in the autumn of 1977 and its epiphytic algal flora was substantially different from the epilithic flora. Diatoms recurred in large numbers in the spring of 1978. In terms of biomass the principal diatoms wereAchnanthes minutissima,Fragilaria virescens,Gomphonema rhombicum,Meridion circulare,Nitzschia fonticolaandSynedra ulna. Regression analysis showed that suspended chlorophyll a concentration was not related to benthic algal biomass (estimated as chlorophyll a) but more closely related to benthic diatom volume or dissolved silicon concentration. The seasonal succession of algae in the channel was very similar to that which occurs in local streams. The initial colonization phase by diatoms bears a striking resemblance to the spring outburst of benthic diatoms in local streams. In the channel (and in natural streams, by inference) this succession cannot be controlled by changes in discharge, water velocity or chemical composition of the water. The onset of diatom growth in the spring in streams is probably largely controlled by light intensity. The absence of large numbers of insect larvae, until mid-April, probably allows the biomass-to develop. After April the impact of insect larvae on diatom populations may be considerable.