Les variations temporelles des concentrations en phosphore, celles des abondances bactériennes et de l'activité de la phosphatase alcaline (APA) ont été estimées in situ en chambre benthique placée en deux points du réservoir Sahela (Maroc) durant les mois de septembre et octobre 98. En période de faible oxygénation, nous avons enregistré des élévations relativement faibles des concentrations en orthophosphates (de 0,020 à 0,035 mg. l-1 au point 1 et 0,015 à 0,025 mg. l-1 au point 2) par rapport au phosphore total (0,080 à 0,100 mg. l-1 au point 1 et de 0,035 à 0,040 mg. l-1 au point 2). À cette période, les abondances bactériennes et l'APA montrent des valeurs maximales (8. 106 bact. ml-1 et 0,323 mmol.PNP l-1.h-1 au point 1 et 6.106 bact. ml-1 et 0,438 mmol.PNP l-1.h-1 au point 2 respectivement). L'apparition des conditions anoxiques et la diminution du pH favorisaient la dissolution du phosphore particulaire et la libération du phosphore réactif soluble. Cette libération s'accompagne d'une élévation des abondances de bactéries anaérobies (de 5.106 à 9,2. 106 bact. ml-1 au point 1 et de 3,8.106 à 7,2. 106 bact. ml-1 au point 2) et une diminution progressive d'APA (de 0,200 à 0,025 mmol.PNP l-1.h-1 au point 1 et de 0,125 à 0,077 mmol.PNP l-1.h-1 au point 2). Ce relargage du phosphore à partir du sédiment est accentué par les rejets domestiques et industriels de la ville de Taounate, ce qui accélère le processus d'eutrophisation de ce réservoir., Temporal variations of phosphorus concentrations, bacterial abundance and alkaline phosphatase activity (APA) were estimated in situ in a benthic chamber. The chamber used has a surface of 0.4m2 and a volume of 90 l ; it resembles those used in oceanography, with a tube connecting the interior of the chamber to the lake surface. The water in the chamber was permanently mixed by an electric agitation system. The chamber was placed at two points in the Sahela reservoir (Morocco). Point 1 was located near Guelta El Haila, a site that receives both domestic and industrial effluent, and point 2 was located in the centre of the reservoir. During the two incubations, eight samples were taken over 24 d in September and October 1998 from point 1, and seven samples were taken from point 2 over 29 d in October. After each sampling, an equal volume of water was injected into the chamber to avoid bubble formation.Under low dissolved oxygen concentrations in the benthic chamber, we noticed a relatively small elevation in orthophosphate concentrations (from 0.020 to 0.035 mg×L-1 at point 1 and from 0.015 to 0.025 mg×L-1 at point 2) in relation to total phosphorus (from 0.080 to 0.100 mg×L-1 and from 0.035 to 0.040 mg×L-1 at points 1 and 2 respectively). The low residual oxygenation of hypolimnic layer allowed the oxidation of iron, manganese and led to their binding to phosphorus released from the interstitial water. Dissociation of calcium-phosphorus complexes as a result of a pH decrease may have contributed to phosphorus release. The orthophosphate concentrations were relatively low, a situation that favours the synthesis of alkaline phosphatase by aerobic bacteria and facultative aerobic bacteria. In this period, the bacterial abundance and APA were comparable and showed the maximal values (8. 106 bact. ml-1 and 0.323 mmol. PNP l-1.h-1 (phosphate nitrophenol) in point 1 and 6. 106 bact. ml-1and 0.438 mmol. PNP l-1.h-1 in point 2 respectively). The lack of a significant correlation between total APA and bacterial abundance at the two sampling points is probably due to the decrease of aerobic bacteria followed by a repopulation with anaerobic bacteria. However, the relative contribution of two bacterial populations that have different sizes, as at points 1 and 2, showed that the majority of total APA was produced by bacteria that are attached to organic matter. Therefore we suggest that attached bacteria contribute more than free bacteria to APA production.The beginning of anoxic conditions and the decrease in pH favored the dissolution of particulate phosphorus and the liberation of reactive phosphorus. We noted an elevation in orthophosphate concentrations (from 0.035 to 0.050 mg×L-1 at point 1 and stabilised at 0.025 mg×L-1 at point 2) and total phosphorus (from 0.100 to 0.150 mg×L-1 at point 1 and from 0.040 to 0.050 mg×L-1 at point 2). This liberation followed an increase in anaerobic bacterial abundance (from 5×106 to 9.2×106 bact×mL-1 at point 1 and from 3.8×106 to 7.2×106 bact×mL-1 at point 2) and the progressive decrease in APA (from 0.200 to 0.025 mmol PNP L-1 ×h-1 at point 1 and from 0.125 to 0.077 mmol PNP L-1 ×h-1 at point 2). The anaerobic bacteria did not activate their alkaline phosphatase and the synthesis of their enzyme was progressively inhibited by a de-repression phenomenon caused by high phosphorus concentrations.We conclude that low oxygen and a decrease in pH favoured the release of phosphorus by dissolution of chemical complexes: calcium-phosphorus, iron-phosphorus, manganese-phosphorus and aluminium-phosphorus. This release is also the result of bacterial phosphatase activity to which fixed bacteria contributed the most. The release of phosphorus from water-sediment interface is further amplified by phosphate import from domestic and industrial waste waters originating from Taounate, which accelerates the eutrophication process in this reservoir.