Résumé: Batch culture experiments using viral enrichment were conducted to test the response of a coastal bacterial community to autochthonous (i.e., co-existing) or allochthonous riverine viruses. The effects of viral infections on bacterial dynamics and activity were assessed by epifluorescence microscopy and thymidine incorporation, respectively, whereas the effect of viral infection on bacterial community composition was examined by polymerase chain reaction-single strand conformation polymorphism 16S ribosomal RNA fingerprinting. The percentages of high nucleic acid-containing cells, evaluated by flow cytometry, were significantly correlated (r2=0.91, n=12, p<0.0001) to bacterial production, making this value a good predictor of active cell dynamics along the study. While confinement and temperature were the two principal experimental factors affecting bacterial community composition and dynamics, respectively, additions of freshwater viruses had significant effects on coastal bacterial communities. Thus, foreign viruses significantly reduced net bacterial population increase as compared to the enrichment treated with inactivated virus. Moreover, freshwater viruses recurrently and specifically affected bacterial community composition, as compared to addition of autochthonous viruses. In most cases, the combined treatment viruses and freshwater dissolved organic matter helped to maintain or even enhance species richness in coastal bacterial communities in agreement to the 'killing the winner' hypothesis. Thus, riverine virus input could potentially influence bacterial community composition of the coastal bay albeit with modest modification of bulk bacterial growth.