This study tests the hypothesis that human impact is an important driver of stream biofilm bacterial population diversity.   The seasonal bacterial composition of biofilm in 4 streams with different levels of human impact was determined over 2 years.    Bacterial diversity derived from 16S rDNA clone libraries, shows both between stream differences and seasonal transitions in bacterial occurrence and population dominance at a class and genus level.   Diversity analysis calculated on pooled seasonal data (class level identification) shows that while composition of the populations are different there is a similar level of both bacterial richness and bacterial diversity in each stream.     Trends in bacterial occurrence suggest that the most degraded stream were dominated by cyanobacteria, the mid range impact streams by aeromonads and gamma proteobacteria, while the unimpacted stream showed both high  diversity and no  dominance by any particular class.