It has been reported that intake of yogurt or powder supplemented with the Bifidobacterium longum BB536 probiotic strain alleviated subjective symptoms and affected blood markers of allergy in individuals with Japanese cedar pollinosis (JCPsis) during the pollen seasons of 2004 and 2005, based on randomized, double-blind, placebo-controlled trials. Furthermore, the 2004 study found that intestinal bacteria such as the Bacteroides fragilis group significantly fluctuated during the pollen season in JCPsis individuals and intake of BB536 yogurt tended to suppress these fluctuations. The present study investigated faecal microbiota to examine whether any changes occurred during the pollen season and whether any influence was exerted by intake of BB536 powder in the 2005 pollen season, which happened to be a heavy season, to confirm the 2004 findings and to evaluate the relationship of microbiota with symptom development. In a randomized, double-blind, placebo-controlled trial, 44 JCPsis subjects received BB536 or a placebo for 13 weeks during the pollen season. Another 14 Japanese cedar pollen (JCP)-specific IgE negative healthy subjects received placebo for the same period. Faecal samples were collected before (week 0), during (weeks 4, 8 and 13) and after (week 17) intervention, and out of JCP season (week 28). Faecal microbiota were analysed using terminal-RFLP (T-RFLP) and real-time PCR methods. Principal component analysis based on T-RFLP indicated distinct patterns of microbiota between healthy subjects and JCPsis subjects in the placebo group, but an intermediate pattern in the BB536 group at week 13, the last stage of the pollen season. The coordinate of principal component 1 at week 13 correlated with composite scores of JCPsis symptoms recorded during the pollen season. Faecalibacterium prausnitzii and the Bacteroides fragilis group were identified as the main contributors to microbiotal fluctuations. Real-time PCR indicated that BB536 intake suppressed increases in the Bacteroides fragilis group compared with the placebo group (P <0.05). These results suggest that faecal microbiota in JCPsis subjects, but not healthy subjects, fluctuate at the end of the pollen season and that BB536 intake plays a role in maintaining normal microbiota.