From November 2004 to April 2005, 5 cases of norovirus (NoV) occurred in Sakai City, Japan. These were all diffuse outbreaks due to infections with genogroup II genotype 4 (GII/4) virus strains. Similar outbreaks occurred throughout Japan; hence, GII/4 was assumed to be the prevalent NoV type. However, a NoV outbreak that occurred at a nursery in May 2005, was caused by infections with GI/4 and GII/6 viruses, respectively, from different children. The time course of newly infected patients showed that this nursery outbreak had a two-peak pattern, with the peak numbers of patients occurring on May 19 and May 22. Virological examination and epidemiological research could not determine whether the GI and GII NoV infections occurred at the same time, or whether there was a time difference in their appearance in the nursery. From this outbreak, it is clear that the timing of obtaining samples and obtaining the minimal necessary number of primary samples are essential for accurate epidemiological information to be obtained. In addition, we detected genotypes that were different from the previously prevalent genotypes, which raises the possibility of more frequent NoV infection or a change in the prevalent NoV genotype in this setting. In conclusion, it is difficult to predict outbreaks of NoV; however, through vigilant and early collection and analysis of later samples throughout an outbreak, it is possible to understand the prevalence and perhaps trace the source of NoV infections.