We conducted a bird census along the same route nearly each week for 14 winters (194 censuses), and compared the mean number of species per station and the total number of species recorded on the census with the length of photoperiod and weather variables. We found significant differences among winters for both indicators of species richness. This result is consistent with previous studies in which abundance of food was measured in the same general area. Both indicators of species richness were negatively associated with the number of days after 1 November. This result is consistent with the hypothesis that wintering species dependent on nonrenewed food resources lose individuals to mortality or emigration. Further, there was a positive relationship between photoperiod and both indicators of species richness. This result is consistent with the hypothesis that the detection of individuals in the early morning hours increases with the amount of daylight they have available for foraging and social behaviors. Wind speed and temperature had negative and positive relationships, respectively, to species richness. The number of species per station was greatest on days when the ground was covered with dew and least on days when snow depth was more than 15 cm. When the 'winters' were divided into four 30-day 'quarters', most of the 61 species were recorded with equal frequency in each quarter. Eight species were detected less frequently at the end of winter than in the beginning. Four species exhibited the reverse pattern. Two species were recorded more frequently at the beginning and at the end of the winter than during the middle. Temperature, wind, photoperiod, successive winter day, year, and species-specific evolutionary history all affect winter bird species richness.