Two 2-km line transects were deployed in Kushiro district, Hokkaido, Japan, one in woodland and one in grassland. Censuses were conducted 22 times along each route during the same breeding season. The relationship between census efforts and the number of species encountered was examined, and suitable census combinations for two-fold census and five-fold census were investigated. Furthermore, seasonal changes in individual bird numbers for major species were examined. Thirty-eight species were recorded along both routes. In woodland, 14 species considered to have established breeding territories along the route and were categorized as major species; 24 other species were categorized as minor species. Nine species were major and 29 were minor species in grassland. For both routes, occurrence rates for major species were more than 40%, while rates for most minor species were less than 20%. To analyze the relationship between census efforts and the number of species encountered, I constructed all possible combinations of census efforts by using the data compiled from 22 visits. The total number of species increased with increasing census efforts in both woodland and grassland. Approximately 90% of major bird species could be recorded in two-fold censuses and 99% in five-fold censuses. Censuses conducted twice in June or once each in May and June can provide census data for most bird species including the majority of major species along the routes, but do not allow discrimination of major and minor species. Five-fold repeat censuses allow discrimination of these two categories by using occurrence rates. Furthermore, the number of individual songbirds appeared to differ among breeding stages. To investigate these seasonal changes, the breeding season should be divided into more than five periods, and a census should be performed in each period. This protocol will also confirm the presence of 99% of major species.