Many studies (e. g., Martin 1983) have indicated a relationship between the number of species and the size of habitat for isolated habitats. However, I did not find a significant correlation between the number of bird species and the area of urban park, but there was a significant correlation between the number of species and the area of vegetation cover. The results of the classification-tree analysis also showed that the area of vegetation cover within an urban park is one of the important environmental factors. This means that the vegetation cover in parks is more important for many bird species than the size of the park. There are many small urban parks in Japanese cities, however most of them have no trees and shrubs.
Vegetation structure and surrounding land use
The result of TWINSPAN and classification-tree analysis showed that the composition of bird species in urban parks was influenced by the propor
tion of woodland within 500 m of the edge of urban parks, the area of vegetation cover within urban parks, the presence of trees (at a height of more than 9 m) and the presence of herbaceous plants and shrubs (at heights of 0.12-3 m).
Many studies have suggested that the land-use pattern around habitats influences bird communities within the habitats, however only a few studies (e. g., Loman and Schantz 1991) have shown the result by field survey. In this study, classification-tree analysis showed that the most important splitting criterion was the proportion of woodland within 500 m from the edge of urban parks. If the proportion falls to less than about 6%, only a few bird species adapted to urban environments can inhabit an urban park. This value is very meaningful, because the proportion of green coverage (except farmland) of most cities in Japan is under 10%.
The relationship between the species richness of birds and vegetation structure has often been studied (e. g., Erdelen 1984; Yui and Suzuki 1987; Ichinose and Katoh 1998). While most studies indicated that forest-interior bird species prefer closed vegetation cover, Peason (1993) showed that the density of the shrub layer was also of similar importance for some wintering birds. In this study, the classification-tree analysis showed that the dense herb and shrub layer influenced the composition of wintering bird species. It is generally known that Cettia diphone and Emberiza spodo- cephala prefer dense shrub habitat. Also the presence of trees (at heights of more than 9 m) was selected as a splitting criterion in the classification – tree analysis. The urban parks of type A and C were divided by this criterion, and no Parus major was recorded in type A urban parks. Because P. major used only the sub-tree and tree layers, the presence of trees is one of the most important factors for P. major habitats.