Material and methods Study area

The study area was the southern part of Nishinomiya City, Hyogo Prefec­ture, central Japan (Fig. 1), located between Osaka and Kobe. Commercial and residential land uses are dominant in the area, though there are many urban woodlands consisting of fragmented forests, shrine and temple for­ests, riparian forests, urban parks and so on. Twenty urban parks with vegetation cover were selected as study sites. The area of the smallest park was 0.19 ha while that of the largest one was 3.4 ha (the median, and first and third quartiles were 0.55, 0.305, and 1.09 ha, respectively).

Bird censuses

Birds in each park were censused from 13 December 1999 to 23 February 2000. This period was the wintering season in this study area. I censused birds three times in each park. I used the point-counts method described by Bibby et al. (1992). I recorded all species observed within a 25 m radius from the standing point and omitted the individuals flying beyond the ob­servation range. Birds were censused for 20 minutes between 0900 and 1500, and the species name, the number of individuals, and the observed position (height from ground and plant species) were recorded.

image59Fig. 1. Location of the study area