The Birds

...

are locally known as ‘‘shola forests’’ (hereafter referred to as shola forests). Altitude (using an altimeter) and habitat type were recorded every 0.5 km, and distance walked measured with a pedometer. Once a sighting (visual or audio) was made, the pedometer reading from the starting point was noted. Altitude and GPS location were also noted for every sighting (whenever possible). The bird was recorded by call only when it could be identi- fied with confidence. While recording the number of individuals using just calls, only the number of individuals heard at the same time were recorded in order to avoid duplication or over counting.

Many species of Turdinae are associated with streams and to quantify this for the shortwing, the distance to the nearest stream from the point of sighting was

Status and habitat preference of White-bellied Shortwing Brachypteryx major 337 recorded. The stream, when present was classified into wet (running or damp)

or dry.

Data analysis

Encounter rate (ER) was calculated as the number of individuals seen per km walked. The sightings were classified into three elevation classes: <1,000 m, 1,000–1,500 m and >1,500 m. The vegetation was categorized into four broad types: evergreen forest, moist-deciduous forest, shola forest and plantations. The three plantation types, pine, wattle and tea were pooled together as ‘‘plantations’’ for the analyses. Mann–Whitney U-tests were used to test for significant differ- ences in encounter rates across various categories of elevation and vegetation. For applying the Mann–Whitney U-test, the data recorded for each category (elevation/vegetation type) were subdivided into sampling units of 0.5 km. To assess the proximity of streams to shortwing sightings, all sightings were classi- fied under the following categories of distance to the nearest stream: (i) under 20 m , (ii) between 20 and 50 m and (iii) over 50 m. The null hypothesis of no difference between the categories was tested for every pair of categories using Chi-square tests. Preference for wet or dry streams was also examined using Chi-squared tests. All analyses were done using Statistica kernel release 5.5 and Microsoft Excel.

Mapping

The Survey of India maps of 1:100,000 were used as base maps. Areas with an elevation of 900 m and 1,500 m were generated from this. The forest cover in different areas was mapped based on Gaussen et al. (1961). MapInfo Professional version 5.0 and Arc View GIS version 3.0 were used for handling the maps.

Results

A total distance of 436.5 km was surveyed across the four broad habitat types: evergreen forest, moist-deciduous forest, shola forest and plantations, covering an elevational range of 60 m to 2,400 m. Four National Parks, seven Wildlife Sanctuaries, two Tiger Reserves and nine Forest Reserves were covered across Kerala and Tamilnadu during the study period (Appendix 2). A total of 214 sightings of the shortwing were obtained during the study.

Vegetation types

The sampling effort in different vegetation types was not equal as the focus was on covering as much forest as possible in different parts of the two states (Figure 1). Only 39 White-bellied Shortwing sightings (18% of the total) were in ever- green forests, even though the highest sampling effort was made here (307 km, or 70% of the total). Contrastingly, the maximum number of sightings (171, 79.5% of the total) were from 72.5 km (17%) of shola forests. Thus, sightings per unit effort were significantly higher in shola forests than in evergreen forests (Mann– Whitney U-test: n = 614, 145; Z = −17.91; P < 0.01). There were no sightings