The small mammal fauna of Sulawesi is of interest to biologists due to its high level of endemicity. At least 8 genera of Murinae with 20 described species are found only in Sulawesi (Corbet and Hill,1992), and these species consist predominantly of forest-adapted mice and rats. There is currently a lack of basic information on the ecology of these species, therefore this study was initiated to quantify the population densities and possible microhabitat preferences of the terrestrial rodents in the Lambusanga and Kakenauwe reserves. In addition to their inherent interest, the small mammal population is also likely to be highly important as a prey resource to civet and reptile species. Investigating microhabitat preferences may also highlight the potential threat posed to the various endemic species by habitat degradation.

The aim of this study was to produce a relative estimate of population density of terrestrial rodents at two forest sites (Lambusanga and Kakenauwe) using a mark-recapture trapping programme. In addition, possible microhabitat preferences were investigated by recording habitat variables at each trap site, which could then be related to the capture data.

The live trapping programme began on the 2^nd of July 2001, and has continued over 8 weeks. Four trapping grids were established, two at Kakenauwe and two at Lambusanga. A pair of grids, one at each site was trapped for a total of four weeks, followed by four weeks trapping at the second pair of sites. Each grid was trapped for five days on alternate weeks, giving a total of eight nights’ data at each grid. Each 10 x 10 grid consisted of 100 Longworth traps with 20m inter-trap spacing. Twenty larger wire-mesh Tomahawk traps were also placed at trap stations distributed throughout the grid. This produced a total of 960 trap-nights of data for each grid. The traps were checked daily, and morphometric data was collected for all captures. In addition, each individual was given a unique fur clip, allowing it to be identified if recaptured. A small tissue sample was also taken from each animal to enable DNA analysis to be performed; this can then be used to advance the knowledge of the taxonomy of these species.

Microhabitat preferences were investigated by setting up 5m quadrats centred at each trap station. Within each quadrat, 13 microhabitat variables were recorded. Canopy cover will be assessed from photographs taken above each trap station. The total number of trees (>15cm circumference) and their circumferences were recorded. In addition, the total number of logs within the 5m quadrat was also recorded, as was their length and diameter. Four 1m quadrats were then set up, one in each corner of the 5m quadrat. Within the 1m quadrats the percentage of bare ground, the percentage of surface limestone karst, and the number of woody stems were recorded. The vegetation density at 0.5m and 1.0m above ground level was also assessed looking from each corner of the 5m quadrat towards the trap at its centre. Other information such as the extent of cleared paths within the quadrat was also recorded.

Over 1820 trap-nights at the two Kakenauwe grids over 70 individuals were captured. The predominant species captured was Bunomys andrewsi, with the next most frequently trapped species being Maxomys hellwaldii. Taeromys celebensis and Paruromys dominator were caught in low numbers. Two Maxomys musschenbroekii were also caught; this is a new record for Buton island. In addition, one rat was caught which was not readily identifiable in the field and therefore was taken as a specimen for identification.

At Lambusanga the capture numbers were lower with only around 40 individuals trapped. In addition, the species composition appeared different; here the predominant species was M. hellwaldii. In contrast to Kakenauwe, B. andrewsi was less frequently caught than M. hellwaldii. No T. celebensis were caught at Lambusanga, however P. dominator were trapped, as was a single specimen of a Rattus species, which was not possible to identify in the field. One M. musschenbroekii was also trapped at Lambusanga. Two Brown Shrews (Crocidura levicula) were also incidentally trapped.

Analysis of the mark-recapture data will take place on return to the UK, where the computer program MARK will be used to produce population estimates for each of the grids.

Four hundred microhabitat quadrats have now been completed and the data should allow some interesting conclusions to be drawn as to the habitat preferences of the rodent fauna. On return to the UK statistical analysis will be carried out to highlight any variation in microhabitats between the different sites, and to investigate which habitat factors which might underlie differing population densities.

A report entitled ‘Relative Population Densities and Microhabitat Preferences of the Terrestrial Rodent Fauna at Lambusango and Kakenauwe Forest Reserves’ will be prepared by Nicola Grimwood, Manchester Metropolitan University by January 2003.