The aim of this research was to investigate the possibility of using soil magnetic susceptibility to differentiate wetland soil from non0wetland soil, thereby enabling the identification of a wetland boundary. The initial methodology to test the viability of using magnetic susceptibility for wetland delineation was carried out at nine sites in three areas of Kwazulu-Natal, South Africa. Changes in vegetation were used to identify the permanently, seasonally and temporarily wet zones at each site to provide a quick indication of the various boundaries. At least one transect was studied at every site, extending from the outer edge of the temporary zone to the water's edge. Magnetic susceptibility readings, soil samples and/or elevations were taken in each zone. It was found that the magnetic susceptibility readings could be used to differentiate between the various wetland zones but boundaries were not identified due to the reconnaissance nature of the work, a specific protocol had yet to be discovered. This methodology was a reconnaissance phase to assess the potential of using soil magnetic susceptibility. A second methodology was designed specifically to identify a wetland boundary. The wetland zones were identified using vegetation indicators from the South African field procedure for delineating wetlands. Magnetic susceptibility readings were done on transects perpendicular to the suspected boundary and a critical value was identified. A magnetic susceptibility boundary was marked according to the critical value and verified by taking readings along transects parallel to it on both sides. The boundary was then identified using South African field procedure according to soil indicates. The soil data showed a similar boundary to the magnetic susceptibility boundary but indicated a 'boundary zone' of approximately l2m in width rather than a specific line. A final methodology was planned to improve the resolution of the magnetic susceptibility boundary. A grid was laid out over a strip through the wetland including the boundary area on both sides. This was to provide accurately spaced points at which to take magnetic susceptibility readings and elevations. The critical value, a value that separates wetland from non-wetland soil, was identified and verified using soil indicators. The field procedure was more difficult to carry out than using the magnetic susceptibility sensor, yet both methods identified the same boundary with a resolution of about 3m. This initial study demonstrates the potential for using magnetic susceptibility for wetland delineation. Although the results at the final site proved the method to be successful, it was not suitable for use at all sites and the results were often difficult to interpret. Limitations include factors such as plinthic horizons close to the soil surface and shallow, rocky soils. Thus further research is required before magnetic susceptibility can be used to delineate wetlands in KwaZulu-Natal, South Africa. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2002.