This thesis describes a systematic study of hillslope runoff processes and stream flow processes from two forms of land use characteristic of the New Zealand steep land environment: indigenous forest (podocarp-hardwood association) and pastoral agriculture. In humid temperate environments, hillslope runoff is now recognised to result from a variety of mechanisms, that are dynamic and spatially variable. In these environments, surface and subsurface hillslope runoff processes are dependent on the soil, topographic and vegetation characteristics of individual catchments. The terms ‘partial and variable source area models’ have been used to describe these concepts and flow processes. The new concepts of hillslope runoff generation have important implications for the way in which land is managed to maintain desirable stream flow characteristics and stream water quality. However, this information has not been used widely as a basis for land management procedures in areas where conflicts of land and water use occur. This situation has occurred because most of the studies of source area hydrology have been qualitative and have not provided information suitable for use in applied management problems. Also, few comparative land use studies have been established since the new concepts of source area hydrology have been developed. The main aim of this study was to clarify the processes of hillslope runoff under indigenous forest and pastoral agriculture, and to interpret the hydrologic implications of these two forms of land use on the basis of the new concepts of hillslope runoff. A comprehensive, integrated study of source area hydrology was established in a steepland catchment supporting the two vegetation types. Although the areas of forest and pasture vegetation were contiguous areas within a single catchment, each vegetation type was treated as a separate hydrologic entity and referred to as an individual catchment. Both ‘process’ scale experiments and some ‘catchment’ scale experiments were established to understand fully the processes of hillslope runoff and the relationships between hillslope flow processes and stream flow processes in the two land use catchments. An experimental design was developed for this study based on the partial and variable concepts of source area hydrology. The experimental catchments were subdivided into three geomorphically distinct sub-catchment units (riparian, midslope and spur regions) on the basis of a preliminary survey of the spatial distribution of soil moisture. Data were collected on rainfall, stream flow, surface and subsurface runoff components and soil moisture, during weekly visits to the catchments over a two year period, July 1979 - July 1981. Quantitative estimates of the spatial variability of surface and subsurface hillslope runoff components were made in the sub-catchment units of each land use catchment. The processes of hillslope runoff were also examined within each sub-catchment unit. Additional studies of forest interception processes, soil permeability and other soil, topographic and vegetation characteristics were completed during this period. The results of this study show that the hillslope runoff regimes are complex in both land use catchments. In each land use, hillslope runoff processes vary spatially and seasonally. The results also show that the differences in hillslope runoff processes are not consistent, either between the sub-catchment units of each land use catchment, or between similar sub-catchment units of both land use catchments. Despite the complexity of the hillslope runoff processes, the integrated stream flow responses from each land use catchment are less variable, and show the differences expected for the two vegetation types. For most storm events, the spatial distribution of hillslope runoff conforms with the patterns described by the variable source area model of hillslope runoff. However, the mechanisms of hillslope runoff and the causes of the spatial distribution of these processes are not described well by this model. Infiltration-excess overland flow is generated above the mineral soil in each land use catchment, despite the intermittent, low intensity, rainfall. Infiltration-excess overland flow occurs because the catchment soils are considerably less permeable than in other catchments in similar environments. The spatial distribution of surface runoff is dependent on the spatial variability of soil permeability in each land use that is, in turn, determined by variations of soil moisture in space and time. Subsurface discharges are also less important in determining the stream hydrograph from the two land use catchments studied, than reported in other studies of similar forms of land use. The catchment soils are so slowly permeable that large volumes of subsurface runoff do not reach the stream channel in time to contribute to storm runoff. An examination of stream flow processes and water balance data was completed to provide an analysis of the integrated runoff response from the two land use catchments. Compared with the forest, greater stream discharges occur from the pasture land use for all flow regimes. However, the time distribution of the storm runoff response is independent of vegetation type or antecedent catchment conditions. The principal differences in the water balance components from each land use catchment are caused by differences in interception losses: similar transpiration rates were observed for the two vegetation types. Selective catchment management, of riparian areas in pasture catchments in the Hapuakohe Range, is suggested as a method of improving general stream flow characteristics from these areas, while the multiple use of the land and water resources is maintained. Estimates of the effectiveness of this management strategy are presented. However, the degree to which riparian re-afforestation causes a satisfactory change in stream flow characteristics, depends entirely on the objective of the land management practices proposed for a given catchment.