Abstract: Abstract: This study aimed to quantitatively evaluate of the movement of rock fragments induced by hoeing at purple mudstone under different rock moisture contents, tillage depths and slope gradients. A physical tracer method was used to determine the effect of hoeing on downslope movement of rock fragments. A total of 30 slope plots were established for the experiment in November, 2017 and March, 2018. Small white gravels were used as tracers to measure the movement of rock fragments and soils caused by hoeing. The mean longest axis of the tracers was 2 cm. At each plot, fourteen holes (6-cm depth and 2-cm diameter) were bored into the bedrock along a 1.4-m contour line with an interval of 0.10 m by an electric drill with 2 cm tungsten carbide steel tipped (TCT) bit, and then the tracers were inserted into those holes. The position of each tracer was recorded with reference to a local system of Cartesian coordinates, and the intersection of the centerline of the first hole on the left side and the surface of plot was chosen as the origin of coordinate system. After these operations, tillage started from the bottom of the field and moved up the slope (i.e., downslope tillage in which rock fragments were translocated into the downslope direction) until the line of tracers. All the experiments were performed by the same operator to maintain the consistent experimental condition. Once tillage operations were completed, three fresh samples of rock fragments were randomly selected for the determination of moisture content and bulk density of bedrock in the laboratory, and the tracers were excavated carefully by a small fork hoe starting at the downslope end and gradually moving upwards across the tillage zone. If a tracer was found, its location as well as its number was recorded so that the displacement distance of all the tracers could be calculated. The experimental data showed that rock moisture content was an important factor affecting the fragmentation of mudstone, and the larger mean displacement distance of rock fragments induced by hoeing tillage was found on the slope with higher rock moisture content. Moreover, when tilling the mudstone with different moisture contents, the relationships between the mean displacement distance of rock fragments to slope gradient and tillage depth were different. For the mudstone with higher moisture content (14.77%), the mean displacement distances of rock fragments were significantly negatively correlated with tillage depth (P<0.001), but no significant relationship was found between the mean displacement distance of rock fragments and slope gradient (P>0.1). For the mudstone with lower moisture content (7.44%), the mean displacement distances of rock fragments were significantly positively correlated with slope gradient (P<0.1), but no significant relationship was between the mean displacement distance of rock fragments and tillage depth (P>0.1). For a given tillage depth, the mean displacement distance of rock fragments decreased rapidly with increasing depth, and greater change occurred in the slope with higher moisture content (14.77%). Above all, this study demonstrated that the process for the movement of rock fragments induced by hoeing tillage was affected by rock moisture content, tillage depth and slope gradient, and the action mechanism of each factor was different in different deep layers. The results present an underlying insight into the mechanism of erosion, weathering and pedogenesis by anthropogenic factor in hilly agricultural regions with mudstone and shale, southwest China.