This thesis is an investigation of bed mobility in gravel-bed rivers. In particular, two hypotheses are addressed: 1) that mobilization (scour) depth in salmon egg nests (redds) differs from that of surrounding bed, and 2) that equilibrium bedload transport rate in gravel-bed rivers can be predicted by assuming equal mobility and accounting for armouring and particle hiding/exposure. To test the first hypothesis, a single spawning reach of Kanaka Creek, British Columbia was monitored for scour depth using a spatially intensive array of wiffle-ball monitors. Monitors were placed in and around egg pockets and tailspills of chum salmon (Oncorhynchus keta) redds. Scour and fill were recorded during each flood event of the 1997/98 flood season. Only one flood event was observed to cause widespread and deep scour and fill. Scour and fill were found to be highly spatially variable, and depended upon local factors such as bed structure, proximity to banks, and the influence of flow obstructions such as large woody debris. Despite the most spatially intensive monitoring to date, no spatial autocorrelation was observed in scour depths. By various nonparametric analyses of variance, mean scour depth of redd egg pockets was not found to differ significantly from that of surrounding bed, but the mean scour depth in tailspills of redds was significantly different from surrounding bed. However, small sample sizes limit the generality of the results. The second hypothesis was evaluated by development of a largely empirical model. Fractional bedload transport rates were estimated through modification of a bedload estimate for median size particles. Armouring was considered through modification of the fractional transport rate by the percent representation in the surface of the size fraction. The effects of relative differences in particle size, due to hiding/exposure interactions and different particle mass, were considered by means of a mobility factor (D/D50)\ Data from ten different studies, including 38 surface/subsurface size distributions and 17 equal transport mobility fractional bedload rate data sets, were utilized to develop the model. Equal transport mobility was assumed to evaluate the exponent x. The exponent x was found empirically to be a function of the surface sorting. The modified transport during a state of equal transport mobility was empirically found to be a function of the nondimensional bed shear stress. Finally, a simple model was proposed to estimate mean scour depth during high flow (equally mobile) events. The scour model combined the new bedload transport model with a fundamental bedload model. Somewhat surprisingly, due to uncertainty in estimates for scour model inputs, the scour model appeared to perform well when tested against the scour measurements.