Rock avalanches are catastrophic phenomena which are not yet exhaustively understood. They consist of rock mass movements of more than one million cubic metres, involving a great amount of energy and travelling farther than expected with a normal sliding friction law. The present study has as main purpose to investigate the propagation mechanisms involved in rock avalanche processes and to identify parameters influencing velocity and deposit characteristics by means of laboratory experiments, i.e. dry unconstrained flows of granular materials and blocks at small scale (bricks) down an inclined board which ends with a horizontal accumulation zone. Two main experimental campaigns have been carried out. The first represents a preliminary study which has been useful to test the experimental set-up, to improve the measuring devices and to assess significant factors governing propagation of granular avalanches. Fall height, volume, materials used (sand or gravel), releasing geometry and the number of consecutive releases have been varied and their influence on front mass velocity and on deposit characteristics has been studied. In the second experimental campaign the varied parameters are fall height, volume, material (aquarium gravel and small bricks), slope inclination, base friction coefficient and the regularity of the pathway (sharp or curved discontinuity at the toe). Bricks are randomly poured into the releasing container before failure (loose mass) or piled orderly (structured mass). Furthermore, it has been possible to compute the morphology of the final deposit and the position of its centre of mass thanks to a new optical technique, the fringe projection method, recently developed and adapted to the laboratory conditions. The analyses of this extensive set of parameters put in evidence the importance of the nature of the released material, the structure of the mass before failure and the topography, i.e. the slope and the regularity of the pathway. Factors causing longer runouts are: larger volume, greater fall height, lower coefficient of friction, higher slope angle, the use of bricks ordered in piles and a smoother discontinuity at the toe of the slope. Morphology is dependent on the type of material used: sand or gravel; gravel deposit seems closer to real cases. There is also a considerable difference in deposit morphology when the event is the consequence of one large volume released at once or of a progressive failure. In the latter case final deposit characteristics depend on the individual smaller volumes. By analysing the velocity of the mass front as it enters the accumulation zone, it is possible to see that a transfer of momentum occurs between the rear approaching part and the front one slowing down ahead, inducing an excessive travel distance. In the case of piled bricks the regime is mainly frictional (energy dissipated essentially by friction at the base) on the inclined panel, where the mass remains relatively structured, and then frictional-collisional (energy dissipated also by friction and collisions within the mass) in the accumulation zone. The abrupt change of flow direction seems to be the cause of the shattering of the mass activating this passage to a different regime. On the other hand, both regimes can be found from the beginning of the slope in the case of loose materials, i.e. gravel and random bricks. The length is independent from the fall height. A marked difference is detected for tests with a curved connection at the toe which show high nondimensional values of the length and of the runout against the cubic root of the volume, closer to the ones of real cases. Statistical analyses confirm the considerations made.