Abstract: We took the photos of the movement parameter of saltation quartz sands whose average diameter is 0.200 mm with HS-16 high velociy cinecamera in wind Tunnel, Institue of Desert Research, Academia Sinica. In total, the movements of 57 sand particles were partly observed under the standard velocity of 500 photos per second. We divide the locus into three parts, i.e. taking off(19 examples), level flying(30 examples) and falling down(8 examples), we also calculate the velocity and acceleration in both horizontal and vertical directions. Based on the photos and calculation, we come to the following conclusions: 1. During the three stages of gand movement, the vertical acceleration and the horizontal acceleration fluctuates a lot, the absolute fluctuation may be 30 times as big as acceleration of gravity, the fluctuation in vertical direction has nothing to do with that in horizontal direction. Thus, the viscous force which caused turbulent fluctuaion may play a great important role in sand saltation movement. 2. The big fluctuation in isotropous is flying plane, its average value is not very big, the vertical acceleration is negative. We deduced that there is no constant buoyance after leave the ground surface, viscous force is not the force which makes the sands take off. The force which makes the sands take off must be one type of the inertial forces, the impact from other particles Or the bounce of impacting particles, bevel flying and vibration. Because the sands revolve with high velocity after they take off, it is reasonable to give up the ideas that sands take off by impact or vibration. In our opinion, saltaion sanbs take off mostly through bevel flying. 3. During the three periods of saltation, the acceleration in horizontal direction is always positive, and that in vertical direction is always negative. Detailed information is as follows: ax ay Taking off(the average of 19 particles) 33.0m/s² -15.3m/s² level flying(the average Of 30 particles) 27.5m/s² -13.8m/s² Falling down(the average of 8 particles) 12.8m/s² -10.6m/s² From the table, we deduce that: 1) Sands were dragged downward by wind drawing force, the stressness of the dragging force is related to the relative disparity in velocities between sands and air flow, the disparity decreases during the sand movement from the taking off to the falling down. 2) The gravity is in action during the whole period of the movement, the acceleration in vertical direction is composed of two parts, i.e. the acceleration of gravity and the resistance in vertical direction. During the period of the taking off, the acceleration is big because it is caused by gravity and downward resistanc;During the period of level flying, vertical acceleration is a little Smaller; During the period of falling down, the vertical acceleration should de smaller than 9.8m/s² because of the action of both gravity and upward resistance, although the figure is -10.6m/s² in the paper, it's still allowed according to the range of calculation error.