Superimposition of text on a picture is usually a caption.This paper describes the changes in reading speed and eye-movement for Japanese text when random-dot kinematograms are used as a background movement.The motion conditions consist of five levels for the speed in the horizontal or vertical directions.The texts are presented in horizonal and vertical format.The reading speed is tha lowest for a 23 deg/s average speed of background motion.Vertical motion is the most significant in reducing the reading speed.There is approximately a 70% in the number of eye-movements for reading is the vertical direction.There is a significant difference in the length of vertical eye-movement between the motion directions.It is found that these changes occur on both format.