This work is concerned with the nature and magnitude of certain transverse aerodynamic forces on cricket balls in flight. The forces considered are those arising from the non-isotropy of cricket balls, that is those occurring mainly because the ball has a prominent seam around its circumference. The forces were measured for a variety of balls in experiments where the balls were either fixed in or projected through a wind tunnel jet. The principal objectives were to provide a set of results for the transverse swing of cricket balls, and to identify and consider the factors on which the effect depends. The experiments supported many, but not all, of the popular beliefs concerning the phenomenon. In particular, cricket balls swing well if the ball has a prominent seam inclined at a slight angle to the direction of projection, and if the surface of the ball is otherwise smooth. A small amount of back-spin imparted along the seam at projection helps to maintain the orientation of the ball during the flight. The experiments did not support the common belief that cricket balls swing more on humid days. Also, it is argued that 'late swing' is an uncontrollable effect. Perhaps the most important conclusion, however, is that the whole phenomenon is easily influenced by transient effects that are beyond control. Therefore swing bowlers should expect considerable variation in swing, even with new balls.