A general procedure is described for treating a movable singularity in an integral. This enables us to change the original integral I0 into GI¡, where G depends only on the parameters of the singularity and /, is a new integral which exists for all values of the parameters. The results are then applied to the particular problem of evaluating f1 f(x)dx .,{(!x2Xl fc2x2)}"2' where / is entire and ...
