Experiments on electronically variable time delay of microwave pulses by transmission through single‐crystal YIG rods are discussed. One type of delay, recently discussed by Olson and Buchmiller and by Kaufman and Soohoo and attributed to magnetostatic waves, occurs with unpolished ends, has rapidly increasing delay with increasingHdc, and has severe pulse dispersion. The second, found here, requires parallel and polished ends, has a slowly decreasing delay with increasingHdc, and usually has at least two pulses spaced by less than the round‐trip time through the rod. Characteristics are similar to those described by Strauss, except that his experiments were of pulse reflection.The two aspects of magnetostatic wave pulse delay to be discussed are: (1) Additional experimental verification of Schlo¨mann's theory of wave excitation by nonuniform demagnetizing fields, by grinding a rod into ellipsoidal shape, and by measure coupling. (2) A nonreciprocal microwave time delay has been constructed utilizing magnetostatic waves in a YIG rod to couple two rectangular waveguides. The coupling is between the positions in the waveguide where the magnetic field is circularly polarized. The directionality of the coupling was found to be greater than 15 dB with delays of 4 &mgr;sec at 2500 Mc/sec.