The development of growth techniques for epitaxial film is shown to yield similar methods for growing bubbly domain material and films of yttrium-iron garnet (YIG) for delay lines. Growth by liquid-phase epitaxy is examined in detail. The development of sophisticated bubble films and YIG delay lines is discussed and shown to be a direct consequence of the ease with which this technique can be implemented. The characteristics of bubble domain films are discussed. The composition of different films which were grown in an attempt to meet optimum bubble characteristics is discussed together with the film defects. The deficiences of popular bubble propagation techniques are examined and a novel propagation technique is introduced and explained. This technique exploits the attributes of a bubble lying in an etched channel. The dynamics of such a bubble are studied in terms of its total internal energy. To identify the primary energy and force requirements which determine the stability of a bubble lying in the etched channel, theory is developed for, and experiments conducted on, bubbles lying in simple etched disc. A theoretical model of the bubble's movement in the etched channel is derived, and experiments conducted to verify this model. A possible manner of extending the etched channel propagation technique into a working device incorporating input/output functions is outlined. The wave equation for spin waves in YIG is derived and its solution shown to yield two types of spin waves - the surface and the volume wave. The dispersion characteristics of those waves is examined and those situations which elicit non-dispersive behaviour are identified. A YIG delay line which exploits this non-dispersive behaviour has been designed and built. The delay of this line is studied experimentally and its characteristic insertion loss measured. To obtain efficient transduction of the input r.f. signal to the magnetic spin wave, it is necessary to reduce the insertion loss. The criteria which have to be met to effect a reduction are shown to be dependent on the radiation resistance of the transducer. A model of the transducer radiation resistance is derived and experiments are carried out to verify this model.