A system consisting of any array of cylindrical, polytopic membrane proteins (or protein complexes) possessed of a permanent dipole moment and immersed in a closed, spherical phospholipid bilayer sheet is considered. It is assumed that rotation of the protein (complex) in a plane normal to the membrane, if occurring, is restricted by viscous drag alone. Lateral diffusion is assumed either to be free and random or to be partially constrained by barriers of an unspecified nature. The dielectric relaxation times calculated for membrane protein rotation in a suspension of vesicles of the above type are much longer than those observed with globular proteins in aqueous solution, and fall in the mid-to-high audio frequency range. If the long range lateral diffusion of (charged) membrane protein complexes is essentially unrestricted, as in the "fluid mosaic" membrane model, dielectric relaxation times for lateral motions will lie, except in the case of the very smallest vesicles, in the sub-audio (ELF) range. If, in contrast, the lateral diffusion of membrane protein complexes is partially restricted by "barriers" or "long-range" interactions (of unspecified nature), significant dielectric dispersions may be expected in both audio- and radio-frequency ranges, the critical (characteristic) frequencies depending upon the average distance moved before a barrier is encountered. Similar analyses are given for rotational and translational motions of phospholipids. At very low frequencies, a dispersion due to vesicle orientation might in principle also be observed; the dielectrically observable extent of this rotation will depend, inter alia, upon the charge mobility and disposition of the membrane protein complexes, as well as, of course, on the viscosity of the aqueous phase. The role of electroosmotic interactions between double layer ions (and water dipoles) and proteins raised above the membrane surface is considered. In some cases, it seems likely that such interactions serve to raise the dielectric increment, relative to that which might otherwise have been expected, of dispersions due to protein motions in membranes. Depending upon the tortuosity of the ion-relaxation pathways, such a relaxation mechanism might lead to almost any characteristic frequency, and, even in the absence of protein/lipid motions, would cause dielectric spectra to be much broader than one might expect from a simple, macroscopic treatment.