Gold band electrodes with widths ranging from 3 to 150 nm are used to investigate the voltammetric response of a series of substituted ferrocenes in non-aqueous electrolyte. The Au electrodes are fabricated by insulating, with an epoxy, all but the edge of an evaporated An film deposited onto a silane-modified glass substrate. Plots of normalized limiting current for each ferrocene derivative as a function of normalized electrode width reveal a negative response from traditional theoretical predictions. The theory used for the band electrodes is based on semiinfinite diffusion of the ferrocenes to a hemicylindrical electrode. The observed negative responses for electrodes less than 50 nm in width are explained by effects that become important when the sizes of the electrode and redox species are comparable. Concerns regarding the real surface area of the band electrodes are addressed by labeling of the exposed An of the nanoband electrode with a ferrocene-tagged alkanethiol and measurements of double-layer capacitance. These studies indicate that roughly 20-500 times more area is exposed than that calculated from the nominal width and length of the bands, even though X-ray diffraction and stylus profilometry determinations of An thickness are in agreement with those found by quartz crystal microgravimetry during evaporation of the An films. Scanning electron microscopy studies of the epoxy \ An \ glass interface point to insufficient insulation of the Au by the epoxy as the result of defects.