Isolation effects were examined when the initial onset of a continuous serial list was or was not disguised. Disguising that onset significantly impaired total-list acquisition and virtually eliminated serial-position effects; isolation effects occurred for the item following the isolated item. In the absence of an explicitly discriminable item in a homogeneous series, subjects established an idiosyncratic starting point on which to base their learning. When either a discriminable beginning in a continuous series or an isolated item in an otherwise uniform series was available, it was used as a starting point.