The picture superiority effect in spatial working memory refers to a better retention of the locations of objects or images compared to locations of words. The present research suggests that such effect extends to situations in which the items are serially presented. Participants were required to reproduce different series of icons or words either according to the original spatial configuration, or following the temporal sequence of presentation, or both these aspects together. Overall, both memory for spatial locations and for the temporal order were better with icons than with words. Furthermore, performance decreased when both the features had to be simultaneously retained, suggesting that we do not automatically integrate spatial and temporal information, but such binding requires attentional processing.