Several studies on cross-modal attention showed that remapping processes between sensory modalities occur in the spatial orienting of attention. One hypothesis that accounts for such links is that spatial attention operates upon representations of common locations in the external space. However, convincing evidence only exists for cross-modal links in spatial orienting, leaving the dynamics of these effects unexplored. Four experiments were designed to cope with this issue by having participants being involved in an endogenous orienting task to visual and auditory target stimuli. Targets on invalid trials were embedded into two different kinds of sequences of stimuli: (1) long sequences, wherein three valid trials in one modality preceded the invalid trial in the other modality; (2) short sequences, wherein only one valid trial in one modality preceded the invalid trial in the other modality. Results revealed modality-specific meridian effects in the short sequences, and a significant decrement of the modality-specific meridian effect in the long sequences. The results of these experiments indicate that cross-modal links in visual and auditory spatial attention are based on representations of common locations in the external space across sensory modalities. Moreover, the results strongly support the hypothesis that representations of space are dynamically built and updated according to task demands.