The integrated use of spatial and temporal information seems to support the separation of two sensory streams. The present study tested whether this facilitation depends on the encoding of sensory stimuli in externally anchored spatial coordinate systems. Fifty-nine children between 5 and 12 years as well as 12 young adults performed a crossmodal temporal order judgment (TOJ) task for simple visual and tactile stimuli. Stimuli were presented either within the same or in different hemifields. Presentation of the two modality inputs in different hemifields improved TOJ only in children aged 10 years and older. In contrast, intramodal TOJ performance (data from Pagel, Heed, & Roder, 2009) was better than crossmodal TOJ performance starting at the age of 6 years. An adult-like level of performance in the crossmodal TOJ task was evident only at the age of 12 years. We speculate that the ability to redundantly code sensory input in modality-specific and modality-independent spatial coordinates facilitates intramodal temporal processing. Further refinement of the processes providing external spatial coordinates then results in the integrated use of space and time to decide whether sensory inputs belong to a common object or to separate events. (Contains 1 table and 2 figures.)