Spatially overlapping media

I am interested in displaying multiple media streams which overlap both 
spatially and temporally, for example video playing within a video, where 
the two video streams have different sources.  While I can't find anything 
in the SMIL 1.0 spec to preclude doing this, I also can't find any 
specification of the semantics for multiple layers of media.  So, can this 
be done?  If so, is there a way to control the order of layering?

-Chase