from reading some an old thread in comp.lang.scheme, it seems like after R4RS, there were a bunch of papers published on macros in scheme- it was a really active time for research in that topic. By the time of R5RS there was already, aside from the R4RS low-level system, syntactic closures, explicit renaming-so maybe it was just a matter of having several low-level systems developed and no consensus on which one to use.