I'm guessing this is referring to the fact that early versions of Fortran stored return addresses in specific memory locations (at the end of the function defin

I'm guessing this is referring to the fact that early versions of Fortran stored return addresses in specific memory locations (at the end of the function definition IIRC) instead of on a call stack. This is why those versions of Fortran couldn't do recursion, because the new return address would overwrite the old one.