FPU expertise/fprem1 anyone?

Can anyone tell me why the following code gives me a remainder of -0.4655 not +0.4655?

Code:

include 'macro.inc'      ; that's mine, based on win32ax.inc
.data
_Const4 dq 1.2
_Const5 dq 0.7345
FltWrk          dd ?
.code
start:
 fninit                                          ;; initialise FPU
   fnstcw word[FltWrk]
 or word[FltWrk],0x0C00          ;; set rounding mode as truncate towards zero (for floor())
 fldcw word[FltWrk]

;if rmdr(1.2,0.7345)!=0.4655 then puts(1,"rmdr(1.2,0.7345)!=0.4655\n") abort(1) end if
 fld [_Const4]
       fld [_Const5]
  @@:
  fprem1
      fstsw ax
    and ah,1
    jnz @b
      fstp ST1

        invoke ExitProcess,0
.end start

Update: I now think I was a bit infatuated with "IEEE754" and the docs saying "all new softare should use fprem1, not fprem". I managed to get the expected result by changing to fprem and swapping the flds.