1: \begin{abstract}

2: The strange quark mass is extracted from a finite energy

3: sum rule (FESR) analysis of the flavor-breaking difference of 

4: light-light and

5: light-strange quark vector-plus-axial-vector correlators, using spectral

6: functions determined from hadronic $\tau$ decay data.

7: We point out 

8: problems for existing FESR treatments associated with potentially

9: slow convergence of the perturbative series 

10: for the mass-dependent terms in the OPE 

11: over certain parts of the FESR contour, and

12: show how to construct alternate

13: weight choices which not only cure this problem,

14: but also (1) considerably improve the convergence of the integrated

15: perturbative series, (2) strongly suppress contributions from

16: the region of $s$ values where the errors on the strange

17: current spectral function are still large and (3) essentially

18: completely remove uncertainties associated with the

19: subtraction of longitudinal contributions to the experimental

20: decay distributions.

21: The result is an extraction of $m_s$

22: with statistical errors comparable

23: to those associated with

24: the current experimental uncertainties in the determination of the

25: CKM angle, $V_{us}$.  We find

26: $m_s(1\ {\rm GeV})=158.6\pm 18.7\pm 16.3\pm 13.3\ {\rm MeV}$ (where the first 

27: error is statistical, the second due to that on

28: $V_{us}$, and the third theoretical).

29: \end{abstract}

30: