Problems with Hydranal water standard?

When performing volumetric KF, we standardize our titrant using either Sodium Tartrate Dihydrate or Hydranal Water Standard.  When we use the latter, we frequently find that the titrator behaves strangely the next day - in fact for several days - even after thoroughly flushing the titration vessel with methanol and/or titrant.  The background titration of drift proceeds to the point where the vessel solution is dark brown.  I assume it's over-titrating, because when I add a mg  or more of water, the drift reading either doesn't react, or reacts only after 5-10 seconds, and slowly creeps up.  Titrations also proceed much more slowly.  The Water standard seems to be composed of xylene and 1-butanol with a quantitative trace of water.  So, two questions:  1) has anyone seen this behavior?  2) can anyone explain it?  Thanks!