On Jun 29, 3:46 pm, Sam > wrote:
> I'd say, this is a potential measuring error asking for attention.

Yes, after rerunning the experiment with closer attention, it was
probably a measurement error associated with a small area ISFET in a
recess where it is easy to retain a small air bubble when the sample
is viscous.
When I went to 200% hydration sample, the problem became easier to
avoid, but the sensor is still very sensitive.
Look he
http://picasaweb.google.com/DocDough...ighGlutenWhite
for a new data set.

The high gluten white starter was mixed at 100% hydration while the
whole wheat was 200%, and I think that both of the excursions from
smooth curves could be reflections of the "bubble" phenomenon. For
each data point, two calibration samples were read (at pH 4.01 and
7.00) after measuring both starter samples. These calibration samples
were used to remove sensor bias drift (all adjustments were in the
range of .02 to .04).

It is still possible that there is a small pH increase after the
initial mixing of starter, water, and flour, but it is on the order of
the sensor bias drift.
There was no apparent scale factor change over the test interval.

See:
http://picasaweb.google.com/DocDough...onsInHydration
for a data set illustrating the stability (at a pH of 6) of the pH of
a sample of flour and water (no starter), the corresponding
calibration measurements,
and the effect of increased sample hydration.
These curves have not been corrected for the sensor offset drift which
can be seen in the top and bottom curves (and also in the flour/water
data).