Rudi: pH before and after development were 8.17 and 8.13, a drop of .04.
PE: You suggested reducing dev-time. I ran another roll this evening (and got the above numbers for Rudi), but with the shorter dev-time of 12:15 minutes. The graphs of XTOL vs the baseline developer:
It's an excellent match with XTOL, and with slightly better performance near the shoulder. Leader densities are 2.70 for XTOL, and 2.75 with the latest roll. Grain in the baseline dev looks slightly finer than XTOL in loupes.
The remaining mystery is: Why does developing an entire roll require less time than just a test-strip to achieve the same densities? The pH drops instead of rises during development, so rising pH from by-products isn't the reason. XTOL doesn't need less time, and the only substantial chemistry differences are (1) the presence of propylene glycol, and (2) ascorbic acid instead of ascorbate.
Alan suggested that the pH would drop, and it did. I wish I had a deep enough understanding of the underlying chemistry to figure out this stuff. I have books, and plan to re-read the O-chem chapter.