The wedding formula has much ascorbic acid, and much metaborate for buffering. I decided to eliminate these variables separately. Following PE's rule of "never do one test", I ran two strips with a formula having less ascorbic acid. Here are their graphs:

Curve11.7-XTOL.jpg and Curve11.15-XTOL.jpg

Here's the formula. Note the addition of citric acid to maintain pH:
Sodium sulfite .................. 45 g
Sodium metaborate .......... 5.4 g (very high to give strong buffering)
Ascorbic acid .................. 4.5 g (a normal amount; AA/Phen ratio=90)
Citric acid ....................... 1 g (brings pH back down to 8.16)
Phenidone ...................... 0.05 g (developer)
Propylene glycol ............. 23.5 ml (amount needed in a concentrate for 30 ml/L)

The good news is the right-side rise seems to be gone. But I notice that both graphs have a bit lower EI and a bit higher contrast compared to XTOL and compared to what I was getting with earlier formulas. These differences are probably too small for folks to notice, but I'd like to know why this is happening. I'm mixing propylene glycol into the water to simulate a concentrate, and I'm wondering if PG reduces EI slightly, or if that's due to the very strong buffering, or if the citric acid could do this. So I'll try reducing metaborate and eliminating the citric acid, unless you have a better idea.

Mark Overton