That was a very interesting article you linked, Miss Wiggle! It makes perfect sense too, since in nature fish constantly swim through thermoclines in the water and get exposed to quick pH and temperature changes every time it rains. But they don't get naturally exposed to high levels of ammonia or sudden changes in CO2.
Logically, it'd follow that the higher the pH of your tank water is, the more dangerous it would be to mix it into the bag (on top of the pH change caused by the loss of CO2). Ammonia gets ~10 times more poisonous when pH goes up by 1 (or more accurately, the proportion of ammonia to ammonium gets ~10 times higher). In an extreme case, going from pH 6 to pH 8, the proportion of toxic ammonia would increase 100-fold. Temperature affects this as well: higher temperature means higher proportion of toxic ammonia, but to a lesser degree (about 50% increase when going from 22C to 28C).
I think this is really worth some consideration. One 20 gram (4-5") fish that's been fed sort of generously could easily produce 12mg of ammonia during a longer transport (assume it has eaten 2% of its body weight the day before, or 400mg; 3% of typical fish food ends up being converted to ammonia). How much water is in a typical bag? Two liters? That'd be 6 ppm of total ammonia+ammonium in the bag; major toxicity if exposed to anything but pretty acidic pH!
Am I missing something? Because looking at those numbers, I think I'll avoid doing any water mixing (and premature bag-opening!) in the future, following the instructions in the article Miss Wiggle posted instead.
By the way, if anyone's interested, I ran into
this handy ammonia/ammonium proportion calculator.