His.
Very pedestrian stuff, sometimes, but I think a decent illustration.
I'm in the oilfield chemicals business. One very common practice used to get more oil out of the ground quicker is called acidizing: you pump a couple of thousand gallons of hydrochloric acid down a well to dissolve some of the (limestone) rock down there and provide easier paths for the oil in that rock to the well and thus to surface and finally our gas tanks.
It was noticed quite a few years ago that this process works better in some areas than in others: in the Panhandle of Texas, for example, you can usually pump acid with almost no additives into a well and get improved flow of nice, clean oil. Down here in the Permian Basin, you can pump the same acid and frequently
decrease production as well as getting big globs of stuff called "sludge" that looks sort of like licorice pudding might look if it existed. In parts of Canada, you can completely plug up wells by acidizing them with just plain acid.
Hypothesis #1, arrived at as recently as 1990 (!) out here, was that acid is somehow reacting with crude oul to make this pasty crud. Lab tests using clean crude and additive-free acid failed to support H#1, though: I can mix up reagent-grade acid and crude from the San Andres formation from Howard County, TX, and they'll separate right out cleanly, with no glop formed.
Hypothesis #2 followed. It was, "maybe we're dissolving enough rust out of the pipe as we pump this acid to change things a bit...." Lab tests confirmed this: if I dissolve some rust into my acid before I mix it with that same oil from Howard Co., I'll get sludge that looks like the stuff that comes out of those wells over there after an acid job! A little more lab work showed that only the ferric form of iron - that in the +3 valence state - caused sludge to form. Ferrous, or +2 state, iron, didn't cause a problem.
Hypothesis #3 followed: we can't get all the iron out of the acid, 'cause it's already 3000 feet down the pipe headed toward that reservoir. But if we could come up with something that we could put in the acid that would change all the +3 iron to +2 iron, would it stop sludge from forming? {Five years of Edisonian trying of stuff and hard work was inserted here before a chemical mixture that 1) did so 2) was affordable and 3) was practical to use in acid was developed.} A godawful-nasty smelling mixture was indeed found that did prevent sludge in lab tests, and then in field jobs. (I eat steak occasionally because of my part in developing this product. The owner of our company eats it even more often.)
Real simple in principle: guess at something, test out your guess, try again if it fails. Repeat. There are oils, uncommonly here but common in California, for one place, that don't fit the "ferric-iron-is-the-baddie" hypothesis. Clean acid alone is enough to make them sludge, and completely different additives are needed to fix their problems. The Panhandle oils I mentioned don't sludge at all, iron or not. But for most Permian Basin oils, additives like the one we have fix the problem.