After my attempt at experimental baking, several people recommended Shirley Corriher’s BakeWise, so I bought a copy and now I’m in love. Corriher is a biochemist by trade, as well as an accomplished chef, and every page of the book has delicious insights. And it’s clarified something I’ve been wondering for ages: why do we let shortcrust pastry rest in the fridge between mixing, rolling, and baking?

Some recipes say that it’s to let the gluten in the flour "relax", suggesting that the gluten is troubled by the hectic pace of modern life. The pastry doesn’t unmix, and doesn’t visibly change, so is anything happening there at all?

To be sure, it's easy to understand one part of putting pastry in the fridge. Most pastries (as Corriher explains) rely on having sheets of butterfat between flour-water layers, so that the layers of flour are kept well separated until they reach the hot oven. That only works if the fat is cold, hence the general emphasis on keeping things chilled.

On the other hand, Corriher and others point out another use of resting: it lets water get distributed properly throughout the dough, which reduces the chance of shrinking and cracks.[1] How could that be, when the dough isn't visibly changing?

At last the penny dropped. It turns out that I found this puzzling because my mental picture of gluten was wrong.

Any cookbook will tell you that when you mix flour with water then stir it around, the wheat proteins get tangled together and form elastic sheets of gluten. I imagined that they were gradually forming a regular structure, like a lattice. Here’s Corriher:

As a chemist, I had a hard time understanding gluten. I wanted a formula—like x number of molecules of glutenin plus x number of molecules of gliadin plus x number of molecules of water equal x molecules of gluten. Gluten doesn’t work that way—no orderly formula. It is just a mess of stuff linked together.

If you take a look at pictures of gluten under a microscope, you can see this chaotic mess. Long parallel strands are visible in pure flour, but it only takes a few folding movements before they’re tangled together.

A wet particle of flour with the starch removed (left), stretched by sliding the microscope slide once (centre), then a second time at right angles (right).

The protein in developed dough looks like a fractal.

A heavily stretched gluten membrane

(All these pictures are from a 1991 paper by Amend and Belitz, now available in pdf form.)

That helps explain the need for resting, because shortcrust pastry isn’t layers of tidy sheets. It’s a tangle of springs and elastic bands. I guess that if it just sits on the shelf in the fridge, some connections will snap under their own weight, allowing others to straighten out and gradually untangle, leading to more connections breaking, and so on.

One last thing—a word on technique. Shortcrust pastry recipes often use a ratio of 4:2:1 flour to butter to water. In my limited experience, that’s prone to shrinkage and cracking, no matter how much it’s rested. Is it possible that authors are too eager for round numbers? Recently I’ve upped the butter to 2.25, which is harder to remember, but nicer to eat.

1. The total amount of protein in the flour has a dramatic impact on how much water it will absorb. Apparently it’s a myth that the ambient humidity does much. And the amounts of different types of gliadins and glutenins determine the dough’s springiness and strechiness. Every flour type is different, so you need to tweak the recipe quantities to get the consistency you want.