There are some very skilled cooks and bakers who swear by this technique, by which you heat milk (and/or cream) to just below the boiling point (to where a skin forms on the surface) then cool it and use it as the basis of a custard mix. People swear it results in a better-textured custard (it makes the interior creamier, the texture more delicate, the surface thinner, etc). The trouble is they can't explain how or why. It reconfigures the proteins so that they, er...it inverts the um...

And in fact food scientists are clueless as to what they're talking about. Not a one, it seems, can either prove what these chefs are saying, or even form a theory as to how scalding might bring these effects about.

Once upon a time scalding did serve a purpose, back before the days of pasteurization. Then, scalding served to kill bacteria and/or deactivate enzymes that could go on to compromise the safety and/or quality of the finished custard. Nowadays however, with ultra-pasteurization et al, there are neither live bacteria nor active enzymes in the milk we buy (or not many at any rate).

Do I scald my milk? Actually I do, but only because a baker and chef like Thomas Keller tells me to. He says it's critical and who am I to argue? Sure, it's possible that he's carrying forward a piece of antiquated technique, one that he picked up from some French pastry chef and now does only out of habit. Maybe Bouchon uses local unpasteurized milk for its quiches and he must (and forgot to note that fact in his cookbook). Or perhaps he simply knows what he's doing. Food science, helpful as it is, is a very young field. There's still so much we don't know about how food substances — especially substances as complex as milk — behave under certain conditions. The fact that science doesn't know why scalding works isn't proof that it doesn't work. That said Thomas Keller is America's chief practitioner of the food arts. I'm inclined to trust him.

There are only a couple of simple rules to remember about water baths. First, that the water you put in them must always be hot, and second that you must never cover them. Why?

While filling a water bath with lukewarm or even cool water seems to make sense (slow warming, right?), it actually leads to broken custard, since the surface of the custard that's exposed to the 325-degree air overcooks by the time the water heats up to cooking temperature. That said it pays to keep a large pot of water simmering wile you prepare your custard. Fill a water bath by sliding a rack out of your oven part-way, placing the roaster with the custards on it, then carefully pouring in the almost-boiling water.

And as mentioned, never cover it. A covered bath prevents evaporation, creating a steamy environment in which the water eventually boils (and you definitely don't want that).

More than a few custard recipes call for the addition of a starch, usually corn starch, as a stop-gap against curdling. You frequently find it in pie recipes, flourless chocolate cakes recipes (which routinely instruct home cooks to bake them for extended periods at higher-than-ideal temperatures), and in custard filling recipes, especially pastry cream.

How does starch protect against curdling? Well if you remember previous posts on bread baking and/or starch gels, you'll recall that heat and moisture have a way of "unbundling" individual starch (carbohydrate) molecules from starch granules like pieces of hay off a hay bale. Inside a heating custard, those starch molecules get between the coagulating proteins, preventing them from bonding to tightly to one another and preserving the delicate lattice. They also absorb some of the oven's heat, preventing the proteins from getting too hot to begin with.

The up side is that you get a nigh-indestructible custard as long as you don't get totally nuts with heat. The down side is that then end product frequently tastes starchy, and isn't nearly as silky and elegant as the custard made without.

So then what are the consequences for a custard when you overshoot the mark, as it were? Well you remember the earlier discussion of how proteins work, the way in which heat unfolds those long string-like molecules, making it possible for their free bonding sites to adhere to one another to create a liquid-trapping lattice (coagulation). Carry that idea too far and those protein molecules begin to bond too tightly to one another. The lattice tightens, squeezing water out like a fist tightening around a sponge. It's this process that turns scrambled eggs to rubber, lemon meringue pies into soppy curds in pools of syrup, and flourless chocolate cakes into chalky chocolate hockey pucks.

Where "stirred" custards are concerned this isn't necessarily the end of the world. A few lumps in your crème anglaise (and there are usually a few, in even the most carefully prepared batches) can simple be strained out. In the case of a still custard however, you're done, finito, S.O.L.. Nothing to do but chalk it up to experience and start again.

Custards, especially "still" or baked custards, have a reputation for being tricky things to make. I suppose that's somewhat true. We've all had our share of soupy, over-cooked quiches, after all. But it doesn't have to be this way assuming you a) follow your recipe and b) know your oven.

For custard baking is a sort of gambit. You're shooting for a very narrow temperature range, the point at which your egg proteins coagulate and the custard sets. In the case of some very delicate custards a mere five degrees can mean the difference between velvet goodness and disaster. Which is why custards are generally baked at low temperatures.

Oh sure, there are some daredevils out there who bake their custards (especially their pumpkin pies) at high heat, but you really have to know what you're doing to pull a stunt like that. For baking custard at high heat is like driving on ice: everything is peachy until it comes time to stop, at which point you end up in someone's front yard. Probably an off-duty cop, whose wife had just left him, but who spent the afternoon laying out all his expensive outdoor Christmas decorations anyway. Oof.

Which is why responsible bakers drive slowly and deliberately down the slippery custard street, blinkers on, hands at the 10 and 2 positions, gently applying the brake as they approach their stopping point. Nice to see you again officer Petry, very sorry to hear about your domestic difficulties.

Much of the time this slow creep is accomplished in a 325 oven, usually in a water bath You can tell a nearly finished custard because it only just jiggles and doesn't slosh when you tap it. Also there is frequently a subtle color difference in the very center. This is the point to take it out of the oven, the point just before its is completely cooked, as residual heat will finish the job as it sits.

If you're an experienced custard maker, you've likely observed a difference in the makeup of firm or "still" custards vis-à-vis pourable "stirred" varieties. That is, that where one typically calls for whole eggs, the other calls for yolks. This has to do with the difference in the properties of yolk proteins versus white proteins that I discussed in my earlier post. White proteins, being very firm and quick to coagulate, are great for building a custard that has to stand on its own. Yet when it comes to a silky type of dealie like a custard sauce...eh, they're not so good. Here yolk proteins are more desirable. Surrounded by lots of fatty low-density lipoproteins (LDL's), they have a hard time holding on to one another, and so create a weaker, more spreadable and/or pourable custard.

There are lots of amazing things about eggs, but this morning I'm most impressed by their ability to thicken things. All by itself a raw egg is a thickener. Add it to some sweetened half-and-half, put the mix in the blender, and in a few seconds they egg's emulsifiers will render a nice thick egg nog. And that's without enlisting the help of those amazing egg proteins.

Egg proteins make up only about ten percent of the total weight of an egg. Heated and coagulated, however, they turn that egg into an extremely rigid gel (say when it's hard boiled). To arrive at a silky and tender custard, more liquid must be added to the matrix. A basic firm custard mix (say for a flan) consists of about one cup of milk to one egg — a six-fold increase in liquid. Yet the finished product will still be firm enough to stand up by itself when you set it on a plate. Too cool.

Here I should point out that there is more than one type of egg protein. Eggs in fact contains dozens of them, from firm and fast-setting white proteins (ovomucin, ovotransferrin) to weaker, slower-setting yolk proteins. Each has their place, depending on the type of custard you're setting out to make.

Custards run the gamut from sauces and flans to cheesecakes and flourless chocolate cakes. What all these preparations have in common is that they are semi-solid gels that use egg proteins to set what would otherwise be rather runny liquids (this is of course less true of something like a flourless, whose ingredients are solids and semi-solids to begin with, though the batter obeys the same basic set of rules).

How do they work? Essentially by applying gentle heat to a mixture of beaten eggs and liquid (usually dairy of some sort). The heat coaxes the long, languid protein molecules in both the egg white and the egg yolk to uncoil by breaking the chemical bonds that ordinarily keep them wadded up like lengths of yarn. Once the egg proteins are "unwound" they're free to re-bond with one another, which they readily do, trapping other types of molecules like water and fat between them. The result is, well, you get it.

What's especially interesting about this process is that it only occurs in the presence of trace minerals. Trace mineral ions, you see, are positively charged, which causes them to collect around the negatively charged egg proteins. This action prevents the proteins from repelling each other and entices them to bond. And while that might sound like an off-the-cuff nugget of geek science, it's actually quite important. Ever ask yourself why custards are made with things like milk and cream and not water? The answer is that dairy products contain minerals. Heat a mixture of beaten eggs and water and the result will egg drop soup, but add a little salt (a mineral) and you get — yes — a custard. Not one I'd like to eat of course, but a custard still. Pretty cool.

Custards come into two basic varieties, firm baked gels (so-called "still" custards) and pourable creams (known as "stirred" custards). Stirred custards include soupy sauces like crème anglaise, ice cream mix and pastry cream. They're made (usually on the stove top) by stirring the egg-liquid mixture as it sets, and action which breaks up the continuous lattice of bonded proteins, but which nevertheless yields a thickened liquid. "Still" custards are baked in the oven (usually in a water bath) until they form a single, continuous gel.

One other interesting thing about egg proteins is that the degree to which they are diluted significantly effects the temperature at which they coagulate or "set". Thus a very light custard mix (like, say, a quiche) must be heated to about 180 degrees before it sets, while a flourless chocolate cake, in which the eggs are essentially undiluted by water, will set at 140.