Saturday, September 17, 2005

Cake Theory 101

At least a year ago, I played around with the idea of writing a few pieces on the theory and practice of making cakes. I decided to have a shot at developing a cake recipe from first principles. I had no idea what the first principles were, but I did have a collection of cook books and a basic knowledge of physics and chemistry to work from. I was inspired, for want of a better word, by a television foodie segment where someone presented an "Angel Food Cake" This confection was basically a meringue with some flour tossed into the mix and a raspberry puree stirred through once the batter was in the tin. The alleged cake used about twelve egg whites but no yolks, which struck me as rather wasteful. The segment should really have been followed by one on zabaglione or some other dish which would use up the twelve left-over egg yolks.

My irritation at the thought of all those wasted egg yolks was minor. What really got me going was the emphasis the presenter laid on the importance of aeration. I pretty much went through the ceiling when she was explaining how sifting your flour helped to add air to the batter. It's precisely this kind of superstitious nonsense that puts people off cookery, I thought, and set out to debunk as many cake superstitions as I could, at least to my own satisfaction.

I began with the recipe books, reading up on the preparation of various sponge cakes, to get an idea of what they all had in common. All of them begin with a foam consisting of very small air bubbles suspended in a mixture of egg-white and sugar or, in the case of the genois which I eventually chose as the basis for my very own original recipe, air bubbles suspended in a mixture of whole eggs and sugar. To this foam you add flour, flavourings and butter. All very simple when you think about it.

For debunking purposes, I decided that the cake was to be flourless. I settled on almond meal as its replacement. In most other respects, the cake would be a standard genois. To add a little moistness, I decided to incorporate a couple of layers of cooked fruit in the cake at cooking time; in the first version of the cake I used apples, lightly sauteed in butter with brown sugar and a little cardamom.

The result was interesting; it rose quite well in the oven, and promptly collapsed once cooking was complete. Unfortunately, the substitution of almond meal for flour wasn't as straightforward as I thought; 100 grams of almond meal didn't do the work of 100 grams of flour. The end result, while edible, was not a cake. It was more like a very thick pancake. It might make an acceptable clafouti, once cherries were in season, but they weren't and that was the end of it.

I did a little thinking about the problem and, in particular, what the flour in a cake is supposed to do. What is flour after all? It's ground up wheat, and mostly it consists of starch. In the process of cooking, the flour absorbs moisture and bulks up. Ah-ha! If I was going to substitute nut meal for flour, I needed to use enough of it to match the bulk of the cooked flour. Or something like that. One thing was clear; almond meal is impossible to sift, so sifting wasn't going to contribute anything to raising the cake. All the lift was going to come from that foamy mixture of air, eggs and sugar.

So, on to version two; I worked out the conversion of 100 grams of flour into cup measure, and had a bit of a think about how many cups of almond meal would equal one cup of flour. I figured I needed at least as many ground up bits of almond in my mix as there would be flour particles in a cup of flour. I settled on doubling the volume measure for starters. Plus, when I came to mix the almond meal into the frothed up eggs, I would have to keep an eye on the consistency of the batter and adjust quantities as I went on. Nonetheless, it was all very simple when you think about it.

Like version one, version two rose quite well in the oven. This time, to prevent the catastrophic collapse, I left it in the oven, with the oven turned off, once cooking was complete. It collapsed anyway, but nowhere near as much as version one had. The collapse didn't bother me - well, not much. On examination, there's not much reason to expect a cake where all the rise comes from the air incorporated into the original batter to stay puffed up once it cools. The cake rises because all those air bubbles expand as they are heated; once they cool down to ambient temperature again, they're bound to contract. It's got something to do with the Ideal Gas Law; as the temperature inside the cake goes down, so does its internal gas pressure. As a result the external atmospheric pressure pushes the cake back to more or less the volume your original batter had. There is no way to avoid this. As long as the cake comes through the process without getting too deformed, you're doing fine.

And version two was fine; quite good enough to cut up into large chunks and fob off on various friends. And good enough to trot out at a recent gathering chez Gummo. And, what the hell I'm going to publish the recipe here, because you never know, there might be somebody out there who might want to give it a try.

Equipment & Materials:

A large mixing bowl (and I do mean large - I use a stainless steel bowl about 30 cm in diameter);
An electric hand mixer
A rubber spatula
A large saucepan
A smaller saucepan
A 13 cm springform cake tin.
Kitchen paper
Oven pre-heated to 180 C


6 medium - large size eggs
3/4 cup sugar
1/2 teaspoon vanilla extract (no I don't mean vanilla essence - it's not the same at all)
2 - 2 & 1/2 cups of almond meal
6 ounces of butter
Punnet of fresh berries (blueberries or raspberries, but definitely not strawberries, OK, suit yourself, see if I care).

Butter and flour the springform tin and line the bottom with a circle of kitchen paper.
(To cut a circle of kitchen paper: fold a square of kitchen paper in half diagonally, and in half again (i.e. into triangular quarters). Fold again into eighths, and once more into sixteenths. Place the point of the paper triangle at roughly the centre of the spring form and cut across the paper where it touched the rim. That's as near as damn it to a circle, and quite good enough for present purposes).

Melt the butter in the smaller of the two saucepans and set aside to cool.

Fill the large saucepan with water and bring it to the boil.

Break the six eggs into the large bowl, add the sugar and vanilla extract. Beat gently until they are all combined.

Warm the egg mixture over the pot of boiling water until it is lukewarm*. Remove from the heat.

With the electric beater, beat the eggs until they form "the ribbon". This is going to take some time; expect to spend at least 10 minutes beating the eggs. They're ready when:
  • They have at least trebled in volume;
  • The colour looks suitable for painting a wall or a picket fence;
  • A little of the mixture dripped off the beaters sits on top of the rest of the mixture without immediately dissolving back into it.
As I said, this takes at least 10 minutes, usually longer. If you're using a hand whisk of any description expect it to take a hell of a lot longer. In fact, you should probably plan on dying from exhaustion well before you're finished.

Once the eggs are thoroughly beaten, fold in the almond meal in stages, using the rubber spatula. Then fold in the melted butter. Do this as quickly as possible.

Pour half of the batter into the cake tin and sprinkle the berries on top. Now pour in the rest of the batter. Smooth the surface with the spatula till it is more or less level. If possible, it should be a little higher in the centre than at the sides. Get the tin into the oven.

Bake for about 45 minutes, then test with a skewer. The cake is cooked once the skewer comes out completely clean. When the cake passes this test, turn off the oven and allow the cake to sit in it for at least another half an hour before taking it out.

Once it's out of the oven, put it on a plate (paper circle and all). Tart it up however you like - apricot glaze would probably work well but I'd lay off the decor mexicain.

* - since writing up the recipe, I've started to entertain doubts about this business of heating the egg mixture until it's lukewarm. One day I'll have to try it with eggs straight out of the refrigerator, to see if this is another inessential step I can dispense with.