SCIENTIFIC INVESTIGATIONS IN THE kitchen have elucidated the mechanisms behind various culinary tricks and offer a means to create novel dishes [see "Chemistry and Physics in the Kitchen," page 66]. Experiments are easy, and all that is needed are shopping trips to the supermarket and to an electronics-parts supply store. An inexpensive microscope that has 400-power magnification is useful but not indispensable.

In the main article, we mention how we quantitatively determined the best way to produce a soufflé. Remarks made at a workshop in Erice, Sicily, indicated that it was not essential to put a soufflé in the oven immediately after folding in the whisked egg whites. It may be left for an hour or so in a hot water bath (a bain-marie, or double boiler) at 40 degrees Celsius (104 degrees Fahrenheit). Moreover, individual soufflé mixtures can be deep-frozen and thawed for cooking-a great help in catering a large dinner party or a banquet.

You can test the hypotheses with individual cheese soufflés. The ingredients consist of 1/3 cup (75 grams) of butter (about 2/3 of a stick), about 2/3 cup (75 grams) of flour, 12/3 cups (400 milliliters) of milk, four egg yolks, four egg whites and about 1/2 cup (50 grams) of a grated cheese such as Gruyere. (Although we provide standard kitchen measures, we advocate the metric units for their greater precision.) Add salt, pepper and nutmeg to taste. The ingredients will fill five individual soufflé dishes that are about 10 centimeters in diameter and 6.5 centimeters high. Be sure the dishes have been buttered and coated with bread crumbs or flour.

First, you need to make the bechamel sauce. Over medium heat, melt the butter in a saucepan, then stir in the flour. Under the microscope, the starch granules in the flour paste will appear tightly packed. Add hot milk to the sauce pan (starch molecules do not dissolve readily in cold milk). Stir until the mixture reaches a thick, creamy consistency. You will see under a microscope that the starch granules have swollen considerably. The swelling is one of the reasons for the high viscosity of the bechamel sauce: the granules cannot move freely. Another reason is that the starch liberates long polymers made of glucose, which help to create a gel structure. The polymers, which are too small to be seen under the microscope, are amylose (linear chains) and amylopectin (branched chains).

Remove the pan from the heat, and stir in the cheese. When the mixture has cooled to 50 degrees C (the sauce should feel tolerably hot to the touch) add the egg yolks. You may add the yolks singly in one experiment, in twos in another and the four of them together in a third to see whether it makes any difference, as some cookbooks state.

Whisk the egg whites in a clean bowl until the foam is stiff (it should be able to support a whole egg). If you examine the foam under the microscope from time to time, you will see that the air bubbles introduced by whipping become smaller and more numerous as the foam thickens. Surface tension explains why smaller bubbles produce a more stable foam. This force, which is the same as that which creates the meniscus at the surface of a glass of water, draws air bubbles firmly against one another and prevents the water from draining out of the bubble walls. In the long run, gravity will eventually destabilize the whisked egg whites.

Gently fold the cheese mixture into the egg foam. Then ladle the mixture into the five soufflé dishes so that it comes about two thirds of the way up the sides.

The experiment on each soufflé consists of measuring its internal temperature during the cooking, the height the soufflé reaches and the rate of collapse after it has been taken out of the oven. These parameters will help you determine how much time one may leave between ladling the mixture into the dishes and placing it in the oven. If your oven has a window, you may want to try measuring the height of the soufflé as it cooks by securing a metal ruler to the soufflé dish with a wire.

Put one soufflé into the oven immediately. Leave the second in a 40 degree C water bath for 45 minutes. Allow the third to stand at room temperature for two to four hours. Place the fourth in the refrigerator for four to six hours; let it warm to room temperature for two hours before placing it in the oven. Store the last in the freezer for 12 to 48 hours, letting it thaw six hours before cooking. In all cases, the oven should be preheated to 180 degrees C (360 degrees F).

The device used to record the temperature inside the soufflé during cooking is called a thermocouple. Essentially, a thermocouple consists of two wires made of different metals joined at a point. It measures temperature as a change in voltage between the two wires.

To make a thermocouple, buy copper and constantan wires that are about 0.2 millimeter in diameter and have insulation that can withstand 200 degrees C. Spot welding is the best means to join the wires, although soft solder will probably do, because the soufflé temperature will not exceed 100 degrees C. One source of thermocouple wire is Willson Scientific Glass (528 East Fig Street, Monrovia, Calif., 91016; telephone: 818-303-1656; fax: 818-3030599). The company will fuse the wires for you and encase them in the appropriate insulation. Although the wires are inexpensive, most companies have a minimum order of $25 to $50. This amount will buy you at least a couple of meters of wire-far more than you will need.

You can also purchase ready-made thermocouples. Look for a copper-constantan or a chromel-alumel thermocouple, which are called K-type and T-type thermocouples, respectively. We recommend, however, that you build your own, because we have found that commercial thermocouples can be too large and too heavy for this experiment.

The thermocouple may be immersed directly in the soufflé mixture. But we prefer to thread the wires through a thin stainless-steel tube, which helps to keep the wires centered in the soufflé. The tube should be about 1.5 to two millimeters in diameter and 40 to 50 millimeters long. A hypodermic needle is a good choice. Anchor the junction of the thermocouple wires to the tip of the needle with a heat-resistant epoxy resin, such as Araldite. Willson Scientific has said it will sell thermocouple wires threaded through a thin glass tube, which would render the needle unnecessary.

To record temperature, connect the thermocouple wires to a voltmeter that can measure to the nearest 50 microvolts. This voltage is proportional to the temperature difference between the junction immersed in the soufflé and the voltmeter. To calibrate the thermocouple, take voltage readings for ice water (zero degrees C) and for boiling water (100 degrees C). If the zero degree C reading produces a negative voltage, reverse the wires at the voltmeter terminal. From these two points, you can easily calculate the corresponding temperature for any given voltage. The output voltage depends linearly on the temperature. For the copper-constantan thermocouple, each Celsius degree rise corresponds to an increase of 42 microvolts. For the chromel-alumel one, it is 40 microvolts. Some commercially available meters can give the temperature directly. But they are usually supplied with probes that may not be suitable for this experiment.

The rising soufflé can move the thermocouple, so you must secure the device to the soufflé dish. Wrap a stiff, metal wire around the dish, bending and looping it high over the top [see Figure 1]. The loop should be at least 100 millimeters high. You can use a short length of metal wire to tie the thermocouple wires to the loop.

To prevent the thermocouple wires from being damaged when the oven door is closed, they should be protected by a glass-cloth sleeve about two to three millimeters in diameter. Such sleeves are available from scientific supply shops. An alternative might be a soft, heat-resistant cloth wrapped around that part of the wires sandwiched by the door and frame.

During the cooking, you will note two temperature steps, one at 60 to 80 degrees C and the other about 100 degrees C. The first corresponds either to the point at which the proteins coagulate or to the rise of the cold layers of the mixture to the thermocouple. The other is the vaporization point of water and corresponds roughly to when the soufflé is done (about 20 minutes).

The soufflé rises in part from the expansion of air in the myriad small bubbles in the whipped eggs. Vaporizing water also aids the rise. The coagulation of the proteins keeps the soufflé high after cooking because the process stiffens the walls of the air bubbles. This coagulation may also play a role in why opening the oven door too soon ruins the soufflé. The bubble walls may become sufficiently rigid so as to resist subsequent rising.

Once the soufflé is done, measure its maximum height. By comparing the heights achieved by the different soufflés and the time it takes for them to collapse, you will conclude that the old recipes are correct when they prescribe cooking the soufflés immediately. The longer the delay between ladling and cooking, the less successful the soufflé is, although even the deep-frozen soufflés are acceptable.

The soufflé experiment quantifies and explains observations made by generations of chefs. But science and technology can also create new dishes, as exemplified by the microwave oven. In traditional cooking, heat penetrates slowly into the food. Taste and consistency depend on the temperature the various parts of the dish reach during the cooking process. In a crisp bread roll, for example, the crust will have reached a temperature of 200 degrees C or more, but the soft inside does not get much above 80 degrees C, because the air bubbles within insulate the dough.

In contrast, microwaves penetrate the food fairly evenly. The waves impart some of their energy to the individual molecules, setting them into motion The friction between these molecules and their surroundings generates the heat. For microwave ovens to function, the food must contain polar molecules. That is, the electrical charges in the molecules must be asymmetrically distributed. Furthermore, the substance being cooked must not be entirely solid, because their molecules, being fixed, cannot move in response to microwaves.

These two conditions can be simply demonstrated. First, compare the behavior of liquid paraffin, composed of nonpolar molecules, and glycerol, made of polar ones. In a microwave oven, glycerol quickly comes to boil, but the paraffin barely heats.

The difference between liquid and solid states is evident in the following demonstration, which amazes even professional scientists. Make a hollow ice block by placing a four-centimeter-diameter beaker into a larger one, about eight centimeters in diameter. Fill the gap between them with water and put them in the freezer. To prevent the beakers from cracking, freeze the water in stages, adding eight- to 10-millimeter layers of water at a time. Once you have frozen a sufficient amount of water, separate the beakers from the ice under warm running water. Place a small beaker filled to three quarters with water into the ice block. The water will come to a boil in a microwave oven in about 30 seconds, but the ice remains so]id and unmelted. Assorted glass containers made from Pyrex will do if you cannot obtain beakers.

The ice-block demonstration is actually the basis for inverting the characteristics of a Baked Alaska. Also called, for obvious reasons, a Frozen Florida, an inverted Baked Alaska consists of a meringue case filled with a concoction rich in alcohol and sugar and covered with chocolate icing. To make the meringue cases, whisk one egg white. When soft peaks form, gradually add about two tablespoonfuls (25 grams) of granulated sugar. Continue whisking until the foam is quite firm (about five minutes), then fold in about 11/3 tablespoonfuls (20 grams) of sifted confectioners' sugar.

Deposit the mixture in four round parts on a baking sheet lined with wax paper (one egg white will yield two Frozen Floridas). Each disk should be 50 to 60 millimeters in diameter and about five to 10 millimeters thick. These disks serve as the lids and bases of the meringue cases.

To make the sides, place the remaining meringue mixture in a forcing bag, used to frost and decorate cakes. The nozzle should be about 10 millimeters in diameter. Spiral the meringue on the two meringue disks that serve as bases. The sides should be about 40 millimeters high [see illustration at left]. Dry the two vol-au-vent cases and the two lids by baking them at about 120 degrees C (250 degrees F) for about two hours.

The filling consists of equal parts jam, sugar syrup and fruit brandy, such as kirsch or apricot (it is best to match it to the type of jam used). Pour the mixture into the dried cases until they are almost full and seal the lids with caramel syrup. You can make the syrup by boiling a mixture of brown sugar and water. Cover with a layer of icing one to two millimeters thick. The icing is made of 1 3/4 ounces (50 grams) of chocolate (preferably 60 to 70 percent cocoa solids) mixed with about a teaspoonful (10 milliliters) of melted, clarified butter. Do not allow the temperature to exceed 65 degrees C (150 degrees F ), because the icing may become lumpy. Make a pinhole in the lids, wrap the confections in plastic wrap and place them in the freezer.

Allow at least 10 hours for the meringues to freeze. Once frozen, they will keep for months. Because of its high alcohol content, the liquid mixture inside will not solidify.

When removing a meringue from the deep freeze, quickly unwrap the plastic film so that water does not condense on the icing. Place the meringue in the microwave set on high and cook until bubbles emerge from the pinhole in the lid (about 30 seconds). The bubbles indicate that the filling has begun to boil. In contrast, the icing remains cold, thanks to the good insulating properties of the meringue. You can increase these properties by coating the inside of the meringue with caramel. The coating prevents the filling from penetrating the meringue and thereby reducing its insulating abilities.

As an alternative to the vol-au-vent technique described, you can try scooping out baked meringue through a small hole (five to 10 millimeters in diameter). Fill with the jam mixture and seal the hole with icing. Such a technique can make bite-size Frozen Floridas. Before serving such small ones, warn any unsuspecting guests about the hot interior.

Of course, you can experiment with the composition of the filling and icing. Freeze samples of identical sizes, place them a few at a time in the microwave oven and determine the time for the filling to boil and for the icing to soften. Perhaps the only drawback to the endless possibilities in experimental gastronomy is that the proof is in the eating. A series of experiments may ultimately compel strict dieting.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.