Baking Science curriculum

Water lets the soda and an acid meet and react, giving off carbon dioxide. Heated with no acid, soda gives off less gas than it can and leaves the cake matrix alkaline, so cake formulas add a leavening acid. Soda is matched to its acid by the acid's neutralizing value: how many parts of soda 100 parts of that acid can fully react with.

This route also yields less carbon dioxide than a reaction with acid, and soda used alone can leave an off color in the crumb, which is why it is normally combined with an acidic salt. Ammonium bicarbonate breaks down differently, into ammonia, water and carbon dioxide, and leaves no residue as long as the product ends up low in moisture.

Their acids come from different places: lactic acid builds up as yogurt cultures ferment, gluconic acid is the main acid in honey, and cream of tartar is a tartrate salt that forms during fermentation. In pancake trials, whey drained from strained yogurt gave less dense pancakes than water did, which the authors put down to its acids reacting with the batter's soda.

Without enough acid, soda also releases only part of its carbon dioxide, while the alkalinity builds as it dissolves. Unreacted soda shows in the pH even before baking: in one pound-cake study, batters whose acid had reacted less during mixing measured pH 7.33 against 6.61. That is why formulas balance soda against their acid, using the acid's neutralizing value.

At low pH the amino groups that start the Maillard reaction are protonated and react less, so fewer brown melanoidin pigments form. Baking soda raises a batter's alkalinity as it dissolves, and in one biscuit study, replacing part of the ammonium bicarbonate with sodium bicarbonate gave a darker surface, which the authors linked to the added sodium.

The starch absorbs moisture so the acid and soda do not react too soon. The active parts are proportioned to react fully: in one pound-cake study, the powder was 41.9% soda and 58.1% acid by weight, set by the acid's neutralizing value. Acids in use include phosphates such as sodium acid pyrophosphate, tartaric acid and cream of tartar.

Monosodium phosphate is one such acid. Leavening acids are grouped by how quickly they dissolve and react with soda, so with a fast acid most of the gas forms in the mixing bowl rather than in the oven. Gas produced before heat sets the batter can escape from the bubbles it inflated.

Sodium acid pyrophosphate is one of the most used slow acids, and its grades are numbered for their early share: SAPP 10 releases about 10% of its gas during mixing and 90% in the oven, SAPP 40 about 40% and 60%. In one pound-cake study, volume more than doubled from batter to cake, which the authors suggested came mostly from gas released in the oven.

In one pancake study, acid whey from strained yogurt was thought to react with the batter's soda and speed its gas release. Soda left without enough acid stays alkaline in the food and can taste soapy. A powder's acid and soda come already blended in proportion by weight, so the powder releases its gas by itself once it is wetted and heated.

In a survey of 59 cocoa powders sold in Europe, every natural sample fell within the pH 5.3 to 5.8 band set for natural cocoa, and every alkalized one within the 6.5 to 7.2 band classed as medium-alkalized. The usual alkalis include potassium carbonate and sodium hydroxide. The two cocoas differ in acidity as well as color and flavor, so use the type a recipe names.

In bread dough, mixing creates the small air cells that set the crumb's structure, and fermentation gas then moves into them. Fat helps hold the air: cake fats aid air incorporation and steady the cells, and milk fat has been reported to aerate dough into finer, more even bubbles than soybean oil. Angel food cake and meringue rely heavily on whipped egg white.

Butter's good results here have been tied to its layer staying plastic and unbroken, with no openings for steam to slip through, and repeated folding usually builds 48 to 256 layers. Steam adds to the rise of leavened goods too: in cookies, water turning to vapor in the oven adds to the thickness that the leavening gases produce.

Yeast needs the right amount of sugar. In one pastry study, sugar at 21% of the flour's dry weight put the yeast under osmotic stress, cutting its sugar use, gas output and the final volume, while a dough with no added sugar ran out of fuel early. Fermentation also makes organic acids and aroma compounds that affect the dough and the flavor of the finished goods.

Ammonium bicarbonate goes into small goods because its ammonia can disperse and escape quickly, and it leaves no residue only where the product ends up dry. In a cake, baking powder releases gas during baking and the batter expands until heat sets it. A gluten dough holds yeast gas through hours of slow rising.

Timing cuts both ways. Gas released well before the structure sets is lost from the bubbles it inflated, while gas produced after setting may no longer expand them and can crack the crumb instead. Gluten is the main gas trap in bread; in gluten-free doughs, hydrocolloid gums are added to thicken the dough and strengthen the bubble walls for the same job.

In one pound-cake study, high levels of a powder whose acid released about 40% of its gas during mixing inflated the bubbles until they merged and burst visibly at the surface, lowering batter volume. In pancake trials, early gas helped a batter as thick as the control, but in the authors' reading it was lost from a thinner batter, which baked up denser.

FSIS notes that above 3,000 feet recipes may need changes in time, temperature or the recipe itself, and that moisture evaporates faster in the drier air. For baked goods it points to egg: extra-large eggs add moisture and structure, more egg strengthens the cell structure of a rich cake, and smaller eggs leave a batter less stable and more likely to fall.

In one pound-cake study, volume rose as powder went from 0.98% to 2.75% of the flour weight, then fell at 4.52%, and the authors described a heavy loss of gas bubbles as the batter set. Gas released after the crumb has set does a different kind of harm: its pressure may crack the crumb and can make the cake collapse.

The two lowest powder levels also gave a coarse, uneven grain of large cells. The authors suggested that, with so little gas, the existing bubbles swelled to their limit and merged before the crumb set. In that study, then, an uneven crumb signaled too little leavening rather than too much.

One study describes batters at about 150% water by flour weight and doughs at about 60%. In pound cake, a more viscous batter has been linked to better air incorporation, and in pancakes early gas helped a batter as thick as the control but was lost from a thinner one. In bread dough, by contrast, a gluten network expands and holds the fermentation gas.

FDA lists pH control among its technical effects. One study of meringue made from salted duck egg white used it at just 0.01% of the egg-white weight, and in angel food cake it helps the foam keep its structure while baking. In one biscuit study, a blend of soda and cream of tartar replaced part of the ammonium bicarbonate without changing the biscuits' structure.

Oven heat breaks it down into ammonia, water and carbon dioxide, and it leaves no residue as long as the product ends up low in moisture. The ammonia smell during baking does not carry into the finished goods. It also promotes acrylamide, and swapping in the sodium salt has cut acrylamide by up to 70% in some studies.

A fine, closed grain needs the leavening gas spread evenly across many bubbles, and when bubbles merge before the structure sets, they are fixed as large cells. Volume and fineness can pull apart: in one bread study, sheeting dough at a 6 mm roller gap instead of 12 mm gave loaves about 5% larger, but with fewer and bigger gas cells.

Protein makes up about 8 to 16% of a mature wheat grain, and gluten proteins account for up to 80 to 85% of the protein in the flour. Among cereals, only wheat, with a small exception for rye, has proteins that form such a gas-holding network when mixed with water.

Chemically, glutenin chains link to one another through disulfide bonds, and mixing helps by working oxygen into the dough. The two protein families develop at different speeds: in one study of isolated fractions, gliadin peaked after about 3 minutes of mixing, while glutenin needed about 31 minutes and kept developing through 60 minutes.

Glutenins bond into huge polymers, from about 80,000 to several million in molecular mass, while gliadins stay as single molecules and act as plasticizers within the network. Silencing the high-molecular-weight glutenins in one wheat line left flour that could not make bread at all, though the authors suggested its weak dough might suit biscuits.

Protein amount and gluten strength are separate traits: a USDA soft-wheat lab notes that a low-protein wheat variety can still have strong gluten. Amount matters on its own too: in that lab's cookie test, cookies spread about 0.4 cm less for each extra percentage point of flour protein.

Milling explains part of it. Milling hard kernels often breaks open their starch granules, which then absorb more water and reach the yeast sooner, a trait the US bread industry wants because most of its bread is yeast-leavened. Soft kernels usually break around the granules, so the flour absorbs less water and blends into a smooth cake batter.

Each trades effort against time. High-speed mixing with oxidants develops a dough in a few minutes at a high energy cost, sheeting does it with a fraction of that energy, and bulk fermentation takes several hours. Bakers also knead by hand or stretch and fold, and judge the result with the windowpane test.

Soft wheat flour gives these products more volume and a more tender texture than hard wheat. In short doughs rich in fat and sugar, texture and spread depend more on starch gelatinization than on gluten, and in cookies gluten acts as a binder rather than the main frame it forms in bread. Acid weakens gluten, while chloride salt strengthens dough.

Its role depends on the product: fats strengthen yeast-leavened bread dough more than cookie dough or cake batter, with fat crystals bracing the developed gluten. Flour's own lipids, only about 1.5 to 3% of it, slow hydration too: in one study, vital wheat gluten with its lipids reached a target consistency after 12 minutes of mixing, and after 8 with them removed.

Bread dough usually carries about 1 to 2% salt by flour weight, and that small amount affects development time, extensibility, yeast activity, color and flavor. At low levels, salt can even stimulate yeast growth. Cutting salt from a bread therefore changes how the dough handles and ferments, not only how it tastes.

Sugar also raises the temperatures at which starch gelatinizes and proteins set, the two events that turn a liquid cake batter solid. Cutting sugar therefore shifts when a cake sets, not only how sweet it tastes, and cake makers are advised to choose sugar substitutes that keep those events properly timed.

Sourdough is the everyday case. Its lactic acid bacteria lower the dough's pH, and in salt-reduction trials the sourdough-fermented breads came out harder than yeast-fermented ones, which the authors put down to the lower pH weakening the gluten and to the bacteria breaking it down.

In one study, the glutenin macropolymer recovered most after a 60-minute rest and then declined with longer rests, which the authors proposed was the flour's own protease breaking it down again. Even at its best, the rested dough held less of the polymer than the dry flour had.

The figure sorts products quickly: one study describes batters at about 150% and doughs at about 60%. Within bread, both extremes cut loaf volume: too little water keeps the gluten network from forming fully, and too much dilutes the proteins so they cannot link into a strong one.

How much water a dough can take depends on what is in it. Bran competes with gluten for water, so bran breads need their water level tuned; in one study, doughs with pre-soaked bran grew softer as more free water was left over, and too much water led to overexpansion and a collapsed loaf.

It is a judgment rather than a measurement, but researchers use it too: one wheat-breeding study scored dough on a 1 to 7 scale for how well it stretched without tearing, and a flour-treatment study kneaded by hand until the dough passed. An early tear suggests the kneading or folding has not gone far enough.

In one study, mixing for 5 minutes instead of 3 cut the glutenin macropolymer by about 20% and left its fine pore structure more collapsed, and no mixed dough held as much of it as the dry flour, even after resting. With bran present, 8 passes through rollers gave bigger loaves and finer crumb than 12, which the authors read as overstretched gluten.

Particle size matters: in one sheeting study, fine and coarse bran hurt loaf volume more than medium bran, and extra gluten development by sheeting offset some of the damage. Pre-soaking bran so it takes up its water before mixing is another fix that has been studied.

It is made by washing the starch out of wheat flour and drying the rubbery protein mass below 131 °F (55 °C), leaving a powder of about 73 to 80% protein. Adding gluten raises a dough's water absorption, and samples differ: in one study, each vital gluten reached its best bread volume once the dough's water was matched to its own absorption capacity.

Chlorine, used on soft wheat flour for high-ratio cakes since the early 1930s, is one such treatment: it whitens the flour by oxidizing its carotenoid pigments, weakens its gluten and raises its water-holding, giving cakes more volume and a finer grain. Chlorine gas raises worker-safety concerns in the mill, and heat-treated flour has partly replaced it in the European Union.

Its built-in leavening is baking powder, the blend of baking soda and acid salts that gives off carbon dioxide once it is wetted and heated, so the flour brings its own lift to a batter or dough. Federal food standards cover it in a regulation of its own, 21 CFR 137.180, apart from the one for plain wheat flour.

Rice flour is the most common base: it is cheap and widely available, and it is white, bland and easy to digest, though its proteins do little for structure. No single ingredient fully replaces gluten; better results come from combining several ingredients with suitable processing.

More is not always better: in one study, loaf volume rose with xanthan up to a point and then fell, apparently because the dough grew too stiff for its cells to expand, and other trials saw no gain. In a rice-flour bread study, Plantago husk, a psyllium product rich in arabinoxylans, raised loaf volume by up to 27% at a 9% share and gave a softer fresh crumb.

Spelt often has more protein than common wheat, but a larger share of it is gliadin, so its doughs stretch easily and its loaves rise less. Buckwheat, a pseudocereal, can replace part of the wheat in bread; one review found breads with up to 25 to 30% pseudocereal flour well accepted, while more reduced quality.

Rye grain holds roughly 2 to 12% arabinoxylans, and they hold a lot of water: in one rye-bread study, adding just 1% raised the water the dough needed and gave bread with more volume and a moister crumb, with the larger, cross-linked arabinoxylans working best. USDA researchers note they strongly affect water absorption in baked products.

Those enzymes free fatty acids that then oxidize, giving rancid, unpleasant smells and tastes. Insects are another risk: in one study, red flour beetles in sealed bottles of infested flour did not increase over three months, while in open bottles they multiplied about 50-fold, and the sealed flour held less oxygen and moisture.

To convert a cup-based recipe, NIST suggests measuring each ingredient once with the cups and spoons the recipe names, then weighing that amount in grams. A scale also lets you tare the bowl to zero and add each ingredient in turn, with fewer cups to wash.

Because everything is pegged to the flour, the figures need not add up to 100, and they hold for any batch size. Typical values read quickly once you know them: bread salt usually sits around 1 to 2%, while the water in a pourable batter can run about 150%, well above the flour.

One bread dough in a published study used 56.5 g water, 3 g yeast and 1.5 g salt per 100 g of flour. For 500 g of flour, multiply each by 5: 282.5 g water, 15 g yeast and 7.5 g salt. Every ingredient changes by the same factor, so the dough's proportions stay the same.

In cookies, the water that sucrose ties up is part of what gives a crisp bite. At baking temperatures some sucrose breaks into glucose and fructose, which then take part in Maillard browning. High-intensity sweeteners can match the sweetness but add almost no body, so swapping them in for sugar changes more than the taste.

That protection can vanish with a sugar swap. In one sponge-cake study, sucrose brought water activity down while two of three stevia-based replacers did not, a sign that low-calorie replacers may shorten the shelf life of moist sweet goods.

In a flour-and-water model with 50% water, adding sucrose moved the starch gelatinization peak from about 142 °F (61 °C) to about 154 °F (68 °C); in a much wetter mix the shift was slight. The delay gives gas bubbles longer to expand before the batter gels. High-ratio cakes, which use more sugar than flour, add extra egg to offset the slowed gelatinization.

In one whole-wheat cookie study, the spread ratio tended to fall as dough viscosity rose; the sucrose dough had the highest yield stress of the sweeteners tested, and sorbitol cookies spread the most. Grain size may matter too: researchers have suggested that fine sugar dissolves early in baking and pushes the dough outward, while coarse crystals dissolve slowly and give smaller, thicker cookies.

Cane molasses holds aconitic acid from the cane itself plus acetic and lactic acids formed during processing, and its amino acids can react with reducing sugars in Maillard browning. An older cookbook leavened molasses mixtures with baking soda alone, the same treatment it gave sour milk.

Honey is mostly glucose and fructose, left after bee enzymes split sucrose. Its crystallizing is put down to glucose being less soluble than fructose, so glucose-rich honeys granulate quickly and high-fructose ones slowly. These free sugars also brown sooner: in one cookie study, sucrose cookies needed longer in the oven to reach the same color. Too much syrup can backfire, leaving hard candy that draws in moisture and turns sticky.

In one cookie study, the granulated sugar's coarsest grains ran about ten times the size of the powdered sugar's, and the powder was estimated to have roughly four times the surface per unit of volume. Fast-dissolving fine sugar is thought to help a cookie spread wider. An older cookbook asked for confectioner's or pulverized sugar in meringue and fell back on granulated only when neither could be had.

In shortened bread dough, fat crystals cling to small bubbles and steady them, and milk fat's fine beta-prime crystals aerate better than soybean oil's beta crystals. Without any fat, dough bubbles tend to merge and bake into a coarse crumb.

One refined milk fat was still about 38% solid at 77 °F (25 °C), and milk fat is fully liquid only near 100 to 104 °F (38 to 40 °C), so at room temperature it is part crystal, part liquid. An old bakers' manual had hard, brittle butter worked until pliable before the sugar went in, and an old cookbook warned that a bowl warmed too much turns butter oily rather than creamy.

Butter is roughly 81% milk fat. A shortening is a formulated fat whose melting point depends on its blend, about 91 to 140 °F (33 to 60 °C) across one set of palm shortenings. Milk fat crystallizes in the fine beta-prime form, while lard and soybean oil usually form coarser beta crystals, and in one comparison solid hydrogenated fat made cookies more tender than sunflower oil did.

Fat and water pull flour opposite ways: in dough, milk fat coats gluten and starch and lowers how much water they take up, while water lets the proteins hydrate into gluten. Water content is a practical limit too, since high-water margarine has limited use in factory cookie and cracker making. In puff pastry, the dough's water turns to steam, and unbroken fat layers trap it so the layers lift apart.

Too little solid and too much both cause trouble. In one review, a fat under about 15 to 20% solid at dough temperature was too soft to serve as a shortening, and a shortening near 47% solid at 68 °F (20 °C) made less acceptable cookies than one near 22%. Butter softens noticeably at that temperature and has to be kept cold to stay plastic for laminating.

In baking, shortening names exactly this ability of a fat to weaken, or shorten, a baked structure. Factory short-dough cookies are generally mixed in two stages, fat, sugar, and water first and the flour last with little mixing, to hold back the flour's contact with water. In one whole-wheat cookie study, cookies grew harder as mixing time rose.

Fat splits aroma compounds between itself and the watery part of a food, and for the oil-loving majority it affects release more than any other ingredient. In flavored yogurts, instruments measured faster, stronger, but shorter-lived aroma release from low-fat samples. Butter also brings aroma of its own, from compounds such as lactones and methyl ketones, and reduced-fat baked goods have generally fared worse on flavor.

In one whole-wheat cookie study, swapping sucrose for sorbitol raised the spread ratio, a width-to-thickness measure, from about 73 to 104. Mixing time has been reported to affect spread too, though in another whole-wheat cookie test, longer mixing made the cookies harder while their size barely moved.

An old household manual mixed cookie dough so soft it was hard to handle, then chilled it for several hours or overnight before rolling, and warned that a thin cake batter should not stand that way. Cold also firms the fat itself: one croissant shortening was 39% solid at 50 °F (10 °C) but 27% at 70 °F (21 °C).

Because one ingredient does all of this, eggs are hard to replace, and soy, wheat, whey, and gum stand-ins have generally fallen short. In a pound-cake study of ten commercial egg replacers, the closest match combined corn starch, an emulsifier, xanthan gum, and pea protein. Several replacer batters set hotter than egg, which the authors tied to plant proteins and gelling gums.

Ovalbumin, more than half of the white's protein, sets into a firm gel as a cake bakes, so whites add structure. Yolk phospholipids and lipoproteins gather where oil meets water and keep a batter's fat dispersed. The two can clash: even a small share of yolk in egg whites cuts their foaming, and one study suggests yolk components may compete with the white's proteins at the bubble surface.

Sugar moves these set points: it raises the temperatures at which egg proteins denature and starch gelatinizes, so a sweet batter firms later and its bubbles have longer to grow. For safety, FSIS advises cooking dishes that contain eggs until the center reaches 160 °F (71 °C) on a food thermometer.

Whipping traps air and lines each bubble with white proteins; in one mixer test, foam volume rose with beating time and speed until it leveled off. One old meringue method starts adding sugar once the whites hold their shape but still drop from the whip. An old manual also said never to let beaten whites stand, since liquid drains from a standing foam and its bubbles can merge.

Hold the sugar back until the whites are already foaming, then add it a little at a time; added at the start, it slows foam formation. An acid such as cream of tartar lowers the pH, and in egg-foam work a low pH improved stability but reduced foaming capacity. Separate eggs carefully and keep fat out of the bowl; one study suggests yolk components may compete with the white's proteins at the bubble surface.

In an old cookbook's terms, beating is what encloses air, while folding alternates a straight downward cut with a turn that brings the bottom of the bowl up, repeated until the mixture is blended. A foam is fragile to begin with: liquid drains from the films between bubbles, and when a film breaks, two bubbles merge into one larger one.

A yellow cake gets its milky taste, tender grain, and deep color from butter and whole egg; a white cake usually swaps in egg whites and shortening. Whites free of yolk whip to a larger foam, and angel food cake is built on beaten whites. Trading one for the other changes the fat and the color at once, not just the egg count.

Those extras are not always a gain. In one bread study, replacing part of the water with milk lowered loaf volume and firmed the crumb, and the authors advised replacing less than 25% of the water. Native whey proteins have also been reported to depress the volume of baked goods. Expect a recipe to change when you swap milk for water.

Buttermilk left from churning sweet cream is not sour, and some sour milks sold as buttermilk were made by adding lactic acid or a bacterial culture to sweet milk. One old manual sifted the soda in with the dry ingredients and added the sour milk with the liquids. In one yeast-bread study, swapping 30% of the water for buttermilk gave the largest loaf, while 60% or more cut volume and sensory scores.

Maillard browning joins reducing sugars with the amino groups of proteins, and milk supplies both. Whey carries more lactose than milk and colors a crust even more. So a yeast dough made with milk in place of water tends toward a darker crust, and its color is worth watching near the end of the bake.

Bread doughs usually carry 1.5 to 2% salt by flour weight. By one measure, 2% salt cut yeast gas production by about a fifth and 4% by about 70%, which is why salt is usually kept out of preferments and away from direct contact with yeast. Reduced-salt dough over-kneads easily and turns weak and sticky, and bread made without salt tastes dull and yeasty.

In one pancake study, acid whey from yogurt in place of water released the soda's gas sooner, which the authors credited to its acids; a thick batter baked lighter, but a thinner one lost the gas and baked denser. In one bread study, swapping some water for milk lowered loaf volume. Milk's lactose also survives yeast fermentation and darkens the crust.

Cake batter is a water phase holding dissolved sugar, proteins, and hydrated flour, with butter or margarine emulsified through it, and its air bubbles are stabilized by the fat. How much of a fat is solid moves with temperature: one croissant shortening was 39% solid at 50 °F (10 °C) and 27% at 70 °F (21 °C).

Pace matters too: an old cookbook warned that liquid poured too fast into a creamed butter-and-sugar sauce gives it a curdled look. Separation is not always fatal. In one study of factory molded cookies, a creaming stage that separated on standing still baked into cookies of similar air porosity, and a tasting panel found no difference in hardness.

An old household manual kept this order for lean mixtures and said a stiff dough should take its fat cold, since melted fat leaves it soft and sticky. Its plain muffins filled greased pans two-thirds of the way and baked about 20 minutes in a hot oven. Mixing wheat flour with water builds gluten, and in one whole-wheat cookie study, longer mixing gave harder cookies.

In a cake, the air cells trapped during creaming add lift and help give a finer crumb. An old bakers' manual added eggs a few at a time, threw in a handful of flour if the cream began to curdle, and ran the butter alone in the machine first, sifting the sugar so lumps would not spoil the cream. Factory cookie makers likewise add the flour last, slowly and briefly, to hold back gluten.

An old household manual made keeping every ingredient as cold as possible the main rule, and asked for a dough as moist as it could be while still handling well. Its pastry chapter makes a parallel point about flakiness: keep the fat solid and in pieces, not rubbed fine or soaked into the flour, which oil tends to do.

Some mixing methods work the fat into the dry ingredients before any liquid goes in; an old household manual used that order for many hot breads, counting the beaten eggs as part of the liquid. In bread dough, milk fat likewise coats gluten and starch and lowers how much water the dough takes up.

A true sponge cake is eggs, sugar, flour, and flavoring, with the eggs supplying all of the rise: whole eggs are beaten until thick and pale yellow, sugar goes in gradually, and sifted flour is folded in last. The foam has to be fully built; in an angel food cake study, short whipping left the meringue unformed and the cakes hard, while longer whipping raised cake volume.

It saves a stage and a rise compared with a sponge method, though a stiff dough is slower to rise than a soft sponge. Set against this direct method, a biga preferment is credited with deeper flavor, a better texture, and a longer shelf life. An early trade manual named ingredient temperature, room temperature, yeast quantity, and flour quality as the points to control.

An older home method put about half the flour into the sponge and stirred in the rest once it had risen, which meant three rises against the straight method's two. An early trade manual advised taking the sponge just as it begins to fall. Food scientists call this an indirect method: an intermediate preferment is built first and then added to the final dough.

In one study a poolish-type sourdough was equal parts flour and water, whereas biga began in Italian bakeries as a way to stiffen dough; one pizza study fermented its biga 24 to 72 hours at about 61 °F (16 °C). Bakers have long reused leftover dough as a starter, and in one test a poolish-type sourdough fermented with a selected lactic acid bacterium kept mold off bread for ten days.

Mixing hydrates the flour, builds the gluten network, and works in air that is thought to seed later gas cells. In one bread study, under-kneaded dough baked into crumb with fewer, larger cells and thicker walls; in noodle dough, over-mixing broke down the gluten and released water it had held. Old manuals judged the end by feel, a smooth, elastic dough, not by a fixed time.

An early trade manual measured the flour and the shop first, then warmed the water so the dough came out near 80 °F (27 °C), using cooler water in a hot shop. In sourdough, warmer fermentation lowers pH faster and shifts the lactic-to-acetic acid balance, and one comparison found 95 °F (35 °C) gave less hard bread than 82 °F (28 °C). A chilled dough recovers slowly, so keep it covered and warm.

Carbon dioxide from the yeast partly dissolves in the dough's liquid and moves into tiny air cells formed during mixing. In one pizza study, longer leavening raised reducing sugars and free amino groups, the raw material of crust browning, and earlier work found that very long leavening breaks down the gluten network until the dough holds gas poorly. An older cookbook let dough double, using a marked glass of dough as a gauge.

The manual cut the rising dough into pieces as large as could be handled, stretched each as long as possible, and stacked them back in the trough. Machine sheeting is a related but separate process: in lab trials, dough rolled thin and folded between passes gained expansion capacity over 12 passes, and earlier work found that very long sheeting can break gluten down.

An early trade manual punched the risen dough down with both hands, folded it over, and repeated until it was firm, then let it rise again; a home cookbook cut the dough down with a knife instead. The manual warned that dough shaped before its excess gas was worked out bakes into a spongy loaf with big, uneven holes that dries out quickly.

An older cookbook kneads each piece until it is smooth and round, stretches the underside that will become the top, and forms a roll as long as the pan and half as high. Pressing the dough well into the corners and ends lets the loaf rise evenly from end to end, and it then rises until it reaches the top of the pan.

Rise speed varies, so an old cookbook gave no fixed time, and one frozen-dough steamed-bread study ended each proof at a set expansion because its doughs rose at different speeds; bread volume peaked at a moderate proof. The cookbook warned that too little final rise makes bread overly fine and heavy, too much makes it coarse, and rising continues in the oven.

In one frozen-dough study of steamed bread, volume rose up to a moderate proof and then dropped sharply, with a harder texture and crust pulling away from crumb; the authors judged over-proofing far more harmful than under-proofing. As proof runs long, gas cells can merge, leaving thinner walls and larger holes. An early trade manual blamed overproofing on bakers chasing a big loaf at the cost of flavor and texture.

In one frozen-dough study of steamed bread, the least-proofed dough gave less volume than the best proof, though its shape stayed acceptable, unlike over-proofed dough. An older cookbook warns that too short a final rise makes bread too fine in texture and likely heavy. Rising continues in the oven, so the final proof can stop a little short of the size you want.

An early trade manual tested a rising dough by pushing a hand into it and took the dough once it began to fall, trying again half an hour later if it held; a sponge was tested by shaking the bowl. A home cookbook gauged the rise with a marked glass of dough set beside the bowl. One steamed-bread study likewise ended each proof at a measured expansion rather than a fixed time.

Refrigerated doughs are usually held at 39 to 45 °F (4 to 7 °C). Slow fermentation at low temperature favors the esters, alcohols, and organic acids that shape bread aroma, and low-temperature processes have scored higher in sensory tests. There is a limit: with long refrigeration, dough turns syrupy and its gluten weakens, so longer is not always better.

Heat expands the water vapor and gas in the loaf, and most of that oven spring comes in the first 6 to 8 minutes of baking. An old cookbook suggested brushing loaves with milk before baking to keep the crust from hardening too fast. Pie bakers cut slits or a design into a top crust so the filling's steam has an outlet instead of pushing the crust up.

Older cookbooks filled muffin and cupcake pans half to two-thirds full and let the rise carry the cake to the rim. One called for a cooler oven for a loaf cake than for layers or small cakes, since the bigger cake needs longer to bake through. Do not move a cake until it has risen and set, or it may fall.

Dough grows less tough and less resistant to stretching in the time after mixing, and in lab kneading tests short pauses let the gluten strands relax and lowered internal tension. Puff pastry in one study rested 30 minutes at 39 °F (4 °C) between folds, which also kept the fat cold. In pizza dough, the first hours of leavening eased the stresses left by mixing and gave the widest disc.

Before baking an empty pie shell, an older cookbook pricked the bottom and sides thoroughly with a fork and noted that a perforated pan does the same job. A two-crust pie is not pricked; its top crust is cut instead to vent steam. An old British bakers' manual also docked its rolled crackers and cookies, which British usage calls biscuits, after rolling the dough to a set thickness.

Heat raises the vapor pressure of the water in the dough and the gas cells swell, with warmer, briefly busier yeast adding gas too. In one baking model, rise stopped once bubble walls ruptured and the pores linked up, between about 160 and 178 °F (71–81 °C), and the modeled loaf ended near 1.9 times its starting volume.

Carbon dioxide dissolved in the dough does come out of solution as it warms. But in that model, vapor reached 80% of the gas in sealed bubbles within 100 seconds and averaged about 70% before the bubbles opened, which the authors say overturns the older view that vapor adds little. Once the walls rupture, the carbon dioxide escapes to the oven.

In batter tests, salt nudged that temperature up slightly by competing with starch for water, and sugar raises it too. In one bread dough, the heat-flow peak tied to gelatinization and evaporation sat near 169 °F (76 °C). The crust is different: water leaves it so fast that its starch never fully gelatinizes, and that partial gelatinization is part of what makes crust crisp.

In one gluten-gel study, the proteins began unfolding at 104 to 122 °F (40–50 °C) and were aggregating, with more intermolecular beta-sheets, by 194 °F (90 °C). In baking dough, starch starts to gelatinize near 149 °F (65 °C), and heat then cross-links the gluten through disulfide bonds until the structure sets. Finished crust keeps that denatured gluten as its network.

That model had the dough stiffening over a wide span, roughly 122 to 230 °F (50–110 °C), as starch gelatinized, proteins denatured and water left. Heat travels inward from the surface, and in the test loaf the edge of the top stopped rising while the center kept climbing for some time longer.

Evaporation holds the crumb at the boiling point, while browning reactions run on the hotter, drier outer surface. Freshly baked, the crust breaks with an audible snap. In storage it goes leathery and loses that snap, a change called crust staling that is thought to come from moisture moving out of the crumb into the crust.

Early steam condenses on the dough, the coldest surface in the oven, and hands it heat until the surface passes the dew point. In a model oven, added water cut the radiant, convective and conductive heat reaching an aluminum test cube as condensation took over part of the work. One industry-aligned toast-bread protocol steamed at the start and opened the vent for the last 15 minutes.

The reaction runs faster at higher temperature and, in general, in alkaline rather than acidic conditions, and its rate typically peaks at a water activity near 0.6 to 0.7. One product of the same chemistry, acrylamide, forms mostly in bread crust. FDA notes that for breads a darker color generally goes with more acrylamide, and that toasting to a light rather than dark brown lowers it.

Sucrose first splits into glucose and fructose, which then break down into brown polymers and aroma compounds such as maltol. Higher heat browns faster but can quickly turn bitter, and the reaction is usually slowest near neutral pH. Where sugar and protein sit together the two browning reactions can run at once, and a bread-baking model places both on the loaf’s surface.

An old cookbook gives the craft rules: beaten egg, or yolk thinned with a little cold water, for a glazed top; milk before baking for a darker crust; melted butter three minutes before the end for a more tender one. In that staling study, those three glazes also kept about 1% more crumb moisture over six days.

In that study, loaves baked cooler and to a lower core temperature came out paler, smaller and more prone to collapse after vacuum cooling. FDA’s acrylamide guidance for industry describes dropping the heat late in a bake, when the product is dry and acrylamide forms more readily, and lists lighter color and a moister, shorter-keeping product among the costs of baking cooler.

Still oven air forms a thin insulating layer at the food’s surface, and fast air thins it. In an air-jet oven study, bread reached comparable or better volume and texture in less time than in a conventional convection oven. FDA notes convection ovens may give bread more acrylamide than deck ovens because they dry the crust more intensely.

In that oven the middle rack gave a more even temperature rise between top and bottom than the lower tray did. Ovens are not uniform inside either: a tested built-in electric oven ran hotter near its walls and cooler near the door, and uneven heat inside an oven can show up as uneven color on the food.

Polished aluminum has an emissivity near 0.04, so it reflects nearly all incoming radiant heat. In that oven, radiation carried about 45% of the heat to the test block, with conduction from the floor plate and convection splitting the rest. The test used metal blocks, not pans; an old household-arts textbook found that loaves took longer to bake in granite pans than in tin, though it judged pan material of no great consequence.

An old cookbook gives the touch tests: a done pan loaf no longer clings and slips out easily, and for most cakes a shrink from the pan is test enough. Pound cakes are the exception; for them and rich fruit cakes the book presses a fingertip into the top, and a done cake feels firm and lets the dent fill back in.

Evaporating water caps a moist crumb at the boiling point. In one toast-bread study, loaves had to reach at least 208 °F (98 °C) in the core to hold their shape through vacuum cooling. For quiche, casseroles and other egg mixtures, FSIS sets 160 °F (71.1 °C) at the center, measured with a food thermometer.

An industry-aligned toast-bread protocol kept cooling loaves until the core fell to 95 °F (35 °C). The starch keeps changing as it cools and ages: amylose forms a gel network early in cooling, while amylopectin reorders far more slowly during storage. An old cookbook advises taking loaves out of their pans at once and cooling them on a wire rack.

Crumb needs enough heat for enough time to gelatinize its starch, or it stays weak. Two other routes lead to a sticky or soggy crumb. Rye flour with very high amylase activity breaks down too much starch and gives a sticky crumb, and an old cookbook warns that bread not risen long enough bakes up heavy and soggy.

Amylose sets into a gel early, while amylopectin recrystallizes far more slowly over days of storage. In one composite-bread study, a heat scan of fresh bread showed no melting peak for these crystals; after four days of storage, one appeared. Commercial bakers add emulsifiers partly to soften crumb and slow staling.

Inside the loaf, water redistributes between starch and gluten and moves from crumb to crust, which is thought to turn the crust leathery. Real water loss adds to the firming too: in one study, uncoated loaves stored in paper bags lost more weight and got firmer over 14 days than loaves with crust treatments.

In one study that peak, from recrystallized amylopectin, was absent in fresh bread and present after four days of storage. Heat is also the old kitchen fix: a household-arts textbook advised freshening bread that has become too dry by moistening the surface and heating it in the oven.

Low storage temperatures can hurry starch retrogradation, which is why one study of how bread hardens kept its loaves at room temperature rather than chilled. Aroma is a separate matter: in one small study the volatile profile of refrigerated bread stayed closer to fresh than frozen bread’s did.

FSIS advises freezing food as fast as possible, because slow freezing forms large, damaging ice crystals, and storing it at 0 °F (−18 °C) or below. Wrap well: air on the surface causes freezer burn, dry leathery spots that are safe but poor in quality. Freezer storage times are about quality only, since food held constantly at 0 °F stays safe.

Water softens crisp crust by acting as a plasticizer, and cookies and puffed snacks gain moisture from the air; a zipper bag with lower vapor permeability delayed that over 20 days at 77 °F (25 °C) and 50% humidity. Bread does the reverse, firming as it dries. An old cookbook applies the same logic while cooling: leave a loaf uncovered for a crisp crust, or cover it with a towel for a soft one.

Many bread doughs hold little free sugar, so the yeast lives largely on maltose that the grain’s own amylase enzymes release from starch, and commercial baking strains handle maltose better than other strains. Yeast also makes glycerol, organic acids and aroma compounds, and its metabolism is reported to be the main source of the variety in bread aroma.

An early-1900s textbook already noted that pressed yeast had to be used fresh while dried yeast kept its cells alive for a long time; today, protective compounds such as trehalose and proline help commercial yeast survive drying or freezing. Both forms have long histories: early compressed yeasts appeared in England, Germany and the Netherlands in the 1780s and 1790s, and dried yeast was first sold in Vienna in 1822.

Rewetting is the risky step: in one proposed model, the cell membranes pass back through a phase change as water returns, and doing that in water below the transition temperature can leave them leaky. For baker’s yeast, cells dried after they stopped growing survive drying and rewetting much better than cells dried mid-growth. An old textbook waits for the yeast sponge to fill with bubbles before adding the rest of the flour.

At 50 to 68 °F (10–20 °C) yeast growth slows and metabolism drops, and near 39 °F (4 °C) yeast can be stored with its viability kept for a long time. Sourdough bakers may use that slowdown on purpose, reaching for ice water and refrigerated flour to rein in a fermentation that runs too fast in summer.

Making glycerol keeps the cells from drying out, but it leaves less carbon for gas, and osmotic stress can show up as longer proofing and smaller loaves. FDA lists excess salt among recipe changes that harm yeast, and cites a study noting that 1 to 2% salt is routinely used in industrial bread.

Yeast prefers glucose and fructose, and in a sweet-dough test most strains’ invertase had split all the added table sugar into those two within two hours. In another study, a dough with 18% sugar, and somewhat less water, had a water activity of 0.95 against 0.97 at 6% sugar. Sugar also raises the temperature at which starch gelatinizes during baking.

In one 23-strain test in dough with 14% added sucrose, the strain sold as an osmotolerant baker’s yeast used the most sugar, more than the standard baker’s yeast, and made the most carbon dioxide. In a separate study, raising glycerol output let a lab strain ferment 50% more in an 18%-sugar dough, but it made no difference to a commercial bakery strain.

Alpha-amylase cuts starch chains at random into dextrins, maltose and glucose, while beta-amylase trims maltose off the chain ends. Swollen, gelatinizing starch is easier for them to reach. Their optimum temperatures sit near 162 to 165 °F (72–74 °C) for alpha-amylase and 140 to 149 °F (60–65 °C) for beta-amylase, which match the conditions used for scalded rye flour.

Malt is rated by diastatic power, so in a rye study a weak malt at 1.5% matched a strong one at 0.5%; recommended doses run from 0.2% to 2%. Added amylase is also used to raise loaf volume and improve crumb structure, and in one crust study, spraying alpha-amylase on the loaves gave a notably darker crust, most likely from extra Maillard browning.

The crust is the exception: its volatiles come from Maillard browning, while the crumb’s come mainly from yeast and from fat oxidation. In one bagel study, compounds such as 2,3-butanedione likely formed during yeast fermentation, and a long cold fermentation raised the strength of several key aromas. Leavening dough cooler for longer has also been reported to increase aroma compounds.

In one bagel study, 24 and 48 hours of fermentation at 41 °F (5 °C) raised a key aroma compound about four and eight times over a same-day control. FDA notes longer fermentation may also lower acrylamide, since yeast use up asparagine, but warns it may weaken gluten and give flatter bread; using more of a low-gassing yeast is one alternative.

The bacteria acidify the starter while the yeasts do most of the leavening. The pairing has been proposed to be mutualistic, but in one lab study yeast populations shrank in the presence of the bacteria. In the same study, maltose-using bacteria paired with the maltose-blind yeast Kazachstania humilis leavened dough as well as bacteria paired with Saccharomyces cerevisiae.

Researchers track that balance as the fermentation quotient, the ratio of lactic to acetic acid, with 2.0 to 2.7 cited as an optimal range. Acetic acid is linked to sourdough’s tang but is thought to inhibit yeast more than lactic acid does. Hydration is often said to matter as well, with wetter starters favoring bacteria, but that pattern is not consistent across starters.

Artisan starters usually live at 68 to 86 °F (20–30 °C) and can be refrigerated between uses. In warm weather, bakers may feed with a smaller share of old starter to offset faster fermentation, and longer gaps between feedings give bacteria more time to build acid. Panettone mother doughs are refreshed every 4 to 16 hours, while refrigerated or dried starters stay usable for about 30 days.

In that bakery, about 28% of the final refreshed sourdough went into the first dough with flour, sugar, yolks, butter and water. Repeated refreshing keeps the microbes in an active state. How much starter goes into the dough matters too: more starter can make a more acidic dough and has been linked to higher yeast counts and more ethanol.

Long gaps between feedings give bacteria time to pile up organic acids, which can hold back microbes that tolerate acid poorly, and yeasts produce more ethanol under acid stress. A starter past its peak also turns pungent and makes fewer bubbles. One review proposes that starters refreshed only monthly, or neglected, will carry acid-tolerant microbes.

Temperature moves the timeline: in one report, sourdough took 11 hours at 77 °F (25 °C) but 9 hours at 95 °F (35 °C) to reach its target pH. Visual signs of a ready starter are poorly described in research, but a starter past its peak is known to turn overly tart and make fewer bubbles.

Wheat’s own protein-cutting enzymes work best at pH 4.0 or below, and in sourdough tests the breakdown of glutenin was first seen after 6 hours and grew stronger by 24 hours. Higher-protein flour, around 10 to 14%, builds a network that better withstands 12- to 24-hour ferments. FDA likewise notes extended fermentation can weaken gluten and give flatter bread.

Mixing traps tiny pockets of air, and carbon dioxide from the yeast later diffuses into them and swells them. The cells can grow because the dough flows, and they are thought to hold because the stretched dough stiffens against more stretching; the balance of the two sets how far the dough expands. A little solid fat also helps: its crystals keep small bubbles from merging or collapsing, for a finer, more even crumb.

A panettone dough, for example, carries large amounts of sugar, yolks and butter. With so little free sugar, yeast in lean dough lives mainly on maltose from starch; in sweet dough it faces osmotic stress instead, and a strain that ferments both kinds well is considered highly desirable. A little solid fat in a dough also tends to give a finer, more uniform crumb.

Solid fat has been found to work better than oil here: fat crystals settle at the surfaces of small air bubbles and, as they melt, help hold the gas in, so the bubbles stay small instead of merging. Schedules can be long; one artisanal sourdough panettone rose 18 hours at 68 to 72 °F (20–22 °C) after its first mix and another 5 to 6 hours before baking.

In one cited bread study, a 6% water roux raised moisture from about 37.4% to 39.1% and cut baking loss from about 5.5% to 3.8%; softer crumb is commonly reported with pre-gelatinized starch. A scald pours 194 to 208 °F (90–98 °C) water over flour and is reported to soften crumb and slow staling. In one gluten-free rice bread, pre-gelatinized flour gave a sticky, wet crumb when baked only as long as the control.

In one crust study, loaves sprayed with 4.9% sodium hydroxide before proofing baked up notably darker, with the smoky smell of a traditional pretzel. The Maillard reaction generally runs faster at higher pH and forms more brown melanoidin pigment there. In a bagel study meant to mimic commercial practice, rings were boiled 2 minutes, flipped at 1 minute, then baked 14 minutes at 450 °F (232 °C).

A lid works by holding in the bread’s own moisture: Somali laxoox cooks on a griddle under a lid that traps condensation so the bread does not dry out, and the trapped steam helps leaven it. Added oven water also changes heat delivery; in a model oven, water vapor absorbed part of the radiant heat.

Somali laxoox, for instance, is a fermented batter spread in a spiral on a hot griddle and covered. Steam from the wet batter, held in by the lid, helps it rise, and the starch sets so fast that bubbles are caught in place as the pockmarks known as eyes. Surveyed cooks put cooking time at two to five minutes and most often judged doneness by light browning underneath or the hiss of condensation hitting the pan.

Some leavening acids react mostly during mixing and others mostly in the oven heat. Soda without enough acid leaves a soapy flavor and yellow color. In cake batter, the carbon dioxide from baking powder partly dissolves into the batter and partly swells the air bubbles beaten in during mixing.

The same cookbook adds that loaves left to rise somewhere too hot show a heavy streak near the bottom. Modern work ties other faults to causes: freezing and frozen storage can damage the gluten and the yeast, giving lower volume and a poorer crumb, and in one study vacuum-cooled toast loaves baked to a lower core temperature were more likely to collapse.

In chocolate, fat is the continuous phase, and fine particles of sugar and cocoa, plus milk solids in milk chocolate, sit suspended in it; ground below about 40 µm, those particles no longer feel sandy. A label's total cocoa solids counts cocoa butter along with liquor and powder, and compounds from the bean such as flavonoids and theobromine give dark chocolate its stronger bitterness.

Typical figures put dark chocolate at 35% or more total cocoa solids, milk chocolate at a minimum of 15 to 25% cocoa solids plus 12 to 14% milk solids, and white chocolate at no less than 20% cocoa butter with at least 14% milk solids. Local and national standards set the minimums, and makers keep wide leeway in their formulas.

Most chocolates hold about 25 to 35% fat. To thin a coating, reach for fat or an emulsifier, never water: in industrial chocolate, lecithin lowers the viscosity of chocolate already flowing and PGPR lowers the threshold needed to start it flowing, while even a little water can make melted chocolate seize.

In one set of measurements the forms melt from about 63 °F (17 °C) for Form I to about 97 °F (36 °C) for Form VI, and a second study put each form a few degrees lower. Chocolate makers aim for Form V, and finding Form VI in place of Form V is the main chemical sign of fat bloom.

In one lab study, cocoa butter chilled to 32 °F (0 °C) without a tempering step held only alpha crystals, the least stable form. Form V alone is not the whole story: in another study, a commercial bar melted and reset by chilling still tested as Form V yet came out much softer, which suggests crystal size and arrangement matter too.

A common cycle heats chocolate above 122 °F (50 °C), cools it to about 81 °F (27 °C), where stable and unstable crystals both form, then rewarms it to 86 to 90 °F (30 to 32 °C) so only the stable ones remain. Seeding instead adds ready-made Form V cocoa butter crystals; one study seeded a nougat filling with 0.5 to 2% of its fat.

The white look comes from big surface crystals scattering light and is attributed to Form V turning into Form VI. Chocolates with nut fillings are especially prone: in one study, oil from a hazelnut filling migrated through the shell and ended up in the bloom. Store chocolate below 72 °F (22 °C) and avoid repeated warming and cooling.

In one study, a drop of water left on the spot for two minutes dissolved sugar bloom, while a spatula warmed on a 104 °F (40 °C) plate melted fat bloom. Warm or humid storage is a named cause of bloom, and factories warm chilled chocolate above the dew point of the packing room before wrapping so no water condenses on the surface.

Seized chocolate cannot be used as plain chocolate again, so keep bowls and tools dry. Much more liquid behaves differently: once water passes about 20% of the mix, it dissolves most of the sugar and the chocolate flows again, which is the principle that lets cream and chocolate become ganache.

Milk and white chocolates make softer ganache because their extra sugar acts like a liquid once melted, and the suggested ratios give them more chocolate per part of cream: about 2.5 or 3 to 1, against 2 to 1 for dark. One method pours near-boiling cream over solid chocolate and stirs until the mixture is smooth.

In a sound ganache, cooling fat crystals bridge neighboring droplets and swollen cocoa particles help hold the structure. A split one tends to carry more nonfat cocoa solids and less sugar, so cocoa particles move quickly into the watery phase and thicken it, which makes the fat droplets harder still to keep apart. In one study of fruit ganaches, mixing for one minute instead of three left the structure incompletely blended.

Dropped into cold water, syrup at soft ball can be shaped into a ball that barely holds, at hard ball into a firm one, and from crack onward it turns brittle. Older candy books differ by a few degrees for the same stage, such as 248 °F (120 °C) or 250 to 255 °F (121 to 124 °C) for hard ball, so pair the thermometer with the water test. Fondant stops at soft ball; hard candy goes on to hard crack.

In one marshmallow method the syrup is cooked to about 80% dissolved sugar at 230 °F (110 °C), while hard candy is boiled to about 293 to 320 °F and ends with only 2 to 3% water. Factories can cook under vacuum to reach that concentration at a lower temperature, with less sucrose splitting into glucose and fructose.

Sugars from corn syrup thicken the syrup, slow sucrose molecules and cling to crystal surfaces. Undissolved sugar on the pan wall can make a whole batch grain, which is why old recipes wipe the sides down with a damp cloth or wet hand. In hard candy, high levels of glucose and fructose can trade graining for stickiness, since they draw in more moisture than sucrose.

In one lab method, fondant syrup boiled to 237 °F (114 °C) was cooled to 120 °F (49 °C) and then mixed for 10 to 12 minutes until thick and opaque. In creamed honey, more crystal nuclei give more and finer crystals, and stirring spreads them evenly. A batch that turns grainy can be reboiled with a little more water.

Caramelization is heat transforming sugar into brown colors and new flavors, and caramel ends up slightly bitter with a burnt-sugar aroma; in lab caramel models, furan compounds formed along the way add to the bitterness. A wet syrup can grain if undissolved sugar is left on the pan sides, and by hard crack the syrup is close to burning, so watch it closely.

That study found the same shift with colder temperatures and faster pulling, and earlier cutting tests had named final moisture as the most critical factor besides cutting speed. Hard candy is cooked until less than 5% water remains, and if it later picks up water from humid air it softens and turns sticky.

One marshmallow method cooks equal parts sugar and glucose syrup to about 230 °F (110 °C), adds dissolved gelatin, then whips the cooled mix. An old white nougat recipe boils honey nearly to the crack stage, stirs in stiffly beaten egg whites and cooks on to crack. In marshmallow, a slip in temperature or sugar concentration can make the foam collapse or weep liquid.

High-methoxyl pectin gels in acidic conditions with a high sugar level, while low-methoxyl pectin gels by cross-linking with calcium instead. Gelatin, made by partly breaking down collagen, gives the more elastic chew, but its gel is thermoreversible, so gelatin candies need cool storage, especially in hot weather.

Gelling continues after the spheres leave the bath: in one study, broth beads made with more calcium, or held longer afterward, came out smaller, denser and harder. Rinse the spheres to remove excess calcium, and treat the time from bath to plate as part of the recipe.

Moisture-driven graining begins at the surface, where a film of absorbed water lets sucrose crystals nucleate, and the grained zone grows as the water works inward. Hard candies high in corn syrup may resist graining yet pick up enough water to turn sticky. Old confectioners advised making fondant on a clear day and keeping hard candies in jars or wrapped.

Cooked icings instead beat a hot syrup into whipped egg whites. In one old boiled frosting, with the syrup taken to the thread stage, beating time decides the result: stop early and it runs, go on too long and it loses its smoothness, though a few drops of boiling water or lemon juice can help. Done right, it sets glossy outside and soft within.

The powdered-sugar kind is easy to make and adjusts like any uncooked icing, with more sugar to stiffen it or more liquid to loosen it. The custard kind is richer: yolks and sugar cooked with milk until the custard coats a spoon, then combined with an equal weight of butter while still lukewarm.

Cut cold, solid fat into the flour with two knives or a fork rather than your fingers, which warm it. Oil saturates the flour more readily, so solid fat gives the flakier crust. Fat worked into a dough also acts as a shortening, weakening its structure so the baked pastry is tender rather than tough.

Wheat proteins form an elastic gluten network once they are mixed with water. An older pastry guide adds ice water in a thin stream while stirring and stops once the dough just holds together. Expect roughly one part water to four of flour by measure, but let the feel of the dough decide.

Fork holes all over the base and sides keep the shell from blistering upward, or you can bake the dough draped over the back of the pan. For a custard pie, a partial bake of the crust lets the filling then cook at the low heat eggs need without curdling.

Pastry bakes best fast in a hot oven, so for a slow filling such as custard, start the pie hot so the crust gets that heat, then lower it for the filling. Even a baked shell softens if it stands long after filling. Strudel makers scatter bread crumbs over the dough under juicy berries to soak up the juice.

Stretched wheat dough springs back; even bread dough measurably retracts as it leaves sheeting rollers. Old practice rests kneaded pastry for a few hours or overnight to lose that elasticity and lays a crust in the pan without pulling it, and one croissant study cut its dough along the last rolling direction to avoid recoil.

In one old set of formulas, plain short paste is just flour, salt, butter and water, the paste for fruit tarts adds sugar and a whole egg, and a drier sugared paste adds more sugar and several eggs. Butter in these doughs acts as a shortening, weakening the structure so the baked pastry is tender.

The fat must spread into thin, continuous sheets between the dough layers, and in the oven those layers hold back steam, which lifts puff pastry and opens a croissant's crumb. Laminated pastries are rich, typically 20 to 35% fat, and croissants can take up to 35% laminating fat by dough weight.

A single turn, folding the dough in thirds, triples the fat layers, and a double turn, folding it in fourths, quadruples them: one double turn plus one single turn gives 12, and repeating that pair gives 144. Repeated folding usually builds between 48 and 256 layers.

Too soft, the butter oozes out as you roll; too firm and brittle, it pokes through the dough layers. Keep dough and butter at the same temperature and consistency, chill between folds, and do not rest the dough directly on ice, which can harden the butter into lumps.

The fat keeps the thin dough layers separate, so steam builds between them instead of escaping. An old pastry guide advises the coldest dough and the hottest oven you can manage, and in one croissant study the oven also got a burst of injected steam at the start of baking.

Both are laminated the same way, by folding a fat block into dough. In one study, shaped croissants proofed for 60 to 90 minutes in warm, humid air until they doubled in volume, while puff pastry is cut, chilled and sent straight into a very hot oven with no proofing.

The more common French method instead wraps one block of fat in dough and folds it into continuous sheets. No single standard exists for these methods, and German texts even give the French name to the reverse, a piece of dough wrapped in fat. Starting from cubes saves time at the cost of gaps in each fat sheet.

If butter pokes through the dough, chill it hard again before rolling on. Roll evenly so the fat is spread the same everywhere: in one croissant study uneven fat layers gave a denser, less airy crumb, and in a puff pastry study 144 fat layers came out too thin and broken, which likely let the dough layers stick together.

Add the flour all at once and stir over the heat until the paste stops sticking to the pan; cooking flour in hot liquid is how its starch gelatinizes. Let the paste cool a little so the eggs do not cook as they go in. In the oven, a protein network in the dough traps steam, and the shell expands.

Older home recipes used four eggs to a cup of flour for cream puffs. One professional formula used 16 eggs per pound of flour for ordinary choux, but only 12, with far less butter, for a plainer paste meant for fritters, and warned against drying that version too much.

One old cookbook gives a simple test: pull a single puff first, and if it holds its shape, the batch is done. If the centers of the baked shells still hold a little moist, doughy paste, scoop it out before you fill them.

The dough is kneaded until it turns elastic, kept warm and covered, then rolled out a little on a floured cloth and brushed with melted butter. With your hands slipped beneath it, palms down, you ease it outward until it is tissue-thin, spread the filling, and raise the cloth so the strudel rolls up on its own.

Both start from eggs, milk and sugar. Stir the stovetop kind constantly over gentle heat, often in a double boiler, and strain it as soon as it coats the spoon. Bake the other kind in a slow oven until a knife comes out clean; more eggs make it firm enough to turn out without cracking.

Egg white starts to lose fluidity and clump near 140 °F (60 °C). In one study, yolk heated below 167 °F (75 °C) formed a plain protein gel, while hotter yolk also aggregated its lipoproteins into a grainy microstructure. In another, sucrose and glucose raised the heat stability of ovalbumin, the main egg-white protein. Overcooked custard can be beaten smooth again but ends up thinner. Egg mixtures are safe at 160 °F (71 °C).

In a starch-bound cream, cook the starch through: one old cream-puff filling cooks flour in milk in a double boiler for 10 to 15 minutes before the beaten egg goes in, and a lab pastry cream held flour and milk at 212 °F (100 °C) for 20 minutes so the starch could gelatinize. Starch can also stand in for some yolk, as in an old swap of one yolk for half a tablespoon of cornstarch.

Bake it in a slow oven until a knife run through the custard comes out clean, and for safety check that the center has reached 160 °F (71 °C) on a food thermometer, the point FSIS gives for egg dishes such as quiche.

The hot-syrup kind, Italian meringue, cooks the sugar to the large-ball stage and pours it into stiff whites in a steady stream while whisking. Sugar added before the whites begin to foam can hold the foam back, and once in, it binds water and steadies the meringue's shape.

One old guide bakes pie meringue in a moderate oven for about 12 to 15 minutes, and FSIS says meringue-topped pies should be safe baked at 350 °F (177 °C) for about 15 minutes. Liquid leaking from a foam is drainage, pulled out of the bubble walls by gravity and surface tension; in one study, added citric acid helped egg-white foam resist it.