Cooking meat

20 Sep, 2026

"Background information" sources include Tornberg, 2004; Baldwin, 2011; and Vaskoska et al., 2020. I won't cite/link every claim traced to them.

Muscle meat (as opposed to "organ meat" like liver or kidney, which we're just going to ignore here today) is made of long cells called fibers, and is roughly 75% water, 20% protein, 3% fat, 2% other stuff. The protein consists of myofibrillar proteins (50–55%, mostly myosin and actin), sarcoplasmic proteins (30–34%, mostly enzymes and myoglobin), and connective tissue (10–15%, mostly collagen). Heat denatures these proteins; as a result, some stick together and become more solid, while others shrink and squeeze out water. This denaturation depends mostly on temperature, secondarily on time.

Temperature

Muscle fibers start shrinking at 35–40 °C (95–104 °F). Shrinkage increases nearly linearly up to 80 °C (176 °F), but in two phases along different directions: transversely (across the fibers) around 40–60 °C (104–140 °F), then longitudinally (along them) above 60–65 °C (140–149 °F).

• Myosin denatures somewhere in 40–60 °C (104–140 °F); actin, somewhere in 66–83 °C (151–181 °F). Sources disagree somewhat on this.
• Myoglobin sticks together around 65 °C (149 °F); most other sarcoplasmic proteins stick together between 40 and 60 °C (104 and 140 °F).
• Collagen denatures around 53–63 °C (127–145 °F), then shrinks; if not stabilized by heat-resistant cross-links (which accumulate as an animal ages, and are part of why older animals have tougher meat), further heating dissolves collagen into gelatin.

Water loss

Shrinking fibers squeeze water out, more as the temperature rises. Vaskoska et al. cooked cuboids of three beef muscles in a water bath for 30 minutes at each of four temperatures, and measured cooking loss¹ from 7–11% at 50 °C (122 °F) up to 34–40% at 80 °C (176 °F). Most transverse shrinkage had already happened at 50 °C (122 °F). Longitudinal shrinkage barely started before 60 °C (140 °F) and did most of its work at 70–80 °C (158–176 °F).

Searing does not change the total loss. Yoo et al. compared steaks that were and weren't finished in the pan and found <1% difference.

Tenderness

Tenderness (or perhaps its inverse, toughness) is often operationalized as the force needed to cut across the fibers ("Warner–Bratzler shear force"). This force decreases from 50 °C to 60 °C (122 °F to 140 °F) and then increases up to 80 °C (176 °F). The explanations for both parts are somewhat disputed?

• For the increase in toughness, the conventional answer (e.g., Purslow, 2018) is that the muscle fibers shrink and stiffen as their proteins denature, but Tornberg says that it's the connective tissue contracting into a denser material.
• For the decrease (i.e., the increase in tenderness), the conventional answer is that it's collagen dissolving, but Tornberg says that it's because there's viscous fluid between fibers, whose viscosity absorbs force and resists cutting, and which is easier to cut if turned solid.

Time

Davey et al. found that prolonged cooking can double tenderness by dissolving all the collagen into gelatin and reducing adhesion between fibers to near zero. At 80 °C (176 °F) this takes about 12–24 hours; cooking for 50–100 hours adds little. The same effect is available at lower temperatures, more slowly: Bouton and Harris, 1981 found tough cuts became most tender between 55 and 60 °C (131 and 140 °F), and 24 hours at those temperatures cut shear force by 26–72% compared with one hour. So, slow cooking at such temperatures can produce a cut that's both fork-tender and medium-rare.

Color

Meat's color mostly comes from the protein myoglobin, which changes color depending on whether its central iron is bound/oxidized and whether it's denatured. Cooking promotes oxidization by unfolding the protein and exposing the iron (Suman & Joseph, 2013). Per the USDA:

• deoxymyoglobin (with unbound iron) is purplish-red;
• oxymyoglobin (with iron bound to oxygen) is red;
• metmyoglobin (with unbound but oxidized iron) is brown;
• denatured myoglobin becomes hemichrome, which is gray (other sources say "dull-brown" or "grayish-brown").

Color only tracks temperature loosely. Over 25% of fresh ground beef patties turn brown before reaching 71 °C (160 °F); low-fat patties can stay pink above it. Meat at pH 6.0 or higher can stay pink at 71 °C (160 °F). Color also depends on the path, not just the endpoint: the faster meat comes up to a given temperature, the redder it is, and the longer it is held there, the paler it gets (Ryan et al. 2006).

Crust

Browning is the Maillard reaction, between amino acids and reducing sugars, which produces hundreds of by-products, and which starts around 150 °C (302 °F). Raising the pH slightly (e.g., by adding a pinch of baking soda) increases browning too. Flavor comes from those Maillard reaction products and from fat. (The same chemistry makes heterocyclic amines, which are carcinogenic in rodents at extremely high doses, but don't have a definitive epidemiology in humans.)

Safety

Pathogens die on a time-and-temperature curve, not at a threshold. All known food pathogens stop multiplying above ~52 °C (126 °F) and begin to die (though some bacteria can transform into (endo)spores that survive even boiling temperatures — they, and the heat-stable toxins some of them leave behind, are why reheating doesn't reset the clock on leftovers). Regulators encode this as equivalent (time, temperature) pairs. The FDA Food Code's table for whole meat roasts runs from 130 °F (54 °C) for 112 minutes, to 158 °F (70 °C) for 0 seconds; poultry gets 165 °F (74 °C) instantaneous. USDA suggests 145 °F (63 °C) plus a three-minute rest for whole cuts of beef, pork, lamb, and veal; 160 °F (71 °C) for ground beef; 165 °F (74 °C) for poultry. USDA is explicit that "appearance and color are not reliable indicators of safety or doneness".

Resting

The traditional case for resting is that it helps juices redistribute, so less is lost on the cutting board; but Prieto et al. tested it and found that, while unrested roasts lost more juices during slicing, rested roasts lost more juices during the rest itself. The result is similar total moisture loss and similar palatability.

Digestion

Carmody et al. (2011) fed mice four preparations of beef (raw or cooked, whole or pounded) and found that cooking raised energy gain, but pounding did not. The mice ate less cooked meat by fresh weight and still lost less weight. They also preferred cooked meat when given the choice.

Cooking makes digestion itself cheaper. Boback et al. measured how much energy pythons spend digesting a meal, and found that cooking and grinding each cut the required energy by ~12%, which almost stacked. However, protein digestibility is close to ceiling either way; Oberli et al. measured rare meat at 94% digestibility and fully cooked at 90%, a nonsignificant difference. For comparison, egg protein is 91% cooked against 51% raw.