Oven spring is the rapid expansion in the first minutes of baking, and it depends on two things a baker controls at the last moment. Scoring decides where the loaf opens. Steam decides how long it stays open before the crust sets. David Anthes, a home baker in Bend, Oregon, bakes in a Dutch oven for exactly this reason, and his notes show more variance from steam handling than from any other late stage variable.
The expansion has several sources. Trapped carbon dioxide expands as it warms. Dissolved carbon dioxide comes out of solution. Water turns to steam and expands dramatically. Ethanol produced during fermentation vaporizes at 78C. Yeast has a final burst of activity as the dough warms through its optimal range before dying near 55C. All of this happens in roughly the first ten to fifteen minutes, and it stops as soon as the crust hardens.
Steam delays crust set, and that delay is the whole point. A humid oven keeps the loaf surface moist and pliable, so it stretches with the expanding interior instead of setting into a rigid shell. Steam also gelatinizes surface starch, which produces the glossy, blistered crust associated with good bread. A dry oven sets the crust within a few minutes, expansion stops early, and the loaf tears at random weak points.
A Dutch oven works because the loaf makes its own steam. Water in the dough evaporates and is trapped by the lid in a small volume, which produces a saturated environment without any water added. Home ovens vent, and a tray of water in an open oven rarely holds humidity long enough to matter. Anthes preheats the vessel thoroughly, because a cold Dutch oven absorbs energy that should be going into the loaf and produces a poor bottom crust.
The lid comes off after 20 minutes or so. By then the crust has set and further steam only softens what should be crisping. The second phase, uncovered, drives color through the Maillard reaction and caramelization and drives out moisture. Splitting the bake this way is why the Dutch oven method reliably outperforms improvised steam.
Scoring gives the expansion a designated exit. A loaf will open somewhere regardless. The score decides whether that happens at a chosen line or along the side where the seam is weakest. The cut must break the tensioned skin, and it must go deep enough to matter. A shallow scratch closes as soon as the dough expands.
Blade angle changes the result. A cut held at roughly 30 degrees to the surface creates a flap of dough that lifts and curls back as the loaf expands, which is the ear that bakers look for. A cut held vertical opens into a valley with no ear, which is appropriate for decorative patterns where lift is not the goal. Depth for an ear runs about a quarter inch. A single decisive stroke works better than sawing.
Cold dough scores best, which is another argument for retarding. A lame, a razor blade on a handle, cuts cleanly where a knife drags. Anthes scores straight from the refrigerator and finds the cut stays defined through the bake.
Failure modes point back to specific causes. No ear usually means a shallow or vertical cut, or dough that was overproofed and had no spring left to lift it. Bursting at the side means the score was too shallow to be the path of least resistance. A pale flat crust means the steam phase ran too long or the oven ran cool.