The cryptodome growing within Mount St. Helens sowed the seeds of its own destruction. Had it been a small thing, it might have become a younger sibling to Goat Rocks. Pacific Northwesterners might have seen a few displays of volcanic fireworks, another dome added to the edifice, and a return to serenity. It hardly would have made the news.
But this dome, unlike Goat Rocks, kept growing, there beneath the surface. It set groundwater steaming. The hydrothermal system driven by its heat caused some pretty spectacular phreatic eruptions. It severely over-steepened the volcano's north flank.
And it kept growing, right up until the end, when an earthquake on the morning of May 18th, 1980, brought the whole thing down.
This is a story of seconds. A lot can happen in a few seconds. A whole mountain can change.
Before 8:32:11 am Pacific Daylight Time, the cryptodome had been following the usual trajectory of masses of magma hanging about under a volcano's skin. The hot dacite was cooling down toward the point where liquids become solids. Various minerals were crystallizing as the magma cooled; toward the outside, where fresh dacite met old volcanic products and water circulated through the hydrothermal system, things had cooled a bit more than in the interior, and probably formed something of a dense crust with only minor, if any, vesicles. Other bits of the dome might have been a little more bubbly, with gasses forming more vesicles than at the relatively solid outer edges. But everything was under pressure - 175 bars of it - and so most gasses would have been dissolved within the melt. Even the water heated by the magma hadn't boiled. Under that much pressure, it couldn't get steamy.
This all sounds very quiet and happy, but gasses, including water vapor, that are under pressure are being forcibly confined to a small volume. They're the kind of things that normally take up quite a bit of space. Give them an opportunity, and they'll proceed to do so rather emphatically.
Opportunity rang at 8:32:11 am PDT in the form of a magnitude 5.1 earthquake. It took about ten seconds for the unstable north face to begin sliding. Old rock, debris, domes, dirt, and glaciers careened down the mountainside, leaving a scarp 700 meters (2,297 feet) high, 1 kilometer (3,281 feet) wide, and just 20° off from vertical. All of that stuff had been keeping the lid on the cryptodome and its hydrothermal system. Now the pressure was off. And all of those gasses that had been kept down were now free to get out.
They didn't go immediately. It took almost twenty seconds for the blast to begin after all that volatile material found itself freed. It's possible the cryptodome and its hydrothermal system weren't yet completely exposed, and expanding gasses might have been working their way through the remains of the north flank towards freedom. But there are also some complicated fluid dynamics to contend with. Even if the cryptodome had been instantly not-crypto, the blast would have taken some time - upwards of at least ten seconds - to form. A rarefaction (expansion) wave had to propagate, water needed to flash to steam, and dissolved gasses begin exsolving (coming out of solution), before an expansion wave involving more than one phase got round to propelling vapor and bits of volcano out.
This is all a complicated (yet greatly simplified) way of saying: water flashing to steam and gasses released from confinement blew the cryptodome apart.
There's still some debate over whether the magmatic gasses within the cryptodome or the steam in the hydrothermal system caused the blast. The fact that so much of the dacite cryptodome got blown out suggests that its own gasses were major contributors. Regardless of which set of volatiles got things moving, once steam, magmatic gasses, bits of dome, and appreciable chunks of Mount St. Helens not stripped off by the landslide blasted through the base of the landslide scarp while more erupted from the top. Mind you, the landslide scarp wasn't an immobile feature: it was slip-sliding down as block #2, leaving plenty of room for the expanding gasses to escape - sideways.
This is not usual behavior at volcanoes. But gasses that have been confined to a small space want to be in a large one, and they are not big on protocol. If a landslide leaves an exit in the side of a mountain, a lateral blast is the way they'll go. Things were merely sonic to begin with, but as the blast reached the open air, it expanded supersonically. Fragments of old rock, hot chunks of the dome, and debris picked up on the way formed a roiling, boiling, dark gray cloud. Going up, gravity worked against it; going down, what was left of the ground interfered - but laterally, it found its way clear. It expanded more than 90° around the vent and sped down the mountain at speeds exceeding 600 km/h (373mph). In the valleys of the North and South forks of the Toutle River, the atmosphere was about to decide the fate of people who had, only seconds before, been enjoying a beautiful May morning.
Lipman, Peter W., and Mullineaux, Donal R., Editors (1981): The 1980 Eruptions of Mount St. Helens, Washington. U.S. Geological Survey Professional Paper 1250.
Special thanks to Helena Mallonee of Liberty, Equality, and Geology for use of her photo; and Lockwood DeWitt of Outside the Interzone for an excellent description of what the gasses were up to before and during the blast.