New Zealand's most active volcano is primed to blow again, just days after scientists thought it had calmed down enough to lower alert levels.
White Island's surprise leap back to life this month, raising fresh concerns of an imminent eruption, again caught our best experts off-guard.
"We actually sat down and met in here on January 14 and were reasonably comfortable that the hazard had decreased," GNS head of volcanology Dr Gill Jolly said.
"Within 24 hours of meeting, the volcano decided to do something different."
Fears of an eruption are now at their highest since August, when the current episode of unrest at the barren Bay of Plenty island quickly picked up.
After more than a decade of peace, the volcano had sounded off with a weak ash eruption and a series of quakes over July and August.
Later, scientists saw the first extrusion of lava in New Zealand in almost 50 years.
A spiky lava dome that quickly appeared also showed magma was being plugged just below the surface.
A few weeks later, activity briefly lulled and aviation alert levels returned to yellow, until the quietwas broken by a flurry of seismic tremors.
The tremor activity has dropped, but is still far from normal levels.
And what are called "hybrid" earthquakes - a quick jolt followed by a period of resonance - have presented a tell-tale symptom of rock fracturing and magma movement.
What comes next could be only more of the same, an end to the cycle, or another eruption that could strike at any moment.
Monitoring offered some hints, said Dr Jolly. Rising magma might be given away by a sudden change in sulphur dioxide levels or extra high steam and lake water temperatures.
But even then, volcanologists could not say an eruption would happen at a certain time, or in a certain place, or even of a certain size.
"With all volcanoes, there is a lot of uncertainty ... it's a bit like forecasting the weather with your eyes closed."
A White Island eruption could hurl rocks about the crater or even into the surrounding sea, as happened at times during nearly four decades of activity between 1976 and 2000.
It could trigger "hot surges" triggering ash avalanches, or push up clouds of ash, said Dr Jolly.
"It's quite hard to tell which of these scenarios is the most likely, and we are obviously collecting more data on a daily basis."
She could, however, say that an eruption would probably be contained to the island itself.
"If the wind was in the wrong direction and we have one of the larger-scale eruptions, then there is the possibility of light ash fall on the mainland, but there is very little in the geological record.
"That suggests that large ash falls are very rare, if they happen at all."
To someone standing on a Bay of Plenty beach, an eruption might appear only as a darker cloud hanging over the island, or a lightning storm powered by ash particles rubbing together.
With episodes lasting between weeks, months and years, it was difficult to say when White Island would end its tantrum, Dr Jolly said.
"They often look like they're building up to something and then nothing happens at all."