The BBC reports that, officials have referred to a possible crack in the suppression chamber of reactor 2 - a large doughnut-shaped structure, also known as the torus, below the reactor housing.
That would allow steam, containing radioactive substances, to escape continuously.
This is the most likely source of the high radioactivity readings seen near the site in the middle of Tuesday.
We're absolutely not looking at another Chernobyl"
Under normal circumstances, the suppression chamber stores a large volume of water that can be used to condense steam produced in the reactor.
The industry newsletter World Nuclear News reports that a "loud noise" came from the reactor chamber, and that "the pressure... was seen to decrease from three atmospheres to one atmosphere after the noise, suggesting possible damage".
This pressure drop is consistent with the chamber cracking and releasing steam.
Once de-pressurised, material would probably be ejected at a slower rate. This is consistent with the swift fall in readings taken around the plant later in the day.
Another possible source was a fire and explosion in the building housing reactor 4.
At the time of the earthquake, reactors 4, 5 and 6 were shut down for planned maintenance.
During this period, fuel rods are withdrawn from the reactor and stored in a pool - looking like a big swimming pool - in the upper level of the reactor building, outside the containment vessel.
The water keeps them cool, and protects workers below from radiation.
The rods may later be placed back in the reactor or taken away for long-term storage and re-processing, depending how long they have already been in service.
An official with the Tokyo Electric Power Company (Tepco), which owns the station, told BBC News that the cause was believed to be a lack of water in the pool.
This caused the rods to overheat, reacting with steam and generating hydrogen - which then ignited, similar to the earlier blasts in reactor buildings 1, 2 and 3 when hydrogen was vented from the reactor vessel.
But at a news briefing in London, academics were perplexed by the idea.
"The fuel pins in the pond should be totally contained, there should be no damage," said Laurence Williams, professor of nuclear safety at the University of Central Lancashire.
"If you did have a fire, you'd have the same issues (as in a reactor melt) - you'd have fission gases in the spaces inside the fuel pin, depending on how long the fuel was in there the iodine might or might not be a hazard... overall, I'm quite mystified."
Kyodo News, the main Japanese news agency, reported that technicians were unable to pump water back into the pool and that helicopters may be used to dump water into the building from the air.
However, according to the US-based Nuclear Energy Institute (NEI), the Japanese Atomic Energy Agency (JAEA) was working on a different theory - ascribing the fire to an oil leak.