For a group of scientists in the little town of Parkfield, California, the Earth finally moved — when an earthquake of magnitude 6 hit last week. After lying in wait for a major rupture in this part of the San Andreas fault for more than 20 years, scientists in the area pounced on their huge array of monitors to examine the data.

The earthquake, which caused little damage and no deaths, will provide a wealth of information for researchers. But it has also thrown up questions about whether major earthquakes come in predictable periods or with predictable characteristics.

Scientists studying Parkfield in the late 1970s hypothesized that an earthquake would hit the area once every 22 years. The current earthquake, striking after a 38-year gap, might make them recalculate. It does seem to add to the evidence that this earthquake zone may be “quasi-periodic”, says David Jackson, an earthquake researcher at the University of California, Los Angeles. But he generally doubts the idea that certain places have earthquakes with specific characteristics.

Previous Parkfield earthquakes have led scientists to expect foreshocks of up to magnitude 5 striking the area about 17 minutes before the main earthquake. Instead they got aftershocks of magnitude 5. The earthquake also moved north to south along the fault, the opposite of what normally occurs in this region. Showing up late, moving backwards and speaking out of order, the delinquent earthquake is posing a few problems. “These are interesting phenomena, but we don't yet have any explanation,” says geophysicist Steve Hickman of the US Geological Survey.

Because of the regularity expected of Parkfield's earthquakes, the area has been intensely monitored for decades. Instruments are used to measure details of structural changes deep below the surface and the movement of the fault. “The data will be copious and sweet,” says Jackson.

The earthquake did not come at a great time for researchers working on the San Andreas Fault Observatory at Depth (SAFOD), also in the Parkfield area. This project aims to drill into the heart of an earthquake-generating zone about 5 kilometres away from last week's rupture to study the fault itself during an earthquake. The project reached a drill depth of 3 kilometres two weeks ago, but seismometers, which would have given unique data, have not yet been installed in the hole. And the project will not drill through the fault itself until next year.

A second, pilot hole a few metres away did have some seismometers in place. But this summer, workers accidentally snipped off 25 seismic stations from the bottom of a string of 32. They have not yet been replaced.

Last week's earthquake has got SAFOD project leaders excited, however. Their part of the fault line is expected to produce smaller earthquakes every two to three years, so they have plenty of time to see more earthquake activity.