The speed of light in a vacuum -- 299,792,458 metres per second - is one of the fundamental principles and outcomes of the theory of relativity, the core constant of the universe and a total, hard limit for the velocity of everything.
Turns out it might be wrong.
Physicist James Franson of the University of Maryland has submitted a paper to the New Journal of Physics in which he claims to have evidence that the speed of light is actually slower than we thought.
Franson observed the supernova SN 1987A, which exploded in 1987, and noted that the photos arrived 4.7 hours later than expected. Scientists at the time said the photos were probably from another source, but Franson argues that the light might actually have slowed as it travelled due to a process called 'vacuum polarisation'.
In that process, photons break down to positrons and electrons for a split second before recombining. Franson argues that process might have a slight impact on the speed of the photon, meaning that over 168,000 light years it could add up to a 4.7 hour delay.
The result of this effect, if true, would not be a change in the fundamental equations of relativity, but a practical consideration when calculating the distances and times between various sources of light and ourselves.
But this alone would be staggeringly important - everything from the distance to the Sun to the distant galaxies would have to be remeasured and recalculated. The very processes that operate the universe would have to be looked at again, from scratch.
The paper is now undergoing peer review.