The pop

Everyone who cracks their knuckles knows the rule without being taught it: crack once, then wait. Roughly twenty minutes pass before the same knuckle will crack again. That waiting period is a clue, and for decades the field read it backwards.

In 1971, Anthony Unsworth and colleagues concluded that the sound was a bubble collapsing. The picture was tidy. Pull the finger, pressure inside the joint drops, dissolved gas forms a cavity in the synovial fluid, the cavity implodes, you hear the pop. Textbooks repeated it for over forty years.

In 2015, Gregory Kawchuk and colleagues put the question inside an MRI scanner. Ten finger joints, pulled by a cable while the machine watched. A cavity appeared exactly at the moment of separation and sound, and then it stayed. It was never seen to collapse. A bubble that survives the pop cannot be the thing that popped. Kawchuk proposed the opposite mechanism: the sound comes from the cavity's formation, a process called tribonucleation, where two surfaces held together by fluid resist, then suddenly let go.

That conclusion had a problem of its own. A cavity forming does not obviously explain how loud the crack is. And the MRI was taking 3.2 frames per second, far too slow to see what a bubble does in the milliseconds around the sound. Catching the event directly would take well over a thousand frames per second, which imaging a living joint cannot safely do.

In 2018, Vinay Chandran Suja and Abdul Barakat replaced the camera with math. They modelled a bubble inside a realistic knuckle joint and computed the sound it would make. A collapsing bubble matches the recorded crack in both loudness and dominant frequency. The catch is that the collapse only needs to be partial. The bubble shrinks violently enough to make the sound, and what is left of it floats on, exactly as the MRI saw.

So the 1971 answer was right about the sound and wrong about the ending. The 2015 correction was right about the bubble and wrong about the noise. Your knuckle crack is a collapse that does not finish.

As for the warning everyone has heard: habitual cracking has been looked at directly. A 2011 study of 215 people found no link between cracking and arthritis in the hand. The crackers' rate came out slightly lower than the non-crackers', well within noise, and neither years nor daily frequency of cracking correlated with arthritis. The bubble is doing drama, not damage.