Somewhere between one in four and one in two people do it. Most of them have been told, usually by a relative, that it will give them arthritis. It will not. But almost everything else you have heard about the sound is either out of date or was never established in the first place.
Start with the part that is genuinely known, because it is stranger than the explanation you were probably given.
A knuckle is a sealed capsule. The ends of two bones sit inside a bag of synovial fluid, and that fluid has gas dissolved in it the way a fizzy drink does. Pull the joint apart and the space inside grows faster than the fluid can fill it. Pressure drops. Dissolved gas comes out of solution and a cavity appears.
This is not speculation. In 2015 a team led by Greg Kawchuk at the University of Alberta put a hand in an MRI scanner and filmed it. A cavity appears at the instant of separation, and — this was the surprising part — it is still sitting there afterwards. It does not vanish.
That finding overturned the textbook explanation. Since 1971 the standard account had been that the pop is a bubble collapsing. Kawchuk pointed out that the 1971 study never actually saw a collapse; it inferred one. So the 2015 conclusion was that the sound is the cavity being born, not dying.
Then in 2018 two researchers at École Polytechnique modelled the fluid dynamics and found something more careful: the acoustic signature matches a partial collapse. Not the whole bubble, just part of it — which would produce the sound and still leave the residual cavity that the MRI saw. Both results can be true at once.
So the honest position, which you will not often see stated: we can see the cavity form, directly, in a living hand. Whether the noise is its birth or a partial collapse a few milliseconds later is still argued. The 2015 study was one man and ten joints. The 2018 study was a simulation, not a measurement. Neither closes the case.
One correction worth making, because it circulates widely: the MRI did not show a bright flash. It showed a dark void. The “like a firework exploding in the joint” line comes from a different study the same year, at UC Davis, which used ultrasound rather than MRI. Two teams, two machines, two very different pictures of the same event.
Crack a knuckle and it refuses to crack again for roughly twenty minutes. This is the oldest measured fact in the whole subject. In 1947, at St Thomas’s Hospital in London, two anatomists tied string around a finger, ran it through a tensiometer and x-rayed the hand at the moment of the crack.
They found the minimum gap between two cracks was seventeen minutes, and never more than twenty-two. They also found the threshold: about six kilograms of pull does nothing. Seven or more and the bones spring apart by one to three millimetres and the noise arrives.
The gas needs time to redissolve, and the joint needs time to settle back. This gives you a free test. If a joint makes a noise you can repeat immediately, as many times as you like, it is not cavitation at all — it is a tendon snapping over bone. That is a different thing, and none of the reassuring research below applies to it.
The famous answer is a man called Donald Unger. For fifty years he cracked the knuckles of his left hand at least twice a day and never his right. At least thirty-six thousand five hundred cracks on one side, zero on the other. He then x-rayed both hands and published the result as a letter in Arthritis & Rheumatism in 1998. No arthritis in either hand, no difference between them. He collected an Ig Nobel Prize for it in 2009.
It is a wonderful story and it is a sample of one, unblinded, judged by the experimenter himself. The actual evidence is duller and better.
The strongest study is from 2011: 215 people aged fifty to eighty-nine, comparing those with hand osteoarthritis on x-ray against those without. Arthritis showed up in 18.1% of knuckle-crackers and 21.5% of non-crackers — a difference well inside noise. More importantly, it was the first study to test dose. Neither total years of cracking nor “crack-years” (daily frequency multiplied by years) related to arthritis at all. Two earlier studies, in 1975 and 1990, found the same absence of a link.
The precise claim is: every study that has looked has failed to find a connection between knuckle cracking and arthritis. That is not identical to “proven harmless,” because all of this work is cross-sectional and relies on people remembering their own habits. But it is a consistent answer across four decades, and the relative who warned you has nothing on their side.
The arthritis worry was never invented out of nothing. The force needed to crack a knuckle has been shown in the lab to exceed the energy threshold that can damage articular cartilage. And cavitation — gas bubbles forming and collapsing in a liquid — is the same physical process that erodes ship propellers. That is exactly why researchers assumed bubble collapse was the sound source in the first place.
The mechanics say it should cause wear. The observations say it does not. That tension is more interesting than either simple answer.
There is one study people cite as evidence of harm, from 1990: three hundred patients, and the habitual crackers had more hand swelling and weaker grip. Its conclusion was blunt — “habitual knuckle cracking results in functional hand impairment.” But it also found that cracking clustered with manual labour, nail biting, smoking and drinking, and it adjusted for none of them. Two later studies, one of them blinded and one using a grip dynamometer on thirty-five heavy crackers, found no grip-strength difference at all. The second of those found crackers had thicker knuckle cartilage than matched controls.
What does show up in the literature is injury from unusual force: a thumb ligament sprain, a tendon sagittal band tear, one case of calcium deposits in the joints. All from vigorous, deviant attempts. There is a real difference between a pop and a yank.
Here is where we have to be straight with you. No study has ever measured pleasure, relief or reward during a knuckle crack. Not one. The confident three-part explanations you will find online are inference dressed as fact.
What is known: joint capsules are densely packed with mechanoreceptors — pressure sensors, four different types, mapped in detail in other joints. Pulling the capsule fires a great many of them at once. That much is solid anatomy. Nobody has ever recorded them firing during a knuckle crack.
The claim that the sound is the reward is the weakest link, and there is evidence pointing the other way. In spinal manipulation, where the same cavitation happens, researchers compared patients who heard a pop against patients who did not. Range of motion, pain scores, disability scores — no difference between the groups. If the noise were the active ingredient, that is precisely where it would have shown up.
There is also a problem with the slot-machine metaphor that gets used for this. A slot machine is compelling because it pays out unpredictably. Knuckle cracking pays out reliably and then refuses for twenty minutes. It is close to the opposite arrangement.
The best-supported reading is modest: pulling a joint apart produces a large, sudden burst of sensory information from a capsule that is usually quiet, and that burst feels like release. Why it is pleasurable rather than merely noticeable, nobody has established. One of the more honest reviews on the subject says so directly — what we do not know yet is all the reasons why people crack their knuckles.
It will not give you arthritis; four decades of looking has turned up nothing. It is not doing maintenance on your joints either — that is a story, not a finding. Do not force a joint that does not want to go. And if you can crack the same spot twice in a row, that is a tendon, not a bubble, and the research above has nothing to say about it.
Roughly two-thirds of knuckles will not crack at all, whatever you do. In one study forty adults tried to crack all four hundred of their knuckles and only sixty-two produced a sound. The commonest reason for failure was identified in 1947 and has not been improved on since: you cannot relax the muscles enough.
Short detours through the parts of being human that nobody explains.
Subscribe on YouTube