Table of Contents (click to expand)
- What Is The Difference Between A Dent And A Crack?
- Why Does A Sharp Crack Tip Multiply Stress So Much?
- Why Does A Dent Stay Put, But A Windshield Chip Does Not?
- Why Did The Comet's Cracks Grow A Little Every Flight?
- How Do Aircraft Stop A Crack, And What Happens When That Fails?
- So, Why Is A Tiny Crack More Dangerous Than A Big Dent?
A dent is metal that moved without coming apart, so a pull still spreads right across it. A crack has a tip so sharp that the same pull piles up on one point, working the metal there dozens of times harder than the metal beside it, and that is enough to tear the crack slightly longer every time the load comes back. The crack that destroyed a de Havilland Comet in 1954 was the width of a human hair when the aircraft left the factory, and only 25 mm long when the cabin gave way.
Picture a parking lot in August. Someone swings a door into the side of your car and leaves a dent the size of a saucer. You can see it from thirty feet away.
Now picture a hairline crack in the same panel, a few millimeters long. You would have to know where to look.
Every instinct says the dent is the damage that matters. Engineers disagree, hard enough to build whole inspection programs around it. A jet can fly for years with a dent in its skin. A crack a few millimeters long can take the roof off.
What Is The Difference Between A Dent And A Crack?
Push your thumb hard into the side of an empty soda can. The metal dishes inward and stays there, and nothing came apart. That is a dent.
Tear the ring pull off and look at the slit it leaves behind. The metal there is no longer joined to itself, and that is a crack.
The difference sounds cosmetic, but it is not. When you stretch a sheet of metal, the pull travels through the material, and engineers track how much of it each small patch carries. They call that stress: force divided by area.
Think of the load as traffic flowing through the sheet. A dent is a hump in the road: the traffic slows, rolls over it and carries on, because no metal is missing. A crack is a closed road, and nothing crosses it, so every bit of load that used to pass along that line has to squeeze around the two ends, and that is where it piles up. Damage is dangerous when it piles the load up, not when it is big.

Why Does A Sharp Crack Tip Multiply Stress So Much?
In 1913, an engineer named C. E. Inglis worked out how bad that pile-up gets. His equation for stress at an elliptical hole fits on one line:
stress at the tip = ordinary stress × (1 + 2√(a ÷ ρ))
Here a is half the length of the flaw, and ρ (rho) is the radius of its tip, a number for how blunt the end is.
Feed a plain round hole into it and you get three: a round hole in a stretched plate carries three times the stress of the metal around it.
Sharpen the tip instead. Take a flaw 2 mm long and give it a tip radius of one hundredth of a millimeter, and the multiplier climbs past twenty. Push rho all the way to zero, which is what a real crack tip is, and the equation predicts infinite stress. Real metal never delivers that, as the next section shows, but the lesson stands. Sharpness, not size, turns a flaw into a threat.
You have run this experiment yourself, on a bag of chips. Pull the sealed top apart and the plastic stretches and mocks you. Find the little molded notch at the edge, pull from there, and it opens at once, because the notch does all the work. It is also why a needle pops a balloon that your whole thumb cannot.

Why Does A Dent Stay Put, But A Windshield Chip Does Not?
Put a dent into the same equation. A saucer-sized dent has a tip radius measured in centimeters, so the multiplier sits barely above one. The dent is ugly, but structurally the metal hardly notices.
Metal has a second trick on top of that. It is ductile, meaning it can change shape for good without breaking. When a sharp flaw in steel or aluminum meets a heavy load, the metal at the tip flows and blunts itself, quietly widening its own rho.
So a smooth dent has no engine driving it. Left alone, it does not grow; it sits there and annoys you.
The catch is the word smooth. The FAA's rule for a dented steel tube sets three size limits, then adds the line that matters: the dent must be "free from cracks, abrasions, and sharp corners." A hard crease is a small rho hiding inside a large dent.
Glass has no blunting trick, so a sharp tip in glass stays sharp forever. Every glass surface is already covered in scratches too small to see, and they alone make glass hundreds of times weaker than its atomic bonds say it should be. A stone chip is a flaw thousands of times bigger, in a pane that flexes over every bump and swells in the sun. Nothing in the material is working to stop it, and plenty is working to help.

Why Did The Comet's Cracks Grow A Little Every Flight?
A pressurized cabin is a balloon that gets inflated and deflated on a schedule. The de Havilland Comet cruised at 40,000 feet (12 km) with its cabin held at the pressure of 8,000 feet, so every flight pumped the fuselage up and let it back down. Engineers call each pump-up and let-down a cycle, and cycles, not hours in the air, are what drive fatigue.
Each cycle nudges a crack tip forward by a tiny amount, and the growth depends steeply on how far the stress at the tip swings. For the Comet's aluminum alloy, it went as the fourth power of the swing. Double the swing and the crack grows sixteen times faster.
Here is the trap. A longer crack is a longer road closure, so more load squeezes past its ends and the pile-up gets worse, so it grows faster, which makes it longer still. Fatigue cracks spend almost their whole lives too small to find, then finish the job in a hurry.
In 1954, one Comet broke up in flight off the coast of Italy and a second followed within three months; the fleet was grounded. The crack that destroyed the first started as a factory defect the width of a human hair and reached 25 mm, about an inch, on flight number 1,286.
Investigators sank a third Comet into a water tank and pressurized it over and over. It failed after about 3,000 cycles, from a crack under 2 mm long at a rivet hole.
You have heard the conclusion: square windows, cracks in the sharp corners, and that is why aircraft windows are round today. The wreckage is less tidy. The fatal crack started at a bolthole more than 50 mm from a window, where the stress was under a quarter of what it was at the window's edge, and the test Comet's cracks started at rivet and boltholes too. Withey's analysis of the accidents puts the blame on "the method of fixing the windows and doubler plates onto the pressure cabin", not on the shape of the cut-outs.

How Do Aircraft Stop A Crack, And What Happens When That Fails?
Round off every corner. A modern cabin window is a rectangle with generously curved corners, and in one real widebody panel each corner is curved to a radius of about 5 cm (2 inches).
Blunt the cracks you already have. When a mechanic finds a crack in a tube, the FAA's repair manual tells them to drill a hole about 2.5 mm across at each end of it. The repair for a crack in an aircraft is to make a hole in the aircraft. It sounds like vandalism, and it is straight out of Inglis: swap a tip radius near zero for a round hole, and the multiplier falls off a cliff.
Then assume you missed one. Fuselages carry tear straps: a growing crack hits a strap and stops, the skin blows open like a flap, and the aircraft stays in one piece. Since 1978, federal rules have required a design to rule out "catastrophic failure" from fatigue, corrosion, defects or accidental damage. The philosophy is not that cracks never appear, but that they must never be able to kill anyone before an inspector finds them.

In 1988, that philosophy was tested on a Boeing 737 flying for Aloha Airlines, when about 5.5 m (18 feet) of its upper fuselage tore away in flight. A flight attendant was swept overboard and killed; the crew landed the aircraft. Where the bonded lap joints had let go, the load moved onto the rivets, and fatigue cracks grew at many neighboring rivet holes at once. In the FAA's words, "many small, hard-to-detect cracks can link up ... to form a long, critical crack."
The tear straps should have caught the rupture, but on this airframe they had disbonded too. The aircraft had flown nearly 90,000 cycles, and the difference between cycles and hours had been lost in the maintenance schedule. We cover the decompression side of that day in its own article.
This is why a maintenance crew will sign off a dent you could see from the gate, then spend an hour running eddy-current probes along a row of rivets that look fine.

So, Why Is A Tiny Crack More Dangerous Than A Big Dent?
Because size was never the question; sharpness was.
A dent is a large change in shape with a gentle tip, so the load rolls over it and stays spread out. A crack is a tiny change in shape with a tip so sharp the load has nowhere to go but the ends. One ruins your afternoon in a parking lot. The other multiplies the stress in a fuselage by a two-digit number, at a spot the width of a hair.
And a dent has no way to grow, while a crack grows using the same stress it piles up. Every cycle makes it longer, every extra millimeter makes the pile-up worse, and the worse pile-up makes the next cycle count for more. It funds itself.
So we round the corners, drill holes at the ends of cracks to blunt them, lay tear straps across the roof, and send people out with probes to hunt damage nobody can see. All of it descends from one line Inglis wrote down more than a century ago.
Next time you catch yourself glaring at a dent in your car door, spare a thought for the crack you cannot find. It is the one doing the arithmetic.
References (click to expand)
- Withey, P. A. (1997). "Fatigue failure of the de Havilland Comet I." Engineering Failure Analysis 4(2), 147–154
- Roylance, D. “Closed-Form Solutions,” MIT 3.11 Mechanics of Materials (MIT OpenCourseWare) — stresses around an elliptical hole, citing Inglis (1913)
- 100 years after Griffith: From brittle bulk fracture to failure in 2D materials — PMC / National Library of Medicine
- MIT OpenCourseWare 3.071 Amorphous Materials, Lecture 8: Mechanical Properties — Massachusetts Institute of Technology
- Lucon, E. (2024). "Review and Consideration of Apparent Negative Crack Growth in Fracture Toughness Tests" (crack tip blunting) — PMC / National Library of Medicine
- AC 43.13-1B, Acceptable Methods, Techniques, and Practices: Aircraft Inspection and Repair — Federal Aviation Administration
- Lessons Learned: Boeing 737-200, Aloha Airlines Flight 243, N73711 — Federal Aviation Administration
- 14 CFR § 25.571, Damage-tolerance and fatigue evaluation of structure — Electronic Code of Federal Regulations
- Chapter 8: Aircraft Fuselage Window, ME 424 course notes — Michigan State University College of Engineering
- Aircraft Accident Report NTSB/AAR-89/03, Aloha Airlines Flight 243 — National Transportation Safety Board







