Terminal Ballistics · Bear Defense

Can a 9mm Stop a Bear?

A 9mm has killed a grizzly. It took seven rounds of hardcast, every hit to the body, one cartridge left in the magazine — and the bear turned and ran rather than dropping. So the question isn't whether a 9mm can kill a bear. It's whether the physics gives you a margin you'd bet your life on.

BallisticEngine · Published 29 August 2026

In 2016 a 33-year Alaska guide stopped a charging brown bear with a compact 9mm. That case gets cited constantly as proof that a 9mm is "enough." It is worth reading the whole sentence: he fired roughly seven rounds of heavy hard-cast lead into the animal's body at point-blank range, the bear broke off and ran, and he had one round left when it was over. It was killed — recovered later — but it was not stopped in its tracks, and nothing touched the skull.

We build a terminal-ballistics engine, so our instinct with a question like this is not to pick a side but to run the numbers. What follows is the energy, the bone, the geometry, and the case record — laid out so you can decide for yourself. This is not a caliber war and it is not advice. It is the physics.

1. The energy budget

Everything downstream starts with kinetic energy:

A 9mm is a superb human-defense cartridge, but it sits at the bottom of the range people carry against big animals. Here is where the common loads land, with the usual "bear calibers" for scale:

All common 9mm loads cluster near 320–410 ft·lbf; grain weight barely moves the number. The gap to a dedicated bear cartridge is not a few percent — it is multiples.

A .44 Magnum carries roughly four times a 9mm's muzzle energy. A .454 Casull, five. The 10mm — the modern compromise — about twice.

2. The material ladder

Energy only matters relative to what it has to defeat. Compressive strength (σ_c) is a useful yardstick for how hard a material resists being pushed through, and it puts the problem in perspective:

Cortical bone runs roughly 130–200 MPa in the biomechanics literature — and stiffens further at ballistic strain rates. That figure is generic long-bone/femur cortical bone, not a bear-specific measurement (more on that in a moment). Even so, the ladder is the point: bone is fifty to seventy times tougher than drywall and lands about halfway to mild steel. The velocity-decay physics behind that "passes through drywall" result is the same drag relationship we derive in How Far Do Bullets Travel Underwater.

A 9mm loafs through drywall and then meets a material two orders of magnitude tougher.

3. The skull problem

You will read everywhere that a grizzly's skull is "an inch to an inch and a half thick." Treat that with suspicion: it has no primary source we could find. The figure circulates through blogs and social posts, and it conflates the overall height and mass of a big skull — with its bony sagittal crest — with the thickness of bone a bullet actually crosses. No measured cranial-bone-thickness study for a grizzly, in millimeters, appears to exist in citable form. That is a genuine gap, and it is more honest to say so than to print a number.

What is measured is the human skull: the frontal bone runs about 6–9 mm. A large bear's is thicker and denser, but "38 mm of solid bone" is not supported by anything measured. And crucially, cranial bone is not a solid billet — it is a sandwich: an outer layer of dense cortical bone, a spongy middle layer (the diploë, often more than half the total thickness), and an inner cortical layer. A bullet crosses that in stages, not as one wall.

The harder problem is geometry. A bear's forehead is convex, so a frontal shot almost never lands square. The bone in the path scales with the impact angle:

A charging bear presents a low, moving head on a rounded skull. The bullet arrives at an angle, the effective bone grows, and the odds of a glance or ricochet climb. There is no published deflection-angle threshold for a 9mm on ursine bone — so we won't invent one — but the direction is clear.

Shooting a charging bear in the head is bad advice not because the skull is impenetrable, but because the geometry is stacked against a clean perpendicular hit.

3½ · aside A note on what we don't know

Two numbers people state with confidence are not actually established: the thickness of a grizzly's frontal bone in millimeters, and the compressive strength of bear cranial bone specifically. Neither turned up in a measured, citable form. The generic cortical-bone range above is the honest stand-in, flagged as generic. Where the data runs out, the right move is to say so — and it's a candidate for a future BallisticEngine bone-target model.

4. Construction matters more than caliber

Whether a handgun round works on a bear depends less on the headline caliber than on how the bullet is built. Three behaviors:

Jacketed hollow point (JHP). Expands on impact — larger frontal area, more energy dumped shallow, less penetration. This is exactly what you want on a human attacker and exactly the wrong behavior on a bear, where you need depth to reach vitals behind hide, fat, muscle, and bone.

Full metal jacket (FMJ). Doesn't expand, but a round nose can skate off bone, and its sectional density is lower than a heavy cast bullet's.

Hard-cast lead flat-nose. Non-expanding, with a flat meplat that cuts a straight path and holds its sectional density — the deepest, straightest penetration of the three. This is why bear loads from makers like Buffalo Bore and Underwood are hard-cast, and why the guide in the 2016 case was running Buffalo Bore hard-cast in his 9mm. Yes, 9mm hard-cast for the woods exists — but it is still a ~500 J 9mm.

Sectional density (bullet mass over frontal area) is the parameter that rewards a heavy non-expanding bullet:

A heavy 147gr 9mm closes some of the SD gap on paper — but only a non-expanding bullet gets to use it, which is why the 9mm bear loads are all hard-cast. We pull sectional density apart in full in Sectional Density, Explained, and the split between penetration and wounding in Game Ballistics vs Real Physics.

The property that makes 9mm JHP excellent on a human attacker — controlled expansion — is exactly what works against it on a bear.

5. The case record

The 2016 case, in full: an Alaska guide, a compact S&W 9mm loaded with Buffalo Bore 147gr hard-cast +P, a charging brown bear at six to eight feet. He landed roughly seven hits — to the side and body, not the head — the bear turned and ran, and he finished with a single round left. It died and was recovered; it was not dropped on the spot. The stop came from body shots with heavy penetrating ammo, not from defeating the skull.

Set beside it a 10mm counter-example: two experienced shooters put a charging interior grizzly down at close range with a Glock 20, center-mass. And the broader pattern across documented handgun stops is consistent — multiple hits, body shots, penetrating ammunition. The rare one-shot stops are central-nervous-system hits: a brain shot through the eye or mouth, or a spine hit. Those are small, moving, protected targets you cannot count on.

Now the part that matters most for reading any of these stories:

Successful stops get written up. The maulings where the gun didn't work do not. A case record is not a success rate.

Every "man kills grizzly with a 9mm" headline is, by construction, a case where it worked. The encounters where a handgun failed to stop a bear — where the person was injured or killed — rarely become shareable anecdotes. Counting the wins tells you the outcome is possible. It tells you nothing about how often it fails.

6. What the data actually says

Two peer-reviewed Alaska studies are the closest thing to real numbers. In Efficacy of Firearms for Bear Deterrence in Alaska (Smith et al., 2012), 269 firearm incidents from 1883–2009 showed handguns succeeding in 84% of cases (31 of 37) and long guns in 76% (134 of 176) — with no statistically significant difference between them. Firearm type and number of shots did not predict the outcome; what did was the bear's species and behavior, whether it charged, and the human's activity. Bears were killed in 61% of incidents.

The caveat is the whole story here: that study does not break success out by caliber, and the handgun sample is just 37 incidents. The 84% figure is every handgun from a .357 to a .454 lumped together. It does not establish a 9mm-specific rate, and reading it as "9mm ≈ 84%" is not what the data supports.

The companion study, Efficacy of Bear Deterrent Spray in Alaska (Smith et al., 2008), found bear spray stopped unwanted bear behavior about 90% of the time across all three North American species. The two studies used different designs, so they are complementary data points, not a clean head-to-head — and we're not going to hand you a verdict on spray versus gun. Here are both figures; the choice is yours.

As for what the state agency says: Alaska's Department of Fish and Game recommends a .300-Magnum rifle or a 12-gauge with rifled slugs if you have to shoot a bear, and adds that heavy handguns such as a .44-Magnum may be inadequate in an emergency, especially in untrained hands. If a .44 Magnum sits on the "may be inadequate" line, that tells you where a 9mm sits relative to the recommendation.

7. So — can it?

Yes. A 9mm can kill a bear; it is documented. But hold the whole picture: the best-known case took seven rounds of hard-cast into the body, ended with one round left, and the bear ran rather than dropped. The physics says why the margin is thin — a bullet carrying around 500 J, meeting bone at 130–200 MPa, on a convex skull that turns most frontal hits oblique, with defensive JHP designs that penetrate less exactly when you need them to penetrate more.

With the right construction (heavy hard-cast) and the right shot (body or central nervous system), it has worked and can work. As a plan — the thing you carry expecting to stop a charge — the same physics is what pushes guides toward more energy, more penetration, and faster follow-up shots. Both of those statements are true at once, and neither one is a slogan.

The question was never whether it can. It's whether you'd bet on it.

Run the barrier and penetration numbers for any load yourself in the BallisticEngine calculator, and decide with the physics in front of you.

8. FAQ