Does Boiling Remove Bacteria?
Yes, and it is finished before the water boils. Pathogens die well below 212°F, which is why a rolling boil is an instruction you can see rather than a temperature you have to hit.
Yes. Boiling kills bacteria, and it kills viruses and protozoa along with them. A rolling boil is a complete answer to anything alive in water.
The more useful fact is that the job is finished before the water boils.
Pathogens die below boiling
New York State’s Department of Health, writing for public health professionals rather than the public, is direct about it: “research shows that water borne pathogens are inactivated or killed at temperatures below boiling (212°F or 100°C).”
They give numbers. Water held at 150°F, 65°C, for 20 minutes “will kill or inactivate those organisms that can cause harm to humans.” And 149°F for five minutes achieves “a 99.999% kill of water borne microorganisms.”
That is pasteurization, which is the same principle applied to milk. Heat plus time. More heat needs less time, and by the time water is visibly boiling it has passed through every one of those thresholds on the way up.
So why does everyone say rolling boil
Because you can see it.
The same New York document says it plainly: “The term rolling boil facilitates communication and assures that an effective pasteurization temperature is reached to kill or inactivate waterborne pathogens.”
A temperature instruction needs a thermometer, needs the person to have one, needs it to be accurate, and needs them to hold a temperature for a period while judging both. A rolling boil needs eyes. It is a deliberately conservative instruction chosen because it cannot be got wrong, and the margin it builds in is why the advice is safe to give to millions of people at once.
Understanding that it is a proxy is what tells you when the margin matters. Which is at altitude, where it matters less than people fear.
Where the minute comes from
CDC says to bring clear water to a rolling boil for one minute. That minute is not the killing time. The killing happened during the climb.
The minute is insurance: against a pot that is not uniformly hot, against someone who called it boiling a bit early, against a thermometer nobody has. It is cheap and it removes the need for judgment.
Where it stops being cheap is fuel. In an outage where you are boiling every drop your household drinks, on a camp stove with a finite canister, the difference between one minute and ten is real. Knowing that the microbiology was done at 149°F is what lets you stop at the boil with confidence instead of boiling hard for five minutes because it feels safer.
What still goes wrong
Cloudy water first. Suspended sediment shields organisms from heat the same way it shields them from chlorine. CDC says to filter it through a clean cloth, paper towel or coffee filter before boiling.
Recontamination after. Boiled water put back into the container it came out of is not treated water any more. Cool it in the pot, covered, and store it in something clean.
Assuming it covers everything. It does not cover anything dissolved, and it makes some of those worse. Boiling answers one question completely and another one not at all, and which question you are in is the thing to settle first.
Forgetting it needs a heat source. A boil-water notice during a power outage is a common pairing, and an electric range is not a plan. This is the case for holding treated water rather than only a way to treat it, and for a chemical method that needs no fuel at all.