Most potent is the protection directed at the H protein. Antibodies for H5, the form of Hpresent in H5N1, are rare in people because it is not found in seasonal flu viruses.

But Michael Worobey, an evolutionary virologist at the University of Arizona, and his colleagues have discovered that someone whose first-ever flu infection was with an H1 or H2 flu virus will probably have some protection against avian H5N1.

This effect is due to an outsize "immune imprinting", probably by the flu strain in global circulation in the year after they were born.

Using historical data on seasonal flu strains, Dr Worobey's team estimates that most people born before 1968 may have this type of immune protection against H5N1.

Most of those born subsequently—and especially in the past decade—will not.

Though this immunity may not prevent infection, there is a good chance that it will shape the severity of disease, says Dr Worobey, and "make a big difference in terms of who lives and who dies".

Similar, though less powerful, immunity may arise from previous exposure to the N1 component, as well as from T-cells primed to attack other types of flu virus.

These, too, may mitigate the severity of a new flu pandemic, says Dr Peiris. It is not only the human body that is better prepared to face H5N1 than covid. Pharmaceutical companies are too.

Existing antiviral drugs used for seasonal flu are likely to work against H5N1 because they target a virus's replication mechanism, which is similar across strains.

Vaccine development is also more advanced. Many countries, including America, have a stockpile of H5N1 vaccines as part of their generic pandemic preparedness plans, reserved for those at the highest risk of becoming infected.

Fifteen European countries are also adding the vaccines to their stockpiles.

Finland, for its part, has started offering the jabs to farm workers most at risk, such as those working with poultry, foxes and mink.

As the vaccines are thought to primarily prevent severe disease, rather than infection, American officials have not followed suit. A pandemic, however, will require billions of vaccine doses.

The fastest way to make them would be to use mRNA technology, which uses a section of the virus's genetic material, rather than the whole virus, to provoke an immune response in the host's body.

These can be made in four weeks, compared with the six months needed for the typical seasonal flu jabs.

The American government has commissioned Moderna, a pharmaceutical company, to develop an mRNA vaccine for H5N1.

GlaxoSmithKline and CureVac, both biopharmaceutical companies, are also in the early stages of developing such a vaccine.

Drew Weissman, who won a Nobel prize in 2023 for developing mRNA technology, is testing a rival mRNA vaccine with colleagues at the University of Pennsylvania.

Dr Weissman also hopes to develop an mRNA vaccine that can be administered to animals in the form of an oral tablet or patch.

This would simplify mass vaccination of poultry and cattle, preventing further spillovers.

At the moment, no H5N1 vaccine has been approved for animal use in America or Europe.

All this vaccine development will take time, as will building up the stockpiles needed to prepare for a large outbreak in humans. Such time would be well-spent.

Should H5N1 become a human pandemic, it is possible—assuming vaccine supplies ran low or existing stocks could not be speedily deployed—that it could kill as many people as covid did.

The lack of aggressive measures to halt the outbreak in cows in America is exacerbating matters.

The risk, says Dr Peiris, is that the virus will become permanent and eventually spread beyond the country. "Then," he adds, "there is no stopping it."