The alarm now sounds with increasing frequency and urgency: the world could be on the brink of an influenza pandemic sparked by the highly virulent avian flu strain ravaging poultry stocks in Southeast Asia, experts fear.

But can that strain -- known as H5N1 -- actually acquire the ability to spread easily to and among people? And if it can, how likely is that dreaded event to occur?

Early in the new year, U.S. scientists will begin experiments that should provide some answers to those questions. In the process, they hope to learn more about why a virus that nature designed to infect migratory water birds has the astonishing capacity to kill mammal species ranging from house cats to tigers to humans.

The work won´t indicate how soon a pandemic might start. And the findings can´t be taken as a guarantee the virus will evolve as the science predicts.

"Like a lot of science, it´s an imitation of nature,´´ explains Dr. Frank Plummer, scientific director of Canada´s National Microbiology Laboratory.

"It doesn´t replicate exactly what happens. But I think it gives you an idea of the propensity of the H5N1 virus to do this thing.´´

The researchers, from the influenza branch of the Centers for Disease Control in Atlanta, will mate H5N1 and human flu viruses in a process known as reassortment. Viable offspring will be tested in animals thought to be good surrogates for humans, to see if the viruses can infect, can be transmitted easily from infected animals to healthy ones and to note the severity of disease each provokes.

In other words, the CDC researchers will be deliberately engineering viruses of pandemic potential. It´s high-risk but crucial work, the influenza community insists.

"It´s a dangerous experiment,´´ admits Dr. Robert Webster, a world-renowned expert on influenza based at St. Jude Children´s Research Hospital in Memphis, Tenn.

Still, Webster has no doubt the work needs doing. Science must gain a better understanding of the menacing H5N1 virus.

"These experiments are fully justified, knowing what we know,´´ he stresses, using a scatological adjective to describe how scared influenza experts are of H5N1.

"This is the worst virus I´ve ever met in my long career.´´

The World Health Organization has been pleading for months for qualified research facilities -- of which there are few -- to undertake this work.

The Geneva-based agency would like to be able to put some kind of odds on how likely H5N1 is to become a pandemic strain and how deadly -- or not -- H5N1 reassortment viruses might be in humans.

If none of the hybrids cause severe disease, the organization might feel comfortable with stepping down its current high level of alert, explains Dr. Klaus Stohr, director of the WHO´s global influenza program.

On the other hand, if CDC researchers easily produce highly transmissible and lethal hybrids, "that would certainly add to our concern.´´

It´s about quantifying risk.

"Currently we do risk-speculation, but we want to do risk-assessment. And that will give us scientific evidence on the possible outcome on the emergence of a pandemic virus,´´ Stohr says.

"It will give us an opportunity to predict the probability because we will have an understanding on the number of reassortment viruses which are viable, the percentage of those that are viable which are then transmissible -- and also on the percentage of those which are viable, transmissible and pathogenic. And how pathogenic they are.´´

The CDC researchers will work in high containment level 3-enhanced labs, says Dr. Nancy Cox, the leading influenza authority who heads the agency´s flu branch. The labs have special features designed to protect both the health of the workers and the world against a viral escape.

"It´s not typical level 3,´´ Cox says, adding the lab workers are "extremely cognizant of the danger of the H5 viruses.´´

"And of course we know we can autoclave these viruses,´´ she adds, referring to a process that kills using high pressure and heat. "So if you find something that is particularly worrisome, you can get rid of it.´´

Still, lab accidents happen. Since SARS was contained in mid-2003, four lab workers in Asia have infected themselves with the potentially deadly disease and one spread it beyond the walls of the laboratory. Earlier this year, a Russian lab worker died after accidentally infecting herself with Ebola virus.

Given that reality, some groups have raised concern that work like this should not be done. But a top virologist from the Netherlands insists it must be, expressing confidence in the CDC´s ability to do the job safely.

"If it´s being done by CDC, then the good thing about that is that safety measures will properly be in place,´´ said Dr. Albert Osterhaus, of the Erasmus Medical Centre in Rotterdam.

The CDC researchers will mix genes from H5N1 with genes from circulating strains of human flu to see which combinations produce viruses that grow and infect. Cox says H3N2 -- the strain that has been responsible for most human flu in recent years -- will be the first priority.

However, Southeast Asia has recently seen activity with H1N1, the mild modern descendant of the strain that caused the Spanish flu of 1918. So the CDC has been sequencing H1N1 viruses recently isolated in Thailand to be able to work with it as well.

"We´re really keeping our minds open. But as long as H3N2 viruses continue to predominate, that will be our first target,´´ Cox says.

Reassortment studies can be performed two ways, she explains. Scientists can use reverse genetics, a procedure that allows them to custom tailor a virus with a predetermined constellation of genes from each parent virus.

The other alternative is what Cox describes as the classical way -- simultaneously infecting tissue culture with H5N1 and H3N2 viruses and seeing what results. That approach is more time-consuming and laborious, but more closely mimics evolution of influenza viruses in nature.

"We will probably be using a combination of the two different methods. I think there are advantages and disadvantages to each,´´ says Cox, who believes her team may have some preliminary results within a couple of months of beginning the work.

Some skeptics have pointed to the fact H5N1 hasn´t yet reassorted to argue it may not be able to do so outside of the artificial confines of a laboratory.

Webster thinks that optimism is misplaced. He calls H5N1 "one of the most promiscuous viruses that we´ve ever seen. It would amaze me if it wouldn´t mate with a human virus, given the opportunity.´´

CDC researchers have already made hybrid viruses with H5N1, using versions of the virus isolated after it first caused human infections in 1997 in Hong Kong.

"Some gene combinations could be produced and others were not,´´ is all Cox will reveal of that as-yet unpublished work.

Even if, in this new round of experiments, CDC´s researchers fail to produce a single viable hybrid virus, that doesn´t prove H5N1 won´t reassort in nature, experts say.

For one thing, flu viruses are constantly evolving. The viruses that the CDC will work with now -- isolated in 2004 -- are not identical to those from the 1997 outbreak. A combination that fails with the 1997 virus might work with a 2004 virus -- or a 2005 version.

"Things in nature can happen that you can´t replicate in the lab necessarily,´´ stresses Plummer, whose lab is working on producing reassortments of H7N3, the strain behind last spring´s avian flu outbreak in British Columbia.

"Can you be sure that all possible virus reassortments have occurred? No. And fairly minor changes in viruses can make really huge differences in their biologic properties.´´