LONDON/DAKAR (Reuters) - A recent sharp drop in new Ebola infections in West Africa is prompting scientists to wonder whether the virus may be silently immunizing some people at the same time as brutally killing their neighbors.

So-called "asymptomatic" Ebola cases - in which someone is exposed to the virus, develops antibodies, but doesn't get sick or suffer symptoms - are hotly disputed among scientists, with some saying their existence is little more than a pipe dream.

Yet if, as some studies suggest, such cases do occur in epidemics of the deadly disease, they may be a key factor in ending outbreaks more swiftly by giving secret protection to those lucky enough to be able to bat the infection away.

"We wonder whether 'herd immunity' is secretly coming up - when you get a critical mass of people who are protected, because if they are asymptomatic they are then immune," Philippe Maughan, senior operations administrator for the humanitarian branch of the European Commission, told Reuters. "The virus may be bumping into people it can't infect any more."

Latest World Health Organization data show new cases of infection in West Africa's unprecedented Ebola epidemic dropping dramatically in Guinea, Sierra Leone and particularly in Liberia.

Most experts are sure the main driver is better control measures reducing direct contact with contagious patients and corpses, but there may also be other factors at work.

So-called herd immunity is a feature of many infectious diseases and can, in some cases, dampen an outbreak if enough people get asymptomatic, or "sub-clinical" cases and acquire protective antibodies. After a while, the virus - be it flu, measles, polio - can't find non-immune people to be its hosts.

But some specialists with wide experience of disease outbreaks are highly skeptical about whether this phenomenon happens in Ebola, or whether it could affect an epidemic.

"There is some suggestion there may be cases that are less severe... and there may even be some that are asymptomatic," said David Heymann, an infectious disease expert and head of global health security at Chatham House.

"But herd immunity is just the wrong term. There could be household immunity developing, but even that is only hypothesis."

Others are more hopeful and are urging researchers in West Africa to seek out and test possible asymptomatic cases with a view to using the secrets of their silent immunity.

Steve Bellan of the University of Texas in the United States argues that if scientists can reliably identify asymptomatic people, they could help with disease-control tasks like caring for patients and conducting burials, reducing the number of non-immune people exposed in these risky jobs.

Bellan points to two studies in particular. One, conducted after an Ebola outbreak in Gabon in 1997, found that 71 percent of "seropositive" people - those with traces of the Ebola virus in their blood - did not have the disease. The other, published in April 2002, found 46 percent of asymptomatic close contacts of patients with Ebola were seropositive.

With the largest Ebola epidemic on record raging through three of Africa's most under-resourced countries, scientists and medics have, understandably, focused all efforts on the sick and dying and not on testing people with no symptoms.

If they do, however, and if they were to find what Bellan and some others suspect, it could prompt a reappraisal of what jolted a relatively sudden downturn in new cases.

Some researchers say they have identified a few cases in the current outbreak with mild symptoms and low concentrations of Ebola virus in the blood. One was a Guinean student who traveled to Senegal and is not known to have infected anyone else, despite having contact with dozens of people.

Ian MacKay, a virologist at Australia's University of Queensland, agrees that possible sub-clinically-acquired immunity is one of many unexplored mysteries of the Ebola virus.

"One thing that this particular outbreak shows is that we really don't know an awful lot about these kinds of infectious diseases," he said. "We tend to think we can answer all the questions, but this is one of those things we may end up being taught by the virus itself."

(By Kate Kelland and Emma Farge; Editing by Peter Graff)