Thomas Eric Duncan had chills and a fever and was using the bathroom often on Sept. 28. They were all symptoms of the Ebola virus that Duncan was eventually diagnosed with at Texas Presbyterian Hospital.

At the time, however, he was staying in a small, two-bedroom apartment in Dallas with four other people.

In the process of cleaning that apartment after Duncan was hospitalized, Dallas officials said that there were three mattresses in the home. And family members reported that Duncan, who died Oct. 8, had contact with all of them.

Yet two-thirds of the way through their quarantine period, all of the people who were living in that apartment -- Duncan's fiancee Louise Troh; her 13-year-old son; Duncan's nephew; and another man -- have shown no symptoms of the Ebola virus. And at this stage in the incubation period, they are unlikely to, Centers for Disease Control and Prevention director Tom Frieden said.

So how could all the people who spent days in close quarters with a man sickened with Ebola manage to evade illness?

One possible explanation that has rarely been discussed is that people could be infected with Ebola without ever showing symptoms. The phenomenon is called "asymptomatic infection," and it is an unusual but potentially very real feature of the Ebola virus in humans.

"Asymptomatic cases are likely to have a little bit of virus for a little bit of time then fight it off," said Steve Bellan, a post-doctoral researcher at the University of Texas at Austin.

Asymptomatic individuals aren't contagious, as Ebola is only transmitted through direct contact with bodily fluids of an infected and symptomatic person. Studies conducted in previous outbreaks have found that between 20 and 60 percent of people in a population can exhibit signs of asymptomatic infection, Bellan said.

These people can come in close contact with the virus and show an immune response in blood tests, but they do not experience typical Ebola symptoms such as fever, diarrhea or vomiting. The virus never progresses to its extreme -- and often deadly -- lengths.

No one knows yet whether people with asymptomatic infections are immune to the virus in some or all cases. But if they are, those people could be recruited to help fight the Ebola epidemic by treating patients.

And the blood of asymptomatic individuals, if they are also immune to the virus, could potentially be used in transfusions to help care for sick patients.

"It's not known whether they have protective immunity," said Bellan. "There's substantial evidence that suggests that there are a lot of people out there who have been asymptomatically infected, and understanding whether they have protective immunity could be very useful."

So far, there has been little or no mention of this phenomenon, because until now, Ebola infections have been relatively rare. Bellan and several other researchers who co-wrote a correspondence published in the Lancet journal this week want to change that by calling for additional research on the subject.

In Bellan's modeling of the current epidemic, his team found evidence that the effect of having a population of people who do not show symptoms even when they are exposed to the disease will start to show over time in Ebola-affected West African countries.

"At the beginning, it doesn't matter," he said. "But eventually what ends up happening is that you get an accumulation of people who are immunized by getting infected but never get sick, that helps bring the infected population down."

However, with the crisis raging in Liberia, Guinea and Sierra Leone with resources still limited, it isn't feasible to test people for Ebola unless they show symptoms. Already, many of whom are treated for the virus or die, can't be tested because of limited laboratory capabilities.

But Bellan said that any vaccine or treatment trial on the ground in West Africa might be the best -- and only -- opportunity to study asymptomatic infections.

"The greatest potential lies in any other studies that might be done with regard to vaccines or transfusions or any targeted interventions to help out people at high risk of contracting the virus," Bellan said. "If you did do a study where you randomly assigned half the people to get vaccinated and the other half do not, in the population that doesn't get vaccinated, you can identify people who didn't get vaccinated but have been asymptomatically infected."

In Dallas, all 48 people who had contact with Duncan after he became ill but before he was hospitalized are still asymptomatic. Bellan believes that researchers in the United States can begin to look at least part of this puzzle there as well.

"In the Dallas case, you can start to get at this symptomatic proportion but you could not figure out whether it's protective," Bellan said.