"Hello? Are you still there?"

Anybody who has experienced a momentary disruption in a cell phone conversation knows that the first task is to verify that the other party is still on the line. Researchers in the pursuit of evidence for extraterrestrial intelligence (SETI) have found themselves on the tantalizing brink of making what appears to be first contact on a number of occasions, only to be unable to verify that they have found a real signal--that is, someone at the other end of the line. (For more on the importance of data in estimating ET civilizations, see "Quantifying ET." For more on other SETI efforts, see "Searching for Life around Other Sunlike Stars." )

Recently the Planetary Society's Megachannel Extraterrestrial Assay (META) found 11 apparent transmissions. Intriguingly, these 11 candidate transmissions tended to be located in or near the plane of the Milky Way's spiral disk, just as one might expect if the transmissions originated from stars in our galaxy [see "Refugees for Life in a Hostile Universe," by Guillermo Gonzalez, Donald Brownlee and Peter D. Ward; Scientific American, October 2001]. In addition, in 1977, as part of its SETI program, the Ohio State University telescope affectionately known as the "Big Ear" picked up what appeared to be a boomingly loud transmission. The operator on duty was so excited, he wrote "Wow!" on the paper record of the observation.

Unfortunately, all attempts to verify these candidates--and others from various SETI programs--have failed. As Carl Sagan wrote, extraordinary claims require extraordinary evidence. Without more evidence than a single detection, SETI researchers are unwilling to cite any candidate as evidence of ET transmissions. Before dismissing these examples too hastily, though, many astronomers--including myself--have suggested that we should also consider whether these might represent real ET transmissions that were not verified because they were corrupted or modified during their journey to Earth.

"Twinkling" Signals

Anybody who has watched the flickering of celestial bodies in the night sky knows that a star can appear to brighten, fade and then brighten again. Suppose that there were ways in which ET transmissions also could brighten and fade. If a possible source of an ET transmission was observed initially when it was bright but then the verification attempts occurred when the signal had faded, one might conclude mistakenly that the ET transmission was not real. Are there ways in which a real ET transmission might brighten and fade?

In fact there are (at least) two ways an ET transmission could brighten and fade. The first is "twinkling." Almost all SETI programs have been conducted in the radio portion of the spectrum. Distant radio sources can twinkle, just like the light from a star seen from Earth's surface. Unlike stars, which brighten and fade because their light passes through Earth's variable and obscuring atmosphere, an ET transmission would change because it would pass through the gas clouds that fill the space between the stars. The typical size of one of these gas clouds is roughly the same as the distance between Earth and the sun. Such twinkling is observed routinely in observations of pulsars. (Pulsars are rapidly rotating neutron stars, the remnants of a supernova explosion of a massive star--stars having masses roughly between 8 and 25 or 50 times the mass of our sun). Also, whereas the twinkling of a star is fairly mild, the twinkling of a distant radio source can be violent. A radio source could double in apparent brightness, then fade to essentially zero and remain faint for a considerable amount of time--hours or even days.

A second possible explanation for ET transmission variability is gravitational microlensing. According to Einstein's theory of general relativity, mass can deflect light. Indeed, one of the early triumphs of general relativity was the observation of the deflection of light by the sun's mass. More recently, stars in the inner part of the Milky Way galaxy have been observed to brighten then fade as massive objects--another star, a white dwarf, a neutron star, a black hole, etc.--pass in front of the background stars, temporarily focusing the light from those background stars.

In 1997 James Cordes, Carl Sagan and I published the first possibility of radio twinkling. We found that the previous attempts to verify the 11 META candidates had not taken sufficient account of this possibility. That is, radio twinkling could have caused a real ET transmission to fade to the point of being undetectable. The META candidates might have represented real transmissions. If that were the case, the original META analysis had shown that--taking into account the fraction of the sky observed in META, the possibility of more distant transmitters, etc.--there could be more than 10,000 ET transmitters in the Milky Way galaxy.

A Disappointment

Between 1997 and 1998 nine of the 11 candidates were reobserved as part of the SETI Institute's Project Phoenix. These were far more sensitive verification attempts than the original META attempts. Jill Tarter, Peter Backus and I have now shown that these new verification attempts were sufficiently sensitive to rule out, effectively, the possibility that the META candidates were transmissions from ETs. More precisely, we were able to show that, even taking radio twinkling into account, we had at least a 97.8 percent chance of detecting the META candidates and probably better than a 99.9 percent chance of detecting them if they were real, intrinsically steady transmitters.

We also considered the gravitational microlensing possibility for the first time. Even at the most optimistic the META candidates imply a galactic ET population not too much larger than 10,000. However, the existing gravitational microlensing observations of stars show that microlensing is extremely unlikely to occur. There were only a small number of META candidates. Yet to explain even this small number as gravitationally lensed ET transmitters requires that the number of transmitters in the galaxy be vast, literally billions and billions.

Noisy Earth

So what were the META candidates? The more mundane, and therefore probably more likely, possibility is that they represent terrestrial transmitters. The META program went to great lengths to try to filter out human-generated transmissions (as have all other radio SETI programs). These efforts may not have been 100 percent successful and some terrestrial transmissions may have slipped through. (Indeed, in the META results contain a similar number of what were thought initially to be ET candidates but later recognized to be terrestrial interference.) In this respect, it is worth noting that the total number of signals analyzed by the META program was in excess of 50 trillion. Only a quite small fraction of terrestrial transmissions would have to slip through the META filters to explain the candidates.

The other, far more speculative possibility is that the META candidates represent real, but intrinsically transient ET transmissions. A terrestrial analog of an intrinsically transient signal is a scientific radar like the one at Arecibo Observatory. As the radar is transmitting to another planet, moon or asteroid, it is sitting on the surface of a rotating planet and moving to track the object being studied. Moreover, the portion of the sky illuminated by the radar is much larger than the object being studied. Any beings in the path of the radar might someday see a brief pulse as the radar beam sweeps past their planet. Similarly, if there are other civilizations with powerful radars, we might occasionally be in the path of their radars. Unfortunately, it is exceedingly difficult to verify that we have detected an ET radar pulse as opposed to a human-generated transmission. Perhaps future radio telescopes will have this capability, though. A couple on the drawing board include the SETI Institute's Allen Telescope Array and the Low Frequency Array project. More distant (perhaps in the next decade) is the Square Kilometer Array.

For now, we must be conservative and conclude that at least nine, and possibly all 11, of the META candidates were not ET transmissions. The original META analysis concluded that if the 11 candidates were real ET transmissions, then there are at least 10,000 ET transmitters in the Milky Way galaxy. Conversely, if none of the ET candidates is real, we must conclude that there can be no more than 10,000 other civilizations (broadcasting on the META frequencies) in the Milky Way galaxy.

Joseph Lazio is a radio astronomer at the Naval Research Laboratory in Washington, D.C. In addition to SETI, his research interests include pulsars and the propagation of radio waves through the interstellar medium. The views expressed here are not necessarily endorsed by the Department of the Navy or of the Department of Defense.