News Archive : 1999

The Case of NEO 1999AN10

Following is a file of information on the near-Earth asteroid called 1999 AN10, discovered by the MIT-USAF LINEAR telescope on 13 January 1999. Dynamicists Andrea Milani, Steven Chesley, and Giovanni Valsecchi carried out an analysis of its orbit, which involves resonances with the Earth and permits close encounters with the Earth over the next several hundred years.

On 26 March 1999, these authors requested several colleagues to look at their manuscript and check the general validity of their calculations of the orbit of this asteroid. They wrote, in part: "The subject of this paper is such that we consider essential that its content be reviewed by the most qualified experts before it is made public. This paper has been submitted to a scientific journal. We do not want the content of this paper to reach the non-scientific media until it has been carefully reviewed. . . . Note that it would be unwise to hurry with a public announcement for three good reasons. First, we have established that there is no risk of impact until 2039, and even then the probability of impact is well below the background level. Second, the asteroid is now almost impossible to observe, and even if it were observed new astrometric positions taken now would not contribute significantly to the improvement of the orbit. Third, the issues raised by this case are indeed very complex. . . Please note that we had no obligation to submit our paper to this highly unusual refereeing procedure: we felt this as a moral obligation. We are asking you to carefully examine our paper looking for every possible fault in our arguments, but with respect for our work and for our scientific priority. . . . We intend to make the paper available on our web server on April 6 unless some of you can point out to some reason not to. Thus you should send us your comments, criticisms, and whatever queries you have, as soon as possible. In particular if there is some fundamental flaw in our arguments we would like to know before making any information publicly available."

Several of the colleagues they addressed responded with detailed technical commentary, but none disagreed with the basic conclusion that this asteroid poses no significant threat of Earth impact for at least the next 40 years. Thus, Milani and his co-authors posted the manuscript on their website early on 6 April, as they had indicated they would do. About a week later the manuscript was circulated to a larger group of experts at the request of the International Astronomical Union. These informal technical referees also agreed with the conclusions concerning the exceedingly low probability of an impact with Earth.

Subsequently to this Web posting, the case of asteroid 1999 AN10 became widely known and has stimulated considerable discussion on the Internet and in the international press. The remainder of this message reproduces some of the commentary related to this asteroid and the mode of release of information. Everything that appears here has already been made public on other websites. The material is drawn together here as a reference on a subject of general public interest dealing with the probability of asteroid impact and of the best way such information should be made available to the public.

David Morrison (19 April 1999)



Andrea Milani, Steven R. Chesley Dipartimento di Matematica, Universitá di Pisa Via Buonarroti 2 56127 PISA, ITALY Giovanni B. Valsecchi IAS-Planetologia Area di ricerca CNR Via Fosso del Cavaliere 00133 ROMA, ITALY March 26, 1999

Abstract: The Earth passes very close to the orbit of the asteroid 1999 AN10 twice per year, but whether or not this asteroid can have a close approach depends upon the timing of its passage across the ecliptic plane. The uncertainty of this timing grows with time: by 2027 it is +/- 12 days. Among the possible orbital solutions there are some that undergo a close approach in August 2027, but no impact is possible. However, the period of the asteroid may be perturbed in such a way that it returns to an approach to the Earth at either of the possible encounter points. We have developed a theory which successfully predicts the 25 possible such returns up to 2040. We have also identified 6 more close approaches resulting from the cascade of successive returns. None of these encounters can result in an impact, except one in August 2039: the probability that the true asteroid actually follows a collision course for that date is less than the probability of being hit by an undiscovered asteroid within any given day. Because of this extremely chaotic behaviour there is no way to predict all possible approaches for more than a few decades after any close encounter, but the orbit will remain dangerously close to the orbit of the Earth for about 600 years


From the IAU website The International Astronomical Union Working Group on Near Earth Objects (WG NEO) provides, as a service to the international astronomical community, voluntary expert review of reports that might have implications for possible future Earth impacts. The review process was first used in April 1999 in the case of newly discovered mile-wide asteroid 1999 AN10. NEOs with orbits that permit close encounters or even collisions with the Earth are of considerable interest to scientists who compute asteroid orbits. As a consequence of their frequent close encounters with the Earth or other planets, it is difficult to predict their orbits with high precision for more than a century or so into the future. One such object is 1999 AN10, discovered by the MIT-USAF LINEAR search program on 13 January 1999. A detailed analysis of the orbit of 1999 AN10 was completed by researchers Andrea Milani, Steven R. Chesley and Giovanni B. Valsecchi in March 1999. Their paper, which has been submitted for publication in a technical journal, includes an examination of the potential risk of 1999 AN10 hitting the Earth in the next several decades. They conclude that, while there is some uncertainty in the exact orbit of this NEO following its next close planetary encounter in August 2027, the chances of its actually hitting the Earth in the next 40 years are minuscule -- the authors estimate that the chance of impact is of order 1 in a billion (1 in a thousand million), which they indicate is 10,000 times less than the chance that the Earth will be struck by some as-yet-undiscovered similar-sized NEO in any one year. The IAU's Working Group on Near-Earth Objects has formed an ad hoc committee, with widely international expert membership, whose members are available to review predictions of impact hazards if so requested. This committee functions similarly to the referees of most technical journal articles in reviewing the predictions, and it also keeps the appropriate IAU officials completely informed about any such predictions. The technical paper by Milani and colleagues has been subject to such informal review during the first two weeks of April 1999, and it is the consensus of the reviewers that the work is accurate and of the highest scientific quality. The IAU reviewers also note that the chances of impact by NEO 1999 AN10 during the time-span considered in this paper are negligible compared to the risks we run continuously of being struck by one of man similar size NEOs that have not yet been discovered. Like asteroid 1997 XF11, which was widely discussed in the press in March 1998, this asteroid does not pose any significant danger to the Earth on the time scale of the next several decades. Astronomers will continue to search for new NEOs and to track the orbits of those already discovered, especially when, like 1999 AN10, their orbits bring them close to the Earth. But this object, as demonstrated in the technical paper by Milani and his colleagues, should not evoke any particular public concern. Thus, the reviewers from the WG NEO agree with the authors in ruling out any danger to Earth from 1999 AN10 in the next forty years. The object will be followed closely over the next several years in order to define the longer-term properties of its orbit more accurately, as will be the case with numerous other, similar objects that will be continue to be discovered over the next several years as NEO searches intensify and orbital computation methods improve.


Imagine a newly discovered asteroid, some one mile in diameter, is on a potential collision course with Earth in just 40 years - and no one is telling you about it. This is exactly what is happening with asteroid 1999 AN10. By pure coincidence, I have come across a research paper by Andrea Milani, Steven R. Chesley and Giovanni B. Valsecchi on the potential risk of 1999 AN10 hitting the Earth in forty years time. Yet instead of informing the interested public about their potentially explosive findings, the authors have hidden away their results on an obscure web page. The asteroid, known as 1999 AN10, was discovered by LINEAR on 13 January 1999. According to the Italian researchers, the object will come particularly close to Earth in August 2027. No impact is possible in that year, but there is a small chance that the asteroid will be perturbed in such a way that it might impact the Earth in 2039. While the chance of an actual collision is small, one is not entirely out of the question. Moreover, the extremely chaotic behaviour of this asteroid makes it impossible to predict all possible approaches for more than a few decades after any close encounter, but the orbit will remain dangerously close to the orbit of the Earth for about 600 years. If this information reminds you of the 1997 XF11 affair, you are spot on. It is in fact only the second time in history that a major impact in the near future cannot be ruled out altogether. And yet there is at one major difference: At least we were informed about 1997 XF11 once a potential hazard became clear. In the case of 1999 AN10, however, it is pure accident that you hear about the information via the CCNet rather than through an official press release. Now, what is really worrying about 1999 AN10 is not the statistically very small impact risk. Nobody needs to lose any sleep due to this object. What is really disturbing, however, is the unnecessary and detrimental secrecy that surrounds this object. There is no reason whatsoever why the findings about 1999 AN10 should not be made available to the general public - unless the findings haven't been checked for general accuracy by other NEO researchers. If, however, no such independent assessment has taken place, the data should not be in the public domain in the first place. Of course, one reason why the authors may have decided to hide their data could be due to the current NASA guidelines on the reporting of impact probabilities by individual NEOs. After all, NASA is threatening researchers with the withdrawal of funding if they dare to publish such sensitive information in any other form than in a peer reviewed medium. Obviously, one's own web site can hardly be considered a peer reviewed journal. One therefore has to wonder why such relevant information is put into the public domain in such a wired and secretive way. The 1999 AN10 'affair', in my view, should be seen as a rather damaging consequence of the over-reaction regarding asteroid 1997 XF11. Moreover, I would argue that the unclear and intimidating NASA guidelines on NEO reporting should be dropped in their present form since they have become part of the problem. Instead, international procedures (which would acknowledge a certain level of scientific uncertainty regarding some particular PHAs) should replace those ill-considered NASA guidelines which were imposed in a rash last year. In order not to repeat last year's mistakes, the discussion should be focused on an international procedure of how future impact risk calculations (and their inherent uncertainties!) should be reported in a satisfactory way.


Andrea Milani, Steven R. Chesley, Giovanni B. Valsecchi Dear Benny- We strenuously object to your characterization of our actions regarding this paper, and to your attempts to sensationalize our work. This whole thing could have been explained easily if you had contacted one of us, but that apparently does not suit your purposes. Instead you released a uninformed report filled with speculation and innuendo. We have submitted the paper in question to a journal, so it is undergoing the usual peer review process even now. It is customary for researchers to make available, through a variety of means, papers which are at all stages of the publication and review process, even work still in preparation. This fosters discussion among the community, and is an essential component of the modern scientific method. In addition to submitting to a journal we have voluntarily sent the paper to more than a dozen international experts and officials for comment and criticism before we made the article available on the Internet. This review has been going on for more than two weeks now, and in fact some technical issues that we raise are still under discussion, but, fundamentally, the content of our report has been well received. (We posted the information to the Internet on April 6, about one week after selectively releasing the paper.) This additional level of review was done voluntarily on our part as we did not want to make available erroneous or misleading information, which many perceive was done in the 1997 XF11 scare. We also wanted to be in a position to issue a controlled release to the public after all issues have been confirmed by our independently convened panel of experts, should that be deemed necessary. In the end a consensus among the experts was quickly reached that this object does not fit any realistic criteria of imminent danger, so we decided to follow normal channels with the paper. This entails submitting the article to a journal and posting to the Internet as a preprint. More than likely, you stumbled upon the paper via Andrea's preprint page (hardly an "obscure web page"): You should note that some of the other papers listed there are also currently under review for publication by scientific journals. Indeed the paper in question is specified as "submitted." It is not our custom to contact the media every time we write a paper, nor is it customary to treat papers undergoing the peer review process as secret. Imagine a newly discovered asteroid, some one mile in diameter, is on a potential collision course with Earth in just 40 years - and no one is telling you about it. This is exactly what is happening with asteroid 1999 AN10. Your opening paragraph (above) clearly indicates that you are trying to spark fear and controversy where none is warranted. You later call our results potentially explosive, but then go on to say that the risk of collision is small enough to be considered negligible. So which is it? This is the fundamental flaw in your claim that we had some obligation to broadcast this work as widely as possible to the general public. Either it is of urgent concern, or it is not. Your claim that "unnecessary and detrimental secrecy" surrounds this object is based on an assumption that this paper presents time critical information which morally obligates us to notify the public, the press, the United Nations, and the commander of the Enterprise. This is clearly not the case here, and so your argument is invalid, even paradoxical. If the important thing is the lack of a press release rather than the risk of collision, as you stated, then there was nothing to release outside of normal scientific channels. You make the point that this case has been handled differently from the XF11 affair, and you are right. In the XF11 scare, a possibility of collision in 2028 was announced, when such possibility did not exist, even based on the then available information. 1999 AN10 is qualitatively similar except that in the cases where there is a non-negligible chance of a very deep encounter (2027 and 2034), we have explicitly stated that collision is not possible. The risk in 2039 is, of course, negligible. Your speculation that we have decided to hide this report for fear of losing our NASA funding is demonstrably false for two reasons. First we have no financial support from NASA, and second, if we were hiding our results you certainly would not have found them published on the Internet. Furthermore, to our knowledge, this proposal was only a proposal, and it was never implemented. We have scrupulously followed normal conventions for the release of scientific data. Aware that this report could be sensationalized, we submitted it to review by the panel of experts. This was done voluntarily, but we also hope to set a precedent with this action. As a result of this case the IAU is moving rapidly to formally establish voluntary guidelines and procedures to be followed in future cases. Their plans are closely modeled on our approach. Under this policy, researchers would submit their results to an ad hoc committee of experts for comment and criticism before going public. How long this delay should last is unclear, but probably would depend on the urgency of the situation, anywhere from 2-5 days. After that period the author could release the information in any manner deemed appropriate, and the officials and agencies first confronted by the press will be able to respond with an informed discussion of the threat. This responsibility to seek confirmation before going public becomes even more critical for more threatening situations. You may object to this policy, but we expect that in the future virtually all researchers in the Earth hazard community will be following it. We are sorry to report that the reason these voluntary guidelines are necessary is to thwart those in the press who would seek to sensationalize reports of potential impact, no matter how carefully we word them. In that sense you are part of the problem, Benny, and a careful and deliberate release of information is only a response to irresponsible actions such as yours. 1999 AN10 itself holds very little relevance to the general public. If our research does have any relevance for the public at large, it is because we have developed a general theory that can rule out impact for some finite period of time, yet it also shows that we can say very little about the possibility of collision for times beyond that point, because each encounter predicted by our theory can spawn more close encounters, a cascade too complex to be analyzed. The good news is that the further down the cascade we go, the lower the probability of impact should a collision solution exist. The essential point is that we feel that this paper should not be used to confuse the general public, and we strenuously object to your accusations that the information was handled irresponsibly. You will have a very hard time to find a scientist who will accuse us of a lack of openness in our research. An important point: We want to avoid the perception of crying wolf when we say in April that a collision is possible, while in July, after more observations become available, we will almost certainly report that a collision is no longer possible. We have verified that the object cannot hit the Earth in the next 40 years, far longer than any threat mitigation would require. And in a few months, when there is less uncertainty in the orbit, the picture will be very different from the one we have now.


The "1999 AN10 Affair" is nothing more than the scientific peer review process at work. The authors have asked scientific colleagues to examine and verify their results prior to issuing any IAU Circular or Press Information Sheet. Better still, the authors intend to publish their results in the refereed literature. The authors are to be applauded for doing it right, that is, they are making sure their results are correct before making any public announcement, and they will provide their full analysis for scrutiny within the professional literature. The timescales involved require no immediate action, hence the weeks (or even months) required for the scientific review process to proceed is of no consequence. Furthermore, there is no reason why this object should merit any extraordinary public attention as the probability falls below that for "undiscovered" objects out there. Richard P. Binzel Professor of Planetary Science Massachusetts Institute of Technology


I regret the misleading wording you have chosen to use (both in your heading and in your text) in announcing the Milani study of 1999 AN10 to CCNet subscribers. You have written that "the chance of an actual collision is small, but one is not entirely out of the question." In the context of the impact hazard, those were the words used 13 months ago to describe a very different situation. When used about 1997 XF11 in an official IAU statement, they implied -- to both the writer, who had a mistaken concept of the distribution of errors (see IAU Circular 6879), and to scientifically literate readers (see Stuart Goldman's sidebar on pg. 33, June 1998 SKY & TELESCOPE) -- a probability of impact of order 1 in 1000. Milani estimates that the chance of impact by 1999 AN10 is of order 1 in a billion (1 in a thousand million), a factor of a million smaller than the mistaken estimate of 1997 XF11's impact probability. He estimates the impact probability as 10,000 times less than the chance that the Earth will be struck by some as-yet-undiscovered kilometer sized object NEXT YEAR! That makes 1999 AN10 a matter of scientific interest, but of no practical interest and hardly meriting the "official press release" you call for. (Unless, of course, you are the sort of person who worries about being killed by snakes while you drive around town chain-smoking and not wearing a seat-belt.) Use of the same wording to describe probabilities that differ by a factor of a million can only serve to confuse a literate, rational understanding of risk. You obviously used this wording in an ironic, argumentative way, rather than as an attempt to confuse. But the point needs to be made, if society is to address risks in a rational way, that the quantitative difference of a factor of a million makes for an *enormous* qualitative difference. That is why the XF11 announcement would have *deserved* the world-wide headlines, had it been true, while the AN10 matter has no relevance to the "man-on-the-street" whatsoever. Andrea Milani has been very responsible in having other experts check his work before posting his results on his public web site. There is no rational reason, however, for Milani to have called a press conference, or offered his results to CCNet, and risked an unwarranted sensation (like you seem to be trying to provoke) from a misunderstanding of his result, which is irrelevant in the larger scheme of things. Such an announcement would attach undue importance to an arcane result. People like yourself, who are interested in the impact hazard, have long been aware of Milani's website, so his work was hardly "hidden" away. NASA is holding no gun to the head of a researcher at an Italian university, so your remarks about NASA intimidation seem to be off-track. There is room, of course, for legitimate debate about how to handle matters of potential practical interest in a responsible way, so that the public isn't misled by faulty, premature results but yet *is* told about potentially important matters in a timely fashion. This is a dilemma long-faced by emergency preparedness officials in communities and nations around the world who have tried to establish responsible protocols, but there are no easy answers. Your apparent belief, Benny, that every infinitesimal threat needs to be announced in a press release seems to me to be a step in the wrong direction. Clark R. Chapman Southwest Research Institute


The current debate about Asteroid 1999 AN10 amongst astronomers and others in the scientific community is mainly concerned with the manner in which information about potentially hazardous asteroids is released to the public. This issue has received considerable attention since April 1998 when a "false alarm" was raised about another asteroid (1997 XF11). A working group of the International Astronomical Union is apparently preparing guidelines for announcement of possible impacts and the 1999 AN10 "incident" should help that group review its work. As a nonscientist who is trying, on a voluntary basis, to get a major NEO search effort re-established in Australia, I would like to offer some comments on this issue. Firstly, I believe the common goal is to protect the Earth from the consequences of impacts by asteroids and comets. This involves: a) detecting near earth objects b) establishing the orbital parameters of newly detected objects c) predicting future orbits d) identifying potential collisions with the Earth e) verifying predictions of collisions to a high degree of accuracy f) implementing measures to avert or mitigate a collision (including tsunami effects) Items (a) to (d) are relatively low-cost, "routine" activities that are well described in the Spaceguard Proposal on NASA's Web site. The entire cost of a ten year, worldwide Spaceguard Survey is about US$100 million -- apparently equivalent to US military expenditure for just two days in the Balkan Conflict! Despite the low cost of this "insurance for mankind," efforts to introduce a worldwide Spaceguard Survey appear to have stalled -- advisors to government don't seem to take the issue seriously (in fact, I have questioned whether Spaceguard is too cheap for its own good). Several groups around the world have been working on this by lobbying politicians and key scientists. "False alarms" don't help this effort and also unfairly undermine the credibility of scientists working in the field. It turns out that the authors of the paper describing the potential hazard of Asteroid 1999 AN10 had submitted it for peer review prior to making a "preprint" available on the Internet. This is entirely appropriate -- the difficulty is deciding at what stage the media should be informed and how such information should be worded in order to not raise undue alarm. Brian Marsden from the Minor Planet Centre has pointed out that 1999 AN10 was added to the list of "Potentially Hazardous Asteroids" on February 16, well before the Italian paper was posted on the Internet. There are media relations precedents in other fields such as the release of economic indicators -- everyone knows that the statistics are analysed by competent people over several weeks but that public release of the results will take place on a certain day. Another (perhaps more relevant) analogy is where a medical doctor detects a potentially cancerous growth on a patient and sends a sample away for pathology tests. The patient is told that it will take a certain time (usually days) before the results of the tests are available. My suggestion is that there be a similar "official" delay in the announcement of the results of NEO impact assessments. People would know that sufficient information had been gathered to enable orbit calculations to be undertaken, that one or more groups were performing these calculations and that an announcement of the results be issued (probably by MPC) on a certain day (say two months after its inclusion on the PHA list). As several people in the NEO field have pointed out, there really is no need for urgency in the release of these results. A related issue is the need for consistency in terminology amongst spokesperson scientists when dealing with the press. Terms such as "potentially hazardous", "especially dangerous" (Brian Marsden's suggestion), "possible impact" and expressions of probability need to be clearly defined and some poorly understood terms need to be avoided altogether. Richard Binzel's suggestion for a scale of impact hazards, similar to the Richter Scale for earthquakes, has merit. Of course, anyone commenting on an impact hazard issue should bear in mind the likelihood that some sections of the media will sensationalise the story. By Michael Paine, News South Wales Coordinator, The Planetary Society Australian Volunteers


By David L. Chandler, Globe Staff, 04/14/99- In a discovery eerily reminiscent of one made just a year ago, astronomers have found an asteroid that will come quite close to Earth in a few decades, and that even has a real but minuscule possibility of an impact. Last year, astronomers made a similar discovery of an asteroid that they said had a slight possibility of hitting Earth in about 40 years. In that case, it was quickly determined that the Earth was safe after all, and astronomers have been arguing ever since about the way the original report was disseminated. The latest asteroid, called 1999 AN10, was described in a detailed scientific paper posted on a Web site by three astronomers. But unlike last year's case, no information has been sent directly to the public and the press. The asteroid is thought to be about a mile in diameter - similar to the one last year - and could possibly come very close to Earth in 2039. There is about a one-in-a-billion chance that it could strike the Earth that year, with devastating consequences. That is less than the risk that an unknown asteroid or comet might hit Earth on any given day, and therefore is not anything to be too concerned about. What might be more worrisome, scientists said, is its long-term potential. For the next 600 years, according to astronomers Andrea Milani, Steven Chesley, and Giovanni Valsecchi, the asteroid could remain very close to Earth, and if it comes close enough to be affected by Earth's gravity its orbit could become chaotic and impossible to predict for more than a decade or two ahead. In that case, the asteroid would require constant, careful monitoring for centuries to guard against a possible impact. This is only the second time in history - or perhaps the first time, depending on whose analysis of last year's discovery you believe - that an asteroid has been discovered that has a small but non-zero possibility of striking the Earth within a few decades. Even though the likelihood is quite small, that makes it an interesting find. Some people have questioned the wisdom of the changes in the way such information is disseminated as a result of what many astronomers considered a serious public embarrassment last year. British anthropologist Benny Peiser, who has written extensively about the effects of past impacts on the Earth, yesterday circulated an e-mail message questioning why this discovery, unlike last year's, has not been announced publicly or shared with news organizations. While acknowledging that the risk of the asteroid hitting the Earth is tiny and not something that anyone should lose any sleep over, Peiser said that he found the lack of public notice disturbing, suggesting that it reflects an overreaction to astronomers' embarrassment about last year's announcement followed by a swift reversal. But astronomers contacted yesterday said that they see this latest case as a perfect example of how such information should be handled. The astronomers who made the calculations of impact probabilities have circulated their unpublished paper to several colleagues around the world who specialize in such calculations, in order to make sure their conclusions are correct. Some astronomers contacted yesterday said that all the comments so far from such specialists have been positive. "I commend them for the process of being careful," said Richard Binzel, an astronomer at Massachusetts Institute of Technology who specializes in asteroids.


Comments on Potentially Hazardous Asteroid 1999 AN10

Asteroid 1999 AN10 made the news recently because, according to a group of researchers in Italy, there is a remote possibility that it could collide with the Earth in the year 2039. Writing in a scientific paper submitted for publication, researchers Andrea Milani, Steven R. Chesley and Giovanni B. Valsecchi say that the chance of a collision in 2039 is exceedingly small, only about one in a billion, but they add that the asteroid's orbit will remain threateningly close to the Earth's orbit for many centuries to come.

Although the threat posed by 1999 AN10 must certainly be taken seriously, the probability of impact for this object is so miniscule that the authors of the paper felt no great urgency to inform the press of the new calculations, and the other NEO scientists reviewing the paper agreed with this policy. To put it into perspective, consider that the probability of 1999 AN10 impacting in 2039 is tens of thousands of times *less* than the probability of an undiscovered asteroid of equivalent size hitting the Earth during the same 40-year period. Furthermore, in just a few months, 1999 AN10 will be observed again, as it moves back into the nighttime sky, and the new data will, in all likelihood, completely eliminate the possibility of impact in 2039. Researchers should then be able to start examining the possibility of impacts after 2039.

As it turned out, the Milani et al. paper was publicized not by the authors, but by a third party who found it accidentally on one of the author's web pages; the authors were not even consulted before their results were publicized. An internet debate ensued on such issues as why the results had not been made public, and whether or not the paper had been peer-reviewed to ensure accuracy. The reasons for not making the results public have already been described: basically, there was no great urgency to publicize a one-in-a-billion-chance impact 40 years from now, when even that remote a possibility will likely disappear in a few months.

On the issue of peer review, Milani and his colleagues followed a commendable course. The authors distributed their paper to qualified experts more than a week before placing the paper on their web page, seeking confirmation of their results. Our group at JPL examined the paper and saw no major flaws. We have also confirmed the existence of the impact scenario for 2039, and we confirm that the probability of impact in 2039 is about one in a billion. Paul W. Chodas NEO Program Office April 21, 1999