They also held, in July, a three-day conference at Houston -- "Concepts and Approaches for Mars Exploration" -- at which over 150 papers were delivered by scientists and engineers.

Many of those papers simply described individual instruments waiting hopefully for a Mars flight; others described individual mission concepts.

But some tackled the central problem of designing an overall Mars exploration program that makes more sense than NASA's previous rush to a hasty sample-return mission stating in 2003 -- a scheme which would have collapsed because of massive technical and cost problems even if the 1998 Mars Surveyors had both succeeded.

As you might expect, there are some considerable differences in approach. Three scientists from France's CNES agency urged that NASA delay its sample-return mission -- in which France was a partner -- only two years; and some American scientists also pointed out the enormous range of important studies that would be possible on returned Mars samples using multitudes of supersensitive Earth-based instruments, some of which weigh tons while others haven't even been invented yet.

But -- as many other scientists pointed out -- sample return from Mars presents unusual problems. It's possible to carry out very satisfactory studies of a world's geology by parcelling out tiny portions of a returned sample weighing only a few kilograms and collected from only a single spot -- but the most important goal in studying Mars samples is to look for minute traces of fossil (or even still-existing) life.

And since such traces are likely to exist only in very small amounts -- and Mars sample-return missions are always likely to each cost one-half to one billion dollars and return only 1 or 2 kg of samples apiece -- it's extremely important to select only the most promising landing sites possible for them, to maximize the chances that such traces of evidence will turn up in that small quantity of returned Martian material.

Thus, detailed reconnaissance of Mars' surface is extremely important before we even try to send any sample-return missions -- a fact that many scientists were indignantly pointing out even before the 1998 Mars fiascos, complaining that the hasty sample-return missions ordered by NASA Administrator Dan Goldin were very likely to come up empty-handed and produce a Congressional backlash against funding any more such missions.

Some other Conference participants, such as Louis Friedman and Bruce Murray, emphasized the importance of initiating manned Mars exploration as soon as possible.

They urged that the U.S. should get going immediately on development of a large centralized "Mars Outpost" at one or two landing sites, which would serve first as a base of operations for unmanned exploration, and later as a pre-prepared operations base for the first manned landing on Mars.

But this idea also has serious problems. It assumes that we'll be able to decide quite quickly on the best possible scientific site for a manned expedition -- a choice which, as I said, is much more likely to require very detailed and sweeping unmanned reconnaissance of all of Mars' surface.

And it also takes for granted that the U.S. will be willing to fund something as hugely expensive as a manned Mars program within the next two decades or so -- which is open to very serious doubt.

It's very unlikely that such a mission would be approved unless previous unmanned probes found proof of past or present life on Mars -- but if such evidence is found, any manned landing at the place where it was found would seriously contaminate the very phenomena it was intended to study, since manned ships cannot be sterilized and would spread Earth germs and organic debris over the immediate area.

(It may very well be that Mars will be declared, by international treaty, to be a sort of nature preserve -- with only sterilized unmanned craft allowed to land there throughout most of this century, although they could be radio-controlled by human crews in Mars orbit. Only after this kind of exploration program had thoroughly studied Martian biology would manned landings be allowed on the planet.)

In one of the Conference's solider talks, UCLA's David Paige -- who had just lost the "MVACS" set of instruments which he designed to be the main payload of Mars Polar Lander -- discussed this problem.

As he pointed out, for about a decade before the (possibly premature) announcement of evidence of fossil microbes in Mars meteorite ALH84001 in 1996 and NASA's resulting sudden decision to try for a quick sample return, one agency study after another had emphasized the importance of a "stairstep" approach, consisting of the following:

• PHASE 1 -- GLOBAL RECONNAISSANCE, focusing on the past and present role of water, and the identification of sites for future detailed study.

• PHASE 2 -- IN-SITU EXPLORATION OF PROMISING SITES, focusing on describing their geologic, mineralogic, elemental, and isotopic characteristics, as well as the abundance and distribution of volatile species and organic molecules.

• PHASE 3 -- DEPLOYMENT OF EXOBIOLOGICALLY-FOCUSED EXPERIMENTS, to provide detailed characterizations of the population of organic compounds, and to search for [fossil] biomarkers of formerly living organisms, and extant [still-existing] life.

• PHASE 4 -- ROBOTIC RETURN OF MARTIAN SAMPLES TO EARTH, to improve the characterization of organic compounds, and to verify any evidence for biomarkers and extant life discovered in Phase 3.

• PHASE 5 -- HUMAN MISSION, providing detailed scientific characterization of sites of unusual biologic interest, or sites that are inaccessible to robotic exploration.

A number of attempts were made to create 'leapfrog' architectures in which Phase 4 sample-return missions came directly on the heels of Phase 1 global reconnaissance missions, skipping Phase 2 and Phase 3 altogether.

The end result of this accelerated approach has led to a series of failed mission concepts for the '01 and '03 opportunities that, in total, will probably end up costing the community about four years and...one billion dollars...

"The Mariner 9, Viking and Mars Global Surveyor orbiter data sets will provide a good deal of the global reconnaissance required in Phase 1[but by no means all of it], and the Viking and Pathfinder landers represent just the beginning of the in-situ analysis required in Phase 2... From a scientific and technological standpoint, we are not at Phase 4 yet.

"We don't know where to go on Mars to get the samples we need to answer the life-on-Mars question, nor do we know how to design and build the [unavoidably complex and expensive] vehicles and systems we need to accomplish a successful sample return mission, especially within the current [economic] resources of the Mars program.

"Putting sample return first is an extremely low-payoff strategy that, in most games, would signal naivete, impatience, dishonesty or desperation...

"The notion that the next site we land at must necessarily be THE site that we go to collect the first set of returned samples has got to be discouraged if we are ever going to explore the true [and great] diversity of the planet and its environmental history.

"Right now, we possess the technology and the resources to do a first-rate job of Martian global reconnaissance and in-situ exploration of a wide variety of sites.

"There will always be scientists with laboratories who will advocate that NASA provide them with Mars samples for them to analyze. The fact is, however, that we don't yet have the technology to do this within acceptable levels of cost and risk.

"Those who are anxious to move the program forward toward sample return have more than enough to do in the areas of basic technological development, risk reduction and testing, wrote David Paige."

The vast majority of Mars scientists seem to agree with Paige, and NASA's new Mars program will follow their recommendation. But what should its details be? In the next part of this report, I'll examine that question.