"Do I dare to eat a peach?" wrote T. S. Eliot in "The Love-Song of J. Alfred Prufrock." The corridors of government are crowded with Prufrocks, imagining dangers on all sides, cautioning against all action. The inability to dare anything is particularly pronounced in the hazy areas connecting science and public policy. Suggestions of risk are easily found in the complex technologies underlying contemporary life. What is not so easy is to make impromptu technological changes on the basis of fleeting fears.

The explosion of the space shuttle Challenger, which stunned the nation ten years ago today and left long tendrils of speculation and second-guessing in its wake, is a particularly twisted tale of the calculation of risk. Most analysis of the event has focused on decisions made the day of the launch: should it have been called off because of cold weather? Who was aware of the risk involved? What information was known and what was suppressed? A new book by the sociologist Diane Vaughan (The Challenger Launch Decision, University of Chicago Press) analyzes the "culture of NASA," attempting to show that the explosion was not a technological problem but a symptom of NASA's style of decision making.

Like many studies of scientific subjects undertaken by non-scientists, this book overlooks some very important facts. The shuttle is a machine, not a social organism, and the technical causes of its failure are readily grasped. The deeper flaws in the shuttle's system of protecting the booster rocket's exterior casing--the O-rings seals and the putty-filled joints which contained the ferociously hot gases produced by combustion of the rocket fuel--are clearly laid out in NASA engineering reports, news reports, and the Report of the Presidential Commission of June 6, 1986. They lead back to a curious point of origin: the 1977 Consumer Products Safety Commission ban on retail asbestos products and the impending ban on all asbestos use by the U.S. Environmental Protection Agency.

Simply put, the Challenger exploded because the putty failed to prevent hot gases from passing through the booster joint and burning the O-rings. This failure was a direct result of a change in the kind of putty used for sealing the joints. For the first nine successful shuttle missions, NASA employed a special asbestos-bearing putty manufactured by the Fuller O'Brien Company of San Francisco. But in the wake of the developing notoriety of asbestos-bearing products and the fear of law suits, Fuller-O'Brien stopped manufacturing the putty that had served the shuttle so well.

Early in 1984, having run through their stock of Fuller-O'Brien putty, NASA engineers turned to another type of asbestos-bearing putty manufactured by Randolph Products of Carlstadt, NJ. The Fuller- O'Brien putty was also used in the Titan 34-D booster rocket joint seals, similar in design to those of the Challenger. However, by 1985 it too had been replaced by a substitute putty. The result of this substitution, following a string of 50 successful Titan launches, was a devastating explosion of the next two Titan rockets, one launched in August of 1985 and the other in April of 1986.

The shuttle engineers became increasingly alarmed about "blowholes" in the Randolph putty and burned O-rings--defects appearing in the boosters used on the 15 space shuttles launched during 1984 and 1985. Gases as hot 6000 degrees Fahrenheit flowed through the blowholes and wrought havoc on the O rings, whose debilitated state showed clearly when the discarded rocket boosters were recovered from the ocean. To quote from an engineering memo of February 28, 1984: "ZCP (type II Randolph zinc chromate putty) failure to provide a thermal barrier can lead to burning both O-rings and subsequent catastrophic failure." The purpose of the putty was clearly explained in NASA review documents. The putty must prevent the hot gases from impinging on the delicate O-rings and also must act as a medium to transmit pressure to the O-rings thus enabling them to seal properly.

In my reexamination of the Fuller-O'Brien and Randolph putties in 1994 it became clear why one failed and the other did not. The Fuller- O'Brien product is very sticky, even at temperatures held for 24 hours at 10 degrees Fahrenheit. It clings tenaciously to the surrounding material: it has something of the same effect as the ooze in the La Brea tar pits, which does not easily let go if you stick your foot onto its surface. The Randolph putty, by contrast, is stiff to the touch. At 10 degrees, it is almost hard: it does not cling. At the near freezing launch temperature it is not surprising that the Randolph putty failed.

The Challenger disaster tells us something more: that measures aimed at lessening risk can actually increase risk, even create risk. The removal of the Fuller-O'Brien product and the hasty substitution of the Randolph putty (which, ironically, also used asbestos) are the single and obvious origin of the tragedy of January 28, 1986. Diane Vaughan suggests that there existed a "culture" in the space shuttle program which led to the Challenger disaster. If indeed there existed a culture of failure then it was part of a larger culture, the regulatory culture devoted to banning products that are perceived to be associated with even small environmental or health risks. The removal of the Fuller-O'Brien putty threw a monkey wrench into the space shuttle program. How many more monkey wrenches are we to throw into our society through faulty risk assessments?

Malcolm Ross is associated with the Science and Environmental Policy Project, Fairfax, VA. He has 41 years experience as a research scientist with the U.S. Geological Survey.