Experiments at the Brookhaven National Laboratory will study collisions bet
ween gold nuclei at unprecedented energies. The concern has been voiced tha
t "strangelets" - hypothetical products of these collisions - may trigger t
he destruction of our planet. We show how naturally occurring heavy-ion col
lisions can be used to derive a safe and stringent upper bound on the risk
incurred in running these experiments. (C) 1999 Published by Elsevier Scien
ce B.V. AII rights reserved.