stood with the aid of nuclear theory and laboratory data, but the telltale signatures of radioactive isotopes also are seen in the expanding shell of debris following a supernova explosion.

When it comes to understanding the origin of elements much heaver than iron, however, scientists can reconstruct much of what must have happened, but the astrophysical factory has not been clearly identified. Intermediate-mass elements are made in a neutron-rich environment in which successive neutron captures occur slowly, and neutron-rich nuclei undergo beta decay back to more stable elements. Still heavier nuclei must have been made by a succession of rapid neutron captures, referred to as the r-process. A dense, highly neutron-rich environment must exist for the r-process to occur. Also seen in the abundances are the traces of other mechanisms, including possible evidence of nucleosynthesis induced by neutrinos. The element fluorine, for example, can be made by neutrinos interacting with supernova debris. In fact, it is strongly suspected that supernovae, once again, must be the place where the remaining elements up to uranium are built, but there is no detailed understanding of how the process occurs. Resolving this problem requires observational data from supernova remnants, experimental data from both nuclear physics and neutrino physics, and the ability to make detailed, fully three-dimensional, theoretical calculations of supernova explosions.

To begin with, theoretical models of supernovae are still incomplete. Simply producing a reliable “explosion” (in the computer) has proven to be an enormous challenge. Recently, the importance of convection driven by neutrino heating from the nascent neutron-star core was confirmed by numerical calculations. The key was to do calculations in two dimensions instead of one (convection in one dimension is impossible). However, not until it is possible to do a full three-dimensional calculation with full and complete physics will the combined role of rotation and convection be clear. A full three-dimensional calculation with proper inclusion of neutrino transport will require the terascale computing facilities that are just now being realized. There is reason to hope that such a calculation will distinguish the site of the r-process and at the same time illustrate the properties neutrinos must have to match what is currently known about the elements, resolving with a single stroke two important questions in modern physics.

To make this step in computational prowess, however, theory will call upon experiment to provide solid ground for the r-process. In equal measure, progress will come from measurements in neutrino physics and in nuclear physics. Neutrino oscillations can dramatically alter the synthesis of the elements in a supernova, because the muon and tau neutrinos made in