Experiments such as these helped pave the way for current satellite missions like ROSAT and CGRO. With a reduced number of satellite opportunities and the emphasis on reducing the size and cost of these missions, reliance on a strong balloon program becomes even more important. For example, gamma-ray instruments tend to be relatively heavy to attain a high efficiency to image weak sources. Consequently, unless the satellite version of the experiment can be conducted on a MIDEX or smaller mission, some vital X-ray or gamma-ray observations may never be conducted from satellites because of the mass constraint. There are numerous astrophysics experiments that can be inexpensively and quickly carried out on heavy lift balloons. Experience with operations in the Southern Hemisphere with long-duration balloon flights of two- to four-week duration clearly demonstrates the capabilities of the balloon platform for long flights. High-energy astrophysics would benefit greatly from further development of Southern Hemisphere long duration flight balloon operations and extension of this capability to the Northern Hemisphere.

The following specific points should be considered in assessing the value of long duration ballooning to high-energy astrophysics.:

• The balloon program is responsive to user needs because it is efficient, low in cost, and is an excellent training ground for young scientists and engineers that are needed for the long-term future of gamma-ray astronomy.

• The suborbital balloon program may be the only way to perform some important astrophysics experiments in the next 20 years (see comments above).

• Free-flying spacecraft are the preferred platform for gamma-ray astronomy if there is no weight or cost restriction. However, because of the restrictions, the balloon program will be needed to verify instrument concepts in the near-space environment and to perform some new frontier science.

• For the foreseeable future there are no alternatives to the balloon program.

3.7 THEORY

As the sensitivity and resolution of NASA's astronomy missions improve, so too must the realism of the theory used to interpret the results and give them meaning. Progress in theory accompanies the progress in experiment. One cannot lead the other for long. Theory is both interpretive and predictive. On the one hand, for data that are relatively well understood, theory builds models to extract the greatest amount of information possible from them.  From these models emerge new predictions that can be tested and improved until the phenomenon can be satisfactorily understood. On the other hand, measurements may uncover surprises that generate great controversy and, if properly understood, offer potential for scientific advancement. It is such phenomena that pose the greatest challenge to theory and require it the most. Many examples of each category could be given. We present just two: nucleosynthesis gamma-ray lines and gamma-ray bursts.

It is widely accepted that the elements heavier than helium are made in stars with supernovae playing a major role. It is also well documented by measurement and understood in theory that the short time scales and high temperatures of supernovae lead to the creation of short and intermediate lived radioactive isotopes. The species ²⁶Al, ⁴⁴Ti, ⁵⁶Co, and ⁵⁷Co have been detected and studied. The role of theory is to obtain quantitative agreement between physical models and the line measurements in terms of flux, line shapes, and angular distributions. This then leads to constraints on models of stellar evolution and supernovae and a better understanding of their nature; improved accuracy in our models for galactic chemical evolution; and a better depiction of massive star formation in our galaxy. Based upon these improved models, theory makes predictions that can be confirmed by subsequent measurements, e.g., a detectable signal from ⁶⁰Fe.

Gamma-ray bursts represent science of a different sort. Although the puzzle has made good headlines for a decade or two, ultimately the goal of science is understanding. Here the theorist is much less constrained, but also highly challenged by an unexpected and poorly understood phenomenon. Quick, qualitative speculations are useful for a time, but ultimately progress requires that speculation be backed up by detailed physical analysis and simulations. Frequently these simulations lead to the death of the model, but that too is progress. Eventually a model, or set of models, will be found that explains what is observed and makes predictions that can be confirmed. Meanwhile theory guides observations in defining the sensitivity of future missions required to see bursts from the halo from Andromeda, for example, or whether X-ray absorption lines should be visible in the spectra of cosmological gamma-ray bursts. Three dimensional general relativistic calculations of neutron star merger can show whether or not relativistic beams can emerge. Calculations of planetesimal accretion on neutron stars in the halo reveal that tidal disruption will likely prevent the intact arrival of an object at the neutron star. Some models predict a large number of hard X-ray bursts for every gamma-ray burst; others predict that enduring hard GeV emission should be a common characteristic, etc. All these predictions can and must be refined and eventually tested.

New observations will drive theory as they always do. It is important however that NASA continue to provide support to theory particularly during these times of constrained budgets. For a comparatively modest investment, NASA ensures the existence of a cadre of trained specialists interested in making the most of the valuable data. Without the activities of these people, the value of the data is greatly diminished.