Abstract
Contemporary economic theories of bargaining depend on aspects of the bargainers' preferences that are difficult to observe. This makes these theories difficult to test in natural environments. It has proved possible, however, to design experiments to test these theories in a controlled, laboratory environment. The results of these experiments reveal shortcomings and incompleteness in the descriptive power of currently available theories of bargaining. However, these results also suggest important regularities in bargaining behavior. Together with recent theoretical developments, these results suggest some directions in which a more descriptively powerful theory of bargaining might be developed.