In many cases cracks leading to fracture occur at interfaces between two different constituents, e.g. a fiber and the matrix in a composite. The early solutions of such problems in the context of linear elastic fracture mechanics (LEFM) revealed the presence of an unsatisfactory behavior: rapid oscillations in the stress and displacement fields, implying the physically impossible phenomenon of interpenetration. In the late seventies two major modifications were proposed: one by Atkinson and the other by Comninou. The Atkinson modification recognizes that the interface between two different materials is almost never sharp, and provides a gradual transition which avoids the oscillatory behavior. The Comninou approach simply resolves the interface crack problem by accepting its inherently unilateral nature (presence of inequalities) and allowing for partial closure at the tips. Both solutions have received high praise and severe criticism, especially since the oscillatory behavior is absent in some, at least, classes of nonlinear materials (Knowles and Sternberg). Current emphasis is placed on numerical models of the elastoplastic behavior of interface cracks. An additional complication is the apparent presence of mixed mode crack tip fields regardless of the type of the applied loading. Valuable as these theoretical efforts may be, it is becoming increasingly imperative to perform experiments to determine the mode of propagation and critical parameters governing interface fracture. Even then, the results must be viewed with caution, because the quantities of interest can only be determined indirectly. The present paper presents an overview of the interface crack problem and describes some preliminary experimental results in the fatigue and fracture of interface cracks.