The objective of this effort was to determine the manner in which small fatigue cracks initiate at notches, extended by cyclic loading, interact with adjacent flaws, and coalesce into a single dominant crack which controls final fracture. The desired product was a predictive scheme capable of analyzing the early stages of fatigue crack growth which are characterized by the growth and link-up of small cracks. Research toward this goal was directed at the following tasks: 1) Crack Growth Predictive Algorithm - A computer program was developed to predict the growth and coalescence of multiple cracks located at notches; 2) Crack Interaction Analysis - Stress intensity factors solutions were computed for multiple cracks located at an open hole. These solutions were required for the multiple crack growth algorithm; 3) Crack Coalescence Experiments - Fatigue tests were conducted with multiply cracked specimens to provide a data base to evaluate the predictive model. The model was verified with large crack results directed toward coalescence aspects of the problem as well as small crack experiments which focused on the initial stages of fatigue crack growth; and 4) Characterization of Small Cracks - This phase of the effort was directed toward the growth and coalescence of physically small cracks, which were expected to behave differently than large cracks.