George Box was fascinated with how we make discoveries. His path-breaking contributions to experimental design made statistics an active partner in the process of discovery. Box introduced us to response surface methods, evolutionary operation, resolution and rotatability, projective properties and design robustness. He developed popular experimental plans like the central composite and Box-Behnken designs. He explored the consequences of imperfect models and derived D-optimal designs for experiments to estimate mechanistic models. Box's ideas grew from close collaborations with scientists and engineers and have been applied successfully in a wide range of disciplines. He has left an indelible stamp on the field of experimental design and on the practice of scientific investigation. Copyright © 2014 John Wiley & Sons, Ltd.