This new and expanded edition of A. W. F. Edwards' classic volume on scientific inference presents his most important published articles on the subject. Edwards argues that the appropriate axiomatic basis for inductive inference is not that of probability, with its addition axiom, but that of likelihood, the concept introduced by Fisher as a measure of relative support among different hypotheses. Starting from the simplest considerations and assuming no more than a basic acquaintancewith probability theory, the author sets out to reconstruct a consistent theory of statistical inference in science. Using the likelihood approach, he explores estimation, tests of significance, randomization, experimental design, and other statistical topics. Likelihood is important reading for students and professionals in biology, mathematical sciences, and philosophy.