Presumably teacher education programs should be guided by theories of learning to teach. This paper reports the results of analyzing documents and interviews about one program (probably fairly typical) from the perspective of two theories of learning. We find some evidence consistent with each theory, more that is inconsistent. Overall, the program does not correspond closely to either theory. It produces confused messages and disappointing results for prospective teachers. Implications for the design of teacher education are discussed. A comment by the program director is appended that explores further the dilemmas discussed in the main paper.