Scientists often use more than the results of experiment to arrive at a result; they use anticipation and analogy to arrive at the results that fit their theories, and sometimes they correct results in the light of analogy. They also need to be clear about the difference between accuracy and precision. They do all this using not only theories, but also apparatus, and the interplay between apparatus and the development of concepts and theories is often crucial. Historians of chemistry (notably including the recent work of Usselman, Rocke, and Holmes) furnish us with plenty of examples of such interplay, and of the selection of data in the light of theory. Lavoisier, Dalton, and Liebig can each teach us a good deal about the way that good scientists arrive at reproducible results.