Where inductive courses have failed, the reason has been most often that the preparatory steps have been omitted by the teacher, and the student has been set to find out something when she has no knowledge of what she has set out to find. Chance discoveries, of course, find their applications later on, but this is not education. The student needs to have clearly in mind the results looked for, before she begins an experiment. This by no means implies _that the result itself should be known, for then interest is dulled. References should be looked up only after the practical work, or its chief value is lost.