SOME years ago, one of my students, upon being asked why the verb was at the end of the clause, replied that the past perfect tense required the verb in that position. Of course, one whose mind worked in this way was far from brilliant; and yet he was not altogether to blame for his answer. The grammar which we were using at the time treated the past perfect tense and subordinate clauses in one lesson and failed to separate the two subjects properly in the exercises. In this respect, the book is typical of a number of grammars which have appeared in recent decades. At the outset, we must admit that textbooks have lately become more and more interesting, stimulating, in touch with actual life. The reading selections are today genuine German and not "grammar Dutch." Furthermore, the presentation of grammatical principles has traversed a long path from the exhaustive, impractical approach to restriction to that which the beginner really needs. In other words, formal grammar has given way to functional. A third component, viz., exercises, also shows improvement in the same direction and for the same general reasons. Today our grammars contain many direct method exercises which stress active practice: German questions on the text, mutation, supplying forms, etc. And as the reading has become more natural, the exercises have followed suit. The exercise method of Ollendorff and Ahn of the last century which elevated mechanical drill on absurd content to a commanding position has been displaced by more normal constructions. But precisely in the trend away from the artificial and toward the practical lies a subtle danger which seems to go unrecognized. (I must state here that I am concerned in this paper with elementary, basic drills and not with advanced exercises.) Long experience with numerous grammars has compelled me to be skeptical of the value of many natural exercises. Let us now examine some cases which I have encountered, with a view to diagnosing a common ailment. I trust that the authors of the grammars from which my evidence is drawn will not feel aggrieved since I thought well enough of their books to advocate using them. One volume employs for drill on passive synopsis a very unwieldy sentence of cultural content taken from the text. From