The six keys to n,ormal occlusion Lawrence F. Andrews, D.D.S. San Diego, Calif. T his article will discuss six significant characteristics observed in a study of 120 casts of nonorthodontic patients with normal occlusion. These constants will be referred to as the “six keys to normal occlusion.” The article will also discuss the importance of the six keys, individually and collectively, in successful orthodontic treatment. Orthodontists have the advantage of a classic guideline in orthodontic diag- nosis, that is, the concept given to the specialty a half-century ago by Angle that, as a sine qua non of proper occlusion, the cusp of the upper first perma- nent molar must occlude in the groove between the mesial and middle buccal cusps of the lower first permanent molar. But Angle, of course, had not contended that this factor alone was enough. Clinical experience and observations of treatment exhibits at national meetings and elsewhere had increasingly pointed to a corollary fact-that even with respect to the molar relationship itself, the positioning of that critical mesiobuc- cal cusp within that specified space could be inadequate. Too many models dis- playing that vital cusp-embrasure relationship had, even after orthodontic. treatment, obvious inadequacies, despite the acceptable molar relationship as described by Angle. Recognizing conditions in treated cases that were obviously less than ideal was not difficult, but neither was it sufficient, for it was subjective, impression- istic, and merely negative. A reversal of approach seemed indicated : a deliberate seeking, first, of data about what was significantly characteristic in models which, by professional judgment, needed no orthodontic treatment. Such data, if systematically reduced to ordered, coherent paradigms, could constitute a group of referents, that is, basic standards against which deviations could be recognized and measured. The concept was, in brief, that if one knew what eon- stituted “right,” ,he could then directly, consisfently, and methodically identify and quantify what was wrong. A gathering of data was begun, and during a period of four years (1960 to 1964), 120 nonorthodontic normal models were acquired with the cooperation 296