To the Editor.— Klein et al 1 offer no hard evidence to support the idea "that the avoidance of agoraphobia is secondary to spontaneous panic," and their assertion that "exposure without imipramine is of benefit only in reducing avoidance" and does not reduce panic is actually misleading (reviewed by Marks 2 ). Regarding the development of agoraphobia, although it is unwise to draw firm conclusions about cause from treatment, as maintaining factors may differ from initiating ones, even treatment data do not sustain the position asserted by Klein et al. As they themselves note, "that drug-induced panic improvement occurred prior to improvement in avoidance was neither substantiated nor invalidated.... our observations were insufficiently fine grained"; nor is there any other systematic data to confirm their view, and they ignore contradictory data. Using Klein's panic scale, Marks et al 3 found that panic began to improve after avoidance during exposure, whether this