FIG. 2. Wave-vector components for minima of curves in Fig. 1. Solid lines are circles corresponding to zeroth-order energy levels according to Eq. (5). Dot size indicates estimated error.tually the far more prevalent situation. As further examples, the [/-0, l] and [j-1, 0] pairs cross at (^ = 45° with no discernible perturbations, as do the I j -0, 1J and [ j -0, Tj pairs at cp = 0°. In fact, such crossings are of great assistance in identifying the minima and determining the energy values needed for the radii of the circles drawn in Fig. 2.Thanks are due to the National Science Foundation and to the Applied Research Laboratory of The Pennsylvania State University for their support of this work, and to Professor Eric Thompson for preprints and discussions of his work in advance of publication.son calculated a splitting of about 6° for LiF, based on a model potential with several empirically fitted parameters. Our experimental results on NaF (Fig. 1) show splittings of the same order of magnitude. However, rather than a simple crossing of two otherwise isolated levels, we see that the situation is complicated by the presence of numerous other zeroth-order levels in the same vicinity. The observed positions of the minima suggest mixing of the following pairs of zeroth-order states: [0-1, Oj and [0-1, T], [0-1, l]and [1-1,0], and [ 1-1,1] and [ 1-1,0] in the notation [j-m 9 n].These pairs are all connected by reciprocal-lattice vectors of the (0, 1) type.On the other hand, the [0-0, 2] level has no discernible effect, as it is connected with the others by longer reciprocal-lattice vectors.