1. C. Oberman and M. Kruskal (unpublished).
2. In agreement with the terminology of Ref. 1, we define a “constant of the equations” to be a quadratic form in the field quantities which is independent of time.
3. Plasma oscillations
4. See, for example, Ref. 3.
5. The choice eikλt is made in order to simplify later formulas.