might expect the emitted radiation to be polarized so that the polarization vector lies in the plane determined by K and % 0 . The polarization must have a nonvanishing component along K so one would expect the intensity of radiation back along -K always to be zero.The above classical analysis shows that a sharply defined optical electric field or a sharply defined notch in such a field can produce bunching in an electron beam. This may partly account for the effect Schwarz and Hora observed. Even though the density of electrons in their experiment was only one electron per fifty "bunches" in the primary beam, there is still time coherence between electron beam charge and current density at the screen and the electric field at the interaction region.However, a quantum mechanical treatment of the problem now being developed seems to indicate that true quantum phenomena may be present. In particular, our wave equations predict that the electron bunching for monochromaticIn an earlier investigation Barmatz and Rudnick 1 (BR) in an effort to determine the thermodynamic first-sound velocity near T x made measurements at the low frequency of 22 kHz. These measurements were sufficiently accurate, and the approach to T x sufficiently close, that attenuation and dispersion effects were measurable. Because of the low frequency, the results were necessarily inaccurate. The purpose of the present investigation was to use essentially the same apparatus over as wide a frequency range as possible, and we report here the results of the attenuation measurements.The frequency range is 16.8 kHz to 3.17 MHz in He II, and 600 kHz to 3.17 MHz in He I. With this frequency range, and microdegree temperature resolution, the measurements yield greater detail about the nature of the attenuation than has previously been reported. In particular, at a frequency a>, the maximum attenuation is unam-electrons obeys the classical Bessel-function relationship only if Hoo Q 2 y/(4eV 0 v)«ir/2. For Hoofy/ (4eV 0 v)»ir/2 and (4eE 0 v/Hou 0 2 ) s in( oo Q d/2v)«1, the bunching appears to vary sinusoidally with distance and thus not decay to zero as y -* °°. If further study supports this finding or predicts other nonclassical effects, the quantum mechanical analysis will be reported in a future paper.The authors are pleased to acknowledge helpful conversations with Professor Helmut Schwarz of biguously shown to occur at a temperature T, below T x , such that a)€"~1 = eonst, where e = |T x ~T|. Recently there has been much theoretical speculation on the nature of first sound at lambda transitions, resulting in a bewildering array of sometimes conflicting results. While there are instances where there is partial agreement between such predictions and our data, we know of none which completely account for the results. In view of this, we make no attempt at a conscientious comparison of the data with existing theory, except as regards the relaxation theory first suggested by Landau and Khalatnikov. 2 Our results, and this situation, point up the need for ...