DYNAMIC PROCESSES IN BE STAR ATMOSPHERES .1. DIMPLE FORMATION IN THE HE-I LAMBDA-6678 LINE OF LAMBDA-ERIDANI

'' Dimples '' are ephemeral absorption features flanked by incipient blue/red emission in the He I lambda6678 line profiles of lambda Eri and several similar mild classical Be stars. They develop rapidly (approximately 15 minutes), last 2-4 hr, and are observed in this line on the average at least once per night, making them the most ubiquitous of the ''spectral transients '' appearing in the line profiles of at least some mild Be stars. Their formation may well be related to the unknown instability-causing mass-loss episodes in Be stars. In this paper we discuss simultaneous IUE and optical (lambda6678) time-serial observations of lambda Eri, together with near-simultaneous UV continuum observations made by Voyager 1, in 1990 October. We find several examples of weakenings of the C IV, N V resonance lines that coincide with the appearance of lambda6678 dimples. The absence of variations in other UV lines and in the UV continuum at the same time (IUE) or nearly the same time (Voyager) argues against dimples being caused by thermal variations from the underlying star. We suggest instead that the resonance line weakenings are caused by non-LTE effects associated with the condensation of high density structures at some elevation over the star. We present a simple model of an opaque (in the line only), essentially stationary slab which backscatters lambda6678 line radiation into a surrounding ''penumbral'' region. Lambda6678 photons are scattered a second time in this region back into the observer's line of sight and in the process acquire the local projected doppler shift from rotation. We examine several attributes of this mechanism and find that slabs typically have a projected radius of about 1 R., an elevation of approximately 0.10R*, and a density of approximately 2 x 10(11) cm-3. Our models can also match dimples resembling '' P Cygni '' features by permitting our model slabs to rise or fall and at the same time to move retrograde or prograde with respect to the stellar surface. Slabs would probably produce too little emission to be easily detected in the Halpha profile. Their detection in strong He I lines seems the best stategy among early Be stars. Slabs contain enough mass (> 10(-14) M.) and develop over such a large area and so rapidly as to greatly limit the number of mechanisms that could form them. Their rapid development spatially over the star precludes a purely thermal origin, e.g., from the propagation of a shock at the speed of sound. Yet, in terms of bulk motion dimples represent an exclusively low vertical-velocity phenomenon. Arguments can be advanced that dimples are not the results of ballistic tossings of blobs or of orbiting circumstellar ejecta. One viable mechanism for their formation requires that dimple-slabs condense from ambient rarefied material via a supersonically propagating instability. If this velocity is understood in terms of an Alfven wave, a case can be made that magnetic fields of greater-than-or-equal-to 100 G can exist in small regions above lambda Eri and many other mild classical Be stars.