THE ASTROPHYSICAL JOURNAL, 460:L49[–]L52, 1996 March 20 © 1996. The American Astronomical Society. All rights reserved. Printed in U.S.A. ------------------------------------------------------------------------------- Next: ACKNOWLEDGMENTS Previous: §3. PHOTOMETRIC OSCILLATIONS Top: TITLE PAGE ------------------------------------------------------------------------------- §4. DISCUSSION The inversely correlated variations of the He I Weq and the NIR light curves seem to be a new cornerstone to understanding the nature of [&eegr;] Car. They resemble those of S Dor variables, although the amplitude of the light curve is smaller and only the higher excitation lines show remarkable variations. The lack of corresponding variations in the UV and the low amplitude in the V band could be explained by light absorption and scattering inside the Homunculus. The 5.52 yr cycle, however, is unusually stable for the standards of the S Doradus class. If [&eegr;] Car is a pulsating star, it would be different from all other known luminous stars because their cycles are much less stable. A binary hypothesis seems to be plausible, in a scenario of coherent cycles. A low-mass secondary star would orbit a M = 120 M[&sun;], R* = 103 R [&sun;] primary one at a distance of about D = 3 R*, in a period of 5.52 yr. An elliptic orbit inside the dense circumstellar region could provide a variable mass accretion onto a compact star, in order to explain the NIR light curve. In this case, however, the photometric peaks would be expected to coincide with the high-excitation phases, contrary to the observations. A critical test is proposed to the 5.52 yr period by the prediction of a new shell phase to occur in 1997 December or 1998 January. The next low-excitation phase will be important to study not only the central star but also the associated nebula. The Homunculus is mainly a reflection nebula, and so the variations of the central star can be traced through it by the [“]light echoes.[”] As can be seen in Figure 1, the larger variations are in the He I Weq rather than in broadband filters. The equivalent width, on other hand, is preserved when the light is scattered and dimmed by the dust. The best season to apply this technique is during the 6 months time interval around the epoch of minimum excitation, when the He I Weq declines and rises again in a rate of nearly 60 Å per month. It can give an independent check of the presently accepted model of a bipolar flow of the Homunculus (Warren-Smith et al. 1979; Hyland et al. 1979; Mitchell et al. 1983; Allen & Hillier 1993; Meaburn et al. 1993; Frank et al. 1995). The most recent spectra show that the He I Weq curve is declining. After a phase of constancy in 1994, the He I Weq decreased by nearly 60 Å from 1995 May to November, with a gradient that is similar to the rate of changes in 1991. This seems an announcement that a new low-excitation event is forthcoming, as predicted by the 5.52 yr cycle. ------------------------------------------------------------------------------- Next: ACKNOWLEDGMENTS Previous: §3. PHOTOMETRIC OSCILLATIONS Top: TITLE PAGE -------------------------------------------------------------------------------