THE ASTROPHYSICAL JOURNAL, 460:L49[–]L52, 1996 March 20
© 1996. The American Astronomical Society. All rights reserved. Printed in
U.S.A.
-------------------------------------------------------------------------------
Next: §3. PHOTOMETRIC OSCILLATIONS
Previous: §1. INTRODUCTION
Top: TITLE PAGE
-------------------------------------------------------------------------------

                        §2. THE 5.52 YEAR SPECTRAL CYCLE

     Spectroscopic observations of [&eegr;] Car were carried out with the 1.6 m
telescope of the National Astrophysical Laboratory (LNA, Brazil) between 1989
and 1995 at the coudé focus with resolution of R = 16,000. During 1992 March
through November, a shell episode was registered in great detail (Baratta et
al. 1993; Damineli et al. 1995). The He I [&lgr;]10830 line disappeared and was
accompanied by a strong fading of other He I lines[—][&lgr;]5876,
[&lgr;]7065, and [&lgr;]6678[—]and of [Ar III] [&lgr;]7135. The lines of
the Balmer and Paschen hydrogen series decreased moderately in intensity, as
was also reported for past shell events. Table 1 shows the He I [&lgr;]10830
line equivalent width (He I Weq) derived from the LNA data and from other
authors (Allen, Jones, & Hyland 1985; McGregor, Hyland, & Hillier 1988).

     Figure 1 displays the He I Weq variations, superimposed on the light curve
of Whitelock et al. (1994), that will be discussed in the next section. The
plot shows a continuously varying line intensity, from He I Weq = 806 Å to He I
Weq = 6 Å. It corresponds to a variation from [≈]1500 to [≈]15 L[&sun;]
in this single line. This behavior is different from that of a quiescent star
perturbed by erratic shell episodes, as is frequently assumed in the
literature. Those shell episodes correspond to phases of short duration around
the minimum of cyclic high- and low-excitation states. I considered the epoch
of the minimum to be the date of the lowest recorded He I Weq, which occurred
on 1992 June 3 (1992.42). I cannot exclude, however, that even zero value could
have been reached in the interval 1992 May 28[–]July 17. The rate of
change in He I Weq [≈] 60 Å per month is a typical value in the time
interval 1992 March[–]November, characterizing a very sharp minimum,
compared to the [≈]10 Å per month scatter during the high-excitation phases.

[Image] Fig. 1

     The maximum value of the He I Weq seems to be decreasing from cycle to
cycle. The exact amount of fading, however, is difficult to derive, because of
the coarseness of the time sampling. For the same reason, we cannot assume that
the first point of the curve is the very minimum or that the second is the peak
of that cycle. Their intensities are only an indication of how high or low is
the state of the line excitation on that particular date.

     There are five shell episodes documented in the literature since 1944,
characterized by faint or absent high-excitation lines (Gaviola 1953; Rodgers &
Searle 1967; Thackeray 1967; Viotti 1969; Melnick, Ruiz, & Maza 1982; Whitelock
et al. 1983; Zanella, Wolf, & Stahl 1984; Allen et al. 1985; Bandiera, Focardi,
& Altamore 1989; Baratta et al. 1993; Altamore, Maillard, & Viotti 1994). The
shell episodes occurred in the years 1948, 1965, 1981, 1987, and 1992. The last
three shell events indicate a cycle of about 5 years, suggested also by the
near-infrared (NIR) data of Whitelock et al. (1994).

     In order to search for a possible coherent variability, I chose as real
minima the shell event of 1992 June 3 (1992.42) and that of 1948 April 19
(1948.30), reported by Gaviola (1953) on the basis of 70 plates, collected
between 1944 and 1951. Eight complete cycles of 2014 days (5.52 yr) would have
elapsed between them. The low-excitation episodes of 1965, 1981, and 1987 fit
quite closely the epochs of predicted minima. O - C values are summarized in
Table 2 (col. [3]). The proposed period was successful not only in fitting all
the registered shell episodes but also in assigning nonshell phases to all the
time intervals in which the [&eegr;] Car spectrum exhibited the high-excitation
lines (Gaviola 1953; Rodgers & Searle 1967; Thackeray 1953; Viotti 1968;
Whitelock et al. 1983; Zanella et al. 1984; Ruiz, Melnick, & Ortiz 1984;
Hillier & Allen 1992; Bidelman, Galen, & Wallerstein 1993; Damineli et al.
1995).

     By adopting the last observed shell event as T0, the epochs of the minima
are given by

[Image]

where JD is the Julian Day and n is any integer. The uncertainty in T0 is
assumed to be the time delay between the lowest recorded He I Weq and the dates
of the two contiguous measures, 6 days before and 48 days after. The next zero
phase is predicted to occur on 1997 December 8, but, taking into account the
uncertainty in T0, it can occur as early as 1997 December 2 or as late as 1998
January 25. Other minima must have occurred around 1975 November, 1970 May,
1959 May, and 1953 November. A search in archival data, spread all over the
world, could reveal some yet unnoticed shell event or perhaps refute the
proposed periodicity.

-------------------------------------------------------------------------------
Next: §3. PHOTOMETRIC OSCILLATIONS
Previous: §1. INTRODUCTION
Top: TITLE PAGE
-------------------------------------------------------------------------------