Astron. Astrophys. 338, 581-591 (1998)

2. Data

2.1. PG3

Four fields were selected in the mid-fifties by Baade and Plaut to search for variable stars (Blaauw 1955; Larsson-Leander 1959). The results from the photographic survey, known as the Palomar-Groningen Variable Star survey, were published in a series of six papers (Plaut 1966, 1968a,b, 1966, 1971, 1973). The centre of PG3 is located south of the galactic centre and skims over the edge of the galactic bulge. Important aspects of this field are the large area covered () and the low interstellar extinction, which is gradually increasing in the direction towards the galactic centre (Wess87).

With emphasis on the RR Lyrae stars, the variable stars in PG3 were re-examined by Wess87, using UKST and Schmidt plates. Fig. 1 shows the Fourier spectral window for the epochs of the plates. The figure shows that a good resolution, as intended, is obtained towards the short period variables, but for the long period variables a deficiency of stars with periods between 320 to 500 days could be possible. The three highest alias peaks correspond in decreasing order with one year, one week, and one synodical month.

Fig. 1. The Fourier spectral window for the plate coverage of PG3 (Wesselink 1987)

The large number of Miras and SRVs discovered in this field makes it very attractive, to subject these stars to a more detailed study. Bl92 studied a sample of Miras and compared them with the IRAS sources in PG3, while we focus on the SRVs.

2.2. The SRV sample

In the GCVS4 catalog (Kholopov et al. 1988) the classification of SRVs is based on the shape and the amplitude of the light curve. Generally, the period of the variations ranges from 20 to 2000 days with an amplitude less than V = 2 Plaut (1971) distinguished in his classification SRa and SRb type variables, while Wess87 made no distinction. Wess87 based his criteria on the and amplitudes, smaller than 2 and periods ranging from 30 - 1000 days. The SRV classification of Wess87 is in most of the cases compatible with a SRV of type `a' (hereafter referred to as SRaV) from Plaut (1971). In general, the lightcurves of the SRaVs resemble those from the Miras. The difference in the classification is merely a consequence of the imposed amplitude limits in the variation.

In this study the SRV stars are selected with the Wess87 classification. Higher priority is given to the observations of stars with a Q = 0 quality flag, indicating that there is no doubt about the classification and the period. The SRVs are selected such, that there is no bias to the brightest SRVs and that the sample covers in 25 days intervals the full period range.

2.3. Near-infrared photometry

Near-infrared photometry (JHKLM) of 78 PG3 SRVs was obtained with the ESO 1-m telescope, La Silla (Chile), equipped with an InSb detector. The observations were carried out under photometric circumstances in the observing seasons 1990 - 1993 (ESO No 49.5 - 011 & 51.7 - 056). Table 1 shows the log of the observing runs. The 1990 - 1991 observations (see Bl92) of the stars S40, S147, S728, S969, S1008, S1016, S1128 and S1204 were carried out as part of the ESO key programme `Stellar evolution in the galactic bulge' (Blommaert et al. 1990; ESO No 45K.5 - 007). The observations were made in a standard way, through a diaphragm with a aperture with the chopping and beam-switching technique. An 8 Hz sky chopping and a beam-switch throw of approximately in R.A. was applied.

Table 1. Log of the observing runs

The JHKLM fluxes are calibrated to the ESO photometric system (Bouchet et al. 1991). The typical errors are 0 in JHK, 0 in L, and 0 in M. For some stars no L and M photometry was obtained, because they were fainter than the limiting magnitude attainable with the telescope. Table 2 gives the JHKLM magnitudes of the PG3 SRVs together with the period (Wess87).

Table 2.^** Near-infrared photometry for the stars in field #3 of the Palomar-Groningen Variable Star Survey (Plaut 1971). Column 1 lists the stellar identifier, adopted from Wesselink (1987); column 2 gives the identification made by Plaut (1971); column 3-7 gives the JHKLM photom- etry, typical errors are 0 in JHK, 0 in L and 0 in M; column 8 gives the observing run identifier (see table 1); column 9 gives the period determined by Wesselink (1987) if available; and column 10 gives the quality flag related with the period and the identification of the star (Q=0: no doubt about the determined period and classification, Q=1: classification is correct but alternative period is possible, Q=2: period determination is correct but the classification is doubtful, Q=3: both period determination and classification are unreliable)

In general the limiting magnitude was 10^m for most of our observing nights, which is due to our relatively short integration time and integration sequence. Photometry for some of the fainter stars were obtained in nights when the photometric conditions allowed it. In this respect the DENIS survey (Epchtein et al. 1994, 1997), with 14^m, should be able to improve on the photometry and reach significantly deeper limits.

2.4. Crowding

The presence of additional stars in the beam cannot be avoided, because we are dealing with crowded field observations. Two cases should be considered: additional stars in the primary or in the background beam. An additional flux contribution in the primary beam would lead to an increase of the flux from the star, while stars in the background beams would give a background subtraction which is too high. As a consequence the flux of the target star will be underestimated. In first approximation, the number of faint stars in both the primary beam and the background beams are similar.

The stars in the background are in general much fainter and not as red as the stars observed. The induced errors from stars present in the background beams are expected to be less than the errors quoted above. A new position for the background subtraction would have been selected, if a bright star was noticed in one of the background beams, but this was not necessary. The flux of the target star will be overestimated with additional stars in the primary beam. Due to stars surrounding S283 (see finding chart in Ng & Schultheis 1997), the magnitudes are slightly too bright and the colours too blue for this star.

2.5. Interstellar extinction

The procedure described by Bl92 is used, to correct for the interstellar extinction. It is based on the PG3 extinction map, constructed by Wess87 from the colour excess of the RR Lyrae stars at minimum light. In this map the extinction is highest in the plate corner at lowest galactic latitude (A() = 1 ) and lowest at the opposite corner (A() = 0 ). The extinction is described with a linear relation, because of the smoothness of the gradient in the extinction map. In first approximation we have A() = , where b is the galactic latitude. For a normal extinction law the corrections for the different infrared passbands can be derived with the standard curve no. 15 of van de Hulst (1949). All the JHK photometry for the PG3 stars discussed in this study are de-reddened with the procedure outlined above. The photometric data in Table 2 are not de-reddened.

2.6. The comparison sample

A comparison is made with other near-infrared photometric studies of SRVs and Miras, in order to obtain a better understanding of the evolutionary status of the PG3 SRVs. We use a sample of PG3 Miras (Bl92), well observed O-rich field Miras (Catchpole et al. 1979) and a magnitude limited sample of field O-rich SRVs (KH94, Kerschbaum 1995). Note, that the field samples mentioned above might not be truly representative in their relative numbers for the local neighbourhood. The field stars were de-reddened with a procedure similar to Feast et al. (1982). The reddening corrections are small, because the visual absorptions ranges typically from 0 We further used the Sgr I Miras from Glass et al. (1995) and the LMC LPVs (long period variables) from Reid et al. (1995) as an extra indication. For the latter we defined Mira & SRV variable groups, following Hughes & Wood (1990) on basis of the I-band amplitudes. We use the Reid et al. data set instead of the Hughes & Wood data set, because the former are in the same photometric system as the Sgr I Miras.

2.7. Photometric transformations

The photometry of the PG3 SRVs & Miras and the field SRVs are in the ESO photometric system (Bouchet et al. 1991, van der Bliek et al. 1996). The photometry for the field Miras was obtained in the SAAO photometric system as defined by Glass (1974). This photometric system is not identical to the SAAO system in which the photometry for the Sgr I Miras, the LMC LPVs, and the period-luminosity & period-colour relations were obtained (Glass et al. 1995). All the photometry discussed in this paper are in the ESO photometric system (Bouchet et al. 1991) or transformed to it from the various SAAO systems. New transformations from Hron et al. (1998) are used, because existing transformations either refer to the old ESO system (Bessell & Brett 1988, Carter 1990, Engels et al. 1981, Wamsteker 1981) or do not cover all SAAO systems (Bessell & Brett 1988, Bouchet et al. 1991, Carter 1990, van der Bliek et al. 1996). Furthermore, these transformations do not include stars with (J-K) 1 Extrapolation of these transformations to the colours of typical AGB stars leads to errors of the same order as e.g. the differences due to metallicity (Hron et al. 1998). The estimated uncertainties in the new transformations are typically 0 in the colours.

© European Southern Observatory (ESO) 1998

Online publication: September 14, 1998
helpdesk.link@springer.de