2 Observations

Figure 1 presents a finding chart of NGC 6231 that shows the eight fields where we obtained UBVI photometry using the 60 cm telescope of the University of Toronto Southern Observatory, Las Campanas, Chile. The observations, that covered an area of 105 squared minutes approximately, were carried out during a 5 day run in May 1995 using a PM $512 \times 512$ METHACROME-II, UV coated chip, with a scale of $0.45\hbox{$^{\prime\prime}$}$/pixel, covering $4^{\prime}$ on a side. Typical exposure times per filter were: from 20 to 600 seconds in V, 10 to 1000 in B, 50 to 1800 in U and 10 to 130 in I including mid exposures. As usual, we took two long exposures per filter to improve the statistics of faint stars. Apart from a long exposure frame disturbed by thin clouds that had to be self-calibrated, the entire run was photometric, with mean seeing values ranging from 1.2 to 1.5 arcsec. The complete reduction techniques are described in Vázquez et al. (1996). Instrumental signatures were removed using bias and a combination of dome flats and twilight flats while instrumental magnitudes were produced via point spread function (Stetson 1987).

  

Figure 1: The finding chart of programme stars in NGC 6231. The size of the symbols is proportional to the star magnitude approximately

The observations were tied to the standard UBVI Cousins system using stars in NGC 5606 (Vázquez & Feinstein 1991) and Hogg 16 (Vázquez et al. 1994) systematically measured every night. The residuals left by these "standard" stars (about 30 each night) average less than 0.025 in magnitudes and colours, a value that we adopt as a measure of the external photometric error of the run. An estimate of the internal errors of our photometry was obtained by comparing photometric values of common stars in the overlapping zones among different frames. We found typical mean differences of about 0.025 mag for V < 16 mag in both, colours and magnitudes.

Table 1 contains the final CCD photometry including star numberings, x and y coordinates and magnitudes and colours for 1060 stars. Errors in colours and magnitudes are plotted in Fig. 2 as a function of the V magnitude and a cross-identification of Seggewis (1968) numberings and ours is shown in Table 2.

  

Table 2: Cross-identification with Seggewiss (1968)

  

Figure 2: Photometric errors in magnitude and colour as a function of V magnitude

A comparison of our photometry with Perry et al. (1991, hereafter PHC91), Seggewiss (1968) and Sung et al. (1998, hereafter SBL98) is shown in Table 3. The agreement with these data sets is excellent although we notice large standard deviations respect of the 157 stars in common with Seggewiss, probably because most of his data is photographic what can yield uncertain photometric values among very faint stars.

Several stars showed $\Delta V$ differences larger than $3\sigma$ respect of PHC91. In some of them the differences could have been produced by star-light contamination of photoelectric measures but in others it would be desirable to carry out further analyses to confirm whether they are real variable stars. All the stars suspected of variability are marked with an ( 1) in Table 1.

To know how much of the cluster we surveyed, we re-determined its size using the Digitized Sky Survey plates (DSS). To construct the cluster stellar density profile, star counts were done in concentric rings around a centre previously determined. Then, this profile was fitted using a Gaussian and the cluster limit was set at the point where the stellar density merges into the level background. The procedure gives a 7 arcmins radius and a total area of the cluster of about 150 square minutes so $\approx$ 70% out of it was observed by us. However, the literature reports two groups of sizes for NGC 6231: small ones, $\approx 15$ arcmins diameter (Shapley 1930; Trumpler 1930), which we favour, and large sizes, $\geq 25$ arcmins (Barchatova 1950; Seggewiss 1968). It is worth mentioning that Raboud et al. (1997) (hereafter RCB97) and SBL98 have found members at more than 7 arcmins from the cluster centre what could be interpreted as a mass segregation effect (Raboud & Mermilliod 1998).

  

Table 3: Comparison of our photometry with other authors in the sense others minus us