2 Observations

The observations were made in 17 photometric nights between 1984 and 1985 with a photoelectric photometer attached to the 0.61 m Lowell telescope of the Cerro Tololo Interamerican Observatory, using a CO₂ cooled ITT FW130 S20 photomultiplier and a photometric system composed of three interference filters: an H$\beta$ narrow, an H$\beta$ wide and an [O III] filter centred at 5000 Å, with passbands of 30, 150 and 70 Å, respectively. In the first run, two additional filters, centred at 4517 and 5320 Å with bassbands of 75 and 100 Å respectively, were used in order to evaluate the continuum emission independently of the H$\beta$ filters. Later on, in view of the compatibility within few percents of the results obtained with and without these two filters we dropped them. The basic observational routine was composed of four exposures of 10 s in each filter, alternating the filters in the sequence 1-2-3-3-2-1. The total exposure time of this routine for the pair object and sky was 480 s (800 s for the first observations with the five filters system). For each object, we have repeated this routine at least twice, taking different positions for the sky subtraction. Each night two or three of the spectrophotometric standard stars $\xi$ Cet, $\theta$ Crt, $\eta$ Hya and $\theta$ Vir were observed for the absolute flux calibration. Due to the presence of strong Balmer absorption lines on the spectra of the standard stars, the absolute fluxes were obtained by comparison between the counts of the H II regions and the standard stars in the [O III] filter. Stellar fluxes published by Kohoutek & Martin (1981) were used. The reduction procedure was the same as described in details in Copetti & Dottori (1989) for an analogous programme. A sample of planetary nebulae was also observed in order to check the quality of the observations and reduction procedure. The results obtained for this control sample, which have been published elsewhere (Copetti 1990), agreed very well with data taken from the literature with a mean difference of 6% for the [O III]/H$\beta$ ratios and 0.03 dex for the logarithmic H$\beta$ fluxes. No evidence of systematic deviations was found.

Figure 1: Comparison of the [O III]/H$\beta$ measurements with mean values from the literature. The squares correspond to the H II regions (Danziger 1974; Esteban et al. 1992; Girardi et al. 1997; Hawley 1978; Heydari-Malayeri 1988; Kwitter 1984; Lortet et al. 1984; Peimbert et al. 1978; Shaver et al. 1983) and the triangles refer to the planetary nebulae of the controlling sample (Acker et al. 1989; Backer 1985; Danziger et al. 1973; Freitas Pacheco et al. 1991; Kaler 1976, 1983; Kohoutek & Martin 1981 Oliver & Aller 1969; Peimbert & Torres-Peimbert 1978; Shaw & Kaler 1989; Webster 1969, 1983). Also shown the identity line