2 Observations and data reduction

 The photometric search method we used to detect Be stars involved taking a CCD image through a narrow-band (15 Å) H$\alpha$ filter and comparing this image with a similar image obtained in the Cousins R band. Stars with strong H$\alpha$ emission should appear brighter in the H$\alpha$ image than in the R image. Given that a typical Be star has H$\alpha$ emission with a full width at half maximum of $\sim$7 Å and peak H$\alpha$ flux 5 times the local continuum flux, this sort of search with a 15 Å filter is readily capable of detecting Be stars.

The observations were obtained using direct CCD imaging on the 1-m telescope at Siding Spring Observatory from September to November 1991. All H$\alpha$images were 900 s exposures. R images were obtained immediately before or after each H$\alpha$ image in order to avoid the possibility of stellar variability affecting our measure of emission strength. A Tektronix 1024$\times$1024 CCD with pixels of scale 0.6$^{\prime\prime}$ was used for these observations, giving us an imaged area of typically 10$^\prime$$\times$10$^\prime$ (compared with 5.7$^\prime$$\times$5.7$^\prime$ in Grebel (1997). V and I images were obtained at the same time for some fields, and in some cases on other nights using a GEC 770$\times$1152 CCD. A log of the observations is presented in Table 1.

The CCD images were processed with IRAF and the photometry of the fields was done using the DoPHOT photometry package (Mateo & Schechter 1989). Standard magnitudes in V and I were derived from the standards used in Sebo & Wood (1994) and from magnitudes given by Walker (1996). The R and H$\alpha$ magnitudes were not standardised and $R-{\rm H}\alpha$ colours have an arbitrary zeropoint.

  

Table 1: The log of observations presented in the current paper