The images were reduced following the standard procedures of bias subtraction and flat-fielding using the IRAF (4) task CCDPROC. The dark current in both detectors used was negligible ( 1count/hour/pixel) and no dark subtraction was applied. In those cases where three or more frames in one band were obtained for the same object the cosmic rays were rejected using the IRAF task IMCOMBINE. When only one or two images were available we removed the cosmic rays interactively using the IRAF tasks COSMICRAYS and CREDIT within the CRUTIL package.
Considering the relatively small number of nights that were photometric (12 out of 49) we decided to use the following strategy in order to flux-calibrate our images. During the non-photometric nights at Palomar and Las Campanas observatories we obtained very deep exposures in R and H. During the photometric nights the adopted follow-up strategy was slightly different at each site. At Palomar we took short (600-900s) B and R exposures of the same objects already observed in R and H, along with images of a large number of photometric standards (Landolt 1992a). Then, using the fluxes of the field stars in the calibrated R-band, deep R and H images along with the filter+detector response functions (see Figure 2) we cross-calibrated our deep R and H images (see Appendix A). At Las Campanas, during the photometric nights we observed both objects and spectrophotometric standards (Hamuy et al. 1992; Landolt 1992b) through all B, R and H filters. During these nights we also took short B and R exposures of the objects already observed in R and H during previous non-photometric nights. The flux calibration of the H images taken during non-photometric conditions was carried out in the same manner as the Palomar case (see Appendix A). The coefficients obtained from the calibration are given in Table 4. This strategy has allowed us to make optimal use of telescope time, achieving relatively small photometric errors (see Table 5). In order to check the reliability of the H calibration based on the use of R-band calibrated data, we compared the results obtained from this method with those resulting from the observation of spectrophotometry standards through the H filters during the Las Campanas photometric nights. The calibration results agreed to within 5%.
Note. - (a) Both the extinction coefficient and the color term are fixed. (b) The color term is fixed. Columns stand for: (1) Date corresponding to the civil time at the start of the observing night. (2) Observatory and telescope (P60: Palomar Observatory 60-inch; C100: Las Campanas Observatory du Pont 100-inch). (3) Band. (4) Photometric zero point in magnitudes. (5) Extinction coefficient in mag/airmass. (6) B - R color coefficient. (7) RMS of the calibration. (8) Number of standard stars used in the fit. (9) Notes.
The continuum-subtracted H images of our sample were obtained from the R-band and line+continuum H images making use of the equations given in Appendix A. Briefly, we compared the flux in counts of a large number of field stars ( > 30) both in the R-band and H images. Then, using the IRAF task GAUSS the image with the best seeing was convolved with a gaussian kernel to match the seeing of the worst-quality image. Finally, we divided the R-band image by the corresponding scaling factor and subtracted it from the line+continuum H image. Residuals in the continuum-subtracted images due to the presence of very bright field stars were removed interactively using the IRAF task CREDIT within the CRUTIL package. For more details about this procedure and the flux-calibration of the resulting continuum-subtracted H image the reader is refereed to Appendix A.
4 IRAF is distributed by the National Optical Astronomy Observatories, which are operated by the Association of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation. Back.