Technical Information on the Data

All of our images were obtained with the HST f/96 FOC (see HST FOC Instrument Handbook) in its "zoomed" 1024 × 512-pixel mode with 0.022''× 0.044'' pixels, giving a field of view of 22''× 22''. The F220W filter was used. This is a broad-band filter with an effective wavelength of ~ 2300 Å and effective bandpass of ~ 500 Å. The exposure time was 10 minutes per galaxy. We will refer to this configuration and exposure time as the standard set-up. The images were processed by STScI's "pipeline" reduction as described in the HST/FOC Data Handbook, after which the pixel scale is 0.0225'' pixel^-1.

The data were obtained before the HST repair mission at the end of 1993, and therefore are affected by spherically aberrated optics. As a result, the point-spread function (PSF) consists of a sharp core of full width at half maximum (FWHM) ~ 0.05'' that contains about 15% of the light, with the rest of the light spread in a complex low-level ``halo'' with a radius of several arcseconds (Burrows et al., SPIE 1567, 284, 1991). In the observing mode we have used, the FOC is limited in its dynamic range to 255 counts (8 bits) per zoomed pixel (128 counts in the de-zoomed pixels after pipeline reduction); additional signal causes the counts to ``fold over'' and start again from 0. Another problem is that the detected count rate becomes nonlinear, gradually saturating for bright sources (see HST Data Handbook). The central pixels of most of the compact bright sources detected in the images may be in the nonlinear regime, and the brightest of them are clearly saturated. In several galaxies where we report the brightness and/or angular size of individual compact sources, our analysis relies mainly on the wings of the PSF, which have low count rates (<= 0.05 s^-1 pixel^-1), using the algorithms described by Maoz et al. (ApJ, 440, 91, 1995; ApJS, 107, 215, 1996). We model the PSF using a well-exposed F220W image of a star observed with the FOC f/96 256×256 pixel mode, which has a large dynamic and linear range (but small field of view). Such empirical PSFs are required for work in the UV. We have not performed corrections for "folding over" and nonlinearity on these data. All images have been rotated so that north is up and east is to the left. To reduce volume, the images have been multiplied by a factor of 100 and converted to integer format. Owing to the limited dynamic range of the original data, this involves negligible information loss.

As in Maoz et al. (1995, 1996), we translate the original FOC counts (i.e., after dividing back by a factor of 100) to a flux density f at 2270~Å assuming 1 count s^-1= 1.66× 10^-17 ergs^-1cm^-2Å^-1, based on the on-line calibration data available from STScI for the FOC and F220W filter, with a 25% increase in sensitivity of the 512 × 1024 zoomed-pixel mode relative to the 512× 512 pixel mode. This calibration assumes a spectrum that is constant in f. As detailed in Maoz et al. (1995), the F220W count-rate vs. f at 2270 Å is weakly dependent on the spectral slope, with a change of only a few percent for a large range in slopes. The uncertainty in the absolute flux is ~ 20% when measuring individual compact sources, but can be as small as ~5 when measuring the UV flux integrated over large areas (e.g. ~ 150 arcsec² Meurer 1995), if the background can be reliably determined. An additional concern in UV imaging photometry is the presence of ``red leaks'' through the F220W filter, which we define as light of wavelength > 3200 Å that may pass the low, but non-zero, transmission of the filter at these wavelengths. Our journal paper analyzes this question in detail, with the conclusion that red leak is small or negligible in these data.

The seven archival images which we also include in this atlas were obtained with different FOC formats, UV filters, and exposure times, providing modified fields of view, bandpasses, and sensitivities. One of these, of NGC 4151, was obtained in 1995, after the first HST servicing mission, and so is free of spherical aberration. We elaborate on these differences for each archival exposure in the notes on individual objects.