3. OBSERVATIONS AND DATA REDUCTION

3.1. UIT Imaging

The ultraviolet images of galaxies presented here were obtained by the UIT, which was one of the three copointed UV telescopes of the Astro observatory. UIT was designed and constructed at NASA-Goddard Space Flight Center by a team headed by T. P. Stecher. The final instrument is an outgrowth of a series of prototype experiments carried on sounding rockets in the 1970s and 1980s for missions of ~ 5 minute duration, from which a number of UV images of galaxies and star clusters with lower spatial resolution were published. UIT flew on two Space Shuttle missions, Astro-1 in 1990 December and Astro-2 in 1995 March.

A complete description of UIT, its performance, and standard data processing and calibration procedures is given in Stecher et al. (1992, 1997). UIT is an f/9 Ritchey-Chrètien telescope with a 38 cm aperture. It contains two cameras, each of which consists of a magnetically focused, two-stage image intensifier with a phosphor output which is fiber-optically coupled to 70 mm film. The mid-ultraviolet (MUV, 1800-3200 Å) camera, designated "A," employs a Cs₂Te photocathode. The far-ultraviolet (FUV, 1200-1800 Å) camera, designated "B," has a CsI photocathode. The photocathode responses combined with the filters yield excellent long-wave rejection; no evidence of "red leaks" has been detected on extragalactic images. The UIT field of view is 40' in diameter. There is some vignetting and reduced image quality in the outer 2' of the field, but most of the images in this atlas are taken from the field center. Fine guidance during exposures is accomplished by an articulated secondary mirror coupled to an external star tracker and gyroscope assembly.

The cameras carried a total of 11 filters, but the images discussed here were taken through only three of these, listed in Table 1. A1 and B1 are the primary broadband filters for the MUV and FUV cameras, respectively. B5 is an FUV filter with a slightly reduced bandwidth to avoid strong airglow emission lines shortward of 1450 Å. This filter was primarily used during orbital daytime exposures. Note that the effective wavelength of the very broad A1 filter depends on the spectral energy distribution of the source. UIT filter sensitivity functions are displayed in Stecher et al. (1992, 1997).

Table 1. UIT Filters

Filter (1)	Substrate (2)	eff (Å) (3)	Peak (Å) (4)	(Å) (5)
Mid-ultraviolet:
A1...	Fused SiO2	2488	2763	1147
Far-ultraviolet:
B1...	SrF2	1521	1443	354
B5...	Crystal SiO2	1615	1518	225

The developed film was digitized using Perkin-Elmer 1010m microdensitometers. Astro-1 data frames were scanned with a 20 µm square aperture and a sample spacing of 20 µm. Astro-2 frames were scanned with a 20 µm square aperture but with a sample spacing of 10 µm; frames used in this study were then box-averaged to a 20 µm pixel. Standard processing steps for the densitometered data included background fog subtraction, linearization, flat-fielding, and flux calibration. Flux calibration and uncertainties are described in more detail in Section 4. Astrometry for most UIT frames was obtained using the Hubble Space Telescope Guide Star Catalog (Lasker et al. 1990).

This standard "pipeline" data reduction process produces 2048 × 2048 images with pixel sizes 1".14 square containing the circular 40' diameter data field surrounded by a section of unexposed film. Calibrated data frames are produced in the observed spatial orientation as defined by the spacecraft roll angle, and in a north-up, east-left orientation, corrected for the geometric distortion induced by the UIT image intensifier. These, and frames at several other steps in the pipeline process, can be obtained electronically from the Multimission Archive at the Space Telescope Science Institute.

Almost all of the data discussed here were derived from the inner 80% of the full UIT field, for which vignetting is negligible. Typical values of the full width at half-maximum (FWHM) of the point spread function in this region are 2".7 and 3".0 for the MUV and FUV cameras, respectively, although individual frames may yield larger FWHMs in the case of pointing system disturbances. The point spread function is approximately Gaussian, with an extended tail at large radii.

3.2. Ground-based Optical Imaging

A primary goal of the UV imaging program was to compare galaxy morphologies across a wide spectral range. Therefore, we carried out an optical band imaging survey of the UIT sample to provide the ancillary data necessary for quantitative analysis of their stellar and interstellar components. Wide-field CCD images in B, V, R, and H were obtained between 1991 and 1994 at Kitt Peak, Cerro Tololo, and Mount Laguna Observatories. Table 2 presents the location and dates of observing runs associated with the data set described in this paper. Given the large angular extent of some UIT targets, we required CCDs with the largest available format. Many of these observations were obtained with the first generation of scientific-grade 2K chips. Characteristics of each observing set are described below. Digital versions of some of the optical imagery were presented by Cheng et al. (1996).

Standard data reductions (bias and overscan correction, flat-fielding, cosmic ray rejection, and cosmetic corrections) for ground-based frames were performed using IRAF. Additional processing, which included astrometic solutions, artifact masking, registration to the UV imagery, and photometric calibration were carried out using software developed by us in IDL. ¹⁶ These tasks are further detailed in Section 4.