1. Introduction

  Near the beginning of the space age, Spitzer & Zabriskie (1959) outlined the significant potential of an orbiting observatory for learning much more about some important properties of the interstellar medium (ISM). Their principal point was that the ultraviolet part of the spectrum held the strong resonance transitions of the more abundant elements in stages of ionization most likely to be found in space. The absorption features arising from these transitions can provide an unparalleled sensitivity to small amounts of gas. Moreover, they pointed out that absorption by the Lyman and Werner electronic bands of molecular hydrogen provides the easiest (and perhaps only) means of detecting this molecule, whose existence in the ISM could only be hypothesized at the time. These predictions provided the principal motivation for building and launching the uv spectrometer that flew on the Copernicus satellite (Spitzer, L. & Jenkins 1975). While the instrument had excellent photometric accuracy, it had two major drawbacks: (1) it had a wavelength resolving power $R=\lambda/\Delta\lambda$ of only 20,000 (corresponding to a doppler shift of 15  km ^-1) and (2) it was a scanning spectrometer and thus not very efficient by today's standards where imaging devices are the norm.

Copernicus was followed by IUE (Boggess et al. 1978a, b), a satellite launched in early 1978 that, remarkably, is still in operation. Although IUE's telescope mirror was of modest size, it's geosynchronous orbit and imaging detectors provided for a great gain in efficiency over Copernicus. IUE was capable of viewing a broad range of astronomical objects, and it had a very high overall scientific productivity (Kondo et al. 1989). However, its value for studies of the ISM was diminished by its somewhat lower R and a limit on the signal-to-noise ratio that was almost always considerably lower than that defined by the photon counting noise.

The pace of discoveries about the diffuse ISM picked up once again when operations of the Goddard High Resolution Spectrograph (GHRS) began on the Hubble Space Telescope (HST) in 1990. GHRS has at its disposal the tremendous light gathering power of HST's 2.4M mirror, and, for the echelle spectrographs, a significant improvement in resolution: R = 80,000. GHRS can detect only 512 spectral elements at a time - a limitation that will be overcome when a more modern instrument, STIS (Space Telescope Imaging Spectrograph), is installed during the next servicing mission (Woodgate & STIS Team 1992).

Against the backdrop of the facilities just mentioned, one might ask ``What is being overlooked?'' The answer lies in an important limitation of IUE and all the spectrographs on HST. As a result of the MgF<sub>2</sub> faceplates on their detectors, these instruments have virtually no sensitivity below about 1150Å. This region of the spectrum, down to about the Lyman limit at 912Å, is often referred to as the ``far uv.'' It has a unique scientific potential for ISM studies Jenkins et al. 1988) because it is the only interval that covers transitions from the ground states of several important constituents, to name just a few, H₂, HD, O VI and most of the Lyman series of hydrogen and deuterium. In order to register any light in this spectral domain, the photosensitive surface within an instrument must be open to the vacuum environment, and the optical system must consist of only reflecting elements. The only major spectroscopic facilities to fly after Copernicus that had any sensitivity in the far uv were the Voyager spectrograph (Holberg et al. 1982), the Hopkins Ultraviolet Telescope aboard the Astro missions (Davidsen et al. 1992), and the ORFEUS telescope (Hurwitz & Bowyer 1991). However, all 3 instruments had values of R significantly lower than those of Copernicus, IUE or HST, thus severely limiting their usefulness for investigations of the ISM.

In short, with all of the aforementioned facilities we come across a gap in capability: Copernicus is no longer functioning, and the major working telescopes of today are incapable of observing in the far uv. To satisfy the objectives in interstellar research that make use of absorption lines at short wavelengths, we must rely on two contemporary instruments that fill the gap: one is the Interstellar Medium Absorption Profile Spectrograph (IMAPS), the subject of this article, and the other is a major observatory called FUSE (Far Ultraviolet Spectroscopic Explorer) (Moos & Friedman 1991) that is now being developed for a launch in 1998. These two instruments are complementary. As discussed in the next paragraph, the principal strength of IMAPS is its very high spectral resolution, surpassing that of even the GHRS echelle spectrograph on HST. FUSE will have a lower, but still adequate resolving power for ISM studies, but it should be able to study much fainter sources.

An important characteristic of the IMAPS optical design is the unusually long focal length of the beam that converges onto the detector. This feature, along with the large incidence and diffraction angles of its echelle grating, renders IMAPS capable of producing a spectrum with R = 240,000. Studies of interstellar absorption lines at high resolution from ground-based telescopes (Wayte, Wynne-Jones, &; Blades 1978; Blades, Wynne-Jones, & Wayte 1980; Welty, Hobbs, & Kulkarni 1994) show that there is considerable fine detail in the velocity structure of the features, even surpassing that which can be resolved by IMAPS. Often, gases with very different properties are found to have differences in radial velocity v that are very small, and successful interpretations must rely on observations with sufficient resolution to separate them. Even when there is no need to separate adjacent features, there is a substantial improvement in accuracy in a column density when one can go beyond just a simple measure of a moderately saturated line's equivalent width and, instead, trace out its optical depth as a function of v. This advantage holds even if the finest details are not resolved and two or more lines of different strength can be observed (