Annu. Rev. Astron. Astrophys. 1991. 29: 499-541 Copyright © 1991 by Annual Reviews. All rights reserved

4.2 Radio Techniques

The 21-cm line of atomic hydrogen is detectable in spiral galaxies, to moderate distances, and with high precision (velocity errors are usually smaller than 10 km s^-1). Tifft (1990) has carefully calibrated various 21-cm systems on a sample of 100 galaxies (see also Tifft & Cocke 1988). Lewis (1987) has reported similar high signal-to-noise, high spectral resolution work. In addition to redshift, the 21-cm profile yields from its integrated flux an estimate of the total H I mass within the galaxy, and from its width, a straightforward measure of the maximum rotation velocity. The discovery of a well-defined relation between a galaxy's luminosity and its rotation velocity (Tully & Fisher 1977) has been used extensively to obtain redshift-independent distances to galaxies, to obtain estimates of H₀, and to attempt studies of deviations from Hubble flow.

The H I line is a good tracer of tidal encounters and serves as an indirect probe of the sweeping of the cold interstellar medium by hot intracluster gas (Haynes et al 1984, Giovanelli & Haynes 1989b). Concurrent with the construction of the major aperture synthesis facilities, the development of modern multichannel spectrometers, low noise amplifiers, and high-gain feed systems has expanded the volume visible to single-dish 21-cm line surveys. To date, 21-cm redshifts have been measured for some 12,000 objects, while in 1975 that number was less than 150 (Roberts 1975).

The maximum velocity cz at which a galaxy of H I mass M_{H I} and velocity width V can be clearly detected, when observed with a single dish of diameter d, a receiver of system temperature T_sys, and an integration time t, is

An interference-free five-minute on-source observation at Arecibo can detect a normal Sb galaxy of H I mass 4.7 x 10⁹h^-2 M at cz ~ 16,000 km s^-1. Upon completion of the upgrade currently under way (see next section), the same observation will be possible in between 1/8 and 1/60 of this time, depending on the source's zenith distance.

Not all galaxies contain significant amounts of H I gas. Most notably, elliptical galaxies, at least the majority of them, do not contain detectable amounts of H I (see Wardle & Knapp 1986). Thus redshift surveys conducted via the 21-cm line concentrate on spiral and irregular objects. Galaxies of type S0 and earlier are underrepresented. Furthermore, spiral galaxies in the cores of rich clusters are severely H I deficient (Haynes et al 1984), presumably as a result of recent environmental alteration. Unseen galaxies, rich in H I but optically invisible do not contribute significantly to the mean mass density of the universe (Briggs 1990) and do not fill the voids (Eder et al 1989). On the other hand, spirals constitute the vast majority of galaxies in the field, and late-type, low surface brightness galaxies are H I-rich and even more easily visible to the 21-cm eye than to the optical one. The 21-cm line surveys are thus biased towards late-type, low-surface-brightness objects that preferentially occupy the field. In contrast, the high-surface-brightness objects typically included in magnitude-limited optical redshift surveys show a higher degree of clustering (Davis & Djorgovski 1985, Giovanelli et al 1986a). Radio and optical surveys are thus, to a large degree, complementary (see also Huchra 1987).

Important and comprehensive compilations of H I data have been presented by Bottinelli et al (1982) and more recently by Richter & Huchtmeier (1989 and refs. therein). Aperture synthesis observations are typically shallower than single dish ones, but the gain in spatial resolution adds critically to understanding the kinematics and distribution of the gas, as shown in recent studies of the Virgo cluster (Guhathakurta et al 1988, Cayatte et al 1990). Multiple-object spectroscopy with single radio dishes is more difficult. The limited held size and large receiver horns in focal planes of paraboloids or, in the case of Arecibo, the long line feeds, do not lend themselves to multiplexing. Focal plane arrays can be built for paraboloids, but their usefulness lies more in mapping extended emission regions or in blind searches than in targeted survey work.