Annu. Rev. Astron. Astrophys. 1978. 16: 103-39 Copyright © 1978 by Annual Reviews. All rights reserved

1. INTRODUCTION

The rotation of spiral nebulae was recognized by Wolf (1914) in M81 and Slipher (1914) in M104 from the inclination of the stellar absorption lines on spectra of the central regions, although the identification of these nebulae as disk galaxies was at that time controversial. That the discovery was made first with absorption lines instead of the more readily observable emission lines can be attributed to observational selection: the central regions of the nebulae were brightest, and the presence there of absorption lines had been known for a long time (e.g. Scheiner 1899). However, we now know that these same central areas are often deficient in the HII regions responsible for the emission lines. The step from rudimentary measurements of rotation in the central regions to a map of the radial velocity field over the whole image of a galaxy has proven to be a laborious one, requiring a combination of modern optical and radio methods of observation. For the same two galaxies in which the original discovery of the rotation using the absorption lines was made 65 years ago, it is only recently that reasonably complete pictures of the distribution of radial velocities have been constructed (e.g. Münch 1959, Rots 1974, Faber et al. 1977).

The first observations to result in a plot of the radial velocity versus radius were made by Pease (1916, 1918) for M31 and M104. It was a tedious business, with exposure times of about 80 hours for each absorption spectrum, but the results on M104 showed that the radial velocities relative to the center increased linearly with radius, reaching more than 300 km/sec at a distance of 2'.5. Pease's measurements along the minor axis of M31 indicated that the radial velocity was nearly constant at all positions, showing that the variation observed along the major axis" .... is without doubt to be attributed to the rotation of the nebula." Although Pease's rotation curve of M104 does not agree too well with recent determinations (Faber et al. 1977, Schweizer 1978), his achievement is remarkable considering the relatively primitive instrumentation available to him at the time.

Emission lines from HII regions in the spiral nebulae were discovered about the same time as "nebular rotation." Pease (1915) recorded [OIII] 5007, H, and H in a 34 1/2-hour exposure of NGC 604 in M33, and Slipher (1915) also detected emission lines in this object. However, for many years emission lines remained of limited use in studies of the rotation of galaxies; although exposure times were significantly shorter, a single spectrum provided at most only a few measured points in the galaxy owing to the discrete nature of the HII regions. In his classical study of the rotation of M31, Babcock (1939) determined 44 velocity measurements in the inner regions from 236 hours of exposure on the absorption lines; an additional 56 hours of exposure on emission lines yielded only 4 more points. These 4 points were, however, very important, for they permitted extension of the rotation curve a factor of 3 further out into the main part of the disk. Later Mayall (1951) increased the emission line observations of M31 to 32 HII regions. Mayall & Aller (1942) managed to observe the emission lines in M33 with up to three HII regions per exposure using a judicious choice of slit orientation; nevertheless, 25 data points required 311 hours of observations. Even the advent of modern image tubes has not led to a very spectacular reduction in the observing time; the detailed study of the HII region velocities in M31 by Rubin & Ford (1970) still required almost 112 hours of exposure for 67 data points.

Galaxies of smaller angular size are clearly more amenable to the trick of placing several H II regions in the spectrograph slit, and the long-slit techniques employed by Babcock (1939) and by Mayall (1948, 1951) were applied to such galaxies in particular by Burbidge and Burbidge (see Burbidge & Burbidge 1975 for a review). The slit was usually oriented along the major axis of the galaxy, so that the H II regions produce emission lines over a large radial extent on a single spectrum. The use of emission lines has the further advantages that the measurements are more accurate and refer to a discrete position in the plane of the galaxy rather than an average along the line of sight. This method has yielded good rotation curves usually from only a few spectra taken near the major axis of each galaxy, often with a check from spectra near the minor axis. The procedure requires 10-20 hours of exposure. It is of course not applicable to the early-type galaxies, which are deficient in HII regions.

Determination of the velocity field over the entire disk of a galaxy would provide important additional information; the amount and extent of noncircular motions would be more clearly revealed; and the orientation parameters could be determined from the velocity field itself rather than from the distribution of broad-band light over the galaxy. The use of modern, fast image tubes (e.g. van der Kruit 1976b) permits long-slit spectra to be obtained in many position angles without requiring exorbitant amounts of telescope time. Fabry-Perot interferometers (Courtes 1960) are beginning to provide detailed data over the relatively larger fields of the more nearby galaxies with only a few hours of exposure (e.g. Tully 1974a), although the reduction of the interferograms is lengthy and complicated in comparison with the relatively straightforward treatment of long-slit spectra.

It has been known for some time that the neutral hydrogen gas in late-type galaxies extends to relatively large radii (e.g. Roberts 1972), so that mapping of the H I velocity field would provide not only independent but also complementary information on the kinematics. The complementarity of the radio-HI observations is further reinforced by two other limitations. First, the relatively lower angular resolution of radio telescopes prevents useful measurements in the central regions of galaxies where the large-velocity gradients are smoothed over the telescope beam. Second, many intermediate-type galaxies (e.g. M31, M81) have little H I in the central regions anyway. The velocity fields in the central regions of most galaxies can therefore be measured at present only by optical means; on the other hand, the HI observations provide information at larger radial distances where much of the total mass and most of the angular momentum is to be found. The complementarity is further useful in more detailed studies of the kinematics of spiral arms: HII regions are often grouped in long, thin structures defining the arms, so that it is difficult to measure emission-line velocities between the arms. However, sufficient neutral gas is usually present between the spiral arms to permit reliable velocity measurements of the HI there.

The first detailed observations of the H I in M31 were made by van der Hulst et al. (1957) using the 25 m Dwingeloo radio telescope with an angular resolution of 0°.6. The measurements resulted in a rotation curve from 0°.6 to 2°.5 from the center, a significant extension beyond the 1.°5 radial distance of Babcock's last point. With a similar angular resolution but improved receiving equipment, Argyle (1965) measured HI profiles over the whole image of M31 and was the first to plot a complete radial velocity field for any galaxy. Modern aperture-synthesis radio telescopes now provide the angular resolution (0'.5-1'.0) needed to separate spiral arms in the largest galaxies out to distances of about 5 Mpc, and are also sufficiently sensitive to study the larger-scale kinematics of many more systems at greater distances.

The radio-synthesis maps that have been published in the last few years have required of the order of 200 hours of observing time per galaxy; however, as is also the case at optical wavelengths, the newest instrumentation now coming into operation promises to reduce the time taken at the telescope by as much as a factor of 10.

In this chapter we review the present status of kinematical studies of intermediate- and late-type galaxies. We do not discuss results on E or SO galaxies, nor' the details of velocity dispersion measurements in the central regions of spirals. The emphasis is on observations that result in extensive maps of velocity fields rather than determinations of rotation curves from major axis observations only. Earlier discussions on several topics covered in this review have been given by Oort (1974) and Allen (1975a), and the interpretation of the kinematical information in terms of galaxy dynamics has been reviewed by Burbidge & Burbidge (1975), de Vaucouleurs & Freeman (1973), Freeman (1975), Roberts (1975b), and Toomre (1977). We believe our literature list to be complete up to September 1977, including preprints that were available to us at that time.