Annu. Rev. Astron. Astrophys. 1995. 33: 581-624 Copyright © 1995 by Annual Reviews. All rights reserved

1. INTRODUCTION

The idea that active galactic nuclei (AGNs) are powered by accretion onto supermassive black holes (BHs) is based on compelling theoretical arguments (Salpeter 1964; Zel'dovich 1964; Lynden-Bell 1969, 1978; Lynden-Bell & Rees 1971; see Rees 1984; Begelman et al. 1984; Blandford 1990; and Blandford & Rees 1992 for reviews), a broad array of circumstantial evidence, and a few incisive observations. The latter include rapid time variability and superluminal jets, which suggest that AGN engines are relativistically compact. Theory predicts BH masses of 10⁶-10^9.5 M. But the BH mass function and the distribution of BHs among galaxies depend on AGN lifetimes and are poorly constrained. This much is clear: quasars were overwhelmingly more numerous at z 2 than they are now, so dead quasar engines should be hiding in many nearby galaxies.

The weak spot in this picture is the lack of proof that BH engines exist. This motivates many searches. But a great deal is at stake. The BH picture has become our paradigm. This situation is dangerous: it is easy to believe that we have proved what we expect to find.

This paper reviews stellar- and gas-dynamical evidence for BHs. Previous reviews are given by Sargent (1987); Richstone (1988, 1993); Filippenko (1988); Dressler (1989); Davies (1989); Gerhard (1992); Kormendy (1992a, b; 1993, 1994); de Zeeuw 1994; and van der Marel (1994c). Gas is very responsive to non-gravitational forces, so the most definitive searches are based on stellar kinematics. Section 2 summarizes the state of the search and lists the eight detections. Section 3 shows why cuspy brightness profiles are not evidence for BHs. After a summary of search techniques, Section 4 discusses the stellar-dynamical BH candidates. An important upper limit is discussed in Section 4.8. Section 5 discusses the detections based on gas dynamics. A preliminary review of BH demographics is given in Section 6. Section 7 considers whether the detected candidates are too inactive to be BHs. Conclusions and a wish list are given in Section 8.

The case is now strong that 10⁶- to 10^9.5 M dark objects have been discovered in a few galaxies. All are objects in which the search is relatively easy because circumstances are favorable (e.g., the center rotates rapidly). In most galaxies, we do not even have limits; when they are available, they are weak. This especially includes the giant ellipticals that motivate us to search because of AGN activity. Despite the difficulty of BH detection, the amount of nuclear dark matter is consistent with theoretical predictions. But the rigor stops here. Suggestions that the dark objects are BHs are based only on indirect astrophysical arguments.

What we need next is a significant iteration in the quality of the data. We need to see whether the BH case gets stronger or weaker as the spatial resolution and modeling techniques improve. We also need better constraints on the radius inside which the dark mass lies to confirm or exclude alternatives to BHs. Improved ground-based instruments and especially the Hubble Space Telescope (HST) are beginning to provide this iteration. Early results are encouraging. But the case is not ready to go to the jury. By themselves, neither the dynamics nor the AGN observations are conclusive. Together, they make a compelling picture.

Our title does homage to A. Pais's (1986) book Inward Bound, an elegant history of twentieth-century physics as it probed to smaller and smaller subatomic length scales. The BH search is similarly a journey. We make incremental improvements in spatial resolution, each expensive in ingenuity and money. This paper reviews the first order of magnitude of the inward journey in radius. But we are still observing at 10⁵ Schwarzschild radii. Surprises are still possible on the way to the center.