The diversity of AGN properties has been a source of confusion for astronomers; however, when they are considered carefully, certain common traits make them easier to understand as a whole. Foremost among these traits are their energy requirements.
The fact that AGNs are observable at all despite their immense
distances imposes stringent constraints on their energy sources.
Nuclear burning is not efficient enough. Even if all available matter
is fused into iron nuclei, less than 1% of the available mass would
be converted into energy. To produce the necessary energy to drive an
AGN over a significant timescale, a substantial fraction of the mass
of the entire galaxy would have to be consumed. Such a process would
leave a high concentration of residual mass near the galactic nuclei,
but there is no evidence of such massive relics. In addition, there's
the problem of delivering this much mass from a galaxy tens of
kiloparsecs across into a region less than 1/4 of a parsec
across [Miyoshi et al., 1995]. The only known process that can reasonably
produce the necessary power is the release of gravitational potential
energy from a relativistic potential well. Theoretically, a rotating
black hole could convert nearly 30% of the infalling mass into
radiant energy. Even if we assume a more modest efficiency of 10%,
the constraints of providing fuel are much more reasonable, and the
remnant masses are consistent with observations. This has lead
astronomers to believe that AGNs are powered by accreting black holes
with masses of 
[Lynden-Bell, 1969].
Numerous observations have provided much evidence (albeit
circumstantial) in support of the supermassive black hole hypothesis.
For instance, the extreme velocity dispersion observed at the center
of M87 is most plausibly explained by the existence of a black hole
with a mass of 
[Young et al., 1978, Sargent et al., 1978]. Stronger evidence is found at the core of
NGC 4258, where Miyoshi et al. have observed water maser emission from
a swirling cloud in perfect Keplerian orbit, whose properties imply a
central mass of 
contained within a radius
of at most 0.13 pc [Miyoshi et al., 1995]. One of the best pieces of evidence
yet found for a supermassive black hole at the center of an active
galaxy is the Fe K- 
feature from the Seyfert 1 galaxy
MGC-6-30-15, where the relativistically broad emission line is
gravitationally reddened and appears to come from the innermost
regions of an accretion disk around a spinning black
hole [Tanaka et al., 1995, Iwasawa et al., 1996]. (Evidence has even been found for the
existence of supermassive black holes at the centers of non-active
galaxies. At the AAS meeting earlier this month, Andreas Eckart of
the Max Planck Institute presented data based upon stellar proper
motions which suggest that an object with a mass of 
lies at the center of our own galaxy.)