New research, funded by NASA and the University of
Tokyo, has shown astronomers may not yet have uncovered the
mystery at the heart of one of the Galaxy’s oldest star
systems, the globular cluster M15.

In September, observations made using NASA’s Hubble Space
Telescope (HST) seemed to indicate the presence of a mid-size
black hole, several thousand times more massive than the Sun,
in the hearts of two clusters of stars. The observations were
the foundation for a claim that intermediate-mass black holes
had at last been discovered. By comparison, stellar-mass
black holes are only a few times the mass of our Sun.
Galactic-center black holes are millions or even billions of
times more massive than our Sun.

New detailed computer simulations, published in the
Astrophysical Journal, show a different way to interpret the
same data. Instead of a new mid-size black hole of unknown
origin in M15, the Hubble data have a more mundane
explanation – a dense clump of stellar remnants, the products
of normal stellar evolution, in the cluster’s core. The
cluster may contain a small black hole or maybe none at all.
The simulations did not address the other star cluster,
called G1, in which evidence for an intermediate-mass black
hole was reported at around the same time as the M15 results
were released.

The question of what lurks at the center of M15 has kept
astronomers busy for at least two decades. At various times,
claims have been made that M15 must harbor a central black
hole. The claims were based on the high density of stars in
the cluster’s congested core and on other tantalizing hints.
Hints like stars near the center seemed to be orbiting faster
than would be expected if only their own gravity were binding
them to the cluster. This would be the telltale gravitational
signature of an unseen mass hiding in M15’s core.

A team of astronomers used the HST for several years to probe
M15’s secrets. A few months ago, the answer to the decades-
long question of the possible black hole in M15 seemed to be
at hand. After a painstaking analysis of HST data, using
earlier model calculations by Indiana University researchers,
the team reported evidence for a central black hole.

Within hours of the team’s announcement, astronomers around
the world were carefully studying the paper and its exciting
conclusions, which were also rather surprising, since the
Indiana group had previously published models that produced
high velocities near the center of M15 without the need for a
black hole.

An international group of researchers, using the world’s
fastest computer, the GRAPE-6 system in Japan, were engaged
in a series of simulations of star clusters that resembled
M15. They used the GRAPE-6 to perform independent tests of
the M15 black hole claim. The GRAPE’s detailed, star-by-star
simulations represent the state of the art in cluster
modeling. Using this unique tool, the team found they could
reproduce the M15 observations without the need for a central
black hole. The GRAPE team’s members are Jun Makino and
Holger Baumgardt, of Tokyo University; Piet Hut, of the
Institute for Advanced Study in Princeton, N.J.; Steve
McMillan, from Drexel University in Philadelphia; and Simon
Zwart, from the University of Amsterdam in the Netherlands.

When the GRAPE researchers notified the HST team of their
results, they learned the Hubble team, along with members of
the Indiana University group, had reached a similar
conclusion. All three groups agreed that a black hole, if
present, had to be significantly smaller than originally
reported. One of the figures in the original paper, published
by the Indiana group, had been labeled incorrectly, throwing
off the later analysis of the Hubble observations.

The GRAPE group’s results appear in the Jan. 1, 2003, issue
of the Astrophysical Journal. The amended Hubble results are
in January’s Astronomical Journal. An addendum to the earlier
paper by the Indiana group will appear in the March 1, 2003,
issue of the Astrophysical Journal.