Steve Willner <swillner@cfa.harvard.edu>

The effects of gravitational waves are ridiculously weak, and direct
evidence for their existence has (probably) not been found with the
detectors built to date.  However, no known type of source would emit
gravitational waves strong enough for detection, so no one is worried.

In the 60's and early 70's, Joe Weber at the University of Maryland
attempted to detect gravitational waves using large aluminum bars,
which would vibrate if a gravitational wave came by.  Because local
causes also created vibrations, the technique was to look for
coincidences between two or more detectors some distance apart.  Weber
claimed to see more coincidences than expected statistically and even
to see a correlation with sidereal time.  Unfortunately, other groups
have used far more sensitive detectors operating on the same
principles and found nothing.

Two new experiments, far more sensitive than those using metal bars, are
being built now.  These are LIGO in the US and Virgo in Italy.  They
will work by detecting displacements between two elements separated by
several kilometers.

An indirect measurement of gravitational waves has been made, however.
Gravitational waves are formed when a mass undergoes change of
acceleration.  They are stronger if the mass is dense and the
acceleration changes rapidly.  One place where this might happen would
be two pulsars circling each other.  A couple of systems like this
exist, and one has been studied actively over the past 20 years or so.
Pulsars make good clocks so you can time the orbital period of the
pulsars quite easily.  As the pulsars circle, they emit gravitational
waves, and these waves remove energy (and angular momentum) from the
system.  The energy released has to come from somewhere, and that
somewhere is the orbital energy of the pulsars themselves.  This leads
to the pulsars becoming closer and closer over time.  A formula was
worked out for this effect, and the observed pulsars match it amazingly
well.  So well, in fact, that if you plot the data on top of the
prediction, there is no apparent deviation.  (It's actually rather
disgusting, none of my results ever come out that well.)  Anyway, Joe
Taylor of Princeton and a student of his, Russell Hulse, shared the
Nobel Prize in Physics for, in part, this work.

Useful references are given in section D.03.

V. M. Kaspi discusses pulsar timing in 1995 April Sky & Telescope, p. 18.

The conference proceedings volume _General Relativity and Gravitation
1989_, eds. Ashby, Bartlett, & Wyss, (Cambridge U. Press 1990) contains
a summary of the aluminum bar results.

_General Relativity and Gravitation 1992_, eds. Gleiser, Kozameh, &
Moreschi (IOP Publishing 1993) contains an article by Joe Taylor
summarizing the pulsar results.

An example of recent pulsar research is the article by Kaspi, Taylor,
and Ryba, 1994 ApJ 428, 713, who give instructions for obtaining their
archival timing data via Internet.

Some references to Weber's work are:
  1969 Phys. Rev. Lett. 22, 1320.
  1970 Phys. Rev. Lett. 24, 276.
  1971 Nuovo Cimento 4B, 199.

Information on gravitational wave detection experiments can be found
on the Web for LIGO <URL:http://www.ligo.caltech.edu/>, VIRGO
<URL:http://www.virgo.infn.it/>, GEO 600
<URL:http://www.geo600.uni-hannover.de/>, and TAMA
<URL:http://tamago.mtk.nao.ac.jp/>.