The 38 slides from Weiss's talk are here in
9 web pages. In the narrative below you can
click on any subject to go to its
page. Click here for a full
PDF version.

pg. 1 GR is one of
several possible covariant theories of gravity. It is only from
observation and experiment that we
conclude that GR is the valid theory. So, Weiss argues, we should base
the teaching of GR on experiment and
observation. To make his points, he uses the historically important
examples of gravitational red shift, bending of light, advance of
perihelion of Mercury, Shapiro test, and
Nordstrom's scalar theory of
gravity, which, though
Lorentz
covariant, fails to predict correctly the precession of perihelion
of Mercury or the gravitational redshift.

pg. 2 You
can motivate the need for curved space by imagining coordinates on a
rotating platform.
Basic phenomena, such as gravitational redshift, can be inferred from
the principle of equivalence.
The difficulty comes when you try to measure something: For the redshift
there are subtle
corrections to be determined and made.

pg. 4 Data from 9
solar eclipses show
how imprecise these measurements were. There are no error bars, but tiny
changes in the position or orientation of the plates or small shifts in
the telescope from thermal variations occurring as the Sun was eclipsed
would have produced large fractional errors in the results. There were
other difficulties
too.

pg. 6 Weiss
contrasts the rate of precession of
Mercury's perihelion with
that of the apsides of the Hulse Taylor binary
pulsar. Where the old measurements were only weak evidence for GR,
modern measurements with precisions of ppm
are strong evidence that GR is right. These data also yield convincing
evidence for
gravitational radiation as well as precise
masses of the pulsar and its
companion. Thirty years of observation continue to improve precision.

pg. 7Shapiro time delay
provides a good context for teaching about coordinate freedom. To test
two different theories you must make complete consistent calculations
with each. Shapiro had to express the results in coordinate free form
(as invariants), and then compare. He found that Einstein's GR predicted
what was observed; Newton's theory did not.

[Cliff Will's book
Was Einstein
Right? gives a nice account of Shapiro's experiments that first
detected this effect. --editor]

pg. 8 Shortly after
publishing his GR theory, Einstein showed that there should be
gravitational radiation. To remind us that there are important
subtleties in this prediction Weiss points out two mistakes that
Einstein made in his first papers on gravity waves: 1) He
erroneously
predicted gravitational radiation from spherically symmetric motion,
which we now understand to be impossible; 2) He made a
factor of 2 error when
calculated the quadrupolar radiation that is the dominant form of
gravitational radiation.

pg. 9 The take-home
message: In the effort to detect gravitational radiation there is much
other physics than gravity
itself. Use it to interest your students. And
motivate them with the elegance
of the technical prowess and instruments needed to examine the
physical consequences of GR.

The table below shows the elaborate analysis needed to show that
GR is needed to account for 42.56 arc sec per century of precession of the
perihelion of
Mercury
The Hulse Taylor binary pulsar 1913+16, on the other hand,
exhibits a GR induced precession rate of 4.2 deg per year. Within 14 years
many of its parameters have been measured with precisions of parts per million.

The above graph on the left shows the measured advance of the
apsides of 1913+16 over 14 years. The line is the GR prediction. The upper
graph at the right shows the residuals between the measured period and the
predictions of GR. The bottom graph on the right shows the residuals when
the GR corrections are omitted. Notice that the scales of the two graphs differ
by a factor of 20. Clearly GR is needed to account for these observations.

Data from 1913+16 also provided the first clear evidence of
gravitational radiation. The graph below shows the cumulative shift of the
times of periastron passage of the pulsar vs. time. The solid line is the
slow-down that GR predicts will occur because of gravitational radiation. This
is the first convincing evidence that there is gravitational radiation.

Combined with GR theory, the measurable properties of the pulsar
binary system constrain the system's masses in different ways. When the
constraints are graphed as on the right above, the masses of the pulsar and its
companion are found to be respectively 1.441 and 1.387 solar masses.