On Mar 15, 12:11 pm, "George Dishman"
wrote:

Henry should enter the values from the above paper
and get his program to calculate the residuals.
Anything else is just handwaving. Now that you have
found the reference, he really has no excuse not to.

Since the link to Gieren (1985) got truncated, I'll repeat
it he
http://articles.adsabs.harvard.edu/c...6A...148..138G

Please refer to Figure 3, the composite radial velocity curve.
An extremely important point to make, is that most of the
scatter evident in this composite curve is -not- due to random
measurement error, but rather represents cycle-to-cycle
variations in the shape and timing of individual Cepheid
pulsations. Random scatter in Gieren's measurements amounted
about +/-0.5 km/s, while random measurement error in Duncan's
1908 curves amounted to somewhat over +/- 1 km/s. The
scatter evident in the Figure 3 composite curve considerably
exceeds anything that can be attributed to measurement error.

The photometric measurement technique available around the turn
of the last century (measuring the photographic density of
deliberately out-of-focus stellar images) was accurate to
several hundredths of a magnitude. Because of this high
accuracy of measurement, random variation in the cycle-to-cycle
timing of Cepheid luminosity curves ("period noise") as well as
variations in cycle-to-cycle maxima/minima ("amplitude noise")
were well established by the time that Shapley wrote his
seminal 1914 article, "On the Nature and Cause of Cepheid
Variation".
http://adsabs.harvard.edu/abs/1914ApJ....40..448S

Prior to 1914, the dominant opinion was that Cepheids were
probably a form of binary star, whose variable brightness might
be due to such effects as, say, tidal influences. (Indeed,
Duncan's 1908 paper dwells at length on the possibility of
Cepheid variation being due to an asymmetric rotating layer of
absorbing material.) After 1914, it became generally recognized
that the binary star hypothesis could not be reconciled with
the existence of period noise and amplitude noise in Cepheid
luminosity curves, and so the binary star hypothesis was
abandoned.

As you have pointed out to Henri, a requirement for stable
orbits places severe constraints on the the types of variation
which could be accommodated in a Cepheid luminosity or radial
velocity curve. One cannot just toss in a third body to try
to explain period noise.

Jerry