DAYTON MILLER ETHER DRIFT EXPERIMENTS IIhttps://www.theflatearthsociety.org/forum/index.php?topic=30499.msg751624#msg751624 (part I)
And there is more...
Dear Tom Roberts,
If I could summarize again:
1) You analyzed an apparently unpublished set of data from one of
Miller's tests in Cleveland, when the most serious data which requires
attention is from his Mt. Wilson experiments. I'm sure one could find
unpublished data from Michelson as well, or from Einstein's work -- it
may have historical significance, but is not the point of discussion if
you wish to refute what provides a foundation for much of new interest
in ether and ether-drift. I have no idea why Glen Deen gave you this
data set, instead of something from the Mt. Wilson experiments. Maybe
he can clarify this.
2) The tests in Cleveland would very likely have produced a signal far
below that of the Mt. Wilson experiments, given the effect of altitude
-- higher altitudes produce higher ether-drift velocities, as
documented by Galaev. Therefore, whatever your critique of the
Cleveland 1927 experiments were, they would not apply, or apply only
less-so to the Mt. Wilson experiments of 1925-26. You cannot presume
to assert the "signal to noise" levels were the same for both sets of
experiments. That's an unproven assumption.
3) Even if we assume, the variance within the measurements for any one
of the four seasonal epochs at Mt. Wilson was large, to rest upon that
observation and go no farther is to miss the forest for the trees.
Larger patterns in data sets often are not apparent or ammenable to
analysis via statistical methodology, but rather require dynamical
methods of analysis, or sometimes graphical or
geographical-astrocartographical methods. For example:
4) I did not mean to imply that low-altitude ether-drift experiments
would yield "no signal" at all. They do, but apparently of a reduced
intensity. Consequently, we might ask if the August 1927 data which
you analyzed yielded a variation over sidereal-clock coordinates? And
if so, is this variation along the same sidereal hour axis as what
Miller noted for the Mt. Wilson experiments, even if the velocity
determination would be at a lower level? If so, that would be in
keeping with his overall theory and findings. Miller's pre-Mt.Wilson
tests in Cleveland DID occasionally show similar vectors, as did the
Morley-Miller and even the Michelson-Morley experiment. Yes, he did a
lot of testing and control experiments, as Einstein was at the time
proclaiming (without evidence) that Miller's work was the consequence
of "thermal artifacts". So he did a lot of work to show, exactly, how
the interferometer would react to both small and large external heating
effects, and precautions were undertaken, such as shielding the
interferometer arms with insulation, and so on. NONE of those
experiments -- Michelson-Morley, Morley-Miller, or Miller in Cleveland
ever produced a fully "null" or "zero" result, which by itself is
significant. But the data was best at Mt. Wilson, and likewise
Michelson-Pease-Pearson also got their best result at Mt. Wilson.
Miller addressed this consideration in the 1933 paper, and
Michelson-Morley were also aware of their own slight positive result,
stating in the 1887 paper the need to perform the experiment over other
seasonal periods -- which they never did. Only Miller did so. The
fact that all four seasonal epochs of the Mt. Wilson experiments
yielded similar sidereal-hour vectors for the axis of drift, and that
this also was the same (though reduced) axis which could be extracted
from the original Michelson-Morley experiment, is THE significant
consideration, even if the velocity determinations were slightly
variable. This is what we call a highly-structured pattern in the
data. The fact that Galaev later found a similar axis of drift in his
work, and the seasonal variations in "dark matter wind" also show a
similar pattern, is "icing on the cake" so to speak.
5) High "signal to noise" ratios plague other data sets from natural
phenomena, such as climate patterns. Daily precipitation is a function
of solar heating and shifting of wind and pressure patterns. But if we
look for variations in precipitation as an indicator of solar heating,
it requires a lot of years of data before we get a climatic curve which
approximates the smooth latitudinal shifting of the sun's location, and
hence, solar heating of the lower atmosphere. Over shorter periods,
rainfall quantities may be extremely variable with large quantities one
day or week, nothing the next day or week, and so on over the years,
with some years very wet, others in drought. If we presume ignorance
of how solar heating works to stimulate rains, we would be hard pressed
to find this pattern in all the "noise" of daily precipitation
variation. We would in fact only find the pattern by recording
precipitation over the year, and then averaging the data by week or
month. Only then, you get a pattern which is valuable, and allows some
degree of confidence and prediction of when a "rainy season" or "dry
season" will occur. Likewise also, I would imagine, with the
determinations of anisotropy in 3-deg.K. in open space -- a lot of
variation, no way to make "statistical analysis" but when it is plotted
on a map -- or along a simple graphical ordination representing
sidereal hour -- it makes a pattern which is important to consider.
Unfortunately, I have no computer-readable data files for Miller. My
role was mostly historical, basically finished after the data sets were
finally obtained, and others set out on that task. I cannot speak to
what Glen Deen and others are doing with the data. My larger interest
today is in the work of Galaev, who developed an elegant and very
simple interferometer using parallel light beams, and seems
potentially easier to use, less afflicted by vibrations, and possibly
could be rendered far more sensitive given current technology. My push
has been, for more experiments to be undertaken, rather than merely to
analyze Miller over and over. I must disagree that your DSP method
will ever critically undermine Miller's findings, if only because my
points above cannot be overcome by purely statistical arguments. If
Miller's four different seasonal epochs had yielded four different
points in the heavens, four different axes of ether-drift, then surely
a rejection of his work would be fully in order and legitimate. But I
encourage you to look again at Figure 2 in my Miller paper.
http://www.orgonelab.org/miller.htmThis shows Miller's data organized firstly by sidereal hour, and
secondly by civil clock time. By sidereal hour, there is a distinct
pattern in the data, one which appears to be robust enough even to
survive your argument about the need for error-bars. However, when the
same data is organized by civil clock time, the pattern vanishes. This
is the issue which you need to address, and it will not be defeated
with DSP methods.
As noted, I do have copies of all of Miller's data sheets, being the
guy who stimilated their re-discovery from dusty storage rooms. You
mention only the one data sheet of Figure 8 from his 1933 paper, which
showed the results of 19 turns of the interferometer over about a
15-minute period. This is like, extracting rainfall records for one
month of one year, exclaiming there is "no solar-related pattern" and
ignoring all the rainfall data from many other months and years. Sure,
look at only one data sheet, and clear determinations may be
insufficient. But really, your DSP analysis was not of that data
sheet, nor of the hundreds of other data sheets from Mt. Wilson.
I have no interests to second-guess Miller's methods, and your claims
really don't suggest any serious reason why one should be concerned.
Nobody including Michelson had any problem with Miller's methods or
findings at the time when he was doing his work, other than Einstein,
who was no expert in the ether-drift methods. In fact Miller was the
student of Morley, and learned the methods as handed down from
Michelson and Michelson-Morley. You presume to have us believe you
know more about it than they did, even though you haven't undertaken an
analysis of the very same published data from which Miller's
conclusions were derived. And all the other validating experiments,
you simply ignore. Sorry to say, this is simply insufficient.
Regards,
James DeMeo
"But we must pause at this juncture to critique Miller’s thinking process, for
he, being a Copernican, is basing his interpretation of data on his belief that
the Earth is moving at least 30 km/sec through space. Interestingly enough, it
is precisely because of this presupposition that Miller runs into some
unexplained difficulty, since his observations begin to conflict with his
mathematical calculations. The one anomaly in all past interferometer
experiments that Miller discovered was the experimenters assumed they knew the
precise velocity of the Earth through the ether in combination with the solar
system’s supposed motion toward the constellation of Hercules, but did they
really know? The geocentrist, of course, would answer that they did not know.
In any case, Miller’s 1925 experiment took into account this “anomaly” and he
made his calculations accordingly. Since he assumed the Earth was moving 30
km/sec, he combined this with the four positions (February, April, August,
September) that he examined of the Earth’s orbit around the sun and then used
Pythagorean geometry to determine the speed of the Earth toward the
constellation Draco, which came to 208 km/sec.[2] In other words, 208 km/sec is
what Miller believed to be the Earth’s absolute speed through the ether. Of
course, being a heliocentrist, Miller is assuming that the ether is motionless
and that the Earth is moving through it. In any case, Miller’s 1933 paper
reveals that his Pythagorean calculations do not match what he observed in the
fringe shifts. As we will recall, his experimental fringe shifts showed a
maximum of 10 km/sec, but this figure is less than his computed value by a
factor of twenty! Miller did not have an answer for this problem, and it is
left as an open-ended question in his 1933 paper. The answer, of course, is
that Miller’s Pythagorean calculations were based on a faulty premise (i.e.,
that the Earth was moving). If that factor were eliminated, his calculations
would be in accord with his observations. The same can be said of recent
experiments performed by Stefan Marinov, in the late 1970s, using
coupled-mirror interferometry.
Miller configured the four interferometer readings in the form of a
parallelogram (February, April, August, September), which assumes the Earth is
in orbit around the sun. The diagonal of each of the four parallelogram points
represents the apex of that period, while the long side represents the motion,
which is coincident with the center of orbit; the short side of the
parallelogram represents Earth velocity of 30 km/sec. Hence, knowing the
direction of the three sides of the triangle, and the magnitude of one side,
allows one to calculate the magnitude of the other sides, which for Miller was
208 km/sec toward Dorado."