Bill Potts wrote:
>
> In October, Ken Alder, the author of The Measure of All Things, gave a
> one-hour talk (including a question and answer period) at a bookstore in
> Winnetka, IL.
....
> He sees the recognition of the error as the beginning of an awareness of the
> concept of uncertainty and points out that, in any case, the very slight
> "lumpiness" of the Earth gives rise to variations in the length of a
> meridian, depending on which meridian is measured.
....
I've read the book and Alder's remarks match his writing. His narrative
takes pains to point out that the main goal of the Delambre-Mechain
survey of the meridian was improved precision. Indeed, it had been
surveyed only a few decades earlier and to a precision that would
astonish the modern reader. (Indeed, that prior survey was argued by
many in the Academy to obviate the need for another, but proponents of
greater precision won the day.)
The motivations behind this quest included a desire to base important
things (such as measures) on natural phenomena instead of human
artifacts and therefore to strive with utmost diligence for greater
precision in observing those phenomena. The newly developed Borda
repeating circle was the tool that the surveyors expected to help them
accomplish this latter goal. (Some models of this circle were calibrated
in degrees and some in grads.) Yet, the impracticality of having every
user of the meter doing his own survey was what led to the manufacture
of the meter-bar standard. And that bar was then expected to be the
prototype (modeled on and representing the Earth) for all
standardization comparisons. At this time, the need for an artifact as a
standard was never doubted as a practical realization; what was
considered necessary was to divest society of anthropomorphic standards
and models.
In fact, the quest for greater precision succeeded all too well, for
the tastes of some. It was so precise that the scientists discovered the
Earth to be not a regular ellipsoid of revolution, with a set and
uniform "flatness" (oblation factor), but in fact a lumpy thing in which
the flatness-value changed with latitude along the survey meridian. The
results of other surveys (such as the one in Peru) supported this
inferrence. Therefor it was expected that similar surveys along other
meridian lines might obtain differing results!
There was an expectation that non-uniforities of mass distribution (as
in mountains and valleys) would influence determinations of "straight
down" by a plumb line.* As Alder pointed out, even Newton knew this and
he had tried to estimate the effects of a nearby mountain on a plumb
line. Plumb lines are important in this survey business because the
latitudes measured along this meridian were determined by celestial
navigation. Since no ocean horizon was available, plumb lines had to be
used to measure zenith distances. Subtracting those from a right angle
gave corrected, observed altitudes of the celestial bodies. But the
discovery that the Earth is in fact a lumpy body (today described by a
geopotential model based on a collection of local geoids) was a
startling discovery. What that means is that, even if one could measure
true sea level at various places over the Earth, one would see a lumpy
ocean due to variations in gravitational attraction forces! This gave
birth to modern theories in planetary geology and the realization that
our view of the Earth must be based on more than a few simple
measurements, such as total mass, volume, and eccentricity.
The other great development was the development of handling
non-identical repeated measurements. At the time of the survey, the
scientist was expected to "throw out" discrepant results. (Even
maintaining a record of such discrepant results and acknowledging their
ejection was a rather new concept not held by many scientists.) The
results of the Delambre-Mechain survey were seminal in providing the
inspiration for the development of the main statistical error analysis
techniques that we today benefit from. When analyzed using modern error
analysis, Mechain made no errors, though he committed the solecism of
trying to cover up what he thought was a discrepancy.
Therefore, we have the survey (and thus the founders of the metric
system) to thank for modern geology's model of the Earth and for
demonstrating the usefulness, indeed the necessity, of statistical error
analysis. In addition, the founders gave us a start in basing our
measurements on natural phenomena, though at times realized through
human artifice, as well as a hunger for ever greater precision. This was
perhaps France's greatest gift to the world and to modern science.
Were the time available in my courses, I would require all my physics
students to read this book.
Jim
* It was already recognized that the perpendicular to the horizon was
along a line that did not pass through the center of the Earth. In fact,
this is the basis for the need to know the flatness factor's value; with
this, corrections could be made to determine geological latitude.
--
Metric Methods(SM) "Don't be late to metricate!"
James R. Frysinger, LCAMS http://www.metricmethods.com/
10 Captiva Row e-mail: [EMAIL PROTECTED]
Charleston, SC 29407 phone: 843.225.6789
