At 3/4/02 12:00 AM, you wrote:
>        The figure "0.217 inch" implies uncertainty of only a few thou.
>What instruments existed in 1740 to measure such a distance within a few
>thou?  Is the distance being measured *defined* that precisely?
>        Spurious precision can arise in convering fractions to decimals.
>One sixth of an inch, for instance ....


Well, looking at Reaumur's figures, I deduced that he must have done what we do, measure ten cells and divide by ten. He got 2 and 2 twelfths inches for ten, which I would represent as 2.17". If this is divided by ten, I would get .217". I don't see how else you would wish to represent this.

Measuring my own natural comb, I got 2 and one eighth inches per ten, or 2.125" which I would represent as .213" per cell.  How else can you show that these two figures are different (which they are)? Would you round Reaumurs's to .22 and mine to .21? Or call them both .2"?

I submit that this *was* done, that the figure was often rounded to .2" and that may account for the difference in the statements made by writers from the 1800s. One *can* obtain very accurate measurements by using more than one cell. This not only increases the accuracy but renders the final result as a *mean* rather than a particular measurement.

Of course, Reaumur was wrong when he said the number was invariable. That is why some writers (Crane) do not refer to the mean but rather, the range. But rather than pick apart the work of others, why don't you *add* something to the discussion?