10

Logic II 10

then, as [Compton?] first made perfectly clear (Phil. Trans.).
that curve must be either the ordinary probability curve
or a distortion of it which still leaves it with a single maximum.
In my memoir 'On the Theory of the Errors of Observations'
(U.S. Coast Survey Report, for [187?], p.200) will be found a
plate giving twenty-four curves each drawn from five
hundred cases, which give one an idea of how much variation
from the theoretical curve can be expected. But now
turn to the admirable examination by Mr. Flinders Petrie
of the weights found in Naucratis (Third Mem. of Egyptian Exploration
Fund, Book I, pp. 69 of seqq. with Plate xxiv.) Here we find, especially
in regard to Egyptian weights of from 9 to 9 3/4 grammes, perfectly
unmistakable evidence of two maxima, showing that the
weights [have?] were indirect copies of two prototypes at least, and
therefore beflong to at least two different classes, and yet they
have some [other?] all intermediate weights. Thereforce, the mere fact of the
existence of intermediate forms does not prove that the division