You have hit on an important problem. There is no standardized way that
weights are reported.
There are thus many different
weights "standards".
There is a theoretical
weight standard at which the coin was struck - say 1/72 of a Libra or
Roman pound, in other words 72 coins should be struck from a
Roman pound of metal. This is problematic given that no one is sure exactly what a
Roman pound weighed (though the variance is not too much as estimates range from circa 322 - 327 grams, thus up to 5 gram difference, or about 1.5%).
There is also a potentially different value for the
weight that the coins on an issue were actually struck at. Did the 1/72 L take into account material losses (2-5%)? Were they deliberately struck underweight at a
mint? With at least late
Roman coins the
weights almost always slide down before a new lower official
standard is introduced.
Both of these
weight "standards" usually have to be guessed at - extrapolated from the existing coins. Though sometimes the coins are marked with the theoretical
weight, i.e.
LXXII for 1/72 L.
Then there are the
weight "standards" derived from measuring existing coins. This can be done by measuring only the coins in one
hoard, or in one
collection, or all the examples checked by the author, etc. More scholarly works will often describe the method used - so some volumes of
RIC will say
average of xx based on yy number of coins from the zz
hoard. Some authors are now doing
meta studies - looking at the results of several
hoards or studies and comparing.
The challenge of course is that measuring existing coins depends on loss due to wear (pre-burial), corrosion (during burial), and cleaning (post-burial). And it is not always loss - some coins that have not been cleaned, or that have certain
types of corrosion or adhesions can gain in
weight.
So you can sometimes see significant changes to averages between large depending on the state they are in. Was one a
hoard buried when new and the other a
hoard of worn coins?
Oh, and of course the issue of
contemporary copies is important. They are usually under
weight. But are they all eliminated or not from the study? In some cases they are difficult to identify.
With some coinage there is a question of different issues being mixed together. The
FEL TEMP REPARATIO was struck at at least four official
weight standards, and possibly several others - whether official, or semi-official. Yet many
weight studies do not break them down by separate issue. If you try to calculate
average coin
weights from a
jar full of pennies, dimes and nickles you are doomed to failure.
Finally there is the question of the math.
Average, median and/or mean? What is the approach to outliers? Some keep in all, some remove the worst outliers. there are various methods in statistics for handling this.
And this is all about averages. If you are dealing with bronze coinage you also
face the fact that they were only originally meant to
average out over a large number of coins - a batch. Thus the
weight of individual coins was far less important.
So you see the problem with "
average"
weights......
Shawn