I recently purchased a 1787 sixpence from Michael Gouby graded at aEF. I don't consider it to be the prettiest of pieces, but was attracted by an interesting detail:
I'm sure it's very apparent to you that there is a funny curved line between the top and bottom of the first "I" in "III". I'm pretty confident that this is an overstrike of sorts. My theory is that the die maker started to strike "Dei" straight after "Georgius" in a momentary lapse of concentration. Immediately noticing his error he corrected it by filling in the curved section of the "D", leaving an "I". In the process of filling he missed a very small fraction running along the edge of the "D"s curve, which resulted in the small, thin raised line you see.
I was encouraged to do some research and so over the course of a couple of days I looked at several hundred images of 1787 sixpences. Basically, I looked at every reference and image to the 1787 sixpence that Google brought up directly and then some. Out of these I found nine definite matches with the same issue. If the sample I looked at was representative of the entire issue, slightly more than 2% have this identifying feature. Not only was the I/D overstrike there, but each coin shared several other anomalies indicating they were all from the same die: Die pitting to the left of the hair ribbon, die pitting in the corner of the mouth, a mark underneath the eyebrow and the precise position of the legend lettering. The 2% figure is consistent with the number of dies produced for the minting of the 1787 shilling (see the Manville/Gaspar article) .
Coincidentally, this die was used to produce a (presumably) very small number of coins on a heavy pure silver flan. Here's one of the two examples I came across:
In one of these two examples it was described as a proof. I don't think this is the case. Although the edge is plain, there is no toothing around the edge as there are in the September proofs struck by Pingo. Moreover, there is bifurcation of the bottom of the lettering. This all indicates that the coin was struck without a collar (proofs were made with a collar) and the edge wasn't milled as you would expect with a circulation coin. An odd hybrid. I can't imagine that a die in poor condition like this would ever be used for a proof either: The pitting of the die; The D/I; The error on the "E" of "Dei" (which seems to have been corrected on the circulated sixpences produced from this die). There doesn't appear to have been any special preparation of the blank's surface either. There's no way that this is a proof.
I'm only very new to numismatics, but here's my theory of the existence of a heavy flan struck with this die. The end of the order was approaching, but the mint was running short on workable dies. They dug out a die from the metaphorical bottom draw that had been put there because of the D/I and "E" errors, but it had accumulated some surface blemishes (small rust pits maybe). Wanting to test whether they could use it, a few test strikes were made using a thicker blank they had to hand so as to differentiate it for the circulation coins. No need to mill the edge as it wasn't meant to be used. Somehow is escaped being melted down again and escaped into the wild.
Do you think my theory has any possibility of holding water?
Anyway, I bought the first coin pictured from Michael for 55 GBP. I was tempted to bid on this sixpence currently listed for the upcoming DNW auction.
It has an estimate of 200-300 GBP, which is out of my reach at the moment. Very happy to have the example I bought from Michael, though.