As I understand it, it was an experiment during Victorian times to try to monitor the output (both quantity and quality?) from individual dies - by numbering the dies it should theoretically be reasonably straightforward to determine if any problems occurred with a given die, though in practice once coins are out in circulation it would be very hard to track them down again. I wasn't aware of their being used on bronze coins before I saw the COIN News article, though.

Here's a detailed link from Tony Clayton's website, which is a must to consult for any queries regarding UK coinage: http://www.tclayton.demon.co.uk/dieno.html

In Ken Bressett's A Guide Book of English Coins, 8th Edition, it states "These numbers were placed on the dies (starting with number one each year) to identify the usage and coinage quality from each particular die."

I skimmed through some of the Royal Mint reports from the 1870s and, while I could not find such a statement, I did see frequent references to the average numbers of coins obtained from dies for particular denominations. It seems that the Royal Mint was somewhat obsessed with this subject at the time, and numerous experiments were conducted to identify the best steel alloys for die making. Since the practice of numbering the dies ended in 1880, I paid particular attention to that year's report, and indeed it states that a satisfactory solution had been found. Again, there was no mention of the tiny numbers punched into dies, but the report suggests that such experiments were no longer needed after that time.