OK, so I’ve used three measuring methods – just out of interest to see the relative accuracies:
1) A strip of card wrapped around the smaller wheel, and its length (circumference) measured with a tape measure, then wrapped around the larger wheel and the difference measured with verniers:

2) Measuring the chord length and its offest from the circumference using verniers:

3) Height gauge:

Assuming the height gauge is the datum (as the most accurate method) Anyone fancy a bank holiday guess at the maximum and minimum deviation from this – on radius, in terms of thou – between the chord and tape methods?
I’m going to use Jason’s spigot method as a double-check on the height gauge readings, and hopefully get it done this afternoon.
Thanks all.