I'd be fascinated to see if these stamps appear on other classes of vehicles. They have the potential to help spot re-stamped engines. I suspect though that worrying about this is rather a motorcyclist's pre-occupation (perhaps because they're so visible ?) Numbers which have been overstamped can of course mislead us in trying to fill in the gaps in information between known contracts.
My main focus is the first wartime Norton contract C.5109 which ran from October 1939 to March 1940 approximately and covered machines W1000 - W6999.
So far, I have photographed the following :-
W1231 is probably the oldest surviving Norton with matching numbers. It turned up in France and carries M^74
W1742 is also a matching numbers bike found in France and also M^74
I have crankcases numbered W4187 - Also M^74
There is certainly consistency here. I'll dig out some later numbers as well.