I've read the history of this somewhere before... Probably the best place to ask this question is over on YORF, as the old timers and old-school rocketry fanatics really get into the history of it all and know more about it than I do...
Basically, the first model rocket motors were manufactured based on fireworks material and fireworks equipment... so naturally they started out using fireworks sized tubes/casings and such... This sort of carried over when Vern Estes made the first model rocket manufacturing machine, Mabel, which really started the consumer hobby of model rocketry. There had been some "weird size" motors (like the old "Coasters", etc) and some manufacturers went with their own sized motors, but since Estes accounted for the VAST VAST majority of the market, they pretty much established the "industry standard" by default, which was the 18 mm standard cased BP motors. Centuri Engineering later came out with "mini-motors" in a 13 mm casing. Estes followed suit later with thier own mini-motors. Centuri came out with the Enerjets, which were larger than the standard motors. The Enerjets didn't last a long time, but the desire for greater power motors was already whetted, and about the same time Estes came out with their "Mighty D" engines (initially the D13, which had a tendency to go boom so was retooled as the D12, and the D11 (which was a bit of an oddball).
Flights Systems, Inc. went their own way and made their motors their own size. They were a couple millimeters off from the "standard" Estes size motors. FSI made motors up to an F100 (which was actually less powerful than that) and their 7 second burn duration F9 "steam machine"... they also offered motors down into the D and E size range as well... not sure about other sizes (but perusing their old catalogs over on the ninfinger site should answer those questions.) IIRC FSI was the ones that came out with the 29mm motor casing size for their F100's.
The main reason for the different casing sizes themselves is that the physical amount of size inside the casing determines the maximum total impulse (power) that can be put into that casing, especially when one includes the delay grain and ejection charge, and nozzle one one end and cap on the other. The casing diameter AND the casing length work together to determine the minimum and maximum burn durations that can be put into that casing. For instance, a narrow diameter casing isn't a good choice for a high-impulse motor with a substantial thrust spike, because the propellant grain will be small in diameter and very long, which means it will be a longer burn duration motor... (and generally speaking, the longer the duration for a given impulse motor, the lower the average thrust). Estes ran up against this problem when they started to produce an "E" impulse motor in the 24 mm casing size... the "D" motor casing, 24 x 70 mm, simply wasn't long enough to contain the propellant grain and delay grain, along with the ejection charge, nozzle, and end cap. So, they elongated the casings to 24 x 95 mm to give enough room in the casing for everything.
When you get into HPR motors, you're in a whole different ballpark, and the sizing for those comes from their own conventions and materials and manufacturing techniques as well I'd suppose. But you'd have to ask someone else about that...
Later! OL JR