Well essentially, it is the distance between pedals. Sheldon Brown calls it “tread” or “Q-factor”, and makes it the distance between pedal attachment points, which is close enough. Distance between pedals on a unicycle is determined by axle length and crank design. Some cranks flare outwards more, increasing their contribution to a given wheel’s Q. Widening a unicycle hub with welded-on flanges not only separates the flanges more, but also separates the pedals more.
On the stock Coker, my feet always wanted to drift outwards, even when using super grippy pedals to try to “lock” my feet in place. This left me fighting the unicycle and was quite unsettling. When I switched to the wide hub, I was noticeably happier on both grippy and non-grippy pedals. Similarly, on my Suzue-based 29er, I had to add pedal extensions to be comfortable. Recent experiments I’ve done with a KH hub-based 29er show me that the pedal spacing for me is pretty good without adding the extensions, since the hub/axle are wider and the cranks have more contribution to Q than the Bike Euros on the other 29er.
Since pedal spacing is a factor in bicycle fit, it is reasonable to conclude it is also a factor in unicycle fit, with the above experiences serving as instances of a general principle. However, a bit of thought about the nature of unicycling and fit shows one that individual body characteristics, seat height, seat type, crank length, riding type, and even terrain type are all complicating factors, so that calculating the ideal Q for a person is a non-trivial task. It is currently not possible to calculate accurately a person’s ideal unicycle Q from body measurements and the like.
However, road riding is the unicycling style that is the most controlled in terms of seat height and body motion, so it may be the one to start with by simple individual experimentation, as I have done.
In general, we currently have control over: seat type, hub width (more precisely, axle length), crank brand, pedal width. These things are worth playing with for a serious rider in order to achieve optimal Q. Note that some hubs have proprietary axle/crank interfaces so that using a different brand of cranks is not possible. Note also that, for some or most cranks, choosing a longer crank also increases that crank’s contribution to Q.
Optimal Q is desirable for: relaxing while maintaining proper foot position on the pedals, even on rough terrain; minimizing side pressure on the saddle; minimizing muscle and tendon interference by the saddle; developing optimal energy transfer through the legs to the pedals; developing higher spinning rates with longer cranks; and the like.
Although some riders dismiss Q as an issue, I believe, partly based on my own personal experience, partly from simple thought experiments, and partly from the bicycling literature, that Q is quite important for the serious unicyclist. As distance riding (both road and off-road, touring or racing) develops, Q-factor will become a major factor in fitting a unicycle to a rider for optimal performance over many miles. I have started developing a unicycle-specific system for quantifying this, and hope to fold it into the unicycles I build.