David Hembrow looks at the correlation between population density and cycling rate for a range of cities. Or rather, he points out the remarkable lack of correlation. Clearly there’s an intrinsic relationship between population (# of people) the distances people need to travel to access that population, and the density of the population, but even for cities with densities comparable to the US (1000-2500 people/km^2) cycling can be a very convenient mode of transport.