I'm looking for classic papers discussing whether the importance of an ecological variable at determining species composition or turnover depends also on the length of the ecological gradient considered.
The relative importance of the mechanisms that generate floristic turnover may vary in relation to the spatial scale of the study and the length of the relevant ecological gradients. When the sampled ecological gradient is long, the effect of resource-driven processes on the community can be expected to be more obvious than when the gradient is truncated by focusing, for instance, on small geographic ranges or just one habitat type, as the latter leads to higher noise-to-signal ratios.