Many aspects of ‘scaling up’ must be considered in the spectrum between promising results of new technologies in experimental plots and wide adoption by farmers. These aspects include extrapolating in time to optimize management decisions, extrapolating in space from small plots to large fields and to other farms and regions, and enlarging the range of (presumed) beneficiaries. Models can help in all these aspects to lay a biophysical foundation on which socioeconomic decisions can be built. We focus here on ‘improved fallow’ systems where trees are planted to restore soil fertility for subsequent food crops. The restoration of soil fertility — based on biomass production, litterfall, and build-up of dynamic soil organic matter pools — depends on total resource capture by the fallow vegetation. Where ‘lateral resource capture’ and ‘lateral resource flow’ play a substantive role in the performance of the fallow, the size (scale) of fallow and cropped plots may influence both the build-up and the decline of soil fertility during a cycle. On small farms, fallow systems should not be seen as pure sequential systems, but as mosaics of spatially interacting fallow and cropped plots. Border effects depend on the lateral spread of the root system of the fallow vegetation, as well as on rainfall and N supply. Scale effects in technology adoption include both positive and negative feedback effects, because the spread of a technology may both accelerate innovations as well as increase threats from pest and disease attack