The authors investigate the relationship between the volume of malignant tumours and their metabolic processes using a large dataset of patients with cancer. They find that cancers follow a superlinear metabolic scaling law, which implies that the proliferation of cancer cells accelerates with increasing volume.

The ability of general scaling models to capture the central tendency or dispersion in biological data has been questioned. In fact, the appropriate domain of such models has never been clearly articulated and they have been supported and challenged using both interspecific and/or intraspecific data. Here, we evaluate several simplifying assumptions and predictions of two prominent scaling models: West, Brown and Enquist’s fractal model (WBE) and a null model of geometric similarity (GEOM). Using data for 53 herbaceous angiosperm species from the Songnen Grasslands of Northern China, we compared both the interspecific and intraspecific scaling relationships for plant geometry and biomass partitioning. Specifically, we considered biomass investment in shoots and leaves as well as related several traits not commonly collected in plant allometric analyses: shoot volume, leaf number, and mean leaf mass. At the interspecific level, we find substantial variation in regression slopes, and the simplifying assumptions of WBE and predictions of both the WBE and GEOM models do not hold. In contrast, we find substantial support for the WBE model at the intraspecific level, and to a lesser extent for GEOM. The differences between our results at interspecific and intraspecific levels are due to the fact that leaf size and stem tissue density vary considerably across species in contrast to the simplifying assumptions of WBE. These results highlight the domain within which simplifying model assumptions might be most appropriate, and suggest allometric models may be useful points of departure within some species, growth forms or taxonomic groups.