In spite of numerous studies on the effect of nutrient levels and/or standing crop on the intensity of resource competition the debate has not been resolved. Field studies that have used natural productivity gradients have generally supported the argument that competitive intensity and resource availability are positively correlated, whereas studies that have used artificial resource gradients have generally refuted the same argument. Here we report the results from study in which both approaches were used within the same system. We studied two species of eucalypt that occupy contrasting parts of the same landscape: Eucalyptus camaldulensis, found mostly along creek lines and in valleys with deep alluvial soils, and Eucalyptus microcarpa, found on hillsides and ridges with shallow soils. We studied the response of seedlings of the two species to the combined effects of competition and manipulated nutrient levels in a glasshouse experiment, and also investigated their responses to removal of neighbouring plants in the field. Eucalyptus microcarpa was less responsive to increased resource availability, which is consistent with one of the principal assumptions of Grime's C-S-R model. In the glasshouse experiment both species of eucalypt responded in a qualitatively similar fashion to the combined effects of resource availability and competition: release from competition resulted in increased growth, but only in pots that received additional resources. In the field we found that neighbouring vegetation could severely limit the establishment of E. camaldulensis but the removal of neighbouring vegetation did not affect the performance of E. microcarpa seedlings. Eucalyptus camaldulensis seedlings suffered high levels of damage from herbivores. Our results thus generally support the predications of the C-S-R model, however, they indicate that the effects of competition and herbivory may be heavily confounded. © 2007 Ecological Society of Australia.