Summary: Metabolic scaling theory predicts that the rate of resource use per unit area is independent of the average mass per individual and that the slope of the log mass–log density relationship should be −4/3. Data were obtained from plant communities along a natural gradient of moisture and latitude in north‐west China to test the generality of this theory. The allometric exponents (slopes of the log mass–log density relationship) for above‐ground biomass decreased with natural moisture levels and plant cover, deviating from the predictions of the energy equivalence theory. Allometric exponents for below‐ground and total biomass were similar among the three sites and were much closer to the predicted value of −4/3. Metabolic scaling theory may be applicable under many growth conditions, but not when restricted to above‐ground biomass under drought stress. The rate of supply of the limiting resource per unit area determines which plant parts behave according to theory.

A standardized neighbor removal experiment was conducted in 12 plant communities located on three continents to test the null hypothesis that competition intensity (CI) was independent of the amount of plant biomass present. Six plots were chosen in each community to cover the range of local variation in plant biomass. In each plot the relative growth rate (RGR) of transplanted Poa pratensis (Poaceae) seedlings was compared in the presence and absence of neighbors. Neighbors were removed experimentally using herbicide. Removing neighbors increased RGR of transplants significantly in most plots. CI increased with an increase in the amount of neighbor biomass present in one community where the range of neighbors biomass was greater than in any other community. In contrast, CI did not change significantly with an increase in neighbor biomass in other communities where the range of neighbor biomass was smaller. For the communities combined, CI was not related to neighbor biomass in a consistent fashion. These results indicate that competition may reduce growth over a wide range of habitat productivity, but the relationship between CI and neighbor biomass may differ among communities.

Seeds of plant species from fire-prone vegetation often show dormancy, and germinate in response to fire-related cues. Seeds of some east Australian Grevillea species were known from earlier studies to germinate in response to fires, but gave mixed results when heat shock was tested as a germination stimulus. In the current study, seeds of seven east Australian Grevillea species were tested for their germination response to smoke, to short heat exposure, and to scarification. Seeds were smoked in a chamber by using a mixture of native woodland vegetation as fuel. The smoking and heating treatments were administered to seeds in bulk: this meant that subsequent germination data for these two treatments were not independent. Cumulative germination in the laboratory was scored for up to 60 days. All three germination treatments, when administered singly, increased germination: smoke increased germination in all seven species, heat in four species, and scarification in a different four species. Pairwise combinations of the three treatments also increased germination in some species e.g. smoke and heat treatments combined led to the highest germination observed in four species, and to increased germination compared with either treatment singly, in another species. Some treatment combinations appeared to interact: heat and scarification combined led to decreased germination in three species. The role of heat exposure and scarification in increasing germination was not due to the breaking of an impermeable seed coat: water uptake of both scarified and unscarified seeds was rapid and equal (after 24 h) in all seven species. The results indicate that seeds of the Grevillea species investigated are responsive to one or more fire-related germination cues.