Photosynthesis-nitrogen relations in Amazonian tree species : II. Variation in nitrogen vis-a-vis specific leaf area influences mass- and area-based expressions.

Abstract

The relationships between leaf nitrogen (N), specific leaf area (SLA) (an inverse index of leaf "thickness" or "density"), and photosynthetic capacity (Amax) were studied in 23 Amazonian tree species to characterize scaling in these properties among natural populations of leaves of different ages and light microenvironments, and to examine how variation within species in N and SLA can influence the expression of the Amax-to-N relationship on mass versus area bases. The slope of the Amax-N relationship, change in A per change in N (μmol CO2 gN-1 s-1), was consistently greater, by as much as 300%, when both measures were expressed on mass rather than area bases. The x-intercept of this relationship (N-compensation point) was generally positive on a mass but not an area basis. In this paper we address the causes and implications of such differences. Significant linear relationships (p<0.05) between mass-based leaf N (Nmass) and SLA were observed in 12 species and all 23 regressions had positive slopes. In 13 species, mass-based Amax (Amass) was positively related (p<0.05) with SLA. These patterns reflect the concurrent decline in Nmass and SLA with increasing leaf age. Significant (p<0.05) relationships between area-based leaf N (Narea) and SLA were observed in 18 species. In this case, all relationships had negative slopes. Taken collectively, and consistent in all species, as SLA decreased (leaves become "thicker") across increasing leaf age and light gradients, Nmass also decreased, but proportionally more slowly, such that Narea increased. Due to the linear dependence of Amass on Nmass and a negative 4-intercept, "thicker" leaves (low SLA) therefore tend, on average, to have lower Nmass and Amass but higher Narea than "thinner" leaves. This tendency towards decreasing Amass with increasing Narea, resulting in a lower slope of the Amax-N relationship on an area than mass basis in 16 of 17 species where both were significant. For the sole species exception (higher area than mass-based slope) variation in Narea was related to variation in Nmass and not in SLA, and thus, these data are also consistent with this explanation. The relations between N, SLA and Amax explain how the rate of change in Amax per change in N can vary three-fold depending on whether a mass or area mode of expression is used.