Oecologia (1993) 96 : 169-178 Oecologia 9 Springer-Verlag 1993 Canopy structure and vertical patterns of photosynthesis and related leaf traits in a deciduous forest D.S. Ellsworth 1, P.B. Reich 2 1 Department of Botany, University of Vermont, Burlington, VT 05405-0086, USA a Department of Forest Resources, 115 Green Hall, University of Minnesota, 1530 N. Cleveland Avenue, St. Paul, MN 55108-6112, USA Received: 6 April 1993 / Accepted: 30 June 1993 Abstract. Canopy structure and light interception were measured in an 18-m tall, closed canopy deciduous forest of sugar maple (Acer saccharum) in southwestern Wiscon- sin, USA, and related to leaf structural characteristics, N content, and leaf photosynthetic capacity. Light attenu- ation in the forest occurred primarily in the upper and middle portions of the canopy. Forest stand leaf area index (LAI) and its distribution with respect to canopy height were estimated from canopy transmittance values independently verified with a combined leaf litterfall and point-intersect method. Leaf mass, N and Amax per unit area (LMA, N/area and Amax/area, respectively) all de- creased continuously by over two-fold from the upper to lower canopy, and these traits were strongly correlated with cumulative leaf area above the leaf position in the canopy. In contrast, neither N concentration nor Am,x per unit mass varied significantly in relation to the vertical canopy gradient. Since leaf N concentration showed no consistent pattern with respect to canopy position, the observed vertical pattern in N/area is a direct conse- quence of vertical variation of LMA. N/area and LMA were strongly correlated with Ama x/area among different canopy positions (r2= 0.81 and r 2 - 0.66, respectively), in- dicating that vertical variation in area-based photosyn- thetic capacity can also be attributed to variation in LMA. A model of whole-canopy photosynthesis was used to show that observed or hypothetical canopy mass distribu- tions toward higher LMA (and hence higher N/area) in the upper portions of the canopy tended to increase integ- rated daily canopy photosynthesis over other LMA distri- bution patterns. Empirical relationships between leaf and canopy-level characteristics may help resolve problems associated with scaling gas exchange measurements made at the leaf level to the individual tree crown and forest canopy-level. Key words: Acer saccharum - Photosynthesis - Forest canopy - Sugar maple - Nitrogen Correspondence to: D.S. Ellsworth Forest canopies are spatially heterogeneous environ- ments. The distribution, size and orientation of leaves in space determines the pattern of light availability within the canopy, controlling such processes as leaf develop- ment, leaf energy balance and water use, and photo- synthesis (Norman and Campbell 1989). The presence of leaves to intercept light in upper portions of the canopy will have a strong effect on all these processes in descend- ing portions of the canopy. These effects should be parti- cularly profound for species displaying large, broad leaves, yet relationships among canopy structure, light interception, leaf morphology and photosynthesis are still poorly understood in broad-leaved forests (Eli/is et al. 1989). Assessing vertical variation in leaf distribution and its relation to patterns of photosynthesis among different canopy positions should provide valuable insight into how carbon and nutrient resources are partitioned within forest canopies (Field 1991). Measurements of the ratio of leaf dry mass to leaf area (LMA; leaf mass per area) and its relationship to variation in area-based photosynthetic capacity between different canopy layers may be an effective means of integrating influences of canopy structure and light environment on leaf photosynthetic performance (Jurik 1986; Oren et al. 1986). Leaves that develop in high light have high LMA as a result of increased leaf thickness and increased mesophyll cell density (Chabot et al. 1979; Witkowski and Lamont 1991). Previous studies have shown that LMA generally increases from the bottom to the top of the canopy in evergreen species (Hollinger 1989; Lewandowska and Jarvis 1977; Schulze et al. 1977). In controlled-chamber studies, photosynthetic capacity of leaves grown in different light levels varies with LMA (Bj6rkman 1981; Nobel et al. 1975), but little such data is available for forest-grown deciduous trees (Reich et al. 1991) Large variation in leaf nitrogen content per unit leaf area (N/area) has also been observed among different canopy positions (Hollinger 1989; Hirose and Werger 1987). Leaf N/area content is closely correlated with photosynthetic capacity per unit area in leaves within