2.1. Incident Light
A Yap habitat was utilized to determine the influence of shade on C. micronesica
leaf nutrients from 12–14 October 2016 (9°32′ N). The soils were formed in residuum derived from green, chlorite, and talc schist (Clayey-skeletal, mixed, isohyperthermic Lithic Tropudalfs) [17
]. The habitat contained more than 1200 trees per ha, and we used this high density to locate seven pairs of trees to compare sun versus shade. We ensured that the height of the two trees in each pair was homogeneous, and the two trees were located less than 20 m apart to minimize soil heterogeneity. The level of midday incident light was quantified with a 0.75 m line quantum sensor (EMS-7, PP Systems, Amesbury, MA, USA), and we only selected shaded plants with 25% to 30% sunlight transmission in order to limit the variation among the shaded replications. Leaflet sampling was restricted to 1100–1300 HR for this study to coincide with the incident light measurements. The height of each tree pair was measured, and ranged from 50 cm to 230 cm.
Each tree was treated as one replication, and we collected leaflets from the youngest flush of leaves, and included leaves growing in the directions north, east, south, and west. From each leaf, we collected leaflets from the base, midpoint, and apex of the rachis. All leaflets for each tree were combined into a single composite sample.
Tissue for nutrient analysis was dried at 75 °C and milled to pass through a 20-mesh screen. Total nitrogen (N) and carbon (C) were determined by dry combustion (FLASH EA1112 CHN Analyzer, Thermo Fisher, Waltham, MA, USA) [18
]. Samples were also digested by a microwave system with nitric acid and peroxide, then phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), iron (Fe), boron (B), manganese (Mn), zinc (Zn), and copper (Cu) were quantified by inductively coupled plasma optical emission spectroscopy (Spectro Genesis; SPECTRO Analytical Instruments, Kleve, Germany) [19
]. We then calculated the C/N, C/P, C/K, N/P, N/K, and K/P quotients.
We also collected 10 leaflets from each of the experimental trees to determine the mass–area relationship for sun and shaded trees. From these, 30 leaflets were randomly selected then a section 2–3 cm in length was cut from the midpoint of each leaflet. The length and diameter of each section was measured, then the area was calculated. The tissue was dried to a constant weight at 75 °C and dry weight was measured.
We attempted to restrict the tree heights for the replications of the shade study to remove the influence of height on leaf traits. We were unable to find seven pairs of trees that met all of our other selection criteria within a limited height range, so we instead employed the use of analysis of covariance (SAS 9.3; SAS Institute, Cary, Indiana) with tree height as the covariate. The influence of shade on leaf traits was not influenced by tree height for any of the response variables for the mass-based or area-based concentrations. Thereafter, we used a paired t-test to determine differences between shade and sun leaves for all response variables. The six stoichiometric variables were log-transformed prior to analysis. We used linear and quadratic regression analysis (SAS 9.3) of the essential element concentrations to determine the influence of leaf age as the independent variable, with 1 designated as the youngest flush and 4 designated as the oldest flush.