Mesophyll anatomy, chlorophyll distribution, and other structural characteristics that could influence photosynthetic CO2 assimilation were evaluated for three ontogenetically and structurally different leaf types found in E. globulus. In addition to Ames/A and chlorophyll concentrations, mesophyll cell surface area and volume per unit volume (Ames/V and Vmes/V, respectively), mesophyll cell volume per unit leaf surface area (Vmes/A), and the chlorophyll a:b ratio were also measured. All of these internal leaf parameters were measured on a whole-leaf basis, as well as incrementally across the entire leaf thickness.
Two Eucalyptus globulus ssp. globulus Labill. provenances, exhibiting different rates of vegetative phase change, were grown in a climate-controlled glasshouse. Night and day temperatures varied between 15–18°C and 21–27°C, respectively, while humidity ranged from 22 to 51%. Additional lighting extended the photoperiod to 15 h. Plants were watered daily to field capacity, and nutrients (NPK) applied as controlled-release fertilizer. Seven-month-old seedlings from St Marys, Tasmania (CSIRO forest research seedlot provenance 16474 sib. CL002) were producing true juvenile leaves at the time of analysis. Saplings from Wilson's Promontory, New South Wales (provenance 16399 Sib. DFC 219) were 2.5 yr old and had transitional and adult leaf types at the time of analysis.
Five representative leaves of each leaf type (juvenile, transitional, and adult) were selected from nodes 18–23, 41–48, and 49–54 from the base of individual plants (N = 3), respectively. Transitional and adult leaves were selected on the basis of petiole length and the ratio of leaf length to width (Table 1). Additional structural characteristics were determined for each of the leaves (Table 1), and stomatal densities were measured (James and Bell, 1995 ).
Leaf sections of 15 mm2 were collected at leaf midlength, midway between the midvein and leaf margin, fixed in 3% glutaraldehyde in a 0.005 mol/L phosphate buffer (pH 7.0) for 24 h under vacuum, dehydrated in a graded ethanol series in 10% steps, and embedded in LR White acrylic resin (London Resin Co. Ltd, Reading, UK) within gelatin capsules (polymerized at 70°C for 24 h). Sections of 1 µm thickness were stained with 0.5% toluidine blue in 0.1% sodium carbonate buffer (pH 9.0). Anatomical characteristics were determined for each leaf from transverse sections and computed as an average of five measurements at 300x magnification (Table 1). Oblique-paradermal sections were cut at an angle of 30° – 85° to the surface of the leaf (Fig. 1), and used to determine mesophyll structure.
The internal anatomy of leaves has previously been determined quantitatively using drawings prepared with the aid of camera lucida (Turrell, 1965 ; Nobel, Zaragoza, and Smith, 1975 ; Smith and Nobel, 1978 ; Thain, 1983 ). This method generates mesophyll layers drawn from paradermal views, while transverse sections are used to construct a three-dimensional model from which all cell lengths and diameters can be determined. Cell surface areas are calculated assuming the palisade mesophyll cells are cylindrical with hemispheres at each end, while spongy mesophyll cells are spherical with certain correction factors applied. Others have developed stereological techniques to quantitatively measure internal leaf structure (Parkhurst, 1982 ; Kubinová, 1991 , 1993 , 1994 ). The current study uses a newly developed image analysis technique to determine mesophyll structure within the Eucalyptus leaves.
The ratio of mesophyll surface area (Ames) to external surface area (A) was determined using an oblique-paradermal section of each embedded leaf (Fig. 1). A consecutive sequence, or transect, of 0.0079 mm2 digital images at 750x magnification was collected from the adaxial to abaxial epidermis using Image-Pro Plus software (Media Cybernetics, Silver Spring, Maryland) and a Javelin CCTV camera (Javelin electronics, Los Angeles, California) attached to a binocular light microscope. The number of images per leaf ranged from four to 15, depending on the thickness of the leaf and the angle of the oblique paradermal section. A single transect was measured for each leaf (N = 5 per leaf type).
Images were manipulated using Adobe Photoshop 3.0 software (Adobe Systems Inc., Mountain View, California) for a Macintosh computer. The intercellular area within each cell was blackened, and contrast was maximized to produce a sharp black-white image. The area of mesophyll cells (Am), intercellular airspace (Aa), and the length of mesophyll cell wall exposed to intercellular airspace (Pi) was determined for each image using the Image-Pro Plus software. Ames/A was then calculated for each image as
and where Ti
is the projected thickness of a given section i
is the mesophyll thickness of the leaf determined from transverse sections of the same leaf, Li
is the length of the image, Ap
is the projected area of the image, Lp
is the projected length of the image, and Wi
is the width of the image. C
is a conversion factor for Pi
, which gives the projected mesophyll perimeter. The sum of Li
for each of the leaf images gives the total transect length of the oblique-paradermal section (Ls
), while the sum of Lp
gives the projected length of Ls
. Where vascular tissue or oil cavities were present within the mesophyll of the image, Ap
was computed as a proportion of the projected area of an image containing only mesophyll. Hence, Ames
for the leaf was calculated as the sum of the individual values for each image or
per unit leaf area (Vmes
) was calculated for each leaf by substituting Pi
in Eq. 1
, and Ames
in Eq. 5
. In the same manner, substituting Am
in Eq. 1
enables calculation for Ames
per unit leaf volume (Ames
Similarly, the volume of mesophyll per unit leaf volume (Vmes
) was calculated by substituting Am
in Eqs. 6 and 7
Relationships between anatomical and morphological characteristics were determined by regression analysis using Minitab Release 10.51 (Minitab Inc., State College, Pennsylvania) for a Macintosh computer. Differences between structural characteristics of the three leaf types were determined using one-way and multivariate ANOVA.
Chlorophyll and carotenoid determination
Four of the five leaves selected as representative of each leaf type were also used to determine leaf chlorophyll content. Six to ten leaf disks 31.2 mm2 in size were sampled from midway along the leaf and paradermally sectioned into 30 µm (juvenile and transitional) or 40 µm (adult) layers using a cold-stage microtome (Spencer Lens Co., Buffalo, New York). Corresponding layers were pooled in a small vial containing 1 mL of 95% ethanol, plus 0.01 mL of sodium ascorbate as an antioxidant. Because the leaf disks separated from the cold stage before the entire thickness of the disk could be sectioned, adaxial and abaxial surfaces were alternately sectioned and tissues pooled separately. Samples were kept in the dark until analysis, and concentrations of chlorophyll a, b, and carotenoids were determined by measuring absorbance at 470.0, 648.6, and 664.2 nm, minus absorbance at 730 nm with a spectrophotometer (Hewlett Packard model 8453, San Diego, California). Extinction coefficients determined according to Lichtenthaler (1987) were used to calculate pigment concentrations. Total chlorophyll content of the leaves was determined by averaging measurements from two leaf disks from the same region. Leaf, palisade, and epidermal thicknesses were also measured using transverse sections of fresh leaves.