The New Zealand Dryland Forests Initiative (NZDFI), aims to establish a sustainable natural durable timber industry in New Zealand. The natural durability of heartwood is highly variable within a species and lower at the centre of a tree. Therefore, it is critical to ascertain that trees meet international standards for durability to ensure a viable industry based on plantation grown naturally durable Eucalyptus. A strategy to ensure that wood from future plantings meet industry requirement is to select superior genotypes in a breeding programme. Several Eucalyptus species have been planted by NZDFI, but the primary focus is on E. bosistoana as it is a class 1 durable timber.

Little is known about the heartwood formation in young trees. It is possible that the young trees do not have true heartwood, but an extended transition zone, the zone between sapwood and heartwood where parenchyma cells remain alive synthesising heartwood extractives. Therefore, the wood quality research of NZDFI needs to ensure that the young trees have formed true heartwood. Understanding heartwood formation is critical for the success of a plantation forest industry aiming to produce ground-durable timber. The objective of chapter-2 was to identify conventional and confocal microscopy methods, which allow the observation of cell organelles and the chemical composition in E. bosistoana parenchyma cells before and after heartwood formation. Nuclei, microtubules and peroxisomes in parenchyma cells of 2-year-old E. bosistoana stems were visualised by confocal microscopy combined with optimised immunolabelling protocols. Iodine/potassium iodide stained starch (amyloplasts), while amido black stained proteins in sapwood. Fluorescence emission spectra confirmed the presence of chloroplasts in xylem parenchyma of 2-year-old E. bosistoana. Fluorescence emission spectral (lambda) scans showed differences between parenchyma and fibre cells as well as sapwood and heartwood. The physiological changes between sapwood and heartwood visualised in parenchyma cells helped the understanding of heartwood formation in young E. bosistoana trees.

As part of an E. bosistoana breeding programme, the hypothesis of prolonged transition from sapwood to heartwood in young E. bosistoana trees, resulting in a wide transition zone has been tested in chapter-3. This needs to be considered when assessing trees for heartwood quantity and quality. Heartwood formation was investigated in radial profiles in cores from bark to bark of 6-year-old trees with conventional and confocal microscopy, and with a range of different staining techniques that visualised the physiological changes taking place in the parenchyma cells. Immunolabelling with antibodies against histone proteins and α-tubulin, histochemical staining using iodine/potassium iodide and fluorescence emission spectral scanning, demonstrated that in heartwood nuclei, microtubules, reserve materials (starch) and vacuoles were absent. The observations revealed that 6-year-old E. bosistoana trees contained heartwood. The loss of water conductivity by tyloses and the death of the parenchyma cells occurred in close proximity resulting in a transition zone of ~1 cm. A key feature of heartwood is the formation of extractives that impart natural durability, a sought after wood property. The durability between trees is highly variable. Apart from the quantity of heartwood extractives in the wood, the chemical composition of the extracts varies. Therefore, the objective of NZDFI is to select young E. bosistoana trees with the most potent extracts for next generation durability improvement.

Chapter-4 describes the development of an antifungal assay to determine the bioactivity of E. bosistoana heartwood extracts against a white rot (Trametes versicolor) and a brown rot (Coniophora cerebella). The most suitable procedure was to spread dimethyl sulfoxide (DMSO) solutions of extract onto solidified agar, inoculate with fungi and calculate the growth rate (cm/h) by fitting a linear regression for the diameter of fungi against time. Controls were needed to normalise the fungal growth, i.e. calculating a relative growth rate, to account for the variation in growth conditions between different runs.

Chapter-5 describes the variability in bioactivity and chemical composition of E. bosistoana heartwood extracts between individual trees grown on two different sites (Lawson and Craven Road). Statistical methods combining the results of the fungal assays and the quantitative gas chromatography (GC) of the extracts allowed the investigation of bioactive compounds. The bioactivity of extracts was assessed against white rot (Trametes versicolor) and brown rot (Coniophora cerebella). Ethanol extracts from E. bosistoana heartwood were less effective on the white rot than against the brown rot. Variability in the bioactivity of extracts against the two fungi was observed between the trees. A small site effect in the bioactivity was found for the white rot but not the brown rot. Bioactivity of the extracts against the white rot was not correlated to that against the brown rot. The absence of such a relationship indicated that the two fungi were affected by different heartwood compounds. Thirty two compounds were quantified in E. bosistoana ethanol extracts by GC, of which six (benzoic acid, hexadecanoic acid, 1,5-dihdroxy-12-methoxy-3,3-dimethyl-3,4-dihydro-1H-anthra[2,3-c] pyran-6,11-dione, octadecanoic acid, polyphenol and beta-sitosterol) were identified. Significant variability in eight compounds (out of the 32) was found between the two sites. Multivariate (PLSR) analysis identified compounds at retention times 10.2 and 11.5 min (hexadecanoic acid) to be most related to the bioactivity of the E. bosistoana heartwood extracts against white rot and brown rot.

Breeding programmes benefit from early assessment, enabling short breeding cycles. While trees only form heartwood when several years old, wounds can be induced at a young age. As wound reaction has some similarities to heartwood formation the objective of chapter-6 was to investigate if the wound reaction can be used as a proxy to assess trees early in a breeding programme for heartwood features. 1.6-year-old individuals from 27 E. bosistoana families with known heartwood diameter and extractive content were wounded and the axial wound reaction was correlated to the known heartwood features. No correlation between wound reaction and heartwood features in E. bosistoana families was found. Wound reaction was under weak genetic control. Therefore, it seems to be not possible to assess heartwood formation early in a breeding programme by measuring the axial wound response in 1.6-year-old E. bosistoana trees.

Furthermore, two families with small and large wound reaction were selected to characterise physiological and chemical variations in woundwood, heartwood and sapwood by microscopy and gas chromatography. Microscopic observations revealed the absence of starch and that vessels were occluded with tyloses in woundwood and heartwood. Morphological features were not discernible by eye between the families with large and small wound reaction. Gas chromatography revealed variation in the chemical composition of woundwood, heartwood and sapwood extracts.