Eucalyptus plantations are one of the largest wood resources, and while they are typically primarily grown for fibre or biomass, there is an increasing interest in using this resource for solid wood products. However, timber from such short-rotation eucalyptus plantations is often ‘difficult-to-dry’, suffering collapse during drying. To make such timber suitable for higher- value solid wood products, this thesis explored two strategies for reducing collapse in plantation-grown eucalypts: reducing collapse through a breeding programme by selecting genetically less-collapse-prone planting stock; and technical solutions to allow the drying of collapse-prone timber with reduced collapse. Additionally, the influence of ions on water transport in wood and collapse was explored.

Chapter 1 provides an overview of timber drying, collapse mechanisms during drying, amelioration of this problem, and different methods of assessing collapse in wood. Review of the literature on collapse suggested that there was no practical method that could solve this problem, even though this issue was identified more than 100 years ago.

Finding a solution for drying collapse is topical since the New Zealand Dryland Forests Innovation (NZDFI) is working to establish a domestic ground-durable eucalypt resource for use in high-value solid wood products.

Chapter 2 investigates the reliability of collapse assessment methods used in tree breeding. This chapter reports on correlations, especially the genetic correlation, between different shrinkage and collapse measurements in E. quadrangulata samples cored from a breeding population. Strong correlations would allow the choice of the least resource-consuming assessment method.

Maximum tangential shrinkage, the most convenient measure tested in this study, was shown to be a reliable predictor of true (recoverable) collapse because of its comparable heritability and strong phenotypic and genetic correlations with the true collapse.

The strong correlation between these traits in E. quadrangulata samples also indicated that most genetic variation in maximal tangential shrinkage was explained by true collapse rather than normal shrinkage when cores were dried at 60°C. Moreover, checking in sawn timber is likely to be related to maximal tangential shrinkage rather than a volumetric measure.

The maximal tangential shrinkage in cores was positively correlated to both recoverable collapse and normal shrinkage. This indicated that selection for lower maximal tangential shrinkage would reduce both normal shrinkage and collapse in a breeding population.

Chapter 3 explored the genetic parameters and amount of genotype by environment interaction for drying collapse, extractive content, basic density, core length, and heartwood diameter in two ~9-year-old E. quadrangulata breeding trials, as well as evaluating the relationships between collapse and other traits.

Collapse was confirmed and was prominent in the heartwood of E. quadrangulata cores. Heartwood collapse and other wood properties were under varying degrees of genetic control, with heritability ranging from 0.19–0.40. In accordance with theory, heartwood collapse was negatively correlated with basic density and positively correlated with extractive content. Significant genetic gain could be expected for heartwood diameter, core length, basic density, and heartwood collapse. However, improving extractive content for this species might be challenging as low heritability was observed, but it should be noted that the low number of families included in these breeding trials partly contributed to low heritability.

Chapter 4 investigated genetic control of collapse and other tree features at mid-rotation age of the emerging plantation species E. globoidea. Using a 14-mm diameter corer, thousands of E. globoidea trees representing 163 families were sampled from a breeding population established at three sites, and genetic parameters for heartwood and sapwood collapse, extractive content, and heartwood diameter were estimated. Heartwood collapse was under genetic control, with a narrow-sense heritability ranging from 0.22 to 0.44. Considering the coefficient of genetic variability of ~13% to 23%, heartwood collapse in E. globoidea can be reduced through selection.

The significant genetic correlation between sites for heartwood collapse (rg = 0.73 to 0.83) suggested low genotype by environment (G × E) interaction. In line with the physical causes of collapse, heartwood collapse was positively correlated with extractive content. Extractive content and heartwood diameter are other traits of interest, since E. globoidea is grown for its ground-durable heartwood. The heritability of extractive content ranged from 0.40 to 0.71. Heartwood diameter was shown to be negatively correlated to extractive content. No significant genotype by environment (G × E) interaction was found for extractive content, while genotype by environment (G × E) interactions for heartwood and sapwood diameter were small. Finally, 12 families were identified with above-average heartwood diameter, extractive content, and below-average heartwood collapse. In summary, this study has shown that genetic selection for collapse and other wood properties of E. globoidea is feasible.

Chapter 5 investigated Joule heating as a technical solution to mitigate drying collapse, an energy-efficient method to heat timber by passing electric current through it. All electric energy is converted into heat inside the wood avoiding losses caused by unfocused electromagnetic fields in microwave heating or the long timeframes and consequent energy losses through imperfect insulation in conventional heating. The effect of Joule heating on collapse, shrinkage, water absorption, and compressive strength of timber was investigated. Tangential and volumetric collapse in E. nitens reduced significantly by ~53% and ~48%, respectively, with Joule heating. Maximum tangential shrinkage was significantly lower, due to the reduction of collapse in Joule-heated samples, because normal shrinkage was not affected by Joule heating. A reduction in collapse was achieved without a statistically significant increase in water absorption capacity or decrease in compressive strength. Further, the estimated costs for Joule heating pretreatment were estimated to be competitive when compared with conventional steam recovery of collapse.

Chapter 6 examined the influence of replacing sap with potassium chloride (KCl) solution or deionised water on drying collapse, wood permeability, and streaming potential of never-dried wood. Ion-mediated response in xylem is considered a major factor in sap-flow regulation. Volumetric and tangential collapse in KCl-treated E. nitens logs decreased significantly by 42%–62% and by 51%, respectively. Improvement in shrinkage properties was dependent on the KCl concentration, with a maximum at a concentration (20 mM) in the range of total ionic strength found in living trees. Logs treated with deionised water showed higher normal shrinkage than the control, without affecting recoverable collapse. Consistent with reduced collapse in KCl-treated logs, E. globoidea stem cores showing low collapse contained significantly more inorganic cations than the high-collapsed wood. The decrease in collapse when treated with KCl solution coincided with increased green wood permeability evident in a higher sap flow rate. Sap conductivity affected the streaming potential, with the polarity of the induced electric potential varying between concentrations and matching literature reports of electric potential measurements in living trees and laboratory experiments.

In summary, this study confirmed that drying collapse was negatively correlated to sap flow, and a potential technical solution to drying collapse in E. nitens could be sap replacement as a pre-drying treatment, and/or nutrient management of the plantations.