wood quality and its biological basiseeb.lu.lv/ftp/grozs/augufiziologijas/augu_resursu... · wood...

Wood Quality and its Biological Basis

Biological Sciences Series

A series which provides an accessible source of information at research and professionallevel in chosen sectors of the biological sciences.

Series Editors:

Professor Jeremy A. Roberts, Plant Science Division, School of Biosciences, University ofNottingham. Professor Peter N.R. Usherwood, Molecular Toxicology Research Group, School of Lifeand Environmental Sciences, University of Nottingham.

Titles in the series:

Stress Physiology in Animals Edited by P.H.M. Balm

Seed Technology and its Biological Basis Edited by M. Black and J.D. Bewley

Leaf Development and Canopy Growth Edited by B. Marshall and J.A. Roberts

Environmental Impacts of Aquaculture Edited by K.D. Black

Herbicides and their Mechanisms of Action Edited by A.H. Cobb and R.C. Kirkwood

The Plant Cell Cycle and its Interfaces Edited by D. Francis

Meristematic Tissues in Plant Growth and Development Edited by M.T. McManus and B.E. Veit

Fruit Quality and its Biological Basis Edited by M. Knee

Pectins and their Manipulation Edited by G.B. Seymour and J.P. Knox

Wood Quality and its Biological Basis Edited by J.R. Barnett and G. Jeronimidis

Plant Molecular Breeding Edited by H.J. Newbury

Biogeochemistry of Marine Systems Edited by K.D. Black and G.B. Shimmield

Wood Quality and its Biological Basis

Edited by

JOHN R. BARNETT School of Plant Sciences

The University of Reading UK

and

GEORGE JERONIMIDIS Department of Engineering The University of Reading

UK

BlackwellPublishing

© 2003 by Blackwell Publishing Ltd

Editorial Offices: Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK

Tel: +44 (0)1865 776868 Blackwell Publishing Asia Pty Ltd, 550 Swanston Street, Carlton, Victoria 3053, Australia

Tel: +61 (0)3 8359 1011

ISBN 1–84127–319–8Originated as Sheffield Academic Press

Published in the USA and Canada (only) by CRC Press LLC, 2000 Corporate Blvd., N.W., Boca Raton, FL 33431, USAOrders from the USA and Canada (only) to CRC Press LLC

USA and Canada only:ISBN 0–8493–2819–5

The right of the Author to be identified as the Author of this Work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use.

Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe.

First published 2003

Library of Congress Cataloging-in-Publication DataA catalog record for this title is available from the Library of Congress

British Library Cataloguing-in-Publication DataA catalogue record for this title is available from the British Library

Set in 10.5/12 pt Timesby Integra Software Services Pvt Ltd, Pondicherry, IndiaPrinted and bound in Great Britain using acid-free paper by MPG Books Ltd, Bodmin, Cornwall

For further information on Blackwell Publishing, visit our website:www.blackwellpublishing.com

Contents

Contributors xiPreface xiii

1 Tree growth and wood quality 1RODNEY ARTHUR SAVIDGE

1.1 Cambial growth 11.1.1 Wood is a biosynthetic end product 11.1.2 Zonation 11.1.3 Bordered-pit development 31.1.4 Secondary-wall lamellae 41.1.5 Microfibrils and lignin 51.1.6 Protoplasmic autolysis 51.1.7 Cambial fusiform cell length and orientation 6

1.2 Perennial cambial growth 61.2.1 Episodic but variable cambial growth 61.2.2 Tapering to the point in form and function 7

1.3 Wood quality in perspective 81.3.1 Defining wood quality 81.3.2 Measuring wood quality 81.3.3 Wood quantity versus wood quality 91.3.4 Stem dimensions and quality 91.3.5 G × E control of wood quality 91.3.6 Variation taraplas within the tree 12

1.4 Wood density 141.4.1 Molecular and anatomical basis 141.4.2 Enhancing wood density through silviculture 141.4.3 Enhancing wood density through tree improvement 161.4.4 Understanding the control of secondary-wall

formation at the level of cell biology 181.5 The larger picture 19

1.5.1 An abundance of wealth 191.5.2 The consumer is always right 211.5.3 The technology of wood-quality assessment

can be flawed 211.6 Discussion: seeing the wood and the trees 22

1.6.1 Philosophical and historical musings 22

vi CONTENTS

1.6.2 Wood quality measured on a three-pan balance 231.6.3 Lignin on the three-pan balance 231.6.4 Looking back 241.6.5 Looking forward 26

References 26

2 Wood anatomy in relation to wood quality 30BRIAN G. BUTTERFIELD

2.1 Wood anatomy 302.1.1 Softwoods and hardwoods 302.1.2 Growth rings 32

2.2 The cell wall of softwood tracheids 342.2.1 Cell wall structure 34

2.2.1.1 The middle lamella 352.2.1.2 The primary wall 362.2.1.3 The secondary wall 37

2.2.2 Cell wall and density 372.2.3 Microfibril angle 38

2.2.3.1 Determination of microfibril angle 382.2.3.2 Microfibril angle variation and its effect

on wood properties 402.3 The cell wall in hardwood fibres and vessel elements 43

2.3.1 Fibres 432.3.2 Vessels 43

2.4 Cell wall pits and perforations 452.4.1 Pits 452.4.2 Perforations 48

2.5 Vessel-less angiosperms 48Acknowledgements 49References 49

3 Wood chemistry in relation to quality 53HELENA PEREIRA, JOSÉ GRAÇA and JOSÉ C. RODRIGUES

3.1 Introduction 533.2 Chemical composition of wood 54

3.2.1 Cell wall structural components 553.2.1.1 Cellulose 553.2.1.2 Hemicelluloses 60

CONTENTS vii

3.2.1.3 Lignin 643.2.1.4 Distribution in the cell wall 68

3.2.2 Extractive components 713.2.2.1 Terpenoid extractives 723.2.2.2 Phenolic extractives 743.2.2.3 Other wood extractives 76

3.3 Variation of chemical composition 763.3.1 Juvenile wood 793.3.2 Heartwood 793.3.3 Reaction wood 803.3.4 Knotwood 81

3.4 Wood chemical quality parameters depending on end-use 81References 83

4 Wood density and growth 87PEKKA SARANPÄÄ

4.1 Importance of wood density 874.2 Density of cell wall material 874.3 Determination of density 90

4.3.1 Water displacement method 904.3.2 X-ray densitometry 91

4.4 What causes variation in density? 924.4.1 Within growth ring 934.4.2 Within a tree 974.4.3 Between sites 101

4.5 Is there a correlation between density and growth rate? 1024.5.1 Effect of fertilisation on growth rate and wood density 109

4.6 Conclusions 113References 113

5 Reaction wood 118JOHN R. BARNETT and GEORGE JERONIMIDIS

5.1 Introduction 1185.2 Early studies of reaction wood formation 1215.3 Induction of reaction wood formation 122

5.3.1 The role of auxin 1225.3.2 The role of ethylene 1245.3.3 The role of gibberellins 1245.3.4 The role of stress 124

viii CONTENTS

5.4 Structure and formation of reaction wood 1255.4.1 Compression wood 1255.4.2 Tension wood 1285.4.3 Opposite and lateral wood 132

5.5 Reaction wood and wood quality 132References 134

6 Growth stresses 137BERNARD THIBAUT and JOSEPH GRIL

6.1 Origin of growth stresses 1376.1.1 Geometric and mass growth: support stress 1376.1.2 Cell differentiation: maturation stress 1376.1.3 Growth stresses 1396.1.4 Role of growth stresses 1406.1.5 General models of growth stresses 140

6.2 Measurement of growth stresses 1416.2.1 In situ peripheral measurement 1416.2.2 Measurement of residual stresses in logs 1426.2.3 Main results for normal maturation strain 1426.2.4 Growth stresses and reaction wood 144

6.3 Consequence of growth stresses for quality 1476.3.1 Log-end cracks 1476.3.2 Lumber distortion 1496.3.3 Reaction wood 150

6.4 Prediction and treatment 1516.4.1 Tree and log morphology 1516.4.2 Consequences of cutting operations 1536.4.3 Observation of reaction wood 153

6.5 Conclusions 154References 154

7 Wood quality for pulp and paper 157DENILSON DA SILVA PEREZ and THIERRY FAUCHON

7.1 Introduction 1577.1.1 Why wood? 1577.1.2 Wood versus non-wood fibres 158

7.2 From wood to paper 1607.2.1 Wood as a raw material 1607.2.2 Wood–pulping process interactions 1617.2.3 Wood–pulp fibre relationships 164

CONTENTS ix

7.3 Resource management and biological decay 1657.3.1 Origin, supply of resources and pulp production 1657.3.2 Biological decay 1667.3.3 Mill specifications and quality control measurement 170

7.4 Wood-quality variability and its consequences for pulp and paper quality 1737.4.1 Wood species and mixtures 1737.4.2 Within- and among-tree property variation 179

7.4.2.1 Within-ring variation 1797.4.2.2 Radial trends 1807.4.2.3 Variations from the base to the top 181

7.4.3 Forestry practice, site index and growth conditions 1837.5 The future 184References 184

8 The mechanical properties of wood 187AUDREY ZINK-SHARP

8.1 Introduction 1878.2 Advantages and disadvantages of wood as a structural

material 1888.2.1 Advantages 189

8.2.1.1 High strength and flexural rigidity in spite of light weight 189

8.2.1.2 Available and renewable resource 1898.2.1.3 Requires less energy to process into

structural material 1898.2.1.4 Ease of fabrication and conversion 1908.2.1.5 Dimensionally stable and durable if used

correctly 1908.2.1.6 Low electrical, thermal, and acoustical

conductivity 1908.2.2 Disadvantages 190

8.2.2.1 Variability 1908.2.2.2 Natural built-in defects 1918.2.2.3 Dimensional instability 1918.2.2.4 Susceptibility to biological attack 1918.2.2.5 Anisotropy 1918.2.2.6 Combustibility 192

8.3 Importance of density 1928.4 Mechanical properties important for structural applications 194

8.4.1 Elastic properties 1948.4.2 Strength properties 196

x CONTENTS

8.5 Creep effects on deformation and fracture 1988.6 Defects affecting mechanical properties 199

8.6.1 Naturally occurring defects: knots and sloping grain 1998.6.1.1 Knots 1998.6.1.2 Sloping grain 200

8.6.2 Processing defects: checks and splits 2038.6.2.1 Checks 2038.6.2.2 Splits 203

8.7 Problems with mechanical joints 204References 209

Index 211

Contributors

Prof. John R. Barnett School of Plant Sciences, The University ofReading, Reading, RG6 6AS, UK

Prof. Brian G. Butterfield School of Biological Sciences, The Uni-versity of Canterbury, Private Bag 4800,Christchurch, New Zealand

Dr Thierry Fauchon AFOCEL, Wood Process Laboratory,Domaine de l’Etançon, 77370 Nangis,France

Dr José Graça Centro de Estudos Florestais, Departa-mento de Engenharia Florestal, InstitutoSuperior de Agronomia, Tapada da Ajuda,1349-017 Lisboa, Portugal

Dr Joseph Gril Laboratoire de Mécanique et Génie Civil,Université Montpellier II (UMR 5508 –CNRS), Sciences et Techniques du Langue-doc, Case 081 – Bâtiment 13, Place E.Bataillon, 34095 Montpellier Cedex 5,France

Prof. George Jeronimidis Centre for Biomimetics, School of Con-struction Management and Engineering,The University of Reading, Reading, RG62AY, UK

Prof. Helena Pereira Centro de Estudos Florestais, Departa-mento de Engenharia Florestal, InstitutoSuperior de Agronomia, Tapada da Ajuda,1349-017 Lisboa, Portugal

Dr José C. Rodrigues Centro de Estudos de Tecnologia Florestal,Instituto de Investigação Científica Tropi-cal, Tapada da Ajuda, 1349-017 Lisboa,Portugal

xii CONTRIBUTORS

Dr Pekka Saranpää METLA, The Finnish Forest ResearchInstitute, P.O. Box 18 (Jokiniemenkuja 1),FIN-01301 Vantaa, Finland

Prof. Rodney Arthur Savidge University of New Brunswick, Forestry andEnvironmental Management – 160 126,New Forestry, Fredericton, New Bruns-wick, Canada E3B 6C2

Dr Denilson da Silva Perez AFOCEL, Wood Process Laboratory,Domaine de l’Etançon, 77370 Nangis,France

Dr Bernard Thibaut Laboratoire de Mécanique et Génie Civil,Université Montpellier II (UMR 5508 –CNRS), Sciences et Techniques du Langue-doc, Case 081 – Bâtiment 13, Place E.Bataillon, 34095 Montpellier Cedex 5,France

Prof. Audrey Zink-Sharp Virginia Tech, Wood Science and ForestProducts, 210 Cheatham Hall (0323),Blacksburg, VA 24061, USA

Preface

Wood is undoubtedly the most versatile raw material available to man. It isburnt as fuel to provide energy (accounting for about 70% of all woodharvested), shaped into utensils and implements of various kinds, used as acost-effective structural engineering material, converted into fibres for mostpaper production, and put to newer uses as a source of industrial chemicals.The steady increase in the demand for wood, resulting from a concomitantincrease in its applications, means that pressure on forests is constantlyincreasing. The need to cut down trees for wood is in direct conflict with theneed to preserve forests for the conservation of biodiversity and as sinks forcarbon dioxide. It is therefore essential that forests are managed sustainably, ifdemand is to continue to be met without detriment to our environment. Thiscan be achieved by developing new forests and replacing trees that are har-vested, while at the same time ensuring that the trees that are grown producewood of good quality.

The problem lies in the definition of wood quality. Wood which mayproduce pulp with good paper-making properties may not be suitable for use inconstruction, for example. The intrinsic variability of wood properties is alsoof concern in relation to quality. In the case of the paper industry, the pulpingprocess is modified and the fibres are blended to produce a uniform end prod-uct. The construction industry relies on the grading of timber at the sawmill toselect those timbers which are fit for purpose. Both processes have importanteconomic implications.

For this reason, selection of seedlings for planting based on their potentialwood properties should depend on their anticipated use. However, it isimpossible to predict what the requirements might be 50–70 years later whenthe tree is ready for harvesting. Because of this, the pulp industry is beginningto look at fast-growing species, such as hybrid poplar, to be harvested in lessthan 10 years. The juvenile characteristics associated with the timber fromthese trees make them unsuitable for other high value purposes, except perhapsas veneers. There are good biological reasons why this juvenile wood developsand is required by saplings, but its presence is currently exercising the minds ofwood scientists concerned about its inferior properties as a raw material.

The quality of wood results largely from the chemical and physical struc-ture of the cell walls of its component fibres. This can be modified in natureas the tree responds to physical environmental stresses, such as wind actingon the growing tree. Internal stresses can accumulate, which are released

xiv PREFACE

catastrophically when the tree is felled, rendering the timber useless, or at leastreducing its value considerably. The quality of timber as an engineeringmaterial also depends on the structure of the wood and the way it hasdeveloped in the living tree.

Thus, tree improvement for quality cannot be carried out without an under-standing of the biological basis underlying wood formation and structure.Wood is what it is because it is made by trees, and the question then is what aretrees doing to wood? The primary aim in preparing this volume was to bringtogether the viewpoints of biologists and physical scientists, to cover the spec-trum from the formation of wood to its structure and properties, and to relatethese properties to industrial use. We have attempted to produce a book whichis different from those concerned entirely with the biological or the engineer-ing aspects of wood, and we hope that it will provide useful insights into bothindustrial and academic aspects of the subject.

We are grateful to all those who have contributed chapters.J.R. Barnett

G. Jeronimidis

1 Tree growth and wood quality Rodney Arthur Savidge

1.1 Cambial growth

1.1.1 Wood is a biosynthetic end product

Cambial (or secondary) growth comprises innumerable phenomena of biophysics,biochemistry and cell biology, and few of these phenomena are yet wellunderstood (Savidge et al., 2000; Savidge, 2001a). There has never beena dedicated resolve on the part of either forestry or biology to develop an in-depthunderstanding of how trees make wood. Consequently, progress has been afunction of the piecemeal efforts of a few individuals and small groups. Under thepressure of increasing demands for wood and wood fibre associated with worldpopulation growth, and in the face of dwindling forest area containing increas-ingly juvenile stock, the need to have greater knowledge of the biological factorscontrolling wood supply, in terms of both quantity and quality, seems obvious.Moreover, the international climate-change community, evidently unawareof how little is understood about wood formation in trees, has identified forestsas important sinks for draining off excess atmospheric carbon dioxide, thecapacity of which can supposedly be readily increased (Savidge, 2001b).

Based on the different wood anatomies of conifers (softwoods) and hardwoods,softwood cambial growth gives the impression of being the less complex. Soft-woods are also of worldwide distribution and importance, making considerationof their secondary growth a logical starting point for summarizing what isknown about life processes underlying wood formation and the control ofwood quality. Some aspects of hardwoods are considered later in this chapter.

1.1.2 Zonation

Under the microscope, the actively growing cambium of a conifer exhibitsseveral developmental zones. The cambial zone (CZ, Fig. 1.1A) is seen tocomprise two cell types, ray and fusiform cambial cells. The former areapproximately isodiametric whereas the latter are many times longer than widewhen measured across either their radial or tangential axes of symmetry.Through periclinal cell divisions, both ray and fusiform cambial cells produceradial files of daughter cells, adding inwardly to the pre-existing wood andoutwardly to the pre-existing phloem (Fig. 1.1A). New fine structural detailsof the periclinal division as it occurs in conifers following cryofixation were

2 WOOD QUALITY AND ITS BIOLOGICAL BASIS

Fig. 1.1 (A) Cross section of the region of active wood formation in Pinus contorta. CZ, cambialzone; RE, zone of primary wall radial expansion; SL, zone of secondary wall formation andlignification, with mature tracheids below. A ray bisects the field of view. Bar = 100 μm. (B) Highermagnification of A showing the region transitional between the RE and SL zones, with a ray bisectingthe field. Bar = 20 μm. (C) Cross section of wood of Abies balsamea. The compound middle lamellaand tripartite secondary wall are evident, as are bordered pits. A ray bisects the field. Bar = 40 μm.

TREE GROWTH AND WOOD QUALITY 3

recently reported (Rensing et al., 2002). Younger cambium in terminal parts oftrees and older cambium at the bases of aged trees generally have shorter fusi-form cells than cambium in the intervening stem region. Fusiform cell lengthis a major factor in determining the final length of conifer tracheids (or fibres);consequently, the longer fibres in uncultivated aged trees are found in themiddle-aged region of the stem.

On the inner periphery of the CZ, cell-division activity is supplanted byenlargement of CZ daughter cells by expansion of the primary cell-wall surfacearea. It occurs primarily in the radial direction, creating a zone of radiallyexpanding and expanded primary-walled cells (RE, Fig. 1.1A). Although celldivision and expansion both contribute to overall increase in tree girth,expansion – thought to be driven by turgor pressure and facilitated by auxin-promotion of cell-wall loosening – is the primary means of moving the CZcentrifugally (Savidge, 1996).

During, but usually near the completion of each cell’s primary-wall expansion,a decision is made to either initiate or not the development of bordered pits(Fig. 1.1D). Those RE cells undergoing bordered-pit development subse-quently begin producing secondary-wall lamellae and lignin, thus generating thesecondary-wall forming layer (SL), a zone of still-living cells (Figs 1.1A–D).

1.1.3 Bordered-pit development

Bordered-pit numbers are typically highest in earlywood and decline to negligiblefrequency in the last-produced latewood tracheids of each annual ring. Borderedpits are usually confined to radial walls, although they occur naturally intangential walls, particularly of latewood tracheids at low abundance (Panshin& de Zeeuw, 1980). Tangential wall bordered pits have been induced todevelop experimentally in large numbers by manipulating auxin concentration(Leitch & Savidge, 1995).

Both the cell biology and biochemistry underlying bordered-pit developmentremain uncertain, but it is clear that the process involves a series of successive

Fig. 1.1 (continued) (D) Radial section of P. resinosa developing earlywood, with mature latewoodon the far right. The arrow points to a bordered pit at an early stage of development of its over-archingborder. The bar (50 μm) is in a ray tracheid having smaller diameter bordered pits. (E) SEM view of acompression wood tracheid showing separations in the microfibrillar matrix of the S2 layer (arrow). Thebar at lower right is 4 μm. (F) Tangential section of Picea glauca showing a microdomain (arrowed) oftracheids with upward to left orientation relative to the surrounding elements. Bar = 200 μm. (G) Radialview of Betula alleghaniensis showing a juvenile core that changes abruptly, at the arrowed location, to adifferent wood. Bar = 15 mm. (H) Cross section at the arrowed point in G showing the interface betweenthe corewood (lower) and the exterior wood (above). The bar at lower right is Bar = 10 μm.


and quite profound modifications to the compound middle lamella betweenadjoining RE cells, as can be seen in Fig. 1.1D. The spatial correspondence ofthe borders in adjoining cells is clear evidence that some form of intercellularcommunication occurs (Savidge, 2001a). After the circular site, or margo, ofthe pit-to-be has become visible and the torus (not found in bordered pits of allspecies but common in the Pinaceae) has begun to form in an RE cell, birefringentputatively cellulosic highly oriented microfibrils begin to be deposited exclusivelyaround the circumference of each margo. This localized deposition of circularlyoriented crystalline microfibrils continues, around and around, progressinggradually inward while winding upward towards an imaginary line at the centreand perpendicular to the plane of the margo. Thus, the over-arching pit borderis formed, leaving an aperture usually about half the diameter of that of themargo. During its formation, the over-arching border can be isolated as a discretering (Savidge, 2000a); however, other evidence indicates that over-archingborders of discrete bordered pits are actually interconnected by thin microfibrillarstrands (Savidge, 1996).

The presence of a spherical organelle tightly appressed to the plasmamembrane, such that the side in contact with the membrane is flattened, maybe the explanation for how the margo and over-arching border arise (Savidge,2000a). Contact between the cell membrane and the flattened organelle isenvisaged to prevent microfibril deposition occurring within the area of themargo and, if the organelle contains lytic enzymes (as supposed from itsevidently vacuolar origin – Bethke et al., 1998; Savidge, 2000a), it couldconcomitantly serve to hydrolyze non-cellulosic constituents of the underlyingcompound middle lamella. The upper, domed surface of the membrane-boundorganelle would obviously also serve as template for the formation of thecircular over-arching border.

1.1.4 Secondary-wall lamellae

After the onset of bordered-pit development, general secondary-wall polysac-charide deposition commences, followed by the initiation of lignification inthe middle lamella, most conspicuously at cell corners. By the completion ofthese processes, three secondary-wall layers appear to be present when viewedunder the light microscope (Fig. 1.1C), although electron microscopy indicatesthat each layer actually comprises a number of sub-lamellae. This structure hasprofound implications for wood properties and utilization, and it is describedin detail in Chapters 2 and 4. The chemical structure of the wall is the subjectof Chapter 3. [The problem of dimensional stability in wood ultimatelyreduces to its chemistry, in particular the relative abundances of differentchemical bonds to resist stress and stabilize the macrostructure. What is reallyneeded in this area is a detailed understanding of the chemical reactionsoccurring during wood formation and how they are controlled.]


1.1.5 Microfibrils and lignin

The essential events underlying differentiation of a cambial derivative intoa woody element are microfibril deposition and lignification. Considerableresearch has been conducted into both over the last century, but at the level ofbiochemistry much remains to be discovered about how these processesproceed (Lewis & Sarkanen, 1998; Atalla, 1998; Delmer, 1999; Brett, 2000;Savidge, 2000a; Savidge & Förster, 2001). To do justice to the continuinguncertainty, it would probably be correct to say that most research so farattempted has suffered from a lack of clear definition, insufficient material forinvestigation and/or a lack of resolution. Compounding the overall problem, nosubstantial and therefore convincing synthesis of cellulose in vitro using a cell-free biological system has yet been achieved, despite a number of reports onputative cellulose synthase genes (Delmer, 1999; Brett, 2000; Taylor et al.,2000; Williamson etal., 2001; Desprez etal., 2002).

A common conclusion from many scientific investigations has been thatmicrofibril orientation is determined by the orientation of cortical microtubules(Chaffey et al., 2000), but no shortage of additional investigations haspresented data indicating the contrary. Many have noted a lack of correlationbetween the two orientations (e.g. Sugimoto etal., 2000a,b; Bichet etal., 2001).It has also been suggested that microfibril production must precede corticalmicrotubule orientation (Fisher & Cyr, 1998). Indeed, some evidence indicatesthat there may be no need for microtubules at all during microfibril deposition(Savidge & Barnett, 1993).

Although the field of microtubule–microfibril correlation analysis seems tobe at an impasse, there is substantial experimental evidence indicating that theorientation of cortical microtubules is altered by the phytohormone environment(Blancaflor & Hasenstein, 1995; Wenzel et al., 2000). Microfibril has becomesynonymous with cellulose, but xyloglucan and glucomannan microfibrils – easilyconfused with cellulose microfibrils when merely imaged at the structurallevel – are well-known constituents of secondary-wall layers and, arguably, areas important as cellulose in determining many of the properties of wood (Jones,1971; Wilkie, 1985; Brett, 2000). Could the explanation be that some micro-fibril polymerizations are, and others are not, linked to cortical microtubules?

The biochemistry of lignification is supposedly far better understood thanthe formation of microfibrils, but the reality is that it also remains an openquestion, far from completely understood (Lewis & Sarkanen, 1998; Savidge& Förster, 2001).

1.1.6 Protoplasmic autolysis

Biological control of protoplasmic autolysis during xylogenesis has receivedlittle attention. It should be noted that sapwood is not, in contrast to popularperception, a dead tissue. A large proportion of cambial derivatives which


become incorporated into sapwood actually remain living, for example as rayand axial parenchyma, and as non-autolyzed fibres, those cells dying onlyyears later in association with heartwood formation (Savidge, 1996). Hard-wood fibres tend to be ambivalent in pursuing a programme of cell death, buttracheary elements (i.e. tracheids and vessel members) in both conifers andhardwoods seem to be committed to apoptosis occurring within at least a year,and usually within a month, after the elements have otherwise differentiated. Asthe newly matured and protoplasmically autolyzed woody elements becomedistanced from the CZ, water begins flowing from one element to anotherthrough bordered pits and, in the case of hardwoods, through the perforationplates separating vessel members. Thus, water is distributed throughout thetree. It is probable that the living component has a role in resisting the onset ofdecay, and it is well established that diurnal and phenological changes inchemistry are normal to the sapwood, a result of their metabolism.

1.1.7 Cambial fusiform cell length and orientation

Fusiform cambial cells usually, but not invariably, give the appearance of beingstretched in a direction more or less parallel to the long axis of the stem (or branch,or root). Fibre elongation in hardwoods is promoted by the phytohormone classknown as gibberellins (Stant, 1961; Eriksson etal., 2000), and these presumablyalso influence the elongated character of fusiform cambial cells in all species(Savidge, 1985; Kijidani etal., 2001). Gibberellin promotion of fusiform cambialcell elongation remains to be unequivocally demonstrated, however. Continuingbasipetal transport of auxin through the cambium was found to maintain thefusiform nature of cambial cells, preventing them from shortening and becomingseptate axial parenchyma (Savidge, 1983; Savidge & Farrar, 1984). As shownin Fig. 1.1F, microdomains of disoriented cells sometimes arise within other-wise oriented populations, and the control mechanisms underlying microdomainformation, although still poorly understood, are believed to be at the heart ofspiral, wavy and interlocked grain formation in trees (Savidge & Farrar, 1984).

1.2 Perennial cambial growth

1.2.1 Episodic but variable cambial growth

Wood can be seen as an engineering material, an aggregate of fibres and fines,or as a mass of different chemical substances and voids in combination, but morefundamentally wood is a biological end product generated during cambialgrowth over successive years. Cambial growth is episodic, restricted in temperatezones to the warm spring and summer months when photoperiod is long, andlimited to periods of water availability in the tropical zones (Savidge, 1993).The outcome of each growth episode can be envisaged simplistically as the


formation of an inverted cone, or layer, of durable, supportive woody elementsdeposited upon the pre-existing structure. Each new layer of wood at thecommencement of its formation becomes chemically cemented to the woodyelements of the preceding layer through covalent bonding between lignin,polysaccharides and other substances. Consequently, wood when green usuallyis a seamless, continuously reinforced material, although there are exceptions(Fig. 1.1G). Vertical or rotational shear-slip at the boundary between increments(Fig. 1.1H) is encountered only following stresses sufficiently severe to crushor otherwise deform the weaker earlywood elements.

1.2.2 Tapering to the point in form and function

Every tree tapers to minute apical meristems supporting primary (or extension)growth. Cambium arises immediately basal to sites of primary growth; thus,cambial age and therefore the time available to increase the girth of the corres-ponding axis necessarily vary along the axis. Sites of primary growth are thelocations within the tree having the highest concentrations of nutriment.Consequently, both solute concentration and water-potential gradients extendlongitudinally over stem and branch axes. In other words, the intrinsic environ-ment experienced by individual cambial cells will vary from one point to thenext, even when the extrinsic environment of the whole tree is maintainedconstant within a controlled growth chamber.

The reality of chemical and physical gradients extending longitudinally, andalso circumferentially and radially, in the cambial region is surely the key factordetermining the variable nature of wood (Savidge, 1996, 2001a). In the finalbiological analysis, individual elements of wood constitute phenotypes arisingthrough interactions between each cambial derivative’s genotype and its sub-cellular environment (Savidge, 1996, 2000a). The cambial genotype throughouta tree can be assumed to be constant, but because the physical and chemicalenvironments experienced by cambial cells vary at different locations, the natureof gene expression and the resulting end products reasonably can also be expectedto vary. Thus, although perhaps not welcome information for wood-processingindustries, within the framework of the biological sciences it is entirely to beexpected that the nature of wood must inevitably vary over the tree.

The above may give the impression that cambial growth, though variable,always occurs, and certainly the concept of annual rings and the physiologicallyunsubstantiated interpretations given to them by the field of dendrochronologyreinforce that supposition. On the other hand, cambial growth in conifers iscommonly suppressed at the bases of a tree’s lower live branches, as well as atthe bole base, in old trees (Meredieu & Caraglio, 2002). Moreover, pronouncedtaper does not necessarily always attend morphogenesis of perennial woodyplants having cambium. Lateral roots, for example, typically extend throughsoil for many metres nevertheless maintaining quite small diameters over the


distance, little or no cambial growth attending their primary growth. Vines alsoare quite capable of focusing their energy and biomass allocation on extensiongrowth, although cambial growth does occur to a limited extent particularlynear the base of the vine. Many hardwood species when grown in tight quartersallocate minimal resources to secondary growth, instead growing in height toproduce long slender stems. Considering the differences between roots, vinesand trees, there appears to be a connection between the occurrence of cambialgrowth and the need of the stand-alone tree to biosynthesize sufficient structuralsupport to resist the force of gravity.

1.3 Wood quality in perspective

1.3.1 Defining wood quality

Wood quality, as understood within both dictionary and practical contexts, hasto do with the degree of excellence – in relation to some preconceived applica-tion(s) – of each log, piece of wood or woody fibre under consideration. Becausequality assessment is multi-faceted and depends on the intended application,there is no absolute measure. Quality assessment by the woodsmen who fell andprocess the trees and by the mill workers who decide how logs should be usedinvolves experienced observation and snap-judgment integration of particularfeatures, based largely on subjective experience.

1.3.2 Measuring wood quality

Some aspects of quality, such as wood density, cellulose, lignin or extractivecontents, can be repeatedly analyzed and quantitatively expressed with highaccuracy and precision, although always within the proviso that the estimate maybe accurate only for the sample actually measured and within the methodemployed. Other measures, such as fibre length, cell-wall thickness, microfibrilangle, bordered-pit number and percentages of the various types of woodyelements coexisting in a wood are more problematic. Upon repeated randomsampling and measurement of the same preparation, a Gaussian distribution ispredictably obtained, and the magnitude of the standard deviation may provideequally or more important information than the mean value. However, thatinformation essentially is no more than a confirmation of what can be readily seenwhen viewing a section of wood in the compound light microscope (Fig. 1.1).Immense variation exists within wood. In other words, the sub-micrometre pre-cision of microscopic imaging and associated measuring of small samples ofwood readily and consistently reveals that so-called accurate estimates on woodsamples obtained through physical and/or chemical analytical procedures in factare low resolution, relatively crude simplifications. They do little justice to theimmense variation existing in wood, from nanometre to higher scales.


1.3.3 Wood quantity versus wood quality

Wood quantity and quality can be contemplated independently of one another,at least in an abstract sense. Whereas wood quantity is concerned with the totalamount of wood (measured as diameter, length, volume, weight or any combi-nation thereof) and can be expressed numerically with accuracy and precision,quality assessment frequently involves subjective estimations. However, bothhave their origin in the cambial region, and the dimensions of the bole usuallyhave a major influence on quality as well as quantity, particularly at the appliedlevel. For example, clear or knot-free wood is obtained from larger diametertrees; lumber dimensions in the mill are decided on the basis of log length anddiameter; slow-grown conifers with smaller diameter stems have more denseand therefore stronger woods than those of fast-grown trees; and more denseand therefore higher yielding pulps come from fast- as compared to slow-grownhardwoods. In other words, fundamental research projects focusing on the con-trol of cambial growth, allocation of photosynthate and nutrients to cambium,metabolism in the cambium, water potential relations across, along and aroundthe cambial sheath, and cambial cell biology in general have direct relevance tounderstanding wood quality as well as wood quantity (Larson, 1969).

1.3.4 Stem dimensions and quality

Stem dimensions are a function not only of how quickly height and girthincrease during any one episode of growth, but of how many growth episodeshave occurred. A massive millenary redwood or a 140-year-old straight-stemmedAmerican chestnut (Figs 1.2A and B, respectively) generally would be expectedto have high quality boles. On the other hand, as Figs 1.2C and D show, despitetheir ages, the boles of a century-old eastern white pine or of an 800-year-oldwhite spruce might fail most quality tests, excepting perhaps those of specialistwood artificers. Thus, aging, although usually an enhancing factor, is of itselfno guarantee that the wood in a tree will be of high quality in relation to themore common uses of wood.

Although the many wood-quality problems attending taper appear to beunavoidable during the youth of the stand-alone tree, trees when approachingtheir maximum height tend to lose taper, secondary growth continuing especiallyin the crown after primary growth has been reduced to production of new leavesor needles with little associated extension growth (Fig. 1.2). Thus, the problemof taper finds a solution in time, and the primary controlling factor ultimately isthe vision forest management can muster in its long-term planning.

1.3.5 G×E control of wood quality

There are both genetic and environmental reasons for time not being the soledeterminant of stem dimensions (Savidge, 1996, 2000a, 2001a). The genetic


Fig. 1.2 (A) Mature Sequoiadendron giganteum bole in Mountain Home Demonstration State Forest,California. For scale, a man at the base of the tree is arrowed. (B) A young Castanea dentata bole ofgood form, Mount Uniacke, Nova Scotia. (C) Pinus strobus, typical of the trees rejected by loggersthroughout New Brunswick, Canada. (D) Picea glauca var. albertiana, severely spiral grained, nearMount Nansen, Yukon Territory, Canada. (E) P. glauca phenotype in the New Brunswick BotanicalGarden. (F) Normal branched and foxtailed Pinus caribea (reproduced with FAO permission fromKozlowski & Greathouse, 1970). (G) Another interesting Picea glauca phenotype in the NewBrunswick Botanical Garden.


makeup (or genotype, G) of the tree influences both competence for growthand the physico-chemical nature of growth. The environment (E) that the treegrows in may serve to accelerate or retard growth competencies and otherwisemodify the physico-chemical attributes arising during growth. What is trulyremarkable about trees is the immense phenotypic plasticity that, if one onlysearches, can be found within every species (Figs 1.2E–G). Again, wood shouldbe seen for what it truly is, a biological phenotype, not merely a raw resourcemined from the forest. Wood is a permanent record of G × E interactions thatoccurred over many years within a highly versatile biological system. Duringthe process of cambial growth and afterwards, as sapwood changes into heart-wood, G ×E interactions occur at the sub-cellular level (Savidge, 1996, 2000a).Thus, sub-cellular level research, although little supported, is where attention mustbe focused if sense is ever to be made of the bigger picture.

A number of traits were identified as worthy of attention early in the devel-opment of tree improvement programmes (Ehrenberg, 1970; Gerhold, 1970;Harris, 1970; Zobel & van Buijtenen, 1989). Ehrenberg (1970) identifiedstem straightness/taper, reaction wood, grain angle, fibre length/width/cell-wallthickness, wood specific gravity and knot characteristics as influenced by branchnumber/diameter/angle/self-prunability. Harris (1970) added extractives,heartwood, the relative proportions of the different woody elements, particu-larly in hardwoods, and the cellulose:lignin ratio to the list. Gerhold (1970)emphasized the need to focus on resistance to damaging insects and microorgan-isms. In an emerging area, Lamlom and Savidge (2003) found that carboncontent in wood varies substantially between tree species. It is one thingto rationalize a wood-quality trait as having a genetic component, quite anotherto be able to fix one which has been perceived as undesirable.

In general, the approach taken by the tree improvement/selection communityto modifying wood quality has been to identify and focus on those traits exhibit-ing a broad range of variation, on the supposition that a substantial portion ofthat variation may not simply have a genetic basis, but also be strongly herita-ble. For example, while trees usually exhibit a branching form, trees havingfew or no branches are well known and demonstrate the possibilities of mini-mizing both knots and taper (Fig. 1.2F). Thus, traits such as height growth,branch angle and wood specific gravity have been given major attention andsuccessfully modified in plantation stock. Conversely, traits yielding a narrowrange of natural variation in wild-type populations have been neglected withintree improvement programmes. For example, although cellulose is, by bio-mass, the single most abundant component of wood and unquestionably isencoded for in the terrestrial plant genome, its percentage content in woodvaries but little between genotypes. Similarly, lignin content, although recog-nized to vary marginally between provenances, and even better understoodthan cellulose in terms of its genetic basis, was essentially rejected by the treeimprovement community as an area where gains were unlikely. Consequently,


the possibilities of modifying cellulose and lignin contents in wood awaited theadvent of genetic engineering (Savidge, 1985; Hu etal., 1999). The tree improve-ment focus on wood density is considered below (Section 1.4.3).

1.3.6 Variation taraplas within the tree

Whether analyzed transversely across radii, around the perimeter or longitudi-nally along the stem (taraplas), wood varies in its chemistry (composition,relative percentages, covalent bonding, etc.), mechanical and anatomicalproperties (Paul, 1963; Megraw, 1985; Sjöström, 1993). Woody fibres typicallyare much longer – not uncommonly 100 times – in the extended (nominally axial)dimension than in their radial and tangential widths. This anisotropy greatlyinfluences the properties of wood, but it is not so much anisotropy per se butthe lack of consistent anisotropic properties combined with many other physico-chemical variables taraplas that make woods so fascinatingly complex. In thislight, it is not surprising that dimensional instability often emerges as a woodquality problem in dry structural members cut from logs. Rather, it is remarkablethat dimensional instability problems are not even more prevalent. Variationtaraplas compounds the physical complexity existing in an already anisotropicsystem, forcing analysis of a host of additional variables in order to detect,model and understand wood quality. The biological reality makes it unlikelythat any but immensely complex mathematical formulae could be developed torepresent variation taraplas accurately, for general modelling purposes.

Major differences in wood quality are readily apparent in the various stemsshown in Fig. 1.3. Physical stresses exist, taraplas, i.e., in the three major axesof stem symmetry as indicated in Fig. 1.3A. These intrinsically normal stressesunderlie much of the defect present within the standing tree and arising sub-sequently, when the log is converted into a pole, cut into planks or cants, orreduced to fragments or fibres (Panshin & de Zeeuw, 1980). Varied grainangles and varying extents of taper, check and shake, decay and frost cracking(Figs 1.3B–F respectively) are just a few of the wood-quality problems commonlyencountered.

In the fresh wood of trees of merchantable diameter, longitudinal, more or lessparallel to the grain, tensile stresses exist at the wood surface (Fig. 1.3A).However, tension in the longitudinal direction declines inwardly with increas-ing radial distance, longitudinal compressive stresses being found near the pith.The stresses existing around the circumference (i.e. perpendicular-to-the-graintangential stresses) tend to be the converse of longitudinal stresses, compressivestresses prevailing at the log surface and tensile stresses existing near the pith.In stems having solid wood (i.e. lacking internal decay or other defect), weaktensile stresses exist across the entire transverse radius (Fig. 1.3A).

In general, the stresses existing in green wood are cumulative, and in theabsence of biological modifications (such as those brought about throughdecay microbes, ants, beetles, etc., Fig. 1.3E) or physical failure (Fig. 1.3F),


Fig. 1.3 (A) Freshly cut logs of Picea rubens indicating by the direction of the arrows the tensile andcompressive forces in the wood of the bole. (B) Pinus contorta logs showing upward to the right (left),straight (centre) and upward to the left grains (right) that developed in trees of the same age on the samesite. (C) Thuja occidentalis logs. Note the pronounced taper. (D) Butt of a T. occidentalis showing thechecking (large arrow) and shaking (arrowhead) that typically occur after the tree is felled as the wood dries.(E) Heartrot in T. occidentalis. The log at lower right also shows an ant cavity. (F) P. rubens showing asubstantial frost crack terminating at a shake (arrowed) that probably developed early in the life of the tree.


stress gradients increment and continue to adjust as the tree grows (Mattheck& Kubler, 1995). The cell-wall softening effect of water, when not limiting,serves to partially relax the strain. Consequently, intrinsic growth stressesalthough always present rarely lead to mechanical failure. When a tree isfelled, the butt develops radial cracks in response to the release of longitudinalstresses and the concomitant induction of transverse tensile stresses (Fig. 1.3D).The tendency for internal failure is exacerbated by the rapid drying occurringfrom the exposed wood surface, which typically results in butt-localizedchecking (Fig. 1.3D). When planks or cants are cut from logs, the physicalequilibrium within the green tree becomes further disrupted, and bowed, warpedand twisted end products commonly result (Panshin & de Zeeuw, 1980).

1.4 Wood density

1.4.1 Molecular and anatomical basis

Wood specific gravity, or density, varies by species and generally is the primewood quality consideration for industry, higher values yielding stronger woodsand more pulp (Stamm & Sanders, 1966). It should not be overlooked, how-ever, that low-density woods can be of high quality and value in service toman. Examples of the latter include very low-density balsa wood (Ochromalagopus) and the low density but relatively strong and beautifully figuredwoods of Paulownia spp.

Figure 1.4 shows some examples of the anatomical basis for the extensive vari-ation in density that can be found in hardwoods. Vessel diameter and frequencytend to be inversely proportional to wood density, particularly in ring-porousspecies having large diameter earlywood vessels. Thus, woods having relativelyhigh fibre:vessel ratios yield more dense woods than those with lower ratios.

1.4.2 Enhancing wood density through silviculture

Paul (1963) and Megraw (1985) provided valuable insight into how wood den-sity and other wood quality features can be manipulated through silviculturaltreatment, and many others have considered the theoretical basis for exploit-ing phenotypic plasticity through silviculture (Sunley, 1963; Jackson, 1965;Wareing, 1966; van Buijtenen, 1969). Irrigation, for example, can greatlyincrease the latewood to earlywood ratio of temperate-zone conifers, evidentlyby forestalling the entry of the cambium into dormancy during the period ofcambial growth in mid- to late-summer (Paul, 1963). Water availability alsodetermines the nature of wood forming in hardwoods (Bissing, 1982). Ingeneral, silvicultural treatments accelerating hardwood growth rate canconcomitantly have an enhancing effect on wood density (Figs 1.5 and 1.6).


Fig. 1.4 Variation in wood. (A, B and C) Fraxinus americana, showing changes in the rate ofdiameter growth and structure of annual rings. (A) Cross section showing transition from wood ofrapid growth and high specific gravity near the centre of the tree (a, lower right) to wood of slowgrowth and low specific gravity. (B) Narrow rings having oven dry specific gravity 0.48, at the pointb indicated in A. (C) Wide rings having oven dry specific gravity 0.65, at the point c indicated in A.(D and E) Acer saccharum, showing pore content per unit area. (D) High specific gravity. (E) Lowspecific gravity. (F, G and H) Carya ovata, relation of wood density to ring width in old growth.A photograph of the disk surface is shown beneath each photomicrograph. (F) Narrow growthrings near the pith, wood of high density; (G) then wider growth rings, wood of high density;(H) subsequent narrow growth rings, wood of low density. (I and J) Liriodendron tulipifera,showing structure of wood produced during periods of rapid and of slow growth in old-growth trees.(I) Wood produced during the early life of the tree while it had growing space, oven dry specificgravity 0.41. (J) Wood produced when the trees were keenly competing for growing space andmoisture, oven dry specific gravity 0.31. (All images in Fig. 1.4 are reproduced from Paul (1963)with permission of the USDA Forest Service.)


1.4.3 Enhancing wood density through tree improvement

Tree improvement programme emphasis has been on enhancing height growthand wood specific gravity simultaneously (Zobel, 1961; Northcott, 1965; Nikles,1970; Zobel & van Buijtenen, 1989). Impressive gains have been made; how-ever, a serious and all too prevalent quality problem (but nevertheless fascinatingbiological development) has emerged with conifers (common for example inPinus radiata). Rapid height growth is conducive to the formation of compres-sion wood in the juvenile core, possibly because cambial auxin is elevated tohyper-physiological levels as a result of the rapid height growth (Savidge, 1983,1986; Timell, 1986). Compression wood, although of relatively high density,tends to be dimensionally unstable and fails without warning. Thus, compression

Fig. 1.5 Growth and specific gravity in two hardwood species. The two photographs on the left are ofRock elm (Ulmus thomasii). The two on the right are Sugar maple (Acer saccharum).

Rock elm: (A) Typical 220-year-old tree that had experienced a long period of crowding, whichdecreased the specific gravity of the wood. As a result of thinning, a rapid growth period followed.(B) A typical tree that maintained a dominant position in the stand throughout its life and at 200 yearsof age was growing rapidly and producing wood of high specific gravity.

Sugar maple: (A) This tree, 134 years old, produced wood of high specific gravity until the period ofmaximum rate of diameter growth was reached. Afterwards, as a result of crowding and reduction ofcrown size, specific gravity decreased. (B) This 100-year-old tree had more growing space, developed alarger crown, maintained rapid diameter growth and produced wood of uniformly high specific gravity.(All images in Fig. 1.5 are reproduced from Paul (1963) with permission of the USDA Forest Service.)

Fig

. 1.6

Que

rcus

alb

a sp

ecif

ic g

ravi

ty v

aria

tion

. (A

) C

ompo

site

dia

gram

of

grow

th a

nd s

peci

fic

grav

ity

for

five

old

-gro

wth

whi

te o

aks,

sho

win

g av

erag

e de

crea

sefr

om c

entr

e to

bar

k. (

B)

Com

posi

te d

iagr

am o

f gr

owth

and

spe

cifi

c gr

avity

for

fiv

e se

cond

-gro

wth

whi

te o

aks,

sho

win

g av

erag

e de

crea

se i

n sp

ecif

ic g

ravi

ty f

rom

cent

re to

bar

k. N

ote

that

the

scal

es fo

r A a

nd B

are

dif

fere

nt. (

All

imag

es in

Fig

. 1.6

are

rep

rodu

ced

from

Pau

l (19

63)

wit

h pe

rmis

sion

of

the

US

DA

For

est S

ervi

ce.)


wood is distrusted as a material to use in structural applications. A relatedecological concern regarding these fast-growing improved conifers is that com-pression wood, although strong in compression, is exceptionally brittle. Treessuch as the foxtailed pine shown in Fig. 1.2F have a high content of compressionwood and are particularly susceptible to stem breakage under loading, e.g., inresponse to hurricane-force winds (Boone & Chudnoff, 1972). It deservesemphasis, particularly in view of the escalation of extreme events attendingglobal climate change, that not only the final quality of wood in service to manbut the health of forest ecosystems hinges on the growing tree not experienc-ing overloading in its pre-harvest lifetime. Compression wood fibres, althoughtolerated in some of the less demanding applications (e.g. W-liner cardboard),are also less than appreciated by pulp-and-paper industries. The lignin of compres-sion wood has a high p-hydroxyphenyl content, making it relatively expensive toremove during pulping. The pulp commonly requires additional bleaching forproduction of fine papers. Moreover, compression wood tracheids being circularin cross section with S2 microfibrils running at a relatively flat angle tend tobehave as reinforced cylinders that do not readily flatten nor fibrillate duringpapermaking (Dinwoodie, 1965). In general, compression wood awaits indus-trial appreciation and innovation for its unique quality attributes.

Enhancing wood density is one thing, achieving uniform wood densityanother. It is unlikely that silviculture, tree improvement, genetic engineeringor any combination will solve the uniformity problem, simply because wooddensity variation taraplas appears to be an unavoidable development during treegrowth (Figs 1.5 and 1.6). It can be suggested, however, that technology torecycle waste fibres (such as those used in newspapers) could be developed toreconstitute uniformly dense artificial woods (logs, lumber, etc.). Combinationplastic – waste wood structural members have already been produced; however,it remains to be seen if they will have the durability of wood. It should be possibleto infiltrate and polymerize a lignin-like material, perhaps in combination withchemicals enhancing other properties, in oriented waste-wood fibres to producehighly uniform end products of varied density and equal or greater durabilityand strength than most woods.

1.4.4 Understanding the control of secondary-wall formation at the level of cell biology

A triplet combination of auxin, cytokinin and gibberellin was found to beessential for occurrence of cell-division activity in the cambium of needle-freeconifer stems, but those phytohormones both individually and in several com-binations were ineffective in promoting development of cambial derivatives togo beyond the stage of radial expansion (Kijidani etal., 2001). Both the initiationof secondary-wall formation and the extent of secondary cell-wall depositionoccurring during differentiation of cambial derivatives into tracheids are


regulated by a tracheid differentiation factor (TDF) derived from mature coni-fer needles (Savidge, 2000b). In addition to this, it was also determined thata conifer root factor was essential for enlarged primary-walled cambial derivatives,not under auxin limitation, to initiate secondary-wall formation and completetheir differentiation into tracheids (Savidge, 2001a). Exogenous application of1-aminocylcopropane-1-carboxylic acid (ACC), an amino acid known to bebiosynthesized in roots, effectively simulates the inductive effect of matureconifer needles on tracheid differentiation in needle-free cuttings (Kijidaniet al., 2001; Savidge & Wu, 2001). Thus, the evidence indicates that ACC isa TDF controlling secondary-wall deposition during wood formation. ACC iswell established to be a precursor of ethylene and was previously confirmed tobe endogenous to developing xylem (Savidge et al., 1983). In vitro workrevealed that ethylene production by conifer cambium required that both auxinand ACC be supplied (Savidge, 1988). However, it remains unclear whether itis ethylene production from ACC, or ACC acting directly, that initiates andcontrols the extent of secondary-wall formation.

In the hardwood Buxus, it was observed that mature leaves were effective ininitiating secondary-wall formation and producing tracheids, as had beenfound with conifer needles (Savidge, 1983). As shown in Fig. 1.7, it has beendiscovered that mature Populus balsamea leaves induce cambial cells to differ-entiate into fibres but not into vessel members, whereas young developing leavesinduce both vessel members and fibres to differentiate (Savidge, 2002).The fibre-differentiation response in the absence of attending vessel memberdifferentiation can be simulated by providing the defoliated cutting with ACC.Differentiation of cambial derivatives into vessel members is well establishedto be under auxin regulation (Savidge & Wareing, 1981) and there is also goodevidence that an auxin–gibberellin combination is conducive to both cambial-cell division activity and differentiation of a complex xylem consisting ofvessel members surrounded by fibre-like but incompletely differentiated cells(Wareing, 1958; Digby & Wareing, 1966). The responses shown in Fig. 1.7corroborate the interpretation that the leaf factor inducing cambial derivativesto differentiate into fibres is neither an auxin nor a gibberellin, rather ACC.

1.5 The larger picture

1.5.1 An abundance of wealth

There are approximately 80 000 tree species on earth, and every one ofthem appears to have a unique kind of mature wood, better suited for somethan other applications. Thus, although all woods exhibit but a few anatomicaltypes of woody elements and contain only three major macromolecularclasses (cellulose, hemicelluloses and lignin), there are at least 80 000 differentcombinations possible. In fact, the number must be much larger, as many


Fig. 1.7 Populus balsamifera. (A) cutting having a single mature leaf harvested from a dormant tree. Thistype of cutting (single leaf) and others with the leaf removed (completely defoliated) by cutting at the baseof the petiole were used, standing them upright in water in lighted warm conditions for three weeks, inthese experiments. Bar = 0.5 cm. (B) Cross section showing the dormant cambial region (arrowed) at thejunction of petiolar traces with the stem following three weeks of incubation of a completely defoliatedcutting. Bar = 50 μm. (C) Cross section showing an extensive zone of new fibres (arrowed near where apetiolar trace joins the stem) induced to differentiate at the trace-cambium junction of a single leaf cutting.Bar = 500 μm. (D) Higher magnification of the new fibres (black arrow) shown in C. The latewoodboundary is indicated by the white arrow. Bar = 50 μm. (E) Another view of fibres (black arrow) induced todifferentiate by a single leaf, showing their Nomarski birefringence. The latewood boundary is indicatedby the white arrow. Bar = 25 μm. (F) Another view of fibres (black arrow) induced to differentiate inresponse to a single intact leaf. This position is distant from the trace-cambium junction, and only scatteredclusters of fibres occur within the cambial zone. The chevron indicates anomalous wall bars (compareSavidge, 2000a; Savidge & Wu, 2001), previously concluded to be indicative of differentiation intotracheary elements. The latewood boundary is indicated by the small white arrow. Bar = 20 μm. (G)Cutting that was defoliated at the start of the experiment but which produced a new leaf from its axillary


thousands of tree species are believed to have become extinct over the precedingcentury, and many also during the passage of paleological time. Unfortunately,despite the immense variation in world woods and the easily appreciated valueof having such extensive variety on earth, an undeniable trend in contemporaryforestry has been to narrow to only a relatively few species, the choices ofwood that will be available to future generations. The current angst aboutpreserving biodiversity in general may be serving to counteract that trend, butpreservation of diverse woods is justifiable in its own right.

1.5.2 The consumer is always right

Whenever woods from a selection of species are available, wood qualityhas economic and aesthetic overtones for the consumer. Wood handbooks,trial-and-error first-hand experience, handed-down traditions and general shopadvice posit particular woods for specific end uses, and those woods tend to beused just so. For major industrial consumers of raw wood, preferred species aregenerally dictated by compatibility and efficiency considerations; conversely,some woods may be abhorred for the same reason. The need to match the woodto the job at hand has long been recognized (Evelyn, 1729). Nevertheless,wood-quality distinctions between species tend to be based on experience,geographically localized and highly subjective judgments rather than decisionsmade by means of a scientifically well-supported classification system.

1.5.3 The technology of wood-quality assessment can be flawed

Alternatively, quality may be evaluated more precisely using measurablecriteria, for example in relation to ranges of tolerances given within timber design,pulp-and-paper or other engineering manuals. The major difficulty with thewood-as-an-engineering-material approach is that quality can vary dramaticallybetween individuals of a species, or even of a clone, when the same height andage in the individual trees are sampled (Figs 1.3–1.6). Moreover, as discussedabove, quality can never reasonably be expected to be uniform over the bole.Wood quality at any height or circumferential position in the bole can varyextensively between individuals of the same species even when they have beengrown on the same site, and the variation exhibited by such a population maybe either amplified or reduced when the same species is grown on different

Fig. 1.7 (continued) bud during the three-week period. Bar= 1 cm. (H) Low magnification cross section ofthe trace-cambium junction of a cutting having a young developing leaf, as shown in G. Note that vesselsare of common occurrence (black arrow). The white arrow indicates the latewood boundary. Bar = 250 μm.(I) Higher magnification of H showing newly produced vessel members (black arrow) with surroundingfibres. The white arrow indicates the latewood boundary. Bar = 100 μm (after Savidge, 2002).


sites or at different geographical locations. Thus, measurements on a few sampleswhen projected to an entire pile of logs, lumber or chips may yield inaccurateconclusions in relation to the envisaged performance of a particular piece ofwood; a serious consideration in relation to certifying the reliability of structuralmembers used in support applications. Acknowledging that every cubic centi-metre of a board, post or beam is likely to have its own unique properties, andthat continuous trends in variation need not necessarily exist either along oracross it, it appears that characterization of the in-service performance of suchheterogeneous material can only be achieved reliably by testing each individualpiece at the full scale of its intended application. Indeed, it is curious that struc-tural testing traditionally has not begun at the log stage, in order to establishbaseline values which would enable discovery of how strength or other engin-eering properties had been affected by subsequent processing. In the face ofgenetic and environmentally induced changes to wood quality, the reliabilityof current engineering values – derived from trees of former times – can only beregarded with suspicion.

1.6 Discussion: seeing the wood and the trees

1.6.1 Philosophical and historical musings

In this age of nuclear energy, gene manipulation, environmental disruption andclimate change, it has become apparent that absolutely everything in ourbiosphere is susceptible to genetic modification, and there are many individualsworking unstintingly to bring about changes. It could be argued that the onlymajor challenge remaining for trees and forests is survival in the face of scientificinterference into their ecophysiological fitness. The paleobotanical recordshows clearly enough that woody (tree-like) vascular plants have been on earthfor more than 300 million years (Kerp, 1996), in contrast to perhaps 0.005 millionyears of plant husbandry (silviculture), 90 years of genetic theory and 30 yearsof recombinant DNA technology. Within the concept of natural selection, andknowing from geology that neither environmental nor climatic change is at allunusual on earth, there is every reason to consider that gene pools in manyextant wild tree species had been thoroughly time-tested and had proven them-selves, not merely fit for survival but also compatible in complex ecologicalassociations, long before the advent of tree improvement programmes set outto fix matters. Selections and concomitant roguing, controlled-pollinationcrosses, up- or down-regulation of gene expression, insertion of foreign DNAand cloning, when not done merely out of curiosity but on massive scales todevelop super trees for satisfying forest management and industrial purposes,may have unforeseen and serious consequences. However undesirable the phe-notypes of the rejected and the accepted-but-for-modification may have beenjudged to be, there can be no doubt that they would have fulfilled an ecological


role just as they were. The fundamental question here is not merely concernedwith sustaining a sufficient buffer of genetic diversity within populations orwhole species such that at least some can survive the once-in-a-thousand-yearcataclysm. That is widely recognized as an important issue, but more funda-mentally, the push to create super races of trees is totally at odds with theholocoenotic principle and the recognition that communities, or associations,of forest life form not because each species is self-reliant, rather because theyhave weaknesses and need one another to be mutually strong.

1.6.2 Wood quality measured on a three-pan balance

None of the components of a tree is of value exclusively to humanity. Threemajor areas, or perspectives, of wood quality deserve co-consideration:

1. Quality in relation to fulfilling the unique physiological needs of eachtree in its individual setting, particularly in relation to the tree survivingand completing its life cycle fruitfully;

2. Quality in relation to the contribution made by wood to the health andfunctioning of its forest ecosystem (or plantation), hence also to thebiosphere and the climate near the ground; and

3. Quality in relation to the usefulness of wood in service to humanity.

The above three areas can be thought of as the pans on a three-beamed balance.The challenge for the forest manager is not to allow any one pan to overbalanceeither or both of the other two, in order that all needs can be optimally met.There is no question that the third or anthropocentric pan has overbalanced theother two since the advent of natural resource management. On the bright side,today’s forest managers generally agree that a less narrow approach, combinedwith vision far into the future, is what is needed for sustainable forestry.

1.6.3 Lignin on the three-pan balance

Lignin in wood can be used as an example of how the three perspectives mustbe balanced. In relation to pulp-and-paper production in particular, there areecological and economic pressures to reduce lignin content in plantation trees.That it can, and therefore should, be reduced is arguable. Both lignin removal(during pulping) and bleaching of residual lignin (during papermaking) requireindustrial chemicals and energy. Those inputs represent major costs and can besources of serious environmental pollution, greenhouse gases and other undesir-able features. Thus, from the industrial and environmental perspectives,a strong argument can be made for minimizing the amount of lignin that has tobe removed. Although the tree improvement community considered it unlikelythat sufficient gains could be made through selection and breeding, geneticmodification of trees has provided a practicable means of accomplishing thatgoal, concomitantly boosting the yield of cellulosic fibre (Hu et al., 1999).


The attainment of genetic/biochemical competence for lignification wasnot merely a milestone in plant evolution but was the key innovation enablingplants to grow into the sizes and forms of trees. Thus, lignin is one of theunderpinnings of forest habitat and the resulting holocoenotic linkages of everyforest ecosystem on earth. Trees within the Magnoliophyta have approximately10% less lignin than trees within the Coniferophyta; therefore, it can be arguedthat it should be possible to reduce the lignin content of conifers some 10%and still have an upright, ecologically acceptable tree (Savidge, 1985, 1986).However, little is known about the ecological and physiological role(s) fulfilledby lignin in forest ecosystems. Certainly, many fungal species thrive on andappear to be obligatorily linked to a lignin diet. Within wood, lignin fulfilsa cementing, structurally reinforcing role. There have also been suggestionsthat lignin serves in defence (e.g. to heighten resistance to microbial decay)and as a means for the tree to cleanse itself of toxins, but both of these ideasremain to be well investigated. Thus, at present, there is no way of knowing towhat extent anthropocentric emphasis on lignin could overbalance the ecologicaland/or physiological sides.

1.6.4 Looking back

What constitutes wood quality has changed with time. Centuries ago, small-diameter logs were deemed of little use in Europe, other than for fencing andfirewood. The sapwood in large diameter old-growth specimens was scorned bytradesmen in deference to dimensionally more stable, more durable and generallymore attractive clear heartwood, which at that time was not in short supply(Evelyn, 1729). Subsequent logging of forests in conjunction with the adventof forest management, and in response to a variety of social pressures (e.g. wars,railroads, cities of wood) and new technologies (papermaking in particular),altered the focus from that of the skilled tradesman (apprenticed, but probablynot otherwise educated) to that of structural and chemical engineers. Thus,new disciplines such as wood engineering, wood science and pulp-and-paperscience emerged, and they soon joined hands with tree improvement efforts.

In retrospect, the progress associated with the industrial revolution markeda historical turning point in man’s relation to the forest. In order to ensure theavailability of high quality materials, the skilled woodworker of yore commonlyselected his standing trees personally before embarking on a project. In contrast,the structural and chemical engineering fields distanced themselves from treesand forestry in general, focusing instead on the physical and chemical proper-ties of wood in relation to the processes of milling, and probably unwittingly insupport of large scale harvesting of forests. It may be suggested that the treeimprovement side became obsessed with satisfying the wants of industry. Thebiological phenomena attending wood formation were deemed scientificallyinteresting but of little real practical concern, evidently on the experiential


assumption that, wherever trees are, there will always be wood, and alsoknowing that the first sort on wood quality necessarily has to occur duringharvesting operations. A few researchers were concomitantly pointing out thatfast-grown trees make weak dimensionally unstable wood, that juvenile sapwoodis inferior to mature sapwood or the once-juvenile core of mature heartwood,that reaction wood and spiral grain were becoming more prevalent, that manyhigh value woods were becoming in short supply, and so forth. But thosedetermining the course of forestry research did not take such expressedconcerns sufficiently seriously. It has often been said that ‘the forest could notbe seen for the trees’, but it is probably more correct to state that the 20th centurywas one where the trees could no longer be seen because of the preoccupationwith wood.

The heterogeneous nature of wood properties within the individual tree, andalso between trees whether of the same or differing species, was well appreci-ated by wood tradesmen centuries before the advent of wood science (Evelyn,1729). During the 1900s, wood-science and tree-improvement researchprogrammes grew hand-in-hand, in the guise of being able to solve problemsof wood quality. From the outset, consortia of both clearly recognized thatwood is an end product of cambial growth. However, neither saw the need toconfront the developmental biology aspects, preferring instead simply to assessend-product outcomes following silvicultural, provenance, controlled pollina-tion and selection trials. In hindsight, heterogeneity in wood properties withinlogs has been marginally if at all reduced. The germplasms selected by treeimprovement have delivered gains but have also introduced problems. More-over, in the face of climate and other environmental changes, there can be littleif any confidence about the wood quality which will exist when the selectedlines are harvested.

It may be asked, ‘Why has the developmental biology of wood formationnot been given greater emphasis in forestry research?’ One difficulty is thatfunding programmes, in particular those connected to industrial applications,are inclined to focus on the feasibility of a research project delivering short-term results (Horn, 1880; Sprague & Sprague, 1976). Ironically, a veritablearmy of forestry researchers has been busy over the last two centuries measur-ing tree rings and physical properties of wood, sometimes correlating those datawith crown transparency, soil nutrient levels, temperature and rainfall records,etc., and invariably arriving at conclusions about what controls wood forma-tion in relation to productivity, forest decline, dendroclimatology and variousother aspects. No important advances in understanding wood formation havecome from those innumerable investigations. There has long been and remainstoday a sizable void in the production of highly qualified personnel able toincrease knowledge about how trees make wood.

Biochemistry, biophysics and cell biology are difficult roads to walk, butthere can be no doubt that they, in conjunction with tree genetics and wood


science capabilities, hold the answers to understanding how trees make wood,thus how to manage forests for wood and all of their other benefits.

1.6.5 Looking forward

It is time for change, for wise decisions to be made about forestry research.Everyone wants forests not merely to be around but to be healthier than ever,whether looking forward one generation or a thousand. A long-lived renewableresource deserves long-lived financial support. Massive support for fundamentalresearch to explain how trees make wood, and to understand secondary growthin general, is more than amply justified. Indeed, in relation to maintaining life-supporting conditions in the terrestrial biosphere and meeting the needs ofhumanity, there is probably no more important or more neglected area ofbiological research.

Should one tree species be researched to the exclusion of others? Certainly,a justifiable case was made for the plant biology community focusing onArabidopsis thaliana (a small ‘weed’ of the mustard family) as a modelflowering plant, but that justification resided in the acknowledgment thatresearch with A. thaliana was unlikely to be useful other than to discoverfundamental developmental principles of general relevance to higher plantdevelopment. There have been proposals that species within the Pinus andPopulus genera could serve as models for understanding tree growth anddevelopment, and in particular wood formation. The difficulty here, however,is that huge diversity is known to exist among trees. At least when analyzedchemically, and usually also anatomically, every tree species produces a uniquewood. Moreover, A. thaliana produces cambium, secondary xylem and phloemsufficiently abundantly to hold its place as a model for discovering the generaldevelopmental principles underlying secondary growth, at least within theMagnoliophyta. On the other hand, approximately 80 million years separatethe appearance of the progymnosperms from that of the angiosperms, and thelatter evidently arose independently of the former. Those two orders could wellhave different physiologies, and indeed there is abundant physiological evidenceindicating that this is the case. Thus, it could be argued that a model speciesis needed to elucidate general principles of conifer growth and development.Having accomplished that, however, there will still remain the need to explainwhat makes the wood of each species different from that of others.

References

Atalla, R.H. (1998) Cellulose biosynthesis in Arabidopsis. Science, 282, 759–773. Bethke, P.C., Swanson, S.J., Hillmer, S. & Jones, R.L. (1998) From storage compartment to lytic organelle:

the metamorphosis of the aleurone protein storage vacuole. Annals of Botany, 82, 399–412.


Bichet, A., Desnos, T., Turner, S., Grandjean, O. & Hofte, H. (2001) BOTERO1is required for normal orienta-tion of cortical microtubules and anisotropic cell expansion in Arabidopsis. Plant Journal, 25, 137–149.

Bissing, D.R. (1982) Variation in qualitative anatomical features of the xylem of selected dicotyledonouswoods in relation to water availability. Bulletin of the Torrey Botanical Club, 109, 371–384.

Blancaflor, E.B. & Hasenstein, K.H. (1995) Time course and auxin sensitivity of cortical microtubulereorientation in maize roots. Protoplasma, 185, 72–83.

Boone, R.S. & Chudnoff, M. (1972) Compression wood formation and other characteristics of plantation-grown Pinus caribea. USDA Forest Service Research Paper ITF-13, Institute of TropicalForestry, Rio Piedros, Puerto Rico.

Brett, C.T. (2000) Cellulose microfibrils in plants: biosynthesis, deposition, and integration into the cellwall. International Review of Cytology, 199, 161–201.

Chaffey, N.J., Barlow, P.W. & Barnett, J.R. (2000) Structure–function relationships during secondaryphloem development in an angiosperm tree, Aesculus hippocastanum: microtubules and cellwalls. Tree Physiology, 20, 777–787.

Delmer, D.P. (1999) Cellulose biosynthesis: exciting times for a difficult field of study. Annual Reviewof Plant Physiology and Plant Molecular Biology, 50, 245–277.

Desprez, T., Orbovi, V., Bellini, C., Kronenberger, J., Caboche, M., Traas, J. & Höfte, H. (2002)Resistance against herbicide isoxaben and cellulose deficiency caused by distinct mutations insame cellulose synthase isoform CESA6. Plant Physiology, 128, 482–490.

Digby, J. & Wareing, P.F. (1966) The effect of applied growth hormones on cambial division and thedifferentiation of the cambial derivatives. Annals of Botany, 30, 539–548.

Dinwoodie, J.M. (1965) The relationship between fibre morphology and paper properties. TAPPI, 48,440–447.

Ehrenberg, C. (1970) Breeding for stem quality. Unasylva, 24, 23–31. Eriksson, M.E., Israelsson, M., Olsson, O. & Moritz, T. (2000) Increased gibberellin biosynthesis in

transgenic trees promotes growth, biomass production and xylem fiber length. Nature Biotechnology,18, 784–788.

Evelyn, J. (1729) Silva: or a Discourse of Forest-Trees, and the Propagation of Timber in His Majesty’sDominions, in Delivered in the Royal Society the 15th of October 1662, on the Occasion of CertainQueries Propounded to that Illustrious Assembly. Honourable and Principal Officers and Com-missioners of the Navy. Facsimile 5th edn, 1979, Stobart & Son, London. ISBN 0 85442 013 4.

Fisher, D.D. & Cyr, R.J. (1998) Extending the microtubule/microfibril paradigm. Cellulose synthesis isrequired for normal cortical microtubule alignment in elongating cells. Plant Physiology, 116,1043–1053.

Gerhold, H.D. (1970) A decade of progress in breeding disease-resistant forest trees. Unasylva, 24, 37–44. Harris, J.M. (1970) Breeding to improve wood quality. Unasylva, 24, 32–36. Horn, W. (1880) Theodor Hartig. Zeitschrift für Forst- und Jagdwes, 12, 292–309. Hu, W.-J., Harding, S.A., Lung, J., Popko, J.L., Ralph, J., Stokke, D.D., Tsai, C.-J. & Chiang, V.L.

(1999) Repression of lignin biosynthesis promotes cellulose accumulation and growth in trans-genic trees. Nature Biotechnology, 17, 808–813.

Jackson, D.S. (1965) Species siting: climate, soil, and productivity. New Zealand Journal of Forestry,10, 90–102.

Jones, D.W. (1971) C. X-ray and electron diffraction. C.1. Structure studies, in High Polymers, Vol. 5,Cellulose and Cellulose Derivatives, Part 4 (eds N.M. Bikales & L. Segal) Wiley, New York,pp. 117–150.

Kerp, H. (1996) Post-Variscan Palaeozoic Northern Hemisphere gymnosperms: the onset to theMesozoic. Revue of Palaeobotany and Palynology, 90, 263–285.

Kijidani, Y., Wu, Z. & Savidge, R. (2001) New insight into phytohormone regulation of woodformation in conifers. Phytomorphology, Golden Jubilee Issue, 185–200.

Kozlowski, T.T. & Greathouse, T.E. (1970) Shoot growth and form of pines in the tropics. Unasylva,24, 6–14.


Lamlom, S.H. & Savidge, R.A. (2003) A reassessment of carbon content in wood: variation within andbetween 41 North American species. Biomass & Bioenergy, 25, 381–388.

Larson, P.R. (1969) Wood formation and the concept of wood quality, Yale University, School ofForestry, Bulletin No. 74.

Leitch, M.A. & Savidge, R.A. (1995) Evidence for auxin regulation of bordered-pit positioning duringtracheid differentiation in Larix laricina. IAWA Journal, 16, 289–297.

Lewis, N.G. & Sarkanen, S. (eds) (1998) Lignin and Lignan Biosynthesis, ACS Symposium, Series697, American Chemical Society, Washington, DC.

Mattheck, C. & Kubler, H. (1995) Wood – The Internal Optimization of Trees, Springer-Verlag, Berlin,Heidelberg, New York.

Megraw, R.A. (1985) Wood Quality Factors in Loblolly Pine. The Influence of Tree Age, Position inTree, and Cultural Practice on Wood Specific Gravity, Fiber Length and Fibril Angle, TAPPIPress, Technology Park, Atlanta, GA.

Meredieu, C. & Caraglio, Y. (2002) External indicators of living branches with missing rings within atree crown of Corsican pine. Forestry, 75, 569–578.

Nikles, D.G. (1970) Breeding for growth and yield. Unasylva, 24, 9–22. Northcott, P.L. (1965) The comparison of tree qualities. Forestry Chronicle, 41, 215–221. Panshin, A.J. & de Zeeuw, C. (1980) Textbook of Wood Technology, 4th edn, McGraw-Hill, New York,

St Louis, San Francisco. Paul, B.H. (1963) The application of silviculture in controlling the specific gravity of wood, USDA

Forest Service, Technical Bulletin No. 1288. Rensing, K.H., Samuels, A.L. & Savidge, R.A. (2002) Ultrastructure of vascular cambial cell cytokinesis

in pine seedlings preserved by cryofixation and substitution. Protoplasma, 220, 39–49. Savidge, R.A. (1983) The role of plant hormones in higher plant cellular differentiation. II. Experiments

with the vascular cambium, and sclereid and tracheid differentiation in the pine, Pinus contorta.Histochemical Journal, 15, 447–466.

Savidge, R.A. (1985) Prospects for manipulating vascular cambium productivity and xylem cell dif-ferentiation, in Attributes of Trees as Crop Plants (eds M.G.R. Cannell & J.E. Jackson), NERCPress, Abbots Ripton, England, pp. 208–227.

Savidge, R.A. (1986) Prospects for genetic manipulation of wood quality, in New Ways in ForestGenetics (eds F. Caron, A.G. Corriveau & T.J.B. Boyle), Proceedings of the 20th Canadian TreeImprovement Association Meeting, Quebec, part 2, pp. 142–151.

Savidge, R.A. (1988) Auxin and ethylene regulation of diameter growth in trees. Tree Physiology, 4,401–414.

Savidge, R.A. (1993) Formation of annual rings in trees, in Oscillations and Morphogenesis (ed. L. Rensing),Marcel Dekker, New York, Basel, Hong Kong, pp. 343–363.

Savidge, R.A. (1996) Xylogenesis, genetic and environmental regulation—a review. IAWA Journal,17, 269–310.

Savidge, R.A. (2000a) Biochemistry of seasonal cambial growth and wood formation—an overview ofthe challenges, in Cell and Molecular Biology of Wood Formation (eds R.A. Savidge, J.R. Barnett& R. Napier), BIOS, Oxford, pp. 1–30.

Savidge, R.A. (2000b) Inductive effects of conifer needles on xylogenesis—auxin not the explanation,in Cell and Molecular Biology of Wood Formation (eds R.A. Savidge, J.R. Barnett & R. Napier),BIOS, Oxford, pp. 237–254.

Savidge, R.A. (2001a) Intrinsic regulation of cambial growth. Journal of Plant Growth Regulation, 20, 52–77. Savidge, R.A. (2001b) Forest science and technology to reduce atmospheric greenhouse gases – an

overview, with emphasis on carbon in Canada’s forests, in Climate Change 2 – CanadianTechnology Development Conference (ed. K.T. Tsang), Canadian Nuclear Society, Toronto,ISBN 0 919784 72 0, pp. 1–19.

Savidge, R.A. (2002) A fibre-differentiation factor distinct from auxin and gibberellin, in Proceedings,29th Ann. Mtg. Plant Growth Regulation Society of America (ed. W.A. Mackay), Omnipress,Madison, WI, pp. 154–159.


Savidge, R.A. & Barnett, J.R. (1993) Protoplasmic changes in cambial cells induced by a tracheid-differentiation factor from pine needles. Journal of Experimental Botany, 44, 395–405.

Savidge, R.A., Barnett, J.R. & Napier, R. (eds) (2000) Cell and Molecular Biology of Wood Formation,BIOS Scientific, Oxford, UK.

Savidge, R.A. & Farrar, J.L. (1984) Cellular adjustments in the vascular cambium leading to spiral-grain formation in conifers. Canadian Journal of Botany, 62, 2872–2879.

Savidge, R.A. & Förster, H. (2001) Coniferyl alcohol metabolism in conifers. II. Coniferyl alcohol anddihydroconiferyl alcohol biosynthesis. Phytochemistry, 57, 1095–1103.

Savidge, R.A. & Wareing, P.F. (1981) Plant growth regulators and the differentiation of vascularelements, in Xylem Cell Development (ed. J.R. Barnett), Castle House Publications, TunbridgeWells, England, pp. 192–235.

Savidge, R.A. & Wu, Z. (2001) Induction of anomalous wall bars in cambial cells, in Proceedings of theMicroscopical Society of Canada, 28th Annual Meeting, MSC, Lethbridge, Alberta, pp. 66–68.

Savidge, R.A., Mutumba, G.M.C., Heald, J.K. & Wareing, P.F. (1983) Gas chromatography – massspectroscopy identification of 1-aminocyclopropane-1-carboxylic acid in compression wood vascularcambium of Pinus contorta Dougl. Plant Physiology, 71, 434–436.

Sjöström, E. (1993) Wood Chemistry, Fundamentals and Applications, 2nd edn, Academic Press, SanDiego, New York, Boston, London, Sydney, Tokyo, Toronto.

Sprague, L.G. & Sprague, C.R. (1976) Management science? Interfaces, 7, 57–62. Stamm, A.J. & Sanders, H.T. (1966) Specific gravity of the wood substance of loblolly pine as affected

by chemical composition. TAPPI, 49, 397–400. Stant, M.Y. (1961) The effect of gibberellic acid on fibre cell length. Annals of Botany, 25, 453–462. Sugimoto, K., Williamson, R.E. & Wasteneys, G.O. (2000a) New techniques enable comparative analysis

of microtubule orientation, wall texture, and growth rate in intact roots of Arabidopsis. PlantPhysiology, 124, 1493–1507.

Sugimoto, K., Williamson, R.E. & Wasteneys, G.O. (2000b) Wall architecture in the cellulose-deficient rsw1 mutant of Arabidopsis thaliana: Microfibrils but not microtubules lose their transversealignment before microfibrils become unrecognizable in the mitotic and elongation zones of roots.Protoplasma, 215, 172–184.

Sunley, J.G. (1963) Factors affecting timber quality. Commonwealth Forestry Review, 42, 129–136. Taylor, N.G., Laurie, S. & Turner, S.R. (2000) Multiple cellulose synthase catalytic subunits are

required for cellulose synthesis in Arabidopsis. Plant Cell, 12, 2529–2540. Timell, T.E. (1986) Compression Wood in Gymnosperms, vols 1–3, Springer-Verlag, Berlin, Heidelberg,

New York, Tokyo. van Buijtenen, J.P. (1969) Controlling wood properties by forest management. TAPPI, 52, 257–259. Wareing, P.F. (1958) Interaction between IAA and GA in cambial activity. Nature, 181, 1744–1745. Wareing, P.F. (1966) The physiologist’s approach to tree growth, in Physiology in Forestry (ed.

R.W.V. Palmer), Supplement to Forestry, pp. 7–18. Wenzel, C.L., Williamson, R.E. & Wasteneys, G.O. (2000) Gibberellin-induced changes in growth

anisotropy precede gibberellin-dependent changes in cortical microtubule orientation in developingepidermal cells of barley leaves. Kinematic and cytological studies on a gibberellin-responsivedwarf mutant, M489. Plant Physiology, 124, 813–823.

Wilkie, K.C.B. (1985) New perspectives on non-cellulosic cell-wall polysaccharides (hemicellulosesand pectic substances) of land plants, in Biochemistry of Plant Cell Walls (eds C.T. Brett &J.R. Hillman), Cambridge University Press, Cambridge, London, New York, pp. 1–37.

Williamson, R.E., Burn, J.E., Birch, R., Baskin, T., Arioli, T., Betzner, A.S. & Cork, A. (2001)Morphology of rsw1, a cellulose-deficient mutant of Arabidopsis thaliana. Protoplasma, 215,116–127.

Zobel, B.J. (1961) Inheritance of wood properties in conifers. Silvae Genetica, 10, 65–70. Zobel, B.J. & van Buijtenen, J.P. (1989) Wood Variation – Its Causes and Control, Springer-Verlag,

Berlin, Heidelberg, New York.

2 Wood anatomy in relation to wood qualityBrian G. Butterfield

2.1 Wood anatomy

2.1.1 Softwoods and hardwoods

Woods are described as softwoods and hardwoods, depending on the type oftree from which they are cut. The terms are rather misleading to scientists andwood engineers because they have nothing to do with the actual softness orhardness of the wood. The terms date from the medieval timber trade, in whichwoods from the cone-bearing trees like the pines were seen to be soft, whereaswoods from the flowering trees like oak were seen to be hard. The distinctionis therefore based on the botanical taxonomic grouping of the tree, and not onthe true hardness of its wood. Softwoods come from the Gymnospermae orcone-bearing plants such as the pines, spruces, firs, redwoods, podocarps andAraucarian pines. Hardwoods come from the Angiospermae or floweringplants and include broad-leaved trees such as the oaks, ashes, beeches andmost of the tropical trees and shrubs. In many cases, softwood trees do indeedpossess low-density woods and hardwoods high-density woods, but this is notalways the case. There are far more hardwood than softwood species in theworld, though the latter are often better known as production forest plantationspecies.

Softwoods and hardwoods are made up of different cell types and as a resulthave different structures. Softwoods have a generally uniform structure withtracheids forming the bulk of the wood (Fig. 2.1). Tracheids are elongatedcells with pointed tips, thick secondary cell walls and no living cell contents atmaturity. They are arranged in radial files reflecting the pattern of the fusiformcambial cells from which they are derived. Their cell tips are arranged in anirregular pattern when viewed tangentially, as a consequence of the pseudo-transverse longitudinal anticlinal divisions in the cambium. This permits longdistance transport of water from cell to cell through pores in the radial walls,a feature that would not be possible if the cell tips did not overlap. It also pro-vides for much greater along-the-grain strength in the timber. Some softwoodsalso have axial resin canals and/or axial parenchyma cells. Axial parenchymacells are not common in softwoods but can be found in most growth rings inTaxodium, Chamaecyparis, Thuja, Cupressus and Sequioa, and sometimesin Tsuga, Larix, Pseudotsuga and Abies. All softwoods have uniseriate, and/orpartly or fully biseriate rays built up of ray parenchyma, and in the case of

WOOD ANATOMY IN RELATION TO WOOD QUALITY 31

Pinus, Picea, Larix, Pseudotsuga and Tsuga, ray tracheids as well. There aremany texts with very good descriptions of the basic anatomies of softwoodsincluding Panshin and de Zeuuw (1970), Core et al. (1976), Walker (1993) andButterfield et al. (2000).

Hardwoods separate the functions of conduction and support into two dif-ferent cell types: vessel elements joined end to end form long conduits termedvessels for long distance water transport, and elongated fibres are modified forsupport. The arrangement of these two cell types within each growth ring,along with axial parenchyma in a range of patterns, leads to an infinite varietyof anatomies in hardwoods (Fig. 2.2). This pattern is further complicated bythe presence in many woods of multi-seriate rays, some extending the length

Fig. 2.1 The wood of Dacrycarpus dacrydioides (Podocarpaceae), a softwood, showing axial tracheidsfor water conduction and support, and uniseriate rays.


of the original internode. Again, many texts, including those listed above, haveexcellent descriptions of the basic anatomies of hardwoods.

2.1.2 Growth rings

Trees and shrubs grown in temperate climates produce growth rings that reflectthe spring onset and autumn cessation of cambial division and cell differentiation.Tropical plants may also have growth rings, but these are usually an indicationof significant climatic events rather than an annual cycle. Wood produced bytemperate trees in spring and summer is referred to as earlywood and thatproduced late in the growing season is termed latewood. The severity of the ringboundary varies both between species and within species grown in different

Fig. 2.2 The wood of Eucalyptus delagatensis (Myrtaceae), a hardwood, showing vessels (V) for thelongitudinal transport of water, fibres (F) for support, and uniseriate rays.


environments. Ring boundaries are therefore described on a scale from indis-tinct to very distinct. In softwoods, the boundary is usually characterized bythick-walled, small-lumened tracheids in the latewood, and thin-walled, large-lumened tracheids in the earlywood (Fig. 2.3a). Growth rings affect softwoodwood quality with the width of the rings and the ratio of earlywood to late-wood within each ring both contributing to wood stiffness. The first few

Fig. 2.3 (a) A growth ring boundary in the softwood Cedrus libani (Pinaceae) showing the thin-walled,large-lumened tracheids of the earlywood (E) to the left of the micrograph, and the thicker-walled,smaller-lumened tracheids of the latewood (L) belonging to the previous year’s growth to the right.(b) A growth ring boundary in the hardwood Alectryon excelsus (Sapindaceae) showing the earlywood(E) to the left of the micrograph and the latewood (L) from the previous year to the right. (c) Large-lumened vessels in the earlywood (E) of the ring porous wood of Quercus robur (Fagaceae). (d) Falserings in the wood of Heimerliodendron brunonianum (Nyctaginaceae) resulting from alternating bandsof fibres and axial parenchyma.


growth rings at any one level in the tree are usually wider than those producedfurther out in most plantation softwoods. This enables the trunk to maintaina generally cylindrical shape despite the tree having fewer growth rings athigher levels. In trees with a large crown of branches held high in the forest orplantation canopy, this is a good method of ensuring that the trunk presentsa similar cross-sectional area at most heights. Tree taper does occur of course,but would be much more pronounced if each growth ring was the same widthat all heights in the tree resulting in a conical trunk form. At low levels in thetree, this means that ring width decreases with ring number outwards from thepith and many narrow rings encase the central core of the trunk. Some treesactually increase in trunk diameter with tree height. This is a feature in speciessuch as kauri (Agathis) where a huge spread of large branches occurs at the topof a single large erect trunk. The pattern of growth rings therefore varies withposition in the tree and is part of the tree’s mechanism to provide a balancebetween the trunk diameter and its compressional load-bearing capabilities.

In most cases, the narrower the growth ring, the higher the proportion oflatewood cells present in each ring. This has the effect of increasing the woodstiffness in the outer region of trunks especially at lower levels in the tree. Vari-ations in the proportion of early to latewood are also a result of species differencesand climatic variations. The different properties of early and latewood tracheidsin softwoods have significant effects on both wood quality and tree physiology.

Growth rings are also a feature of most temperate hardwoods (Fig. 2.3b).The variety of cell types present in hardwoods increases the complexity of thecell pattern within the rings. Vessels can be distributed throughout the ring ina diffuse porous arrangement or grouped in the earlywood in a ring porouspattern. The latter is a common feature of many deciduous species (Fig. 2.3c).Axial parenchyma and fibres are also arranged in a bewildering array ofpatterns within each growth ring in different hardwoods (Fig. 2.3d). Fulldescriptions of these patterns can be found in most texts on wood anatomy(e.g. Panshin & de Zeuuw, 1970). These cell patterns give character to thecross-sectional view of hardwoods and are useful in wood identification. Itwould be fair to say, however, that the implications of these various patternsfor the physical properties of hardwoods are not well understood. An excess ofthin-walled parenchyma cells as in Entelea (Fig. 2.4a) significantly reduces thewood density, whereas an excess of very thick-walled libriform fibres in manymembers of the Myrtaceae (Fig. 2.4b), for example, increases the wood density.

2.2 The cell wall of softwood tracheids

2.2.1 Cell wall structure

The basic structure of the tracheid cell wall was determined by Bailey usingthe compound light microscope, and confirmed by Preston and others using


polarized light microscopy and X-ray techniques. Both scanning and transmis-sion electron microscopy have confirmed these observations and added to ourknowledge of the wall (Harada, 1965; Abe et al., 1991, 1992; Kataoka et al.,1992), as have the atomic force microscope (Hanley & Gray, 1994, 1999), theconfocal laser microscope and image-processing techniques (Donaldson, 1998,2001). The tracheid wall comprises an amorphous middle lamella, a primarywall and a multi-layered secondary wall overlaid in some species with varioustypes of thickenings and warty depositions (Figs 2.5a and b). The orientation ofthe cellulose microfibrils in these various wall layers varies considerably andstrongly affects the wall properties.

2.2.1.1 The middle lamella Strictly speaking, the middle lamella is an intercellular region and adheresthe walls of adjacent cells together. In enlarging and differentiating cells, thiszone comprises largely pectic compounds. In fully differentiated tissues, lignincomprises a significant percentage of the middle lamella. The middle lamellazone is very strong and cells subject to load usually fail by trans- or inter-wallfailures (depending on the direction of the load) rather than by cell separationat the middle lamella. It is, however, reasonably easily digested in the laboratoryusing cell maceration techniques or during industrial processing using chemicalseparation.

Fig. 2.4 (a) A transverse surface of the wood of Entelea arborescens (Tiliaceae). The wood of Enteleahas a very low density on account of having very thin-walled fibres (F) and bands of parenchyma (P).The vessels are marked (V). (b) Transverse surface of the wood of Metrosideros umbellata(Myrtaceae). This wood has a very high density on account of having thick-walled libriform fibres andvery small numbers of axial parenchyma cells.


2.2.1.2 The primary wall The primary or outermost layer of the cell wall is normally very thin (0.1–0.2μm)and remains plastic prior to the deposition of the secondary wall. It is oftendifficult to distinguish from the middle lamella using microscopy and hencethe term compound middle lamella is often applied to both regions. The primary

Fig. 2.5 (a) The structure of axially elongated cells (tracheids and fibres) showing the dominantorientation of the cellulose microfibrils in the various cell wall layers. ML – middle lamella,P – primary wall, S1, S2, S3 – secondary wall layers, HT – helical thickening (when present),W – warty layer (when present). (b) Tracheids in Pinus radiata (Pinaceae) showing the dominant S2secondary wall layer overlaid on the lumen side by the thinner S3 layer. (c) Helical thickeningsoverlying the secondary wall in a compression wood tracheid of Taxus baccata (Taxaceae). Thewinding angle of the helices follows the steeper left-handed angle of the S2 wall microfibrils overwhich they are laid. Helical thickenings overlying the S3 wall layer in normal wood tracheids are right-handed and generally steeper. (d) Callitroid thickenings bordering the inter-tracheid pit apertures inCallitris glauca (Cupressaceae). Note also the heavy warty layer overlying the S3 wall.


wall is capable of permanent extension during cell expansion and as a conse-quence its predominant cellulose microfibril angle is random except at the cellcorners. The microfibrils are bound into the matrix complex of hemicellulosesand pectinaceous material. The primary wall also becomes lignified followingsecondary wall deposition.

2.2.1.3 The secondary wall The secondary wall is laid down after the primary wall and usually shows threedistinct layers, S1, S2 and S3. The cellulose microfibrils are highly ordered inthese layers, and lie in parallel helices with consistent winding directions andangles. The S1 layer is about 0.1–0.3 μm thick. Sometimes it is possible to seea series of thin concentric lamellae with microfibrils arranged predominantlyin a left-handed- or S-helix at about 50–75° to the long axis of the cell. Someright-handed- or Z-helix microfibrils also occur, though the lamellae are betterdeveloped in the S-helix zones. The S2 layer is usually 1–5 μm in thickness,with the cellulose microfibrils in a right-handed- or Z-helix at angles of around10–30° in normal tracheid. S-helices are not present. Considerably larger anglesoccur in compression and core wood cells. The S2 wall layer is by far thethickest and dominates the tracheid secondary wall especially in latewood cells.The innermost or S3 layer is again very thin – 0.1 μm thick with its cellulosemicrofibrils arranged in a left-handed- or S-helix within the 60–90° range.Additional or tertiary wall layers such as helical thickenings (Fig. 2.5c) anda warty deposition (Fig. 2.5d) are features of the tracheids of some woods.

Because of its significance for timber properties, the S2 wall layer has beenthe subject of considerable research. Two models have been proposed to inter-pret the cellulose microfillar structure of this layer. One proposes that themicrofibrils are arranged in lamellae parallel to the lumen surface (Kerr &Goring, 1975), while the other suggests that they lie in radial lamellae (Sell &Zimmermann, 1993). Computer modelling by Donaldson (2001) suggests thatthe most likely arrangement is one where the microfibrils are in a randomarrangement with varying amounts of tangential or radial alignment perhaps inweakly defined clusters.

The structure of the tracheid cell wall determines the mechanical andphysical properties of softwoods. The polylaminate structure is ideal for com-pressional loading. Density, as influenced by cell wall thickness, has long beenknown to affect stiffness in wood, and the microfibril angle in the dominantS2 wall layer governs two major wood properties, namely axial stiffness(Cave, 1968, 1969) and longitudinal shrinkage (Harris & Meylan, 1965).

2.2.2 Cell wall and density

The density of oven-dried cell wall material for all woody plants is about1500 kg m–3. Wood with thicker-walled cells has a higher density than wood


with thinner-walled cells of comparable size. Density, however, is also deter-mined by the diameter of the cells and the presence of extractives. The amountof extractives can vary from 1% in the sapwood to 10% in the heartwood of theoven-dried mass. Cell wall thickness is species related but variations betweenearly and latewood tracheids, and between corewood and outerwood cellsfrom the same position within each growth ring can often be very significant.Latewood cells frequently have much thicker walls than earlywood cells andthe earlywood cells of outerwood rings may have thicker walls than those inthe corewood. The relationship of the corewood density to the outerwooddensity, however, is not clear and can be highly variable. Predicting futureouterwood density from seedling wood is therefore generally unreliable.

The significance of wood density for cell stiffness has long been seen asa major factor influencing stiffness in wood, leading many to make excessiveclaims as to its intrinsic value in the assessment of wood quality (Bunn, 1981;Bamber & Burley, 1983). The preoccupation with density as the over-ridingdeterminant of wood quality is the consequence of two factors: firstly it is validas a generalization, and secondly it is a relatively cheap and easy parameter tomeasure (Walker & Woollons, 1998). There can be no doubt that Bamber andBurley (1983) were correct when they claimed “. . . Of all the wood properties,density is the most significant in determining end use. It has considerableinfluence on strength, machinability, conversion, acoustic properties, wear-ablility, paper yield properties and probably many others.” However as Walkerand Woollons (1998) have demonstrated, density by itself is not a self-sufficientindex of strength in wood.

2.2.3 Microfibril angle

The cellulose microfibril angle of the S2 layer is a critical factor determiningthe mechanical properties of wood (Bendtsen & Senft, 1986). Microfibrilangle is defined as the angle between the most probable cellulose microfibrilorientation and the long axis of the cell (Evans, 1998). The S2 layer carriesmost of the axial loading in tracheids with the S2 mean microfibril angle havingan inverse relationship to the axial stiffness of the cell. The Cave (1968) rela-tionship (Fig. 2.6) is now universally accepted, with cell wall stiffness increasing5-fold with a downwards shift from 40 to 10° in mean microfibril angle.

2.2.3.1 Determination of microfibril angle Microfibril angle can be determined by a number of techniques includingmeasuring the alignment of iodine crystals (Bailey & Vestal, 1937), cell wallchecking (Marts, 1955), the angle of the inner pit apertures (Donaldson, 1991),confocal microscopy (Verbelen & Stickens, 1995) and X-ray diffraction (Cave,1966; Meylan, 1967; Cave & Robinson, 1998a; Evans, 1998; Evans et al.,1996). An excellent summary of these techniques is presented in Huang et al.


(1998). The correlation between microfibril angle and iodine staining is high(Huang et al., 1998), but latewood tracheids are difficult to measure by thistechnique on account of their very thick walls. The angle of the pit apertures isa reliable guide to microfibril angle in these cells but is less reliable when usedfor earlywood tracheids where the pits are usually rounded in outline.

X-ray diffraction has received a lot of attention in recent years as a tool formeasuring microfibril angle (Cave, 1997a,b; Cave & Robinson, 1998a,b). It isfast and reliable, but gives only a relative figure for the microfibril angle,called the T-angle using the Cave interpretation of the diffraction pattern.The relationship of T to the true angle must be determined by other techniques.Meylan (1967) determined microfibril angle to be 0.612T + 0.843 for radiata

Axi

al s

tiffn

ess

of th

e ea

rlyw

ood

cell

wal

l (G

Pa)

60

50

40

30

20

10

00° 10° 20° 30° 40°

Mean microfibril angle

Fig. 2.6 Axial stiffness of the cell wall as a function of microfibril angle (after Cave, 1968).


pine, while Megraw et al. (1998) showed microfibril angle to be 0.93T + 10.71for loblolly pine with slightly differing equations for early and latewoodsamples. Standard X-ray diffractometery requires the use of small wood sam-ples not exceeding 1.5 mm in thickness in order to produce a good diffractionpattern, and the cutting and processing of many hundreds or even thousands ofsmall samples necessary for a good statistical result can be very time consuming.Development of the SilviScan-2 by Evans and others at CSIRO, Australia(Evans etal., 1995, 1996; Evans, 1998) has dramatically improved this techniqueby enabling entire cores from pith to bark to be continuously X-ray scanned.This instrument not only provides data on microfibril angle but also on otherparameters including density and fibre cross-sectional dimensions.

2.2.3.2 Microfibril angle variation and its effect on wood properties Microfibril angle varies considerably within the trunk of a tree with largeangles common in the corewood and small angles in the outerwood (Preston,1948; Preston & Wardrop, 1949; Cowdrey & Preston, 1966; Bendsten &Senft, 1986; Donaldson, 1993; Donaldson & Burdon, 1995). Microfibril anglealso varies with height of the tree, generally declining with height in ringsof comparable age (Donaldson, 1998). Microfibril angle is an inheritable traitin radiata pine (Donaldson & Burdon, 1995) and is significantly influencedby physiological age (Donaldson, 1996). Large microfibril angles are alsoa feature of compression wood (Evans, 1998).

There can be no doubt that the large microfibril angles in the first fewgrowth rings of a vertical stem, at whatever height, are the major cause of lowstiffness in the corewood. Cave and Walker (1994) attributed low stiffness injuvenile wood to the large microfibril angles in radiata corewood and thispattern appears to be common in most softwoods grown in open plantations.It is less clear that the same phenomenon occurs in coniferous trees that grownaturally under an established forest canopy leading to the possibility thatwind flexing the young leading stems is a major determinant of large micro-fibril angle. Studies on glasshouse grown seedlings (Butterfield & Li, 2000)indicated that tied clonal plantlets of radiata pine produced fewer compressionwood tracheids, with smaller microfibril angles in their first growth rings, thandid free standing clonal plantlets. This suggests that large microfibril anglesare a natural feature of small diameter leading shoots and occur in associationwith increased amounts of compression wood. Once established, the meanmicrofibril angle declines with radial growth of the stem unless furthercompression wood is formed as a result of asymmetric stem loading caused bylateral branches or other forces such as a strong prevailing wind from onedirection or tree inclination.

Selecting clonal plantlets that have smaller microfibril angles in their firstgrowth rings would clearly provide the timber industry with trees of superiorwood quality, as microfibril angle always gets lower with increasing distance


from the stem centre. But obtaining a reliable figure for the mean microfibrilangle in young clonal plantlets is difficult. In a study on the microfibril angleof radiata pine clonal plantlets (Butterfield & Li, 2000), the wood from theupper (tension) side of plantlets force grown at 45° to the vertical was used inan attempt to find stem wood that was totally free from compression wood.

It is now universally accepted that coniferous trees grown in fast-rotationplantation conditions possess a central core of wood with physical features thatare inferior to those found further from the pith. One of the characteristics ofcorewood in trees is the presence of tracheids with large microfibril angles(Walker & Butterfield, 1996). Corewood is generally described as the wood ofthe first few growth rings that shows the juvenile characteristics of shorttracheid length and large microfibril angles. In radiata pine, corewood occu-pies the first 10 or so growth rings. As this core ascends the trunk of a tree asa cylinder, corewood occupies an ever increasing percentage of the stem crosswith increasing height in the tapering trunk of the tree (Cown, 1992). In manytrees corewood can occupy 50% of the merchantable total volume of the maintrunk. It stands to reason, therefore, that improving the quality of the corewoodcan add a greater value to the log than attempts to improve the outerwoodwhere microfibril angle has stabilized.

Trunk corewood is a consequence of several inter-related factors includingseedling disturbance during planting, wind-inflicted stress on thin stems andcompression wood induced by lateral branches. The first, obviously, only affectswood produced low in the trunk but the latter two factors affect the leadingshoot at all growth heights. Prevailing winds, especially during the springmonths when cambial divisions and differentiation rates are high, will causeexcessive flexing of the stem and result in tracheids of large microfibril angleand low stiffness. The weight of laterals relative to the small diameter ofthe leading shoots is also highly significant in producing large amounts ofcompression wood in stems of small diameter. Once established, these largemicrofibril angles appear to take several years to decline before cells of highstiffness are produced.

The other major wood quality trait under the influence of microfibril angleis longitudinal shrinkage. Longitudinal shrinkage increases with microfibrilangle but in a highly non-linear relationship (Fig. 2.7). Longitudinal shrinkageis always less than transverse shrinkage. Tangential shrinkage in turn is around1.5 to 2.5 times that of radial shrinkage (Walker, 1993). Cellulose microfibrilsdo not change in length or cross section with changes in moisture content, butthe amorphous matrix of the cell wall does show a tendency to change. Thecellulose microfibrils therefore act to restrain the matrix from shrinkage inthe direction parallel to the microfibril angle. Barber and Meylan (1964)developed a very good model of the cell wall based on the microfibril–matrixinteraction showing that where the microfibril angle lies between 0 and 45°,changes in width will exceed the change in length and this is supported by


experimental observations. Both juvenile and compression woods showunusually high longitudinal shrinkages which is in keeping with their theory.The most serious effect of asymmetric shrinkage is warp in timber. Timberthat includes tracheids with differing microfibril angles will cut straight whenwet but will shrink unevenly when dried. This is a particularly serious problemwhen boards are cut through the pith as this zone includes the most significantrange of microfibril angles. Tracheids with microfibril angles in excess of 35°will shrink by 0.5–5.0%, whereas those with smaller angles will shrink lessthan 0.5% (Andrews, personal communication). Timber that crosses the core-wood where microfibril angles are changing significantly will therefore displaysevere warp on drying.

Although the influence of microfibril angle on wood quality has beenunderstood for several years now, it has always been very difficult to actually

TANGENTIAL

LONGITUDINAL

Pinus jeffreyi

Per

cent

shr

inka

ge7

6

5

4

3

2

1

00° 10° 20° 30° 40° 50° 60°

Microfibril angle (θ)

Fig. 2.7 Shrinkage as a function of microfibril angle (after Meylan, 1968).


measure this parameter in the field or at the mill site. The advent of modernacoustic tools for measuring wood stiffness of entire logs is now beginning tochange this (Andrews, personal communication). Acoustic technology is nowable to relate stiffness (as influenced by microfibril angle) to actual woodquality at the mill allowing for rapid segregation of logs into those suitable forhigh quality timber and those more suitable for mechanical chipping.

2.3 The cell wall in hardwood fibres and vessel elements

2.3.1 Fibres

Although most research relating cell wall properties to wood quality has beenundertaken on softwood tracheids, most of the findings are equally applicableto hardwood fibres. The fundamental difference between softwood tracheidsand hardwood fibres is that the latter do not play a major role in conductingwater up the trunk of the living tree. As a result, the cell walls of hardwoodfibres are frequently much thicker and their pits reduced in size (Figs 2.8aand b). Hardwood fibres, while produced from derivatives of the cambialfusiform initials in much the same way as softwood tracheids, differ in thatthey undergo significant elongation by cell tip intrusive growth prior tosecondary cell wall deposition. This intrusive elongation of the fibre tips isparticularly important when the fibres are derived from a storied cambium.The effect of this tip growth on the subsequent physical properties of fibres isnot fully understood. It is obviously important in maintaining a dense interlacingpattern of cell tips – the effect being obvious by the variety of fibre cross-sectional sizes in a transverse view of wood (Fig. 2.8a) as well as beingobvious in longitudinal views.

The hardwood fibres of some woods also differ from softwood tracheids intheir ability to lay down an extra (G or gelatinous) wall layer comprisinga low-angle cellulose-rich wall deposited as a substitute to the normal S3 layeror occasionally on top of it (S1-S2-G or S1-S2-S3-G). This layer showsa higher longitudinal shrinkage on drying than the larger angled S2 wall layer(Fig. 2.8c) (Clair & Thibaut, 2001). This appears at first instance to refute thewidely held theory of Barber (1968) but may simply suggest that hardwoodfibres are in a state of stress. This stress is released when the fibres are cut anddried, allowing the G-layer to separate from the S2 layer and to displaya greater shrinkage than the rest of the fibre wall (Fig. 2.8d).

2.3.2 Vessels

Hardwood vessels have a wall structure rather different from both tracheidsand fibres. Here the microfibrils of the S2 wall layer are aligned at close to 90°


to the long axis of the vessel element (Fig. 2.9a), and the cell ends are perfor-ated with membrane free openings in various arrangements (Butterfield et al.,2000) in order to facilitate the rapid movement of water (Fig. 2.9b). The verylarge microfibril angle is presumably an adaptation to the fact that the vesselsare under internal water tension.

Fig. 2.8 (a) Libriform fibres seen in cross section in Hoheria angustifolia (Malvaceae). Note theapparent different sizes of the fibres. (b) Libriform fibres exposed by longitudinal cut in Beilschmiediatawa (Lauraceae). The brooming of the left-hand wall of each fibre in this micrograph, together withthe angle of the pit apertures, clearly shows the left-handed winding helix of the S2-layer microfibrils.(c) Gelatinous fibres in the tension wood of Populus sp. (d) The cell wall cut longitudinally in Populus sp.showing the gelatinous (G) layer overlying the S2 layer. The G-layer shows a higher shrinking than theS2 layer.


2.4 Cell wall pits and perforations

2.4.1 Pits

The walls of most wood cells are punctured by pits. Strictly speaking, twocontiguous pits connecting adjacent cells form a pit-pair, but the term pit isnow often rather loosely used to describe the connection between adjacent cells.Pits arise during cell differentiation and form important canals for cytoplasmic

Fig. 2.9 (a) A vessel wall in Knightia excelsa (Proteaceae) split to reveal the S3, S2, S1 layers and theinter-vessel pit membranes. Note the high microfibril angle of the S2 microfibrils. (b) Four openingsfrom a long scalariform perforation plate in the end wall of a vessel element in Hedycarya arborea(Monimiaceae), some with remnants of the digested former perforation partition in the corners.(c) Inter-tracheid coniferous bordered pit-pairs in Pinus nigra (Pinaceae). (d) A simple fibre to fibrepit-pair in Aralidium pinnatifidium (Araliaceae).


continuity between cells. In this role they clearly function in providingdifferentiating xylem cells with material for cell wall deposition. Afterdifferentiation, the pits in tracheary cells (tracheids in softwoods and vesselsin hardwoods) function in water conduction allowing long distance watertransport to occur by way of the empty cell lumens in wood. The pit-pairsconnecting fibre cells in hardwoods would appear to be of limited use aftercell autolysis since these cells play a very minor role in conduction.

The pits of softwood tracheids, hardwood vessels and most hardwood fibresare bordered, meaning that the secondary wall overarches the pit membrane.In the case of the softwood tracheid (a cell type that functions in both conduc-tion and support), this allows for the maximum pit membrane surface area forthe transference of water with the minimum interruption of the secondary wall(Fig. 2.9c). In the case of the hardwood fibre, the pit membrane surface area isvery small in keeping with the cell’s main function of support (Fig. 2.9d).

The basic structure of the softwood inter-tracheid pit is now well under-stood (Liese, 1965; Bauch et al., 1972; Fukazawa & Ishida, 1972; Imamura &Harada, 1973; Sano & Nakada, 1998; Sano etal., 1999), the pit-pairs connectingearlywood tracheids having a highly specialized pit membrane divided intoa secondarily digested margo and a central undigested torus (Fig. 2.10a).Water flow between tracheids occurs through the margo of the pit membrane,though fine pores have occasionally been reported in the tori of some species(Sano et al., 1999). The deflection of the pit membrane to one side, known aspit aspiration (Fig. 2.10b), is an irreversible process that accompanies cavita-tion. The pit membranes between tracheids in the latewood are not normallydifferentiated into a torus and margo in keeping with the transition in functionof the latewood from conduction to primarily support. Softwood inter-tracheary pits occur mainly on the radial walls except in the last few cells of thegrowth ring in some woods. This is a logical situation as pits on the tangentialwalls only allow water transference between cells derived from the samecambial initial in a radial file, and do not therefore facilitate long distance axialtransport of water.

Hardwood inter-vessel pits also play a vital role in long distance transport.The observation that most vessels are comparatively short (Skene & Boladis,1968; Zimmermann & Jeje, 1981; Middleton & Butterfield, 1990) and form ananastomosing network within wood, means that water must move sidewaysfrom one vessel to another in order to ascend stem distances that are greaterthan the length of the longest vessel (Zimmermann, 1983). This is achieved bylateral or sideways movement of water from one vessel to another through theinter-vessel pit-pairs. Inter-vessel pit membranes are not differentiated intoa torus and margo but are made up of a mat of randomly arranged microfibrilsthat are porous on account of the loss of their matrix wall components(Fig. 2.10c). This process has been described and illustrated by O’Brien(1970), Yata et al. (1970) and Bonner and Thomas (1972). The hydrolyzing


enzymes released just prior to cell autolysis reduce the pit membrane toa residue of fine cellulose microfibrils allowing vessel to vessel water movement.Although the pit membranes appear substantial under the scanning electronmicroscope (Fig. 2.10d), they have been demonstrated to be permeable(Bonner & Thomas, 1972; Petty, 1981). Vessel walls are commonly prolificallypitted with most of the area of primary wall–middle lamella complex beingconverted to pit membranes. This means that the secondary wall of the vessel

Fig. 2.10 (a) Part of a coniferous type bordered pit membrane in Pinus radiata (Pinaceae) showing theundigested torus to the left and the digested margo to the right. Water flow occurs largely unobstructedthrough the margo traversed by a web of cellulose macrofibrils. (b) An aspirated pit in Agathisaustralis (Araucariaceae) showing the pit membrane deflected back against the pit border of the cell tothe rear of this micrograph. (c) Inter-vessel pit membranes exposed by wall separation in Populus sp.Note the lack of a torus and margo. (d) Inter-vessel pit membranes exposed by wall fracture inKnightia excelsa (Proteaceae) showing the close packed nature of the membranes and the randomorientation of the cellulose microfibrils within each membrane.


element is attached to the primary wall over a very small percentage of its totalarea and may partly explain the need to wind S2 wall microfibrils at sucha large angle to the long axis of the cell in order to attain hydraulic efficiency inthe vessel.

2.4.2 Perforations

The wall hydrolysis that digests the pit membrane matrix wall components isalso responsible for the opening up of the perforation plates in the end walls ofeach vessel element. This process leaves a single large opening in the end wallin the case of simple perforation plates, or a number of openings variouslyarranged in the case of scalariform or reticulate perforation plates. It isnot uncommon to find remnants of the cellulosic web still present in theopenings of multiple perforation plates especially traversing the smaller openings(Fig. 2.9b). Long distance water transport in hardwood vessels thereforeoccurs from vessel element to vessel element through the perforations in theend walls and laterally through the inter-vessel pit-pairs.

2.5 Vessel-less angiosperms

Some plants have an interesting wood structure that is intermediate betweensoftwoods and porous hardwoods. Members of the Winteraceae and three

Fig. 2.11 (a) The wood of the vessel-less angiosperm Pseudowintera colorata (Winteraceae)comprises only axial tracheids, axial parenchyma and broad multiseriate rays. (b) The inter-trachearypits connecting the tracheids in the vessel-less woods are separated by thin pit membranes that lack thetorus and margo characteristic of the inter-tracheary pit membranes of softwoods. This view is alsoP. colorata.


other minor families are technically hardwoods since they are angiosperm(flowering) plants. Their woods, however, while possessing multiseriate rays,lack vessels and rely entirely on tracheids for long distance water transport andsupport in much the same way as the softwoods (Fig. 2.11a). However, unlikesoftwood tracheids, their inter-tracheary pit membranes are not differentiatedinto a torus and margo (Fig. 2.11b), and resemble more closely the pit mem-branes of vessels (Meylan & Butterfield, 1982). The woods of these familiesare of no commercial value, but they have provided us with useful informationon pit membrane differentiation.

Acknowledgements

All the micrographs are the work of the author and Dr Brian A. Meylan (nowretired) with whom the author collaborated for nearly 20 years. A few micro-graphs have been previously published elsewhere: Figures 2.5a and b arereproduced from Three Dimensional Structure of Wood, B.G. Butterfield andB.A. Meylan, Chapman & Hall, London 1981; Figures 2.8b and 2.10b arereproduced from B.A. Meylan B.G. Butterfield Structure of New ZealandWoods NZ Government Printer 1978; and Figures 2.9a and 2.10d from papers byB.G. Butterfield and B.A. Meylan published in the Journal of the InternationalAssociation of Wood Anatomists. Figure 2.6 is adapted from Cave (1968) andreproduced by kind permission of Wood Science and Technology, and Figure 2.7from Meylan (1968) with kind permission of the Forest Products Journal.

References

Abe, H., Ohtani, J. & Fukazawa, K. (1991) FE-SEM observation on the microfibrillar orientation in thesecondary wall of tracheids. Bulletin of the International Association of Wood Anatomists, 12,431–438.

Abe, H., Ohtani, J. & Fukazawa, K. (1992) Microfibrillar orientation of the innermost surface of conifertracheid walls. Bulletin of the International Association of Wood Anatomists, 13, 411–417.

Bailey, I.W. & Vestal, M.R. (1937) The orientation of cellulose in the secondary wall of tracheary cells.Journal of the Arnold Arboretum, 18, 185–195.

Bamber, R.K. & Burley, J. (1983) The wood properties of radiata pine. Commonwealth AgricultureBureau, Slough, 83pp.

Barber, N.F. (1968) A theoretical model of shrinkage of wood. Holzforschung, 22, 97–103. Barber, N.F. & Meylan, B.A. (1964) The anisotropic shrinkage of wood. Holzforschung, 18, 146–156.Bauch, J., Liese, W. & Schultze, R. (1972) The morphological variability of the bordered pit

membranes in gymnosperms. Wood Science and Technology, 6, 165–184. Bendtsen, B.A. & Senft, J. (1986) Mechanical and anatomical properties in individual growth rings of

plantation grown eastern cottonwood and loblolly pine. Wood and Fibre Science, 18, 23–38. Bonner, L.D. & Thomas, R.J. (1972) The ultrastructure of the intercellular pathways in vessels of

yellow polar (Liriodendron tulipifera L.), Part 1: Vessel pitting. Wood Science and Technology, 6,687–698.


Bunn, E.H. (1981) The nature of the resource. New Zealand Journal of Forestry, 26, 162–199. Butterfield, B.G. & Li, G. (2000) Wood properties of glasshouse grown clonal radiata plantlets. Report

to the MultiClient Seedling Group, University of Canterbury. May 2000, 12pp. Butterfield, B.G., Meylan, B.A. & Eom, Y.G. (2000) Three Dimensional Structure of Wood. WIT

Consulting, Seoul, 149pp. Cave, I.D. (1966) Theory of X-ray diffraction method for the measurement of microfibril angle in

wood. Forest Products Journal, 16, 27–42. Cave, I.D. (1968) The anisotropic elasticity of the plant cell wall. Wood Science and Technology, 2,

268–278. Cave, I.D. (1969) The longitudinal Young’s modulus of radiata pine. Wood Science and Technology, 3,

40–48. Cave, I.D. (1997a) Theory of X-ray measurement of microfibril angle, Part 1: The condition of

reflection. Wood Science and Technology, 31, 143–152. Cave, I.D. (1997b) Theory of X-ray measurement of microfibril angle, Part 2: The diffraction pattern.

Wood Science and Technology, 31, 225–234. Cave, I.D. & Robinson, W. (1998a) Measuring microfibril angle distribution in the cell wall by means

of X-ray diffraction, in Microfibril Angle in Wood (ed. B.G. Butterfield), Proceedings of theIUFRO/IAWA International Workshop on the significance of microfibril angle to wood quality,pp. 81–93.

Cave, I.D. & Robinson, W. (1998b) Interpretation of the (002) diffraction arc by means of the minimalistmodel, in Microfibril Angle in Wood (ed. B.G. Butterfield), Proceedings of the IUFRO/IAWAInternational Workshop on the significance of microfibril angle to wood quality, pp. 108–115.

Cave, I.D. & Walker, J.C.F. (1994) Stiffness of wood in fast-grown plantation softwoods. ForestProducts Journal, 44, 43–48.

Clair, B. & Thibaut, B. (2001) Shrinkage of the gelatinous layer of poplar and beech tension wood.Journal of the International Association of Wood Anatomists, 22, 121–131.

Core, H.A., Cote, W.A. & Day, A.C. (1976) Wood Structure and Identification. Syracuse UniversityPress, 169pp.

Cown, D.J. (1992) Corewood (juvenile wood) in Pinus radiata – should we be concerned? New ZealandJournal of Forestry Science, 22, 87–95.

Cowdrey, D.R. & Preston, R.D. (1966) Elasticity and microfibrillar angle in the wood of sitka spruce.Proceedings of the Royal Society London, 166, 245–272.

Donaldson, L.A. (1991) The use of pit apertures as windows to measure microfibril angle in chemicalpulp fibers. Wood and Fibre Science, 23, 290–295.

Donaldson, L.A. (1993) Variation in microfibril angle among three genetic groups of Pinus radiata.New Zealand Journal of Forestry Science, 23, 90–100.

Donaldson, L.A. (1996) Effect of physiological age and site on microfibril angle in Pinus radiata.Journal of the International Association of Wood Anatomists, 17, 421–429.

Donaldson, L.A. (1998) Between-tracheid variation in microfibril angles in radiata pine, in MicrofibrilAngle in Wood (ed. B.G. Butterfield), Proceedings of the IUFRO/IAWA International Workshopon the significance of microfibril angle to wood quality, pp. 206–224.

Donaldson, L.A. (2001) A three dimensional computer model of the tracheid cell wall as a tool forinterpretation of wood cell wall ultrastructure. Journal of the International Association of WoodAnatomists, 22, 213–233.

Donaldson, L.A. & Burdon, R.D. (1995) Clonal variation and repeatability of microfibril angle in Pinusradiata. New Zealand Journal of Forestry Science, 25, 164–174.

Evans, R. (1998) Rapid scanning of microfibril angles in increment cores by X-ray diffraction, inMicrofibril Angle in Wood (ed. B.G. Butterfield), Proceedings of the IUFRO/IAWA InternationalWorkshop on the significance of microfibril angle to wood quality, pp. 116–139.

Evans, R., Downes, G., Menz, D.N. & Stringer, S.L. (1995) Rapid measurement of variation in tracheidtransverse dimensions in a radiata pine tree. Appita, 48, 134–138.


Evans, R., Stuart, S.A. & Van Der Touw, J. (1996) Microfibril angle scanning of increment cores byX-ray diffractometery. Appita, 29, 411–414.

Fukazawa, D. & Ishida, S. (1972) Study on the pit of wood cells using scanning electron microscopy.3. Structural variation of the bordered pit membrane on the radial walls between tracheids inPinaceae species. Mokuzai Gakkaishi, 19, 413–420.

Hanley, S.J. & Gray, D.G. (1994) Atomic force microscope images of black spruce wood sections andpulp fibers. Holzforschung, 48, 29–34.

Hanley, S.J. & Gray, D.G. (1999) AFM images in air and water of kraft pulp fibers. Journal of Pulpand Paper Science, 12, 196–200.

Harada, H. (1965) Ultrastructure and organization of gymnosperm cell walls, in Cellular Ultrastructureof Woody Plants (ed. W.A. Cote), Syracuse University Press, pp. 215–234.

Harris, J.M. & Meylan, B.A. (1965) The influence of microfibril angle on longitudinal and tangentialshrinkage in Pinus radiata. Holzforchung, 5, 144–153.

Huang, C.-L., Kutscha, N.P., Leaf, G.J. & Megraw, R.A. (1998) In Microfibril Angle in Wood (ed.B.G. Butterfield), Proceedings of the IUFRO/IAWA International Workshop on the significanceof microfibril angle to wood quality, pp. 177–205.

Imamura, Y. & Harada, H. (1973) Electron microscope study on the development of the bordered pit inconiferous tracheids. Wood Science and Technology, 7, 189–205.

Kataoka, Y., Saiki, H. & Fujita, M. (1992) Arrangement and superimposition of cellulose microfibrilsin the secondary walls of coniferous tracheids. Mokuzai Gakkaishi, 38, 327–335.

Kerr, A.J. & Goring, D.A. (1975) The ultrastructural arrangement of the wood cell wall. CelluloseChemistry and Technology, 9, 563–573.

Liese, W. (1965) The fine structure of bordered pits in softwoods, in Cellular Ultrastructure of WoodyPlants (ed. W.A. Cote), Syracuse University Press, pp. 271–290.

Marts, R.O. (1955) Fluorescent microscopy for measuring fibril angles in pine tracheids. Stain Technology,30, 243–248.

Megraw, R.A., Leaf, G. & Bremer, D. (1998) In Microfibril Angle in Wood (ed. B.G. Butterfield),Proceedings of the IUFRO/IAWA International Workshop on the influence of microfibril angle towood quality, pp. 27–61.

Meylan, B.A. (1967) Measurement of microfibril angle by X-ray diffraction. Forest Products Journal,17, 51–58.

Meylan, B.A. (1968) Cause of high longitudinal shrinkage in wood. Forest Products Journal, 18,75–78.

Meylan, B.A. & Butterfield, B.G. (1982) Pit membrane structure in the vessel-less woods of Pseudo-wintera (Winteraceae). Bulletin of the International Association of Wood Anatomists, 3, 167–175.

Middleton, T.M. & Butterfield, B.G. (1990) Vessel-length distribution in the stems of three species ofNothofagus. Wood Science and Technology, 24, 17–22.

O’Brien, T.P. (1970) Further observations on hydrolysis of the cell wall in the xylem. Protoplasma, 69,1–14.

Panshin, A.J. & de Zeuuw, C. (1970) Textbook of Wood Technology, McGraw-Hill, New York, 705pp.Petty, J.A. (1981) Fluid flow through the vessels and intervascular pits of sycamore wood.

Holzforschung, 35, 213–216. Preston, R.D. (1948) The fine structure of the walls of the conifer tracheid III. Acta, 2, 370–383. Preston, R.D. & Wardrop, A.B. (1949) The fine structure of the walls of the conifer tracheid. IV.

Biochimica et Biophysica Acta, 3, 549–559. Sano, Y. & Nakada, R. (1998) Time course of the secondary deposition of incrusting materials on

bordered pits of Cryptomeria japonica. Journal of the International Association of Wood Anato-mists, 19, 285–299.

Sano, Y., Kawakami, Y. & Ohtani, J. (1999) Variation in the structure of intertracheary pit membranesin Abies sachalinensis, as observed by field-emission scanning electron microscopy. Journal ofthe International Association of Wood Anatomists, 20, 375–388.


Sell, J. & Zimmermann, T. (1993) Radial fibril agglomerations of the S2 on the transverse fracturesurfaces of tracheids of tension loaded spruce and white fir. Holz als Roh- und Werkstoff, 51, 384.

Skene, D.S. & Boladis, V. (1968) A study of vessel length in Eucalyptus obliqua. Journal ofExperimental Botany, 19, 825–830.

Verbelen, J.P. & Stickens, D. (1995) In vivo determination of fibril angle orientation in plant cell wallswith polarization CSLM. Journal of Microscopy, 177, 1–6.

Walker, J.F.C. (1993) Primary Wood Processing: Principles and Practice. Chapman & Hall Ltd,London, 595pp.

Walker, J.C.F. & Butterfield, B.G. (1996) The importance of microfibril angle for the processingindustries. New Zealand Journal of Forestry, 40, 34–40.

Walker, J.C.F. & Woollons, R. (1998) Cell wall organisation and the properties of xylem – a speculativereview, in Microfibril Angle in Wood (ed. B.G. Butterfield), Proceedings of the IUFRO/IAWAInternational Workshop on the significance of microfibril angle to wood quality, pp. 13–26.

Yata, S., Otoh, T. & Kishima, T. (1970) Formation of perforation plates and bordered pits in differen-tiating vessel elements. Wood Research, 50, 1–11.

Zimmermann, M.H. (1983) Xylem structure and the ascent of sap. Springer-Verlag, Berlin, 143pp. Zimmermann, M.H. & Jeje, A.A. (1981) Vessel-length distribution in the stem of some American

woody plants. Canadian Journal of Botany, 59, 1882–1892.

3 Wood chemistry in relation to quality Helena Pereira, José Graça and José C. Rodrigues

3.1 Introduction

Wood properties result from the combination of

(i) macroscopic morphology – presence, extent and distribution of differenttypes of wood tissue, e.g. reaction wood, heartwood, knots, wound wood,growth rings

(ii) anatomy – types of cells and biometry, and their proportion (iii) chemical composition – cell wall components and extraneous materials.

Anatomical and chemical characteristics are the ultimate factors that determinethe overall properties of wood as a material and distinguish it from other prod-ucts (cellular polymeric solids, ceramics, etc.). Different combinations of cellstructure and chemical composition, resulting from the effects of environment orgenetics, are also responsible for the differences between wood species andwithin a species (i.e. hardwoods and softwoods).

Anatomy and chemical composition also distinguish the different types ofwood within a stem. Reaction wood (compression or tension wood) has specificcharacteristics of cell structure and chemical composition that clearly differ fromthose of normal wood. Heartwood differs chemically from sapwood, especiallyin relation to the extractive components. Traumatic tissues have a differentanatomical organization and accumulate protective chemicals. Knot woodcontains reaction wood and wound wood. Juvenile wood differs in some anatom-ical and chemical features from mature wood.

In the end, a large proportion of the variation found between specific woodsamples is an expression of macroscopic morphological features, whichthereby largely affect the end-use properties and quality of the material. Forexample, the proportion of heartwood is linked to the amount of extractives,which may increase wood durability (a positive factor for timber use) butdecrease pulp yield (a negative factor for pulp production).

However, the underlying factors controlling wood properties are essentiallythe result of its chemical composition at three levels:

1. chemical features of the molecules that constitute the cell walls (struc-tural components) and those contained within the cellular structure(extractive components), in terms of monomeric composition, molecularstructure and intermolecular organization


2. distribution of the chemical components in the cell structure (theorganization of the cell walls) and

3. relative proportions of the different chemical components in the woodcells and tissues.

The specific properties of wood may be traced back to a combination ofthese aspects, and wood utilization and end-product related quality shouldthereby link directly to its chemical characteristics. However, the complexityof the microscopic and macroscopic organization of wood and the conflictinginfluence of different factors do not allow the individual effects to be separatedin most cases, and the measured properties are the combined result of chemistry,anatomy and morphology.

In this chapter, the first part will deal with the chemical aspects of wood: themain chemical components that constitute the cell wall (cellulose, hemicellu-loses and lignin) and the extractive components will be characterized, as wellas their distribution in the cell wall, calling attention to the features that havean impact on properties. The second part will focus on the relevance of chemicalcomposition and its effect on end-use, and on chemically related quality factors,particularly for the two main areas of wood utilization, solid timber, and pulpand paper.

3.2 Chemical composition of wood

The chemical components of wood may be classified as:

(i) Structural components – These components make up the structure ofcell walls and are responsible for the form of cells and for most of thephysical and chemical properties of wood. They are polymeric insol-uble macromolecules. Removal of a structural component from the cellwall requires chemical or mechanical treatments to depolymerize, atleast partially, and solubilize it. The cellular characteristics and thewood properties are substantially altered in this process. The structuralcomponents of wood are cellulose, hemicelluloses and lignin.

(ii) Extractive components – These are non-structural components thatare contained in cell lumina, cellular voids or channels. The organicextraneous components are largely soluble and may be removedfrom wood by use of solvents with adequate polarity, without appre-ciably changing the cellular structural characteristics. They arecommonly referred to as extractives. Extractives include a largevariety of chemical compounds, in general of low molecular mass,and only a few are polymers. Inorganic components are also present,usually to a small extent (<1%), commonly referred to as the ashcontent of wood.

WOOD CHEMISTRY IN RELATION TO QUALITY 55

The chemical composition of wood ranges broadly between 40 and 50%cellulose, 20 and 30% hemicelluloses, 20 and 35% lignin, 0 and 10% extract-ives. Variation of chemical composition is high between different species and,to a more limited extent, also within a species, as a result of environmental andgenetic factors. Within a tree, age, growth and stress factors also influence thechemical composition. Although the structural components may vary inquantity and, in some cases, in composition (e.g. lignin and hemicelluloses),most of the chemical variability both in structure and concentration is found inextractives. Wood chemistry has been a subject of research over the years, andthe literature contains numerous specific publications and various reviews(Sjostrom, 1981; Rowell, 1984; Fengel & Wegener, 1989; Hon & Shiraishi, 1991;Lewin & Goldstein, 1991).

3.2.1 Cell wall structural components

3.2.1.1 Cellulose Cellulose is the main component of wood and the skeletal polysaccharide of cellwalls. It is a long polymer chain of β-D-glucose molecules in the pyranose formlinked together by β(1-4) glycosidic bonds. Two adjacent glucose moleculesbind by elimination of one molecule of water, thereby making the anhydro-glucose units that constitute the repeating chemical entities in the cellulose mole-cule. Due to the β-conformation of the glucose molecule (the –OH group on C1opposite to the group on C4), the C1�O�C4′ glycosidic bond between the twomolecules requires them to be rotated by 180° (flipped vertically) while main-taining the same equatorial plane, making up one cellobiose unit (Fig. 3.1).

The cellulose molecule is built up by the repetition of cellobiose resi-dues, in a long chain containing several thousands of anhydroglucose units,with the molecular formula of (C6H10O5)n. In wood, the degree of polymeri-zation is about 10 000. The dimensions of a single cellulose molecule havebeen estimated as 5 μm for length and 0.5 nm and 1 nm for cross-sectionaldiameters, respectively perpendicular to, and in plane of, the pyranose ring.The molecular chain of cellulose is linear but the successive glucose units

O

H

O

H

HO

H

HOHH

OH

OH

O

H

HO

H

HOHH

OH

O

H

O

H

HO

H

HOHH

OH

O

H

O

H

HO

H

HOHH

OH

O

Anhydroglucose unit

Cellobiose unit

Fig. 3.1 Representation of the cellulose molecule, showing the anhydroglucose monomeric unitderived from β-D-glucopyranose, the glycosidic 1–4 inter-monomeric link, and the structurallyrepeating cellobiose unit.


show a bent conformation to eliminate stereochemical interactions betweenhydrogens on C4 and C1′ and between O2 and C6′ (Gardner & Blackwell,1974), as shown in Fig. 3.2.

The supramolecular structure of cellulose is characterized by a highlyordered arrangement with densely packed molecules, building up a fibrous-like

OO

O

O

O

C2

C3C4

C1'

C6C5

O1

C1

O2

O3

O1'

O2'O5

O5

O6

'

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

Fig. 3.2 Intramolecular and intermolecular H-bonding in cellulose. Adapted from Gardner andBlackwell (1974).


rod structure called a microfibril. This structure is based on the alignment ofthe cellulose chains parallel to each other and in the same direction, buildingup planar layers. Intramolecular H-bonds between adjacent monomeric units(O3–H ⋅O5′ and O2–OH ⋅O6′) and intermolecular H-bonds (O6–H ⋅O3′)between adjacent chains act as structural anchoring and tighten the structure(Fig. 3.2). The different layers are arranged parallel to each other and linked byvan der Waals’ forces (Fig. 3.3).

The regular arrangement of the cellulose molecules forms what is calledcrystalline cellulose (in wood known as cellulose I) for which the repeatingunit cell is proposed to approximate a monoclinic two-chain model (Fig. 3.4).In this model, four cellulose chains make up the corners of the lattice, seen as alozenge in cross section, with a central chain located in the middle and stag-gered longitudinally in relation to them (two chains = one central chain + fourcorner quarter-chains) (Gardner & Blackwell, 1974). This is a densely packedlattice without any empty space between the cellulose chains (Fig. 3.5).

OO

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

H-bonds

van der Walls

OO

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

OO

O

O

O

Fig. 3.3 Schematic organization of cellulose molecules in parallel sheets. In-plane cellulose chains showintramolecular and intermolecular H-bonding; van der Waals’ forces are acting between the planes.


Cellulose microfibrils are not totally crystalline, and regions with a less-ordered arrangement of the cellulose molecules occur along the chain length aswell as across the chains. In wood it is estimated that about 70% of thecellulose is crystalline. The dimensions of microfibrils, of eventual submicrofi-brillar units (elementary fibrils) and of the crystalline regions within them havebeen a matter of controversy. It is accepted now that the microfibril is ofindefinite length and estimated to include crystalline regions with cross-sectional dimensions of about 2–4 nm, rather square, that are separated byhemicelluloses and that could be associated in higher systems (Fig. 3.6). Thecrystalline regions have a length of about 30–60nm. This means that a single chainwill pass through regions of high crystallinity as well as regions of low or nocrystalline structure in which the molecules are in a looser association with

OO

O

O

O

OO

O

O

O

O

O

O

O

O

O

OO

O

O

O

OO

O

O

O

OO

O

O

O

O

O

O

a

b c

Fig. 3.4 Model for supramolecular structure of cellulose (parallel two-chain model) showing the unitcell (a = 0.817 nm, b = 0.786 nm, c = 1.038 nm, the central chain is staggered by 0.276 nm).


Fig. 3.5 Space filling schematic representation of the cellulose crystalline lattice using the spatialdimensions of glucose according to the atomic van der Waals’ radii. Adapted from Fengel andWegener (1989).

Non-crystalline

Crystallite30–60 nm

Hemicellulose

Non-crystallineHemicelluloseCellulose

Fig. 3.6 Tentative model for a longitudinal and a cross-sectional representation of the cellulosemicrofibril.


each other (Fig. 3.6). Immediately surrounding the fibrillar crystalline core is anarea of low crystallinity comprising non-crystalline cellulose and hemicelluloses.

In spite of the hydroxyl groups that are present in the monomeric buildingunits of anhydroglucose, the establishment of H-bonding (Fig. 3.2) and theclosely packed crystalline lattice (Fig. 3.3) make cellulose insoluble. The struc-ture of cellulose is that of a linear polymer, at both the molecular and thesupramolecular level. It is strongly anisotropic, with the chain length directionbeing mechanically stiffer and stronger. In addition to the fact that most of thecovalent bonds in the backbone of the cellulose molecules are located along thechain, the high tensile strength of cellulose fibres when strained in chain directionis given by: (a) the direct H-bonding of polymer chains within crystallites, (b) theprobable interlinking of adjacent crystallites through sharing of polymer chainswithin a fibril and, (c) the pinning of chain ends within the crystalline andamorphous regions by H-bonding and physical entanglements (Emsley & Stevens,1994). Flexibility of cellulose in directions perpendicular to chain length isallowed by plane slipping of the parallel-oriented sheet layers of cellulose mol-ecules (Fig. 3.3). Reaction on the hydroxyl groups and glycosidic bonds onlyoccurs in moderate conditions in the amorphous regions of cellulose or after dis-ruption of the crystalline lattice. This also applies to sorption of water molecules.

3.2.1.2 Hemicelluloses Hemicelluloses are non-cellulosic polysaccharides that comprise variouscompounds of different chemical composition and molecular structure. Theyare heteropolymers, usually classified according to the main types of sugarresidues that are present. The most important hemicelluloses in wood cell wallsare xylans and glucomannans. Other hemicelluloses may be present, usually inlow proportions, although in more significant amounts in some species, forinstance, arabinogalactan in larchwood.

Hemicelluloses differ from cellulose in many aspects which may be sum-marized for the main wood hemicelluloses as follows:

• They are heteropolymers built up of 2–3 monomers. • The molecular structure consists of a linear backbone chain with short

side branching of usually one monomeric unit (in larchwood arabino-galactan, three side units), with possible branching of the backbone chain(1–2 branches).

• The OH groups of the sugar units may be partially substituted withacetyl groups.

• The degree of polymerization is much lower, up to about 200. • They are not crystalline. • Hemicelluloses do not have a uniform molecular composition and within

each type of hemicellulose the monomeric composition, the degree ofpolymerization and the substitution and branching characteristics mayvary with species and wood type.


The monosaccarides that are found in hemicelluloses include pentoses(β-D-xylose, α-L-arabinose), hexoses (β-D-mannose, β-D-glucose, α-D-galactose)and uronic acids (β-D-glucuronic acid, α-D-4-O-methylglucuronic acid andα-D-galacturonic acid) (Fig. 3.7). α-L-Rhamnose and α-L-fucose may be presentin small quantities.

Hardwoods and softwoods differ in the amount and type of hemicellu-loses. In hardwoods, the predominant hemicelluloses are xylans (O-acetyl-4-O-methylglucuronoxylans) accompanied in much lower amounts byglucomannans. In softwoods, the main hemicelluloses are galactogluco-mannans (O-acetyl-galactoglucomannans) with some xylans (arabino-4-O-methylglucuronoxylans).

The chemical and molecular structural characteristics of the main woodhemicelluloses are described below and the corresponding chemical structuresare schematically shown in Fig. 3.8 (a–e).

Pentoses

O

H

H

H

H

OHH

H

OH

O

OH

H

H

H

H

OHH

H

OHH

OHH

H OH

HO H

O

H

HO

H

β-D-Xylose(β-D-Xylp)

α-L-Arabinopyranose α-L-Arabinofuranose

Hexoses

O

H

HO

H

HO

H

H

OHH

OH

OH

O

H

HO

H

HO

OH

H

HHOH

OH

O

OH

H

H

HO

H

OH

OHHH

OH

(α-L-Arap) (α-L-Araf )

HOHO

HO

β-D-Glucose β-D-Manose α-D-Galactose

Uronic acids

O

H

HO

H

HO

H

H

OHHOH

O

H

H

HO

H

H

OHH

HOOCHOOC HOOC

OH

O

OH

H

H

HO

H

OH

OHHH

β-D-Glucuronic acid α-D-Galacturonic acid

De-oxy-hexoses

O

OH

H

OH

H

H

H

OHCH3

H

OH

O

H

HO

OH

H

OH

H

HCH3

H

OH

α-L-Rhamnose α-L-Fucose

(β-D-Glcp) (β-D-Manp) (α-D-Galp)

(α-D-GalpA)(α-D-Me-GlcpA) α-D-4-O-Methlglucuronic acid

(β-D-GlcpA)

(α-L-Fucp)(α-L-Rhap)

H3CO

Fig. 3.7 Structural formulas of the main monosaccharide units in hemicelluloses.

→ 4

- β

- D

– X

ylp

-1 →

4 -

β -

D –

Xyl

p -1

→ 4

- β

- D

– X

ylp

-1 →

4 -

β -

D –

Xyl

p -1

→ 4

- β

- D

– X

ylp

-1 –

3

α-L-

Ara

f 1 ↓

→ 4

- β

- D

– X

ylp

-1 →

4 -

β -

D –

Xyl

p -1

→ 4

- β

- D

– X

ylp

-1 →

4 -

β -

D –

Xyl

p -1

→ 4

- β

- D

– X

ylp

-1 –

A

c ↓ 2 →

4 -

β -

D –

Man

p -1

→ 4

- β

- D –

Glc

p -1

→ 4

- β

- D

– M

anp

-1 →

4 -

β -

D –

Man

p -1

→ 4

- β

- D

– G

lcp

-1 -

→ 4

- β

- D

– G

lcp

1→ 4

- β

- D

– M

anp

-1→

4 -

β -

D –

Glc

p -1

→ 4

- β

- D

– M

anp

-1→

4 -

β -

D –

Man

p –

1 -

β - D

– G

alp

1

11

↓↓

↓6

66

1 ↓ 6

β -

D –

Gal

p1 ↓ 6

6 6

66

6

β -

D –

Gal

p

11

11

1 ↓

↓

↓↓

↓

→ 3

- β

- D

– G

alp

1→ 3

- β

- D

– G

alp

1→ 3

- β

- D

– G

alp

1→ 3

- β

- D

– G

alp

1→ 3

- β

- D

– G

alp

1→ 3

- β

- D

– G

alp

1

Ac ↓ 3

A

c ↓ 3

Ac ↓ 2

↓ 2

α -

D –

Me

- G

lcpA

1

α -

D –

Me

– G

lcpA

↓ 21

β - L

- A

rap

1 ↓

α-L-

Ara

f1 ↓ 33

(a)

(b)

(c)

(d)

(e)

1 ↓ 6

A

c ↓ 3

1α

- D –

Gal

p

↓ 6

β -

D –

Gal

pα-

L-A

raf

β -

D –

Gal

pβ

- D –

Glc

Ap

β -

D –

Gal

p

Fig

. 3.8

(a)

Rep

rese

ntat

ion

of t

he s

truc

ture

of

O-a

cety

l-4-

O-m

ethy

lglu

curo

noxy

lans

. (b)

Rep

rese

ntat

ion

of th

e st

ruct

ure

of a

rabi

no-4

-O-m

ethy

lglu

curo

noxy

lans

.(c

) R

epre

sent

atio

n of

the

str

uctu

re o

f O

-ace

tylg

alac

togl

ucom

anna

ns.

(d)

Rep

rese

ntat

ion

of t

he s

truc

ture

of

gluc

oman

nans

. (e)

Rep

rese

ntat

ion

of t

he s

truc

ture

of

arab

inog

alac

tans

.


Xylans

O-Acetyl-4-O-methylglucuronoxylans. These hemicelluloses are the majorhardwood hemicelluloses (usually representing 15–25% of wood). They havea linear backbone of β-D-xylopyranosyl units linked by β-(1-4) glycosidic bondswith units of 4-O-methyl-α-D-glucopyranosyluronic acid attached by α-(1-2)glycosidic bonds to every tenth xylose, on average, resulting in irregularlydistributed short side chains. About half of the xylose units are acetylated onC2 or C3 carbons (Fig. 3.8a). Small amounts of rhamnose and galacturonicacid are also present in the xylan chain. The number-average degree of poly-merization ranges from 100 to 200. The xylan chain may be branched, probablywith one short branch per molecule.

Arabino-4-O-methylglucuronoxylans. These hemicelluloses are present insoftwoods (5–10% of wood). They are similar to hardwood xylans with somedifferences in composition and molecular structure. They have a linear back-bone of β-D-xylopyranosyl units linked by β-(1-4) glycosidic bonds withunits of 4-O-methyl-α-D-glucopyranosyluronic acid attached by α-(1-2)glycosidic bonds to every five or six xylose units and α-L-arabinofuranosylunits linked by α-(1-3) glycosidic bonds to every 6–10 xylose units (Fig.3.8b). Small amounts of rhamnose and galacturonic acid are also present.The number-average degree of polymerization is about 50–185. The xylanchain is branched, at C2 of the β-D-xylopyranosyl unit, with one or twobranches per molecule.

Glucomannans

O-Acetylgalactoglucomannans. These are the main hemicelluloses in softwoods(20–25% of wood). They have a linear backbone of β-D-mannopyranosyl andβ-D-glucopyranosyl units linked by β-(1-4) glycosidic bonds in a ratio ofabout 3:1, randomly distributed along the chain. Side units of α-D-galacto-pyranosyl are attached by α-(1-6) glycosidic bonds to mannose and glucoseunits, in variable ratios. In a galactose-rich fraction, the ratio of galactosyl toglucosyl to mannosyl units is 1:1:3, while in a low galactose fraction the ratiomay be 0.1:1:4. About half of the mannose units are acetylated equally on C2or C3 carbons, probably with an irregular distribution (Fig. 8.3c). The number-average degree of polymerization is in the range 40–100.

Glucomannans. These hemicelluloses are present in hardwoods (3–5% of wood).They have a linear backbone of β-D-glucopyranosyl and β-D-mannopyranosylunits linked by β-(1-4) glycosidic bonds in a ratio of about 1–2:1, with a randomdistribution and with a possible branching of the chain (Fig. 2.8d). The number-average degree of polymerization is about 40–70.


Arabinogalactans Arabinogalactans are softwood hemicelluloses, usually present in small amountswith the exception of larch wood where they may represent up to 35%. Theyhave a backbone of β-D-galactopyranosyl units linked by β-(1-3) glycosidicbonds with side chains at C6. These side chains consist of β-(1-6) linkedβ-D-galactopyranosyl residues, usually with two residues, and 3-O-β-L-arab-inopyranosyl-L-arabinofuranosyl residues or L-arabinofuranosyl units (Fig. 3.8e).The ratio of galactose to arabinose residues is about 3–10:1. Small amounts ofD-glucuronic acid units are also present. The degree of polymerization is vari-able for different fractions of the arabinogalactans from about 100 to 600.

As regards their supramolecular structure, the hemicelluloses in the cell wallare amorphous. Xylans cannot have a strong intermolecular association byH-bonds due to the irregular presence of side groups and branching along thelinear chain. An orderly lattice of repeating xylan molecules is therefore notpossible, at least over a long distance. In comparison to a hexosan chain, axylan is more flexible due to the lack of the primary hydroxyl and carbon (C6) thatstiffens the molecule as a result of stereochemical hindrance and intramolecularH-bonding. In a similar way, the O-acetyl groups also prevent a chain orientationin galactoglucomannans and the presence of galactose side groups and branchingdo not allow packing into a lattice.

The hemicellulose molecules may however present a loose fibrillar arrange-ment with some orientation of the backbone chain parallel to the cellulosemolecules. The presence of hydroxyl groups allows, for instance, H-bondingto available positions in neighbouring cellulose molecules, and to bondingwith lignin.

Chemical and enzymatic reactivity of hemicelluloses is high due to theavailability of numerous hydroxyl groups and glycosidic bonds as well asof ester bonds in xylan acetyl groups. The hemicelluloses are alkali solubleand the glucomannans with a high galactose content are water soluble. Thehighly branched arabinogalactans are also an example of water solublehemicelluloses.

3.2.1.3 Lignin Lignin is an aromatic polymer that corresponds to about 20–30% of the woodcell wall material and is its most complex structural component. Lignin is highlyheterogeneous and, in spite of continuing investigation, several chemicalfeatures are still under discussion regarding in situ cell wall lignin structure. Morecorrectly, the term lignins should be used since composition and structure differwith species, type of wood and localization in the cell wall. However, usagehas coined the word lignin for this family of molecules and it will be followedalso here.

Lignin is a macromolecule formed by the polymerization of three phenyl-propane monomers (C9 units), the p-hydroxy-cinnamyl alcohols p-coumaryl


alcohol, coniferyl alcohol and sinapyl alcohol, which only differ in the degreeof methoxyl substitution in C3 and C5 (Fig. 3.9). The aromatic rings of thesealcohols are named respectively p-hydroxyphenyl (H), guaiacyl (G) andsyringyl (S) on which is based the designation of the different types of lignins.The higher order of variation in lignin composition, which was used forlignin classification (Sarkanen & Hergert, 1971; Faix, 1991) is associated withbotanical origin.

The proportion of monomers participating in the construction of the macro-molecule depends therefore on the type of wood: hardwoods have a ligninmade up mainly of coniferyl and sinapyl alcohols (guaiacyl-syringyl lignin,GS-lignin) and softwoods of coniferyl alcohol (guaiacyl lignin, G-lignin).

The polymerization is initiated by the enzymatic dehydrogenation of thealcohol group forming phenoxy radicals. These are resonance-stabilized struc-tures with radical character not only on the phenolic oxygen atom, but alsoon the ring carbons 1, 3 and 5 and on the β carbon of the aliphatic chain(Fig. 3.10). The most reactive position is the phenoxy oxygen because it hasthe highest π-electron density.

The reaction of these radicals occurs by random coupling to form dimericstructures (dilignols). Bonding may occur at various positions as ether andC–C bonds of various types, such as β-O-4, α-O-4, β-β, β-5, 5-5, 4-O-5, β-1(Fig. 3.11). Probability of coupling depends on the reactivity of the variouspositions, with β-O-4 as the most frequent type of linkage in the lignin molecule,as well as on steric hindrance, which prevents coupling on methoxylsubstituted C3 and C5 ring positions and disfavours addition to the C1 ringposition. Coupling proceeds via quinone methide intermediates to yield thephenolic dimeric structures by addition of water (in β-O-4 linked dilignols),intramolecular arrangement (in 5-5, β-β, β-5 linked dilignols), or loss of thealiphatic chain (in β-1 linked dilignol).

OH

CH

CH

CH2OH

OH

CH

CH

CH2OH

OCH3

OH

CH

CH

CH2OH

OCH3H3CO

p-Coumaryl alcohol Coniferyl alcohol Sinapyl alcohol

Fig. 3.9 Chemical structure of the monomeric building units of lignins.

OH

CH

CH

OC

H3

OC

H3

OC

H3

OC

H3

OC

H3

OC

H3

αβδ

12

34

56

O.

CH

CH

. OCH

CH

.. O

O

.

O

I

I

I

III

I

V

V

–e,

–H

+

CH

2OH

CH

2OH

CH

2OH

CH

CH

CH

2OH

CH

CH

CH

2OH

CH

CH

CH

2OH

Fig

. 3.1

0R

eson

ance

str

uctu

res

of th

e ra

dica

l obt

aine

d by

the

enzy

mat

ic d

ehyd

roge

nati

on o

f co

nife

ryl a

lcoh

ol.


The polymerization proceeds further by the formation of a dilignol radical thatmay react either with monomeric or other dilignol radicals, yielding tri- or tetral-ignols. The construction of the molecule continues by this random coupling ofmonolignols and oligolignols. The final molecule has a 3D structure with a networkof different linkages between the monomeric units. It is a highly branched andcomplex structure of phenylpropanoid units linked by various types of covalentbonds (C–O and C–C). The structure is space filling, and on average, isotropic.

Different functional groups are present in the macromolecule: aromatic andaliphatic hydroxyls, benzyl alcohol and ether groups, carbonyl and methoxylgroups. Methoxyl groups are a characteristic feature of lignins, since theyderive from the initial building units: softwood lignins contain about 12–18%,

CHOH

OH

OH

HC

CH2OH

O

OCH3

OCH3

CH

CH

CH2OH

OCH3

HC

HC

H2C

OH

H3CO

CH

CH

CH2

O

O

HC

OH

HC

CH2OH

OCH3

O

OCH3

CH

CH

CH2OH

I + II + H2O II + II II + III β-O - coupling β-β coupling β-5 coupling

CH2

OH

HC

CH2OH

OCH3

OCH3

OH

CH

OH

H3CO

CH

OH

OCH3

CH

CH2OH

CH

CH2OH

II + IV III + III β-1 coupling 5-5 coupling

Fig. 3.11 Main types of linkages in dilignols obtained by the coupling of two radicals derived fromconiferyl alcohol.


hardwood lignins about 15–22%. Average molecular formulas have beencalculated for lignin based on elemental analysis and determination of functionalgroups such as C9H7.92O2.88 (OCH3)0.96, in spruce wood, and C9H7.93O2.95(OCH3)1.46, in beech wood (Fengel et al., 1981).

Chemical bonding exists between lignin and hemicelluloses, mostly as benzylester and ether linkages as well as phenyl glycosidic linkages.

The size of the macromolecule is a matter of discussion, and determinationsdepend largely on the method of isolation. Lignins are polydisperse polymers andaverage molecular weights (Mw) of a few thousands to more than 80 000 havebeen reported, corresponding to molecules with just 20 units up to more than400 units. The visualization of the structure of the lignin macromolecule isfacilitated by the use of models that summarize the main linkages and functionalgroups that occur in the polymer, as given in Fig. 3.12 for a softwood lignin(Sakakibara, 1980) and in Fig. 3.13 for hardwood lignin (Nimz, 1974).Computer-simulated synthesis also offer insights into the special developmentand chemical properties of lignin (Jurasek, 1997).

Lignin is amorphous and does not show an organized supramolecular struc-ture. Under the microscope lignins present a more or less spherical appearancewith globular particles in the range 10–100 nm (Fengel, 1976). The aromaticrings give bulk and rigidity to the structure, while the aliphatic chains allowflexibility into various conformations and packing arrangements between thehemicelluloses in the available space in the cell wall. Lignin is the last structuralcomponent to be incorporated into the cell wall, and therefore the limited spacebetween the polysaccharide chains may cause a preferential orientation of thephenylpropanoid units in parallel with the microfibrils. As a result of its chemicalstructure, lignin is a rigid and hard polymer with strong covalent bonds distributedas a 3D network, responsible for the stiffening of cell wall and for the woodresistance to compression. It is mostly hydrophobic and its water absorptionis low.

3.2.1.4 Distribution in the cell wall The microfibrillar structure of the cell wall layers has been described in detailin Chapter 2. Figure 3.14 represents schematically the cell wall organizationin terms of interconnection between the structural polymers: a microfibrillarskeleton of cellulose enveloped by hemicellulose molecules organized as succes-sive ribbon-like structures, with lignin occupying the empty spaces leftbetween the hemicellulose molecules.

Regarding the chemical composition of the different cell wall layers, theprimary wall has the highest concentration of lignin and the lowest ofcellulose, with the middle lamella being made up mostly of lignin. Thecomposition of spruce tracheids may be used as an example: the compoundmiddle lamella contains about 60% of lignin, 14% cellulose and 27% hemi-celluloses; the S1 layer, 29% lignin, 36% cellulose and 36% hemicelluloses;


the S2 and S3 layers, 27% lignin, 58% cellulose and 15% hemicelluloses(Fengel, 1976).

Many properties of wood relate directly to the microfibrillar organization ofthe cell wall. Unlike most homogeneous and isotropic materials, wood issignificantly stronger in longitudinal tension than in longitudinal compression.The highest strength values are obtained in longitudinal tension (along thegrain), a direct result of the high density of covalent bonds of the cellulosemolecules in that direction. The lowest strength values are obtained in tangential

OHC CH CH2OH

OH3CO

HC

HCOCH3

O

CH

H2C

OH

CH

OCH3

OCH

CH2OH

OH3CO

CO

CH

CH2OH

OH3CO

HCOH

HCOH

CH2O

O

OCH3

HC

HC

CH2OH

HCOH

OH

OCH3

HC

H2C

O

O

CH2

CH

CH

H3CO

O CH

CH2OH

CH

OH

OCH3

OH3CO

HCOHC CH2OH

CO

CH

CH2

CH

OHC

CH2OH

OHH3CO

O

H3CO

HOCH

CH

CH2OH

OH3CO

CH

HC

O

CH2OH

HC

H3CO

CH

CH3

CarbohydrateOH

H3CO

CH

CH

CH2OH

OCH3

OH

O

HCOH

CHO

CH2OHH3CO

HC

CH

CH2OH

OH3CO

HO

CH2

HC

HCOH

O

OCH3

OH

OCH3

HC

CH

CH2OH

O

OCH3

CH

CH

CHOOCH3

HCOH

HC

CH2OH

O

CH

HC

O

CH2OH

OH3CO

OCH3

HC

CH

CH

CH2OH

HO

H3COCH

CH2OH

O

H3CO

CH2

CH2

CH2OH

CH2OH

1

2

11

3

5

4

8

7

9

19

12

18

21

13

22

23

17

1514

24

28

25

26

16

6

27

20

OHC

CHO

CH2OHHCOH

O

10

Fig. 3.12 Model for a softwood lignin, containing 28 phenylpropanoid units. Variations may occur forinstance in units 1, 10–11, 19–16–15, 18. Adapted from Sakakibara (1980).


and radial tension (perpendicular to the grain, about 40 times weaker thanlongitudinal strength) due to the much weaker horizontal bonding between thecellulose molecules and to the layered structure of the cell wall. In longitudinalcompression, the microfibrillar structure is dislocated and the cellulose crystal-lite orientation is altered, resulting in the formation of cross running kinks(Dinwoodie, 1968).

Much of the variation in wood strength and fracture morphology dependson the microfibrillar angle of cellulose in the S2 layer, and weak regions arelocated in the transition between the S1 and S2 layers: fracture under longitu-dinal tension occurs either across the cell walls or between S1 and S2 layers,

CO

H3CO

O

1

CH

CH2OH

2

CH

HC

CH2OH

O

OCH3H3CO

3 O

OCH3

OCH3

CH

HC

CH2OH

O

4

O

OCH3

HC

CH2OH

CHO

5

H3CO OCH3

HC

HC

CH2OH

O

CH2

CH

HC

6

OH

OCH3H3CO

CH

H3CO

OH

7

HC

CH2OH

8 O

OCH3

CH

CH

CH2OH

9 10

O O

CO

CH

H2CO CHO

CH

HC

CH2OH

O

11

H3CO OCH3

12 O CH

HC

CH2OH

13

O

O

CH

HOH2C CO

OCH2

CH

HC

O

O

20

O

CH

HOH2C CHO

18

19C

OCH3

HC

CH2OH

O

CH2OH

CO

16

H3CO

O CH

HC

CH2OH

17

OH

14

HC

HC

CH2OH

O

15

CH

CH

CHO

21

H3CO OCH3

HC

HC

O

H2C

O

CH

CH

CH2

22

O

OCH3H3CO

CH

23

OH

OCH3

CH

CH2OH

O

24

H3CO

HC

HC

CH2OH

O

25

OCH3

CO

HC

CH2OH

OCH3

OCH3

OCH3

OCH3

H3CO

OCH3

H3CO

OCH3H3CO

HCOH

CH

CH2OH

Fig. 3.13 Model for a hardwood lignin containing 25 phenylpropanoid units. Variations are proposedfor units 5–6, 9–10, 24–25. Adapted from Nimz (1974).


and failure under longitudinal shear also appears within the cell wall, usuallybetween the S1 and S2 layers.

Dimensional variation with moisture also reflects the cell wall organization.Water molecules are absorbed only on amorphous cellulose and on hemicellu-loses, resulting in relatively small swelling and shrinkage in the longitudinaldirection and larger changes in the transverse direction (the wood tangentialand radial sections).

3.2.2 Extractive components

Wood includes in its composition a variety of compounds of low and mediummolecular mass that may be removed by solvent extraction, and therefore arecalled extractives. Most of the extractives are secondary metabolites, compoundsthat play other roles in the tree, besides those involved in growth and celldevelopment namely the protection of the tree against pathogens or otherbiotic attacks. Their presence is responsible for the natural durability of solidwood.

In the living tree, the secondary metabolites are preferentially deposited inthe inner part of the stem, in heartwood which is no longer involved in waterconduction and fulfils only a structural role. The synthesis and deposition ofthese protecting compounds is carried out by the living parenchyma cells,mainly the ray parenchyma, at the time of the heartwood formation. Many of

radial

tangential

Cellulose

Cellulose

Hemicellulose Hemicellulose

Lignin

Lignin

LP-linkage

a

b

Fig. 3.14 Schematic representation of the organization of the structural components in the secondarywall of softwood tracheids and hardwood fibres. Adapted partially from Fengel and Wegener (1989).


these compounds are coloured, imparting a different and sometimes conspicu-ous colour to the heartwood. Heartwood is therefore the wood region wheremost of the extractives are found. In the outer part of the stem, the sapwoodincludes recently divided cell layers, where molecules used in the metabolicsynthesis of the cell wall are found as extractives. These are mostly monomericor oligomeric carbohydrates and fatty reserve molecules. Some trees havespecialized defence structures to protect them from attack, either physical orbiotic. This is the case of the resin canals in softwoods, particularly developedin pines, that produce resin, a mixture of terpenoid compounds that appear asimportant extractives in the wood of these trees.

As a rule, extractives represent a small proportion of wood, under 10%,with the exception of tropical woods, where this value can be higher. In spiteof the usually low quantities in a single species, extractives may include hundredsof different molecules. The extractives can be classified in several ways, basedfor instance on their different polarity or the solvent in which they aresolubilized, or organized by families of common biosynthetic pathway or chemicalstructure. The latter approach is followed in the summary below on the extract-ives found in woods.

3.2.2.1 Terpenoid extractives Terpenoids are a large family of different compounds, which have in commonthe fact that they can be seen as structural multiples of isoprene. Isoprene(2-methyl-1,3-butadiene) is a five-carbon structure that upon condensationcan produce the terpenoid skeletons, although it has no role in the biosynthesisof these compounds. According to the number of the isoprene units linked, theterpenoids are divided into hemiterpenoids (1 isoprene unit), monoterpenoids(2), sesquiterpenoids (3), diterpenoids (4), triterpenoids (6), polyterpenoids(more than 8), and others of more rare occurrence (Fig. 3.15). All kinds ofterpenoids, up to the triterpenoids, can be found in softwoods, namely in theresin, which is basically a solution of diterpenic acids dissolved in monoterpenes.In hardwoods, except for some tropical species, only higher terpenoids are found(Fengel & Wegener, 1989).

Monoterpenoids include many of the fragrant and flavoured substancesfound in plants, and in wood they comprise the more volatile fraction of theconifer softwood resin. The more common wood monoterpenoids are monocycliclike limonene (Fig. 3.15a) or bicyclic, as α-pinene and β-pinene (Dev, 1989).The pattern of monoterpenoid composition is genetically and environmentallycontrolled, and widely variable compositions can be found in different popula-tions of the same species.

Sesquiterpenoids (Fig. 3.15b) are also part of the volatile fraction of resins,in smaller proportions compared to monoterpenoids. Sesquiterpenoids areuncommon in hardwoods from temperate zones, but present in some tropicalwoods, like the commercially very valuable α- and β-santalol in sandalwood.

(a)

Lim

onen

e(b

) F

arne

sene

CO

OH

(c)

Abi

etic

aci

d

HO

(d)

β-A

myr

in

HO

(e)

β -S

itost

erol

Fig

. 3.1

5T

erpe

noid

ext

ract

ives

fro

m w

ood

(a)

Mon

oter

peno

id, (

b) S

esqu

iter

peno

id, (

c) D

iter

peno

id, (

d) T

etra

terp

enoi

d an

d (e

) St

erol

.


Wood diterpenoids include as their more common representatives the tricyclicditerpenic acids of softwood resin (Fig. 3.15c), which in pines can amount to0.2–0.8% of wood dry weight (Fengel & Wegener, 1989). Upon the evaporationof the monoterpenoid fraction, these diterpenoids solidify into a thermoplasticvitreous material, interesting as a source for speciality chemicals, but a problemfor many wood uses.

Triterpenoids are widespread in hardwoods, and also in Pineaceae in softwoods.They are based on many skeletal types, most of them pentacyclic. Oleananesare the largest class of triterpenoids, including β-amyrin (Fig. 3.15d) one of themost widely occurring (Dev, 1989). A tetracyclic group of triterpenoids is thesteroids, which include β-sitosterol (Fig. 3.15e), the major sterol found in woods,as well as in many other plant tissues (Nes, 1989). Triterpenoids are also foundesterified to fatty acids, or linked to glycosidic residues, the latter being calledsaponins, which are present in some tropical woods.

Polyterpenoids, in the form of acyclic isoprenoid polymers, including gutta-percha and caoutchouc, are found as exudates in some woods. In a few species,like in the rubber tree, they are produced in sufficient quantity to allow theircommercial exploitation and industrial use (Barlow, 1989).

3.2.2.2 Phenolic extractives The other major group of wood extractives includes compounds with phenolunits in their structure. Phenol as such is not found in extractives. The colourfound in heartwood, as well as the toxic or repellent properties for bioticattackers, come mostly from phenol-based compounds. Some of these phenoliccompounds, namely the group of tannins, can reach relatively high molecularmasses and even condense after their synthesis. Many of these phenolic extract-ives are found as glycosides, linked to glucose and other sugars.

One of the main groups of phenolic extractives, present both in softwoodsand hardwoods, is the lignans. Lignans consist of two linked phenylpropanoidunits, often similar to the dimeric structures found in lignin. Some of the mostcommon lignans have β-O-4 and β�β linkages between the monomers, makingthe tetrahydrofuran ring, as in the pinoresinol from softwoods and the syrin-garesinol from hardwoods (Fig. 3.16a).

Stilbenoids (Fig. 3.16b) are found mostly in the heartwood of pines and areone of the types of extractives that have been proven to inhibit the growth offungi and be formed as a response to insect attack (Gorham, 1995). Thepresence of stilbenoids is a problem for some wood uses, being responsiblefor the light-induced darkening of wood and also difficulties in producing pulpfor paper.

Another important group of phenolic extractives found in wood is the fla-vonoids. Flavonoids are C6–C3–C6 three ring structures, and the structure of thecentral ring defines different classes of flavonoids: flavones, flavanes, flavanones,isoflavones, chalcones, aurones. Substitutions of hydroxyl and methoxyl groups

HC

HC

CH

CH

2

O

H2C

OC

H

OH

H3C

O

OH

OC

H3

(a) S

yrin

gare

sino

l

C H

H C

OH

HO

O

OH

OH

OH

OH

HO

(b)

Pin

osyl

vin

(c)

Cat

echi

n

OH

HO

OH

CO

OH

(d)

Gal

lic a

cid

OH

HO O

CO

OC

O

OH

OH

(e) E

llagi

c ac

id

H3C

OO

CH

3

Fig

. 3.1

6Ph

enol

ic e

xtra

ctiv

es f

rom

woo

d (a

) L

igna

n, (

b) S

tilb

enoi

d, (

c) F

lavo

noid

(fl

avan

-3-o

l), (

d an

d e)

Hyd

roly

zabl

e ta

nnin

phe

noli

c un

its.


in the two aromatic rings define the individual flavonoids. Catechin, a flavane,is one of the more widespread flavonoids (Fig. 3.16c). The colours found in theheartwood are in many cases related to their flavonoid extractives. Flavonoidsexist in wood as such, as glycosides and also in oligomeric and polymeric forms(Harborne, 1989).

Flavanes like catechin (flavan-3-ol) and leucocyanidin (flavan-3,4-diol)condense in dimeric forms, as biflavonoids (proanthocyanidins), or in higherdegree to form polyflavonoids known as condensed tannins. The other type oftannins present in wood is the hydrolyzable tannins, named as such becausethey are hydrolyzed to monomers with acids. They are esters of gallic acid(Fig. 3.16d) and of its dimers, digallic and ellagic acids (Fig. 3.16e), with sugars,usually glucose. Wood tannins, known for their tanning properties of animal skin,can have adverse consequences in gluing wood and pulp production.

3.2.2.3 Other wood extractives Other types of extractive are found in wood, such as fats and waxes, sugarsand alkaloids. Fats are esters of glycerol with long-chain fatty acids and waxesare esters of long-chain alkanols with fatty acids. These are low-polarity com-pounds extracted with non-polar solvents and have been found as extractivesin woods in small quantities. The hydrophobic properties of fats and waxes caninterfere with wood processing by affecting its permeability.

Carbohydrates are present in polar solvent extracts, as monosaccharidessuch as glucose and fructose, disaccharides such as sucrose, and some solublehigher polyoses. These compounds are found mostly in the sapwood, particularlyin the outer part close to the cambium.

Small quantities of nitrogen-containing compounds, like amino acids andproteins can also be found as extractives in wood. Alkaloids are aromaticcompounds with nitrogen as a heteroatom in some rings, and are found in smallquantities in the wood of tropical hardwoods, and in higher quantities in otherplant organs. Alkaloids are compounds of potent biological action in animalslike humans, and some of the better known alkaloids, such as quinine, berberineand strychnine have been detected in some wood extractives (Fengel &Wegener, 1989).

3.3 Variation of chemical composition

Wood chemical composition varies with geographical origin, genus and species.In general softwoods have higher lignin content (25–35%) and their hemicellu-loses contain galactoglucomannan (15–23%) and arabinoglucuronoxylan (7–10%),while hardwoods have less lignin (18–30%) and their hemicelluloses containacetylglucuronoxylan (15–30%) and glucomannan (2–5%). The chemical


composition of different woods from different geographical origins has beencollected over the years (Fengel & Grosser, 1975; Pettersen, 1984) and examplesare in Tables 3.1 and 3.2. However, differences in analytical procedures andsample history in addition to the complexity of the chemical analysis involved,require caution whenever comparisons are made.

An important variation in composition is found in the heterogeneity of ligninwhich has a large impact in the pulping industry. Conifers have basically a Gtype lignin with a few H or S units and show less variation in lignin compositionthan hardwoods. Hardwoods have a more complex lignin composed ofsyringyl (S) and guaiacyl (G) units in varying ratios with a minor percentage ofH units (Sarkanen & Hergert, 1971). The S/G ratio can vary widely from aslow as 0.51 for Acer macrophyllum (Chang & Sarkanen, 1973) to 5.2 forEucalyptus maculata (Bland et al., 1950). Species of the same genus can alsoshow a large variation in the S/G ratio: in eucalypts, a ratio of 5.2 was foundfor E. maculata and E. diversicolor (Bland etal., 1950) and 0.7 for E. tereticornis(Kawamura & Bland, 1967), and in Acers 0.4 for A. negundo (Towers &Gibbs, 1953) and 3.3 for A. rubrum (Creighton et al., 1944).

Table 3.1 Range of variation in the chemical composition of 49 hardwoods and 35 softwoods fromUSA determined using Tappi standards in percent oven-dry wood (Hillis, 1991)

Softwoods % o.d. wood

Hardwoods % o.d. wood

Ash 0.1–0.5 0.1–1.4 Ether solubles 0.1–5.5 0.2–2.1 EtOH-benzene solubles 1–14 2–7 Hot water solubles 2–11 2–15 Klason lignin 25–34 18–30 Holocellulose 55–71 59–85 α-cellulose 37–49 37–52 Pentosans 5–13 14–23

Table 3.2 Range of variation in the chemical composition of 39 hardwoods from southwestern USAin percent oven-dry wood (Hillis, 1991)

% o.d. wood

Extractives 1.1–13.2 Lignin 17.4–30.9 Cellulose 33.8–48.7 Hemicelluloses 22.4–37.7 Glucomannan 1.0–5.0 Acetylglucuronoxylan 16.4–31.9 Arabinogalactan 0.7–2.2 Pectin <0.1–1.3


Tree to tree variation is believed to contribute to lignin heterogeneity althoughthere are few reported data. In eucalypts, the S/G ratio varied from 1.5 to 2.6 inE. globulus (Rodrigues et al., 1999), from 0.68 to 2.22 in E. tereticornis andfrom 1.37 to 2.01 in E. camaldulensis (Kawamura & Bland, 1967), the lattertwo examples being associated with a wide geographical area of distributionfrom tropical to temperate zones. Differences in lignin composition were alsofound between different parts of the tree as in E. botryoides, with S/G ratios of0.7 (foliage petiole), 2.0 (root wood), 1.5 (heartwood) and 1.6 (sapwood), andin E. regnans with 3.8 (heartwood) to 4.6 (sapwood) (Sarkanen & Hergert,1971). For E. globulus and E. camaldulensis a decreasing radial trend in theS/G ratio was found and, additionally, a decreasing trend with tree height forE. camaldulensis (Ona et al., 1997). Lignin heterogeneity also occurs amongtissues in relation to the time of deposition in the cell wall (Terashima et al.,1993). For instance, in Betula papyrifera the lignin in the middle lamella andsecondary walls of vessels is mainly the G type, whereas secondary wall offibres and ray parenchyma is mainly S type (Fergus & Goring, 1970).

The components showing the highest variation are the extractives, with thequantity and types of extractives varying widely between species, within differentsites for the same species and within the same tree. Hardwoods usually havemore extractives than softwoods, besides having different patterns of extractivecomposition. The total quantity of extractives, as a percentage of dry-weightwood, can be less than 1%, as in poplar, more than 10%, as in redwood, andabove 15%, as in some tropical woods such as iroko and obeche. For most ofthe woods from temperate zones the total amount of extractives is around 5%(Tsoumis, 1991).

The effects of site, environmental conditions and genetic variation on thetype and content of extractives in wood are mixed and frequently difficult toseparate. The amount of the stilbenoid pinosylvin was found to vary in theheartwood of Pinus sylvestris from south to north in Sweden (Erdtman etal., 1951).In this same species, the monoterpene composition of the resin was studiedthroughout the area of distribution, including all of Europe and most ofSiberian Asia. The proportion of all seven compounds studied was found tovary widely in the monoterpene mixture. For instance, α-pinene was foundto vary from 5 to 70% of the mixture, 3-carene from 0 to 65% and limonenefrom 0 to 26%. Genetic and environmental factors together are probablyinvolved in this variation (Tobolskii & Hannover, 1971). In Pinus elliotii, thequantity of resin produced, and also the pattern of monoterpene composition,was found to be under strong genetic control (Hillis, 1987).

The occurrence of physical wounds, mechanical stresses, or attack by bioticorganisms induces many trees to produce traumatic tissues, synthesizing anddepositing protecting compounds in the areas affected. Most of these compoundsare extractives and appear as such in high amounts in the parts of the woodinvolved. This is the case for the traumatic resin canals of softwood conifer trees.


Also, both softwoods and hardwoods, particularly from tropical and sub-tropicalareas, produce gums in response to wounding which are of carbohydrate orpolyphenolic nature, such as kino in Eucalyptus. These appear as extractives inaffected parts of the wood.

Important aspects of variation in chemical composition, which are particu-larly relevant to the overall variation in wood and to end-use quality, relate tothe occurrence of different types of wood within the tree stem: juvenile andmature wood, heartwood and sapwood, reaction wood, and knotwood. Theircharacteristics will be described separately.

3.3.1 Juvenile wood

Juvenile wood is produced by a young cambium. It is therefore found in theinner part of the tree cross section and in larger proportion in the higher partsof the stem. In hardwoods the chemical composition shows little change frompith to bark and from base to top, while softwoods have wood cores with lowercellulose and higher lignin, but with little or no significant difference inhemicelluloses for either softwoods or hardwoods. However, the study of theradial variation of chemical composition is complicated by the presence of heart-wood and the possible occurrence of reaction wood, especially of compressionwood in softwoods.

As a consequence of the chemical differences, juvenile wood has a muchgreater longitudinal shrinkage and lower longitudinal tensile strength than maturewood. Further differences in properties (e.g. basic density) are linked to simul-taneous differences in anatomical features such as shorter fibres and tracheids,and these have been extensively studied (Zobel & van Buijtenen, 1989) anddiscussed in Chapter 4.

3.3.2 Heartwood

The central portion of tree boles is altered mainly chemically (to some extentalso structurally) by infiltration and encrusting with extractives that are largelypolyphenolic, derived from the conversion of starch, sugars and other organicextractives present in the sapwood parenchyma. The parenchyma cells usuallydie and their contents infiltrate cell walls, encrust pit membranes with a strongtendency to aspiration of bordered pits and can plug vessels. The natural per-meability of heartwood to liquids and gases decreases in relation to sapwood,and the moisture content is lower.

The formation of heartwood is the main cause of variation in the type andquantity of extractives within the tree. Significant amounts of extractives aredeposited in the heartwood, up to two to three times more than in sapwood.The vertical gradient found in the content of extractives, higher in the lowerparts and lower near the top, is due to the lower proportion of heartwood in thelatter.


Heartwood may have a distinctive colour (brown, yellow, orange, red, etc.)due to the presence of extractives. These will have an effect on wood reactivityor treatment processes, e.g. finishing, preservation, gluing, production ofpolymer/wood composites. In certain species, heartwood extractives are toxicto some extent and render the wood more resistant to decay from the attack ofmicroorganisms and insects.

3.3.3 Reaction wood

Reaction wood is formed as a biological reaction of the tree to external forcessuch as gravity and wind forces, either in the zone of compression (in soft-woods) or in the zone under tension (in hardwoods), as discussed in Chapter 5.Reaction wood is generally associated with elliptical stem cross sections.Compression wood forms on the underside of tree stems growing out of thevertical, on their windward side, in the lower part of trees planted on a slope,and on the underside of branches. Tension wood forms on the upperside ofleaning tree stems and of branches.

Compression and tension wood differ from each other and from normalwood in anatomy and chemical composition. The occurrence of both reactionwoods is a very troublesome problem for wood utilization and one major con-tribution to xylem non-uniformity.

Compression wood has a relatively high lignin content with a com-position characterized by more H-units and a more condensed macromole-cule leading to a classification of compression wood lignin as a GH-lignin.Cellulose content is lower (about 10% less), less crystalline and the micro-fibrillar angles in the S2 wall are flatter than in normal wood (up to 45°). Thecompression wood hemicelluloses also show some differences with anincrease in D-galactosyl units. Some hemicelluloses found in compressionwood include a galactan formed by 1-4 linked β-D-galactopyranosyl unitswith a small proportion of 1-6 linkages and with 1-6 linked β-D-galacturonicacid, as well as an acidic glucan formed by 1-4 and 1-3 linked β-D-gluco-pyranosyl units with branching with D-glucuronic and some D-galacturonicacid residues.

Tension wood has fewer and smaller vessels and the fibres have an additionalcell wall layer, the gelatinous layer (G), either substituting the S2 and S3 or inaddition to them. The G layer consists of concentric lamellae of cellulose fibrilsof high crystallinity aligned with the fibre axis. The chemical composition oftension wood differs from normal wood by a lower content of lignin andhemicelluloses and a higher content of cellulose.

Tension wood is exceptionally weak in compression parallel to the grain,but slightly stronger in tension and tougher than normal wood. It has abnormallyhigh longitudinal shrinkage, probably because of the G layer but, as yet, notfully understood; tangential shrinkage is also rather greater than normal. The


cut surfaces of tension wood tend to be woolly since the fibres tend to bepulled out rather than cut clean, because of the high content of cellulose.

3.3.4 Knotwood

Knotwood is extremely dense and hard. It contains a high percentage of reactionwood and, in some softwoods, it is very resinous. The stem regions below andabove knots are also high in reaction wood and exhibit severely distortedgrain. The presence of knots is generally considered detrimental to processesinvolving wood machining, finishing, gluing and to most strength properties,particularly bending strength.

3.4 Wood chemical quality parameters depending on end-use

The chemical components and their assembly in the cell wall are relateddirectly to the properties of wood, and their effect may be either positive ornegative, depending on end-use. In timber, the role of lignin is associated withcompressive strength and that of cellulose with tensile and bending strength,and changes in their contents or topochemistry are bound to affect theseproperties. However, most of the variation found in solid timber is due to thepresence of different types of wood, namely reaction wood, with which thelargest differences regarding the cell wall structural components are associated.

For solid wood the most important chemical factors affecting quality are theextractives, since their presence affects the processing and use of wood. The majorcontribution of extractives in the use of solid wood is certainly in the naturaldurability they impart. Exposed in use, either in the soil, above groundor immersed in water, wood is susceptible to attack from a number of fungi,insects and marine organisms (Zabel & Morrell, 1992). These xylophagousorganisms are, at least in part, deterred from degrading wood mainly by thetoxicity and repellence of the terpenoid and phenolic extractives. Many com-pounds of these two groups of extractives, like the monoterpenes, diterpenic acids,lignans, flavonoids, stilbenes and hydrolyzable tannins, have been shown tohave anti-microbial activity (Hart, 1989). Because these protecting extractivesare accumulated mainly in the heartwood, the natural durability of the latter isgreater than the sapwood of the same species. Major variations exist in the naturaldurability of woods from different species, but the heartwood of some of themcan be very durable, surviving more than 25 years without xylophagous attacks.

Large amounts of extractives, particularly when they are located inside thecell walls, can increase the density and diminish shrinkage and swelling, atleast of the heartwood, with beneficial consequences for wood utilization(Tsoumis, 1991; Walker, 1993). The combined effect of higher durability andlower dimensional variation makes heartwood a valuable material for structuralapplications (Faherty & Williamson, 1995).


In contrast, the presence of traumatic resin and gum pockets produceslocally high concentrations of these extractives, making those parts of woodunusable. Resin-rich softwoods, like pines, spruces and larches, can exuderesin to the surface of the wood. This happens for instance under the hightemperatures of kiln drying, and it affects the finishing operations, such assanding or the application of paint and varnishes (Walker, 1993). Also, theaccumulation of extractives within the cell walls, and particularly in the pitmembranes, makes the wood less permeable, making the drying of wood or itspreservative treatment more difficult.

Extractives can affect the wettability of wood surfaces and thereby theapplication of paints and adhesives (Uprichard, 1993). The polymerization ofsome adhesives can be inhibited by extractives, such as non-polar hydrocarbonsor hydrolyzable tannins, leading to problems in gluing operations. The presenceof high quantities of extractives in some wood parts, like knots, can damagethe paint films applied in those areas (Imamura, 1989).

Extractives impart colour to wood and their alteration by external factors,such as light or water, can lead to discolouration or more dull and less attractiveappearance. Plant extractives are bioactive compounds and many wood extract-ives have found uses as pharmaceuticals. These are mostly alkaloids, but alsoothers such as lignans and triterpenes are used (Beecher et al., 1989). In a fewcases, wood extractives were shown to be hazardous to health (Hausen, 1981):terpenes and phenolics were found to have allergenic action, tannins are suspectedto have carcinogenic activity and some alkaloids, like coniine from the poisonhemlock, can be lethal (Beecher et al., 1989; Swan, 1989).

The chemical composition of wood has a major impact on its utilization forpulping, both regarding the structural components and the extractives. Extract-ives and heartwood are definitely an obstacle to pulping. Impregnation ofheartwood with the pulping liquors is more difficult due to its lower perme-ability, and this can lead to larger amounts of screened-pulp rejects. The presenceof extractives increases the consumption of pulping chemicals, decreases thepulp yield, the solubility of lignin and increases the colour of the pulp leadingto a more difficult bleaching. During pulping, most extractives are removedbut a residual pitch fraction may remain; it agglomerates in particles in theequipment or appears as a contaminant in the final pulp where it can lead toproblems related to brightness reversion, reduced pulp wettability and self-sizing in paper (Hillis, 1972; Parham, 1983; Kai, 1991; Holmberg, 1999). Thenegative correlation between pulp yield and content of wood extractives hasbeen reported for several species (Turner et al., 1983; Hillis, 1991; Raymondet al., 1994; Wallis et al., 1996; Miranda & Pereira, 2001). Extractives alsodecrease the lifetime of equipment due to enhanced corrosion (Hillis, 1991;Kai, 1991).

As regards structural components, the amount of cellulose is positively cor-related with the pulping yield whereas lignin is negatively correlated (Amidon,


1981; Wallis et al., 1996). The lignin monomer composition is referred to aslignin quality, as a perception of its importance for pulping in view of deligni-fication rates, chemical consumption and pulping yields. It is common know-ledge that hardwoods are easier to pulp than softwoods, and this is attributed tothe differences in lignin composition with higher reactivity in kraft pulping ofS-lignins in comparison with G-lignins (Fergus & Goring, 1969; Chang &Sarkanen, 1973; Tsutumi et al., 1995). The delignification rate does not dependon the accessibility of lignin but rather on its chemical structure (Fergus & Goring,1969) and it is directly proportional to the syringyl to guaiacyl ratio (Chang &Sarkanen, 1973). High S/G ratios were correlated with low alkali requirementsfor pulping for a series of Papua New Guinea woods including Eucalyptustereticornis and E. deglupta (Tsutumi et al., 1995). For E. globulus wood theS/G ratio was moderately correlated with pulp yield but the correlation waspoor for E. nitens wood and E. camaldulensis (Collins et al., 1990; Rodrigueset al., 1999).

The different types of wood, i.e. reaction and knotwood, will behave inpulping according to their specific chemical features; their presence and extentare important factors in establishing the pulping quality of a wood assortment.Compression wood is an inferior source of pulpwood material, with lower pulpyields and delignification degrees due to the higher content of a more condensedlignin (Timell, 1986; Lohrasebi et al., 1999; Hortling et al., 2001). The highmicrofibrillar angle of cellulose, its low crystallinity and a difficult externalfibrillation during refining also combine to give papers of low strength anddimensional stability. Tension wood, in contrast, has a high cellulose contentand it is desirable for pulp yield and pulp brightness. However, the presence ofthe gelatinous layer hinders fibre collapse and lowers fibre strength, giving thickand porous sheets (Isebrands & Parham, 1974; Parham et al., 1977). Knotwoodwill produce excessive pulp screen rejects and paper sheets with inferiorstrength, light absorption and surface properties (Allison & Graham, 1988;Sahlberg, 1995).

References

Allison, R.W. & Graham, K.L. (1988) Reject material in kraft pulp from radiata pine, Part 1: Effect ofknotwood. Appita Journal, 41, 197–202.

Amidon, T.E. (1981) Effect of the wood properties of hardwoods on kraft paper properties. TappiJournal, 64, 123–126.

Barlow, F. (1989) Rubber, gutta, and chicle, in Natural Products of Woody Plants: Chemicals Extraneousto the Lignocellulosic Cell Wall (ed. J. Rowe), Springer-Verlag, Berlin, vol. 2, pp. 1028–1050.

Beecher, C., Farnsworth, N. & Gyllenhaal, C. (1989) Pharmacologically active metabolites, in NaturalProducts of Woody Plants: Chemicals Extraneous to the Lignocellulosic Cell Wall (ed. J. Rowe),Springer-Verlag, Berlin, vol. 2, pp. 1059–1164.

Bland, D., Ho, G. & Cohen, W. (1950) Aromatic aldehydes from the oxidation of some Australian woodsand their chromatographic separation. Australian Journal of Scientific Research, Series A3, 642–648.


Chang, H. & Sarkanen, K.V. (1973) Species variation in lignin. Effect of species on the rate of kraftdelignification. Tappi Journal, 56, 132–134.

Collins, D., Pilotti, C. & Wallis, A. (1990) Correlation of chemical composition and kraft pulpingproperties of some Papua New Guinea reforestation woods. Appita Journal, 43, 193–198.

Creighton, R.H., Gibbs, R. & Hibbert, H. (1944) Studies on lignin and related compounds. LXXV.Alkaline nitrobenzene oxidation of plant materials and application to taxonomic classification.Journal of the American Chemical Society, 66, 32–37.

Dev, S. (1989) Terpenoids, in Natural Products of Woody Plants: Chemicals Extraneous to theLignocellulosic Cell Wall (ed. J. Rowe), Springer-Verlag, Berlin, vol. 2, pp. 691–807.

Dinwoodie, J.M. (1968) Failure in timber, Part 1: Microscopic changes in the cell wall structureassociated with compression failure. Journal of the Institute of Wood Science, 21, 37–53.

Emsley, A.M. & Stevens, G.C. (1994) Kinetics and mechanisms of the low-temperature degradation ofcellulose. Cellulose, 1, 26–56.

Erdtman, H., Frank, A. & Lindstedt, G. (1951) Constituents of pine heartwood. XXVII. Svensk Papper-stidning, 54, 275–281.

Faherty, K.F. & Williamson, G.T. (1995) Wood Engineering and Construction Handbook. 2nd edn,McGraw-Hill Inc., New York.

Faix, O. (1991) Classification of lignins from different botanical origins by FT-IR spectroscopy.Holzforschung, 45, 21–27.

Fengel, D. (1976) Structure of native cellulose. Holzforschung, 30, 1–6. Fengel, D. & Grosser, D. (1975) Chemische Zusammensetzung von Nadel- und Laubholzern. Holzals

Roh- und Werkstoff, 33, 32–34. Fengel, D. & Wegener, G. (1989) Wood. Chemistry, Ultrastructure, Reactions, Walter de Gruyter,

Berlin, 602pp. Fengel, D., Wegener, G. & Feckel, J. (1981) Beitrag zur Charakterisierung analytischer und technischer

Lignine.1. Chemische Analysen. Holzforschung, 35, 51–57. Fergus, B.J. & Goring, D.A. (1969) The topochemistry of delignification in kraft and neutral sulfite

pulping of birch wood. Pulp and Paper Magazine Canada, 19T, 314–320. Fergus, B.J. & Goring, D.A. (1970) The distribution of lignin in birch wood as determined by ultra-

violet microscopy. Holzforschung, 24, 118–124. Gardner, K.H. & Blackwell, J. (1974) Structure of native cellulose. Biopolymers, 13, 1975–2001. Gorham, J. (1995) The Biochemistry of the Stibenoids, Chapman & Hall, London. Harborne, J. (1989) Flavonoids, in Natural Products of Woody Plants: Chemicals Extraneous to the

Lignocellulosic Cell Wall (ed. J. Rowe), Springer-Verlag, Berlin, vol. 2, pp. 533–570. Hart, J. (1989) Role of wood exudates and extractives in protecting wood from decay, in Natural

Products of Woody Plants: Chemicals Extraneous to the Lignocellulosic Cell Wall (ed. J. Rowe),Springer-Verlag, Berlin, vol. 2, pp. 861–880.

Hausen, B. (1981) Woods Injurious to Human Health. A Manual, Walter de Gruyter, Berlin. Hillis, W.E. (1972) Properties of eucalypt woods of importance to the pulp and paper industry. Appita,

26, 113–122. Hillis, W.E. (1987) Heartwood and Tree Exudates, Springer-Verlag, Berlin. Hillis, W.E. (1991) Eucalyptus: chemistry, uses. Appita Journal, 44, 239–244. Holmberg, M. (1999) Pitch and precipitate problems, in Papermaking Chemistry (ed. L. Neimo), Tappi

Press, Finlandia, pp. 223–239. Hon, D.N.-S. & Shiraishi, N. (eds) (1991) Wood and Cellulosic Chemistry, Marcel Dekker Inc., New

York, 1020pp. Hortling, B., Tamminen, T. & Pekkala, O. (2001) Effects of delignification on residual lignin-

carbohydrate complexes in normal pine wood and pine wood enriched in compression wood.I. Kraft pulping. Nordic Pulp Paper Research Journal, 16, 219–224.

Imamura, H. (1989) Contribution of extractives to wood characteristics, in Natural Products of WoodyPlants: Chemicals Extraneous to the Lignocellulosic Cell Wall (ed. J. Rowe), Springer-Verlag,Berlin, vol. 2, pp. 843–860.


Isebrands, J.G. & Parham, R.A. (1974) Tension wood anatomy of short rotation Populus spp. beforeand after kraft pulping. Wood Science, 6, 256–265.

Jurasek, L. (1997) Molecular modelling of fibre walls, in 9th International Symposium on Wood andPulping Chemistry, June 9–12, Montreal, pp. E5-1-4.

Kai, Y. (1991) Chemistry of extractives, in Wood and Cellulosic Chemistry (eds D.N.-S. Hon &N. Shiraishi), Marcel Dekker Inc., New York, pp. 215–251.

Kawamura, I. & Bland, D.E. (1967) The lignins of Eucalyptus wood from tropical and temperate zones.Holzforschung, 3, 65–74.

Lewin, M. & Goldstein, I.S. (eds) (1991) Wood Structure and Composition, Marcel Dekker Inc., NewYork, 488pp.

Lohrasebi, H., Mabee, W.E. & Roy, D.N. (1999) Chemistry and pulping feasibility of compressionwood in black spruce. Journal of Wood Chemistry and Technology, 19, 13–25.

Miranda, I. & Pereira, H. (2001) Provenance effect on wood chemical composition and pulp yield forEucalyptus globulus Labill. Appita Journal, 54, 347–351.

Nes, W. (1989) Steroids, in Natural Products of Woody Plants: Chemicals Extraneous to the Ligno-cellulosic Cell Wall (ed. J. Rowe), Springer-Verlag, Berlin, vol. 2, pp. 808–842.

Nimz, H. (1974) Beech lignin-proposal of a constitutional scheme. Angewandte Chemie, 13, 313–321. Ona, T., Sonoda, T., Itoh, K. & Shibata, M. (1997) Relationship of lignin content, lignin monomeric

composition and hemicellulose composition in the same trunk sought by their within-tree vari-ation in Eucalyptus camaldulensis and E. globules. Holzforschung, 51, 396–404.

Parham, R.A. (1983) Wood variability, in Pulp and Paper Manufacture. Properties of Fibrous RawMaterials and Their Preparation for Pulping (eds M.J. Kocurek & C.F.B. Stevens), TechnicalSection Canadian Pulp & Paper Association, Canada, vol. 3, pp. 55–65.

Parham, R.A., Robinson, K.W. & Isebrands, J.G. (1977) Effects of tension wood on kraft paperfrom a short rotation hardwood (Populus Tristis No.1). Wood Science and Technology, 11,291–303.

Pettersen, R.C. (1984) The chemical composition of wood, in The Chemistry of Solid Wood (ed. R.M.Rowell), Advances in Chemistry Series, 207, Am. Chem. Society, Washington, DC, pp. 57–126.

Raymond, C.A., Balodis, V. & Dean, G.H. (1994) Hot water extract and pulp yield in provenances ofEucalyptus regnans. Appita Journal, 47, 159–162.

Rodrigues J., Meier, D., Faix, O. & Pereira, H. (1999) Determination of tree to tree variation insyringyl/guaiacyl ratio of Eucalyptus globulus wood lignin by analytical pyrolysis. Journal ofAnalytical and Applied Pyrolysis, 48, 121–128.

Rowell, R.M. (ed.) (1984) The Chemistry of Solid Wood, Advances in Chemistry Series, 207,Am. Chem. Society, Washington, DC, 614pp.

Sahlberg, U. (1995) Influence of knot fibres on TMP pulp. Tappi Journal, 78, 162–168. Sakakibara, A. (1980) A structural model of softwood lignin. Wood Science and Technology, 14,

89–100. Sarkanen, K.V. & Hergert, H.L. (1971) Classification and distribution, in Lignins: Occurrence,

Formation, Structure and Function (eds K.V. Sarkanen & C.H. Ludwig), Wiley Interscience,New York, pp. 43–94.

Sjostrom, E. (1981) Wood Chemistry. Fundamentals and Applications, Academic Press, New York. Swan, E. (1989) Health hazards associated with extractives, in Natural Products of Woody Plants:

Chemicals Extraneous to the Lignocellulosic Cell Wall (ed. J. Rowe), Springer-Verlag, Berlin,vol. 2, pp. 931–952.

Terashima, N., Fukushima, K., He, L.F. & Takabe, K. (1993) Comprehensive model of the lignifiedplant cell wall, in Forage Cell Wall Structure and Digestibility (eds H.G. Jung, D.R. Buxton, R.D.Hatfield & J. Ralph), Am. Soc. Agronomy Inc., Crop Sci. Soc. America Inc., Soil Sci. Soc. AmericaInc., Madison, pp. 247–270.

Timell, T.E. (1986) Compression Wood in Gymnosperms, vol. 3, Springer-Verlag, Berlin. Tobolskii, J. & Hannover, J. (1971) Genetic variation in monoterpenes of scotch pine. Forest Science,

17, 293–299.


Towers, G. & Gibbs, R. (1953) Lignin chemistry and the taxonomy of higher plants. Nature, 172, 25–26. Turner, C.H., Baloddiss, V. & Dean, G.H. (1983) Variability in pulping quality of E. globulus from

Tasmaniaaan provenances. Appita Journal, 33, 371–376. Tsoumis, G. (1991) Science and Technology of Wood: Structure, Properties, Utilization, Van Nostrand

Reinhold, New York. Tsutumi, Y., Kondo, R. & Imamura, H. (1995) The difference of reactivity between syringyl lignin and

guaiacyl lignin in alkaline systems. Holzforschung, 49, 423–428. Uprichard, J.M. (1993) Basic wood chemistry and cell wall ultrastructure, in Primary Wood Processing:

Principles and Practice (ed. J.C.F. Walker), Chapman & Hall, London, pp. 56–67. Walker, J.C.F. (ed.) (1993) Primary Wood Processing: Principles and Practice, Chapman & Hall,

London. Wallis, A.F.A., Wearne, R.H. & Wright, P.J. (1996) Analytical characteristics of plantation eucalypt

woods relating to kraft pulp yields. Appita Journal, 49, 427–428. Zabel, R.A. & Morrell, J.J. (1992) Wood Microbiology – Decay and its Preservation, Academic Press,

Inc., New York. Zobel, B.J. & van Buijtenen, J.P. (1989) Wood Variation. Its Causes and Control, Springer-Verlag,

Berlin.

4 Wood density and growthPekka Saranpää

4.1 Importance of wood density

Density is one of the major technical properties of wood. It is relatively easy todetermine and is well correlated to many other physical properties of wood,like strength, stiffness and performance in use. The shrinking and swellingbehaviour of wood is also affected by density, although the relationship is notas direct as in the case of strength properties. For example, bending strength ofNorway spruce increases linearly with increasing density and the correlation isstrong (e.g. Okstad & Kårstad, 1985; Seeling, 1999; Saranpää & Repola, 2001).Density is used also as a wood quality value, i.e. it is related to the suitabilityof wood to different end-use purposes. Structural timber needs a high densityand strength. Low-density wood may be more suitable for pulp and paper prod-ucts than for construction. Density is a useful indicator of pulpwood qualitybecause of its relationship to certain wood and fibre properties such as thick-ness of cell wall. The yield of pulp is directly related to density. One can learna great deal about the nature of a wood sample simply by determining its density.

The mass density of a substance is its mass per unit volume. The structureof wood can be simplified into solid material (cell walls) and air space (celllumens) or void volume with the result that the wood structure determines itsdensity. Wood density is also very variable and large differences in densityoccur between tree species (Fig. 4.1). Owing to the wide variation within andbetween trees, accurate average values by species are not available. The basicdensity of common wood species is 330–600 kg m−3, and hardwoods usuallyhave higher density than softwoods. Figure 4.1 shows variation of basic densityfrom the stem base to the tree top of Nordic softwood and hardwood species.

4.2 Density of cell wall material

The most abundant material of a woody stem is α-cellulose. Conifer cell wallsconsist of 40–50% cellulose, 20–35% hemicellulose and 15–35% lignin (Panshin& de Zeeuw, 1980; Sjöström, 1993; Walker, 1993). Together, these three majorcomponents determine the density of cell wall and wood tissue. Softwoodshave more lignin than hardwoods. Mean α-cellulose, hemicellulose and lignin(gravimetric lignin + acid soluble lignin) concentrations of the stem woodof Norway spruce have been reported to be 48.1 ± 3.5%, 21.2 ± 4.1% and


28.9 ± 0.8%, respectively (Anttonen et al., 2002). In comparison, Eucalyptusglobulus has 51.3% cellulose, 25.2% hemicellulose and only 21.9% lignin(Sjöström, 1993).

The unit cell of cellulose crystal lattice is composed of cellobiose residuesfrom five cellulose molecules. According to Sugiyama et al. (1991) the unitcell of cellulose Iα is P1 or triclinic, whose parameters are a-axis = 0.674 nm,b-axis = 0.593 nm, c-axis = 1.036 nm, and the angles α = 117°, β = 113° andγ = 81°. The volume of this unit can be calculated with the formula (Cullity,1978):

(4.1)

The former parameters give V=339.5Å3. The unit contains 12 carbon, 10 oxy-gen and 14 hydrogen atoms (Finkenstadt & Millane, 1998), and the total massof atoms is

M = 12 × 12.0107u + 10 × 15.9994u + 14 × 1.0079u = 318.23u1u = 1.660538 × 10−24 kg (4.2)

Thus, we get a calculated density of cellulose, ρ = M/V = 1556.5 kg m−3. The density of cell wall substance has been reported to increase when the

cellulose content and its degree of crystallinity increase (Kellogg et al., 1975).The crystallinity has been reported to increase from ring 2 to ring 10 from the

Fig. 4.1 Variation of basic density in Finnish softwoods and hardwoods from the stem base to the treetop. Redrawn from Hakkila (1998).

V = abc 1 cos2α cos2β cos2γ 2cosαcosβcosγ+–––

WOOD DENSITY AND GROWTH 89

pith and remains constant after ring 10 in Norway spruce (Andersson et al.,2003), ranging from 58 to 79%.

Density of various lignin preparations analysed by a liquid suspensionmethod was found to vary between 1333 and 1376 kg m−3 (Stamm, 1969).Density of hemicellulose for four softwood and two hardwood species deter-mined by the same method varied between 1457 and 1798kgm−3 (Beall, 1972).

If we use the calculated density for cellulose (1557 kg m−3), a reported valuefor spruce lignin preparation (1347 kg m−3; Stamm, 1969) and a reported valuefor softwood hemicellulose (1622kgm−3; Beall, 1972), and calculate the densityof cell wall of Norway spruce, which contains 48.1% cellulose, 21.2% hemi-cellulose and 28.9% lignin, we get an average density of cell wall matter of1509 kg m−3 for dry wood.

The density of the actual cell wall is believed to be similar to that of solidcell wall substance when both are measured in oven-dry condition. Micro-cavities in the cell wall increase the volume with increasing moisture content(MC) of the wood below fibre saturation point. This results in a decrease incell wall density below that of solid wood substance. The density of soft-wood cell wall tissue is slightly higher than that of hardwoods. According toKellogg and Wangaard (1969), the density of cell wall matter in softwoodsvaries between 1517 and 1529 kg m−3 and in hardwoods between 1497 and1517 kg m−3.

The density of cell wall tissue can be accurately determined with a pycno-meter (e.g. Weatherwax & Tarkow, 1968; Kellogg & Wangaard, 1969; Walker,1993). To determine the volume of a solid, the pycnometer is weighed empty,filled with liquid of known density and with the liquid plus a known weight ofthe solid. Organic liquid like silicone can be used as a displacement fluid.However, organic liquids do not swell and penetrate the cell wall and there isthe possibility that some submicroscopic pores exit within the cell wall, whichremain inaccessible, and non-swelling fluids may give a too high value of thevolume of the cell wall material and give a too low value of density. A value of1465 kg m−3 has been reported for Picea sitchensis using silicone as the dis-placement fluid (Weatherwax & Tarkow, 1968). However, when water wasused, the density of the cell wall was ca. 5% greater than the previous value(1546 kg m−3; Weatherwax & Tarkow, 1968). Water is absorbed within thecell wall and therefore its density is slightly reduced. The volume of the cellwall material is underestimated and density slightly overestimated. Thus, anaverage cell wall density of 1500 kg m−3 could be used for oven-dried woodtissue (Walker, 1993).

Density of cell wall tissue has been reported to be quite constant withina stem and not affected by growth rate. Voids in dry cell walls constitute a verysmall volume varying between 1.62 and 3.27%. Differences in void volumebetween tree species or the effect of growth rate on void volume has beenreported to be insignificant (Tsoumis & Passialis, 1977).


4.3 Determination of density

Basic density is widely used in scientific literature dealing with wood. It isdetermined as oven-dry (0% MC) mass per green volume (kg m−3), i.e. it tellshow much dry material is contained in a cubic metre of fresh wood. The termbasic is used since both green volume and oven-dry mass are nearly constant,and reproducible measurements can be obtained with wood.

The term specific gravity is also used, especially in American literature. It isalways calculated using oven-dry weight or mass. Volume can be determined atany moisture content, but the moisture content must be specified. Specificgravity is the ratio of density of a substance to the density of pure water (4°C).Like all ratios, specific gravity is a dimensionless quantity. In the metric system,the density of water can be considered to be equal to 1 g cm−3. Thus, the spe-cific gravity of a substance equals numerically its density, assuming the sameprinciple of determination. For example, oven-dry wood with a specific gravityof 0.5 (SG at 0% MC) has 0.5g of dry wood substance per cm3 or 500kgm−3.

Weight density, or weight of wood per unit volume, is calculated on thebasis of both the weight and the volume of the piece taken at the same moisturecontent (Panshin & de Zeeuw, 1980). For example, mechanical tests, like thosedetermining modulus of elasticity of wood (MOE), are usually done at 65%relative humidity (RH), corresponding to a moisture content of the wood of ca.12%. The density is usually determined under the same conditions. Air-dry densityis also sometimes used, especially for sawn goods or plywood. Weight and volumeare determined at the moisture content of the wood at the time of experiment,or weight can also be determined on oven-dry basis. Such values can be used forrelative comparisons within an experiment, but care must be taken whencomparing such data with other experiments.

One must consider that wood is a hygroscopic material, which swells andshrinks with changes in moisture content. As moisture content increases fromthe dry condition up to the fibre saturation point (the moisture content at whichthe cell walls are fully saturated and the cell cavities are free of water), theweight increases and, as a result of swelling, so does the volume. If the weightof a wood sample is determined as oven-dry, the increasing moisture contentresults in a volumetric swelling and a decrease in density. Since swelling doesnot take place with increasing moisture content above the fibre saturationpoint, the density remains constant at higher moisture content, if the measure-ment is based on moisture free weight.

4.3.1 Water displacement method

The green volume of a piece of wood is commonly determined by waterdisplacement according to the principle of Archimedes. The samples should befirst soaked in water to ensure that water is not taken up during the immersion


process. After determining the green volume, the samples are oven-dried at103°C. The drying time depends on the size of the samples and also on thecapacity of the oven. It is a good method for large stem disks but it can also beapplied to incremental borings, even down to a sample volume as small as0.0075 cm3 (Olesen, 1971). With a sample of 1000 g, a standard error of lessthan ±0.001 specific gravity unit has been reported (Megraw, 1985).

Wood cells are formed in an aqueous environment and exist in the livingtree in the green or maximum swollen condition. Unfortunately, once a piece ofwood has been dried, the original green volume is not recovered again simplyby re-soaking it. This is due to the incomplete rehydration of the water sorptionsites in the cellulose once drying has taken place (Stamm, 1964). Thus, ifre-swollen volume is used to calculate basic density, it is important to state it.However, the difference between re-swollen and original green volume forloblolly pine (Pinus taeda) has been reported to be small (<1%) but also veryvariable (Megraw, 1985).

Extractive content increases during heartwood formation. Siberian larch(Larix sibirica) may contain up to 20% of extractives (mainly arabinogalactan)in the outer heartwood and it is necessary to treat the samples with hot water,to remove the extractives, before determining the basic density. Norway sprucenormally has a low extractive content (2–4% dry weight) and does not normallyrequire extraction prior to the determination of the basic density by the waterdisplacement method. Scots pine has slightly larger amounts of extractives,especially in the heartwood. When using small samples, such as incrementcorings, care must be taken that the sample is not resin-soaked which maycause a substantial error in basic density values.

4.3.2 X-ray densitometry

X-rays have been used for a long time to study density variations in wood (e.g.Harris & Polge, 1967; Polge & Nicholls, 1972; Hoang & McKimmy, 1988;Koubaa et al., 2002). X-ray densitometry measures average density and severalwithin-ring density characteristics in wood cross sections. These include averageearlywood and latewood density, minimum earlywood density, maximumlatewood density, earlywood and latewood width, and latewood percentage.A thin wood sample, for example 5mm wide and 5mm thick, extending from thepith to the bark is removed, placed on a film and X-rayed parallel to grain. Toensure that the X-rays are parallel and normal to the specimen, the distancebetween the sample and the source of radiation must be relatively large, i.e.over 2 m. Alternatively, the radiation can be collimated through a narrow slit(Echols, 1973).

The X-ray method can be calibrated using standards made by pressingwood into different densities, and exposing them together with the unknownsample (Sauvala, 1979). The films produced are scanned to get a density


profile (Fig. 4.2). The method can be applied to green, air-dry or dry wood.Direct scanning microdensitometers (e.g. Silviscan – Evans etal., 1997; Itrax – CoxAnalytical Systems) are also available for X-ray densitometry. X-ray analysisis an accurate and efficient technique for determining density. A series of 20independent X-ray evaluations of the same ring have been reported to producea standard deviation of about ± 0.005 specific gravity units (Megraw, 1985).

4.4 What causes variation in density?

If we assume that the density of cell wall is constant, the density of wood ismainly determined by the amount of cell wall substance, i.e. by the size ofwood cells or their lumens and by the thickness of cell wall. Density is wellcorrelated with the cell wall area and cell wall thickness (e.g. Diaz-Váz et al.,1975; Quirk, 1984). Latewood percentage has been proved to be a good pre-dictor of wood density in conifers (e.g. de Kort et al., 1991; Wimmer, 1995). Ina normal range of latewood percentage (19–50%), earlywood wall thickness,either radial or tangential, has a high potential to alter wood density. However,

Fig. 4.2 Density profile of Norway spruce (Picea abies) from the pith to the bark. The trees have beenfertilised annually after age 12 (arrow) which has resulted in wider growth rings and decreasingdensity (courtesy of Kari Sauvala).


radial diameter and wall thickness of latewood tracheids have the most influenceon wood density (Wimmer, 1995). Latewood percentage can explain up to 60%of the density variation in Douglas fir (Pseudotsuga menziesii; de Kort et al.,1991). Growth rate, tree age and heredity are the major factors determining thedensity of a stem (e.g. Lewark, 1982; Zobel & Jett, 1995; Lindström, 1996;Haapanen et al., 1997; Beets et al., 2001; Wodzicki, 2001).

4.4.1 Within growth ring

Most tree species (excluding Araucaria sp. and diffuse-porous hardwoods likepoplar and birch) show a large variation of basic density across one annual ring.According to Megraw (1985), the greatest variability of density occurs withineach annual ring in softwoods. This is due to seasonal climatic changes and theformation of latewood. Figure 4.3 shows the gradual change of thin-walledearlywood tracheids into thick-walled latewood tracheids in Norway spruce(Picea abies). According to Mork’s (1928) definition, latewood in softwoodsincludes tracheids in which the common wall between two cells is exactly halfor over half the radial width of lumen. In other words, the double wall thickness(measured from lumen to lumen) is equal to or greater than the half-width ofthe cell lumen. Although this definition was originally used for spruce, it canbe applied to other softwoods also. However, there are different interpretationsof Mork’s original rule. Some have interpreted as saying that ‘latewoodtracheids are those in which the width of their common cell wall in the radialdirection is equal to or greater than the width of either cell cavity’. This rulemay be more useful for species which have relative thick cell walls throughoutthe growth ring (Denne, 1989).

Figure 4.4 shows an example of the formation of one growth ring in a 42-year-old Norway spruce in southern Finland (61°21′ N, 24°59′ E, alt. 60m a.s.l.). Budbreak took place on 8 June 1994 and cambial cell divisions at breast height tookplace soon after that. The rather late bud break in that spring (1994) was due toa cold period at the end of May and the beginning of June as compared to the long-term average. The length of the growing season was 133 days. The growth ringhad on average 60 tracheids in one radial row. The average diameter of earlywoodtracheid lumens was 30μm and the average double cell wall thickness was 6μm.Latewood formation (according to Mork’s definition) occurred during the firstweek of August and there were only ca. 15 latewood tracheids present. The changefrom earlywood to latewood was rather gradual and the lumen diameter startedto decrease in the middle of the growth ring, i.e. in the middle of July (Fig. 4.4).False rings, thin latewood bands followed by further earlywood production occurrandomly in softwoods. These are generally associated with temporary interrup-tion of terminal height growth or terminal needle elongation due to moisturestress, low temperature or other short-term environmental factors (Larson, 1969).Night frosts during early summer cause false rings at high latitudes.


Owing to the difference between earlywood and latewood, the densityvariation within one growth ring is usually far larger than the variation withinthe stem. Harris (1969) has reported latewood density as high as 870kgm−3 andearlywood density as low as 170kgm−3 for Douglas fir (Pseudostuga menziesii).Norway spruce grown in northern conditions shows a smaller differencebetween earlywood and latewood (Fig. 4.5). Black spruce (Picea mariana) hasan average earlywood density of 376kgm−3 and a latewood density of 569kgm−3

(Zhang & Morgenstern, 1995). Norway spruce also normally shows a gradualtransition from earlywood to latewood (Fig. 4.3), but Scots pine (Pinus sylvestris)

Fig. 4.3 A cross section of Norway spruce (Picea abies). Juvenile wood has wide annual rings, smalldiameter tracheids and a low percentage of latewood (left). Mature wood has narrow growth rings and ahigh percentage of latewood (right). Transition from earlywood to latewood is gradual. Scale bar is 1 mm.


Fig. 4.4 Formation of a growth ring in 1994 in a 42-year-old Norway spruce stem in southern Finland.Cell lumen diameter (•) decreases and cell wall double thickness (�) increases from the earlywood to thelatewood. Cambial activity started at the beginning of June soon after bud break, and latewoodformation (according to Mork’s definition) occurred during the first week of August and there were onlyca. 15 latewood tracheids present (arrow). The change from earlywood to latewood is rather gradual andthe lumen diameter started to decrease in the middle of the growth ring, i.e. in the middle of July.

Fig. 4.5 Variation of earlywood and latewood density from the pith to the bark at breast height inNorway spruce based on X-ray densitometric analysis. Unextracted weight density was determined at12% moisture content (ρ12).


and especially Douglas fir (Pseudostuga menziesii) and larch (Larix sp.) havean abrupt change from earlywood to latewood (Fig. 4.6).

Average ring density at breast height has been found to be related to ringwidth, and even more closely to the proportion of latewood in Norway spruce(e.g. Mäkinen et al., 2002a). The steady increase of ring density in maturewood of black spruce (Picea mariana) is also due to increase in proportion oflatewood (Koubaa et al., 2002). The percentage of latewood is normally low inthe first growth rings around the pith and increases towards the bark. In themature wood of Norway spruce, it increases from ca. 20% in ring 20 from thepith to 35% in ring 110 (Hakkila, 1968). In black spruce, the average increasein latewood proportion from the transition zone (ring 8–12 from the pith) tomature wood (ring nos. 18–25) was 41.0% compared to an average increase inearlywood density of 4.5% and an average decrease in latewood density of2.0% (Koubaa et al., 2002).

Olesen (1977) has reported that climatic differences do not change the orderof density levels of the innermost eight annual rings in Norway spruce. Thisis due to the large difference in the number of tracheids per unit area betweenthe annual rings, especially in the innermost part of the juvenile wood. Thedifference in basic density levels between the innermost annual rings is mainlya result of a rapid decrease in the number of tracheids per cross-sectional area.

Fig. 4.6 Cross sections of Siberian larch (Larix sibirica, left) and Douglas fir (Pseudotsuga menziesii,right) showing an abrupt change from earlywood into latewood and large differences in growth ringwidth due to differences in annual climatic conditions.


Mäkinen and co-workers (2002a) have also indicated a close relationshipbetween wood density and fibre properties (cell diameter, cell wall thickness,latewood proportion) in Norway spruce. This relationship is in fact to beexpected because the amount of cell wall material is a strong determinant of thewood density of spruce, which has a low extractive content, even in the heart-wood (Hakkila, 1968). Furthermore, the density of the cell wall material itself issupposed to vary only slightly compared to other sources of wood densityvariation (Kellogg & Wangaard, 1969; Skaar, 1972). According to Hannrupand co-workers (2001), genetic control is strong for both radial and tangentiallumen diameter of earlywood tracheids, moderate for latewood proportion butlow for cell wall thickness in Scots pine (Pinus sylvestris). They found thatearlywood lumen diameter (both radial and tangential) and latewood proportionhave the strongest correlation with wood density in mature wood. With multipleregression analysis, they could explain up to 73% of the variation in basic density.

Latewood tracheids form when secondary wall thickening is favoured overradial expansion (Larson, 1969). The increase in secondary wall thickening, whichleads to the formation of latewood, occurs when the requirements of the majormetabolic sinks within the crown have been met and the current-year needlesbegin exporting photosynthates to other parts of the tree, primarily the stem.This phase of wall thickening begins at the stem base and progresses upwardsas the season advances (Larson, 1969).

4.4.2 Within a tree

Juvenile wood is the wood first laid down by the cambium near the centre ofthe tree (Fig. 4.7). There is no absolute shift from juvenile to mature woodwithin one year but the change occurs over several years. Nearly all woodproperties are very variable from ring to ring within the juvenile wood zoneand much more constant within the mature zone. A tree produces juvenilewood at every age. Since the cambium is a continuous sheath round the stem,it produces juvenile wood near the top of a large stem, and mature wood at itsbase (Thomas, 1984; Zobel & van Buijtenen, 1989). According to Larson andco-workers (2001), the term juvenile wood is an unfortunate misnomer. Theystate that true juvenile wood is formed only during the first three years ofgrowth. Thereafter, similar but not identical wood is produced in the centralcore of wood at all height levels in the stem. This wood has been referred tomore appropriately as corewood. It has also been referred to as crown-formedwood (Larson, 1969; Larson et al., 2001) because it is produced either in theliving crown or in its proximity, close to the physiological processes emanatingfrom the living crown.

The presence of juvenile wood has been suggested as being one of the greatestcauses of variation in wood properties of conifers. Cell length, diameter and cellwall thickness increase rapidly from the pith outwards (Thomas, 1984; Zobel


& Sprague, 1998; Saranpää et al., 2000). Large microfibril angles, short fibreswith thin walls and low percentage of latewood in the annual rings close to thepith may all contribute to the lower quality of juvenile wood products. Juvenilewood is also characterised by lower density but higher longitudinal shrinkagethan mature wood (Boutelje, 1968; Pearson & Gilmore, 1971; Bendtsen, 1978;Thomas, 1984; Zobel & van Buijtenen, 1989; Zobel & Sprague, 1998). A highamount of compression wood has been reported to be present in the juvenilewood of fast grown trees (Bendtsen, 1978). However, it has been stated thatcompression wood has little influence on juvenile wood properties (Thomas,1984).

In many tree species, especially in pines, density increases gradually fromthe pith to the bark. In Norway spruce the density is much more variable(Saranpää, 1994). Figure 4.8 shows the variation of basic density from the pithto the bark at breast height in Norway spruce in southern Finland (60–64° N).Sample trees were selected both from natural stands and from plantations withvarious spacings. Altogether, 240 stems from 48 different stands were studied.Sample disks were sawn at 1, 2, 4, 8, 12 and 16 m heights in the stem. From

Juvenile wood

Heartwood

Sapwood

Fig. 4.7 Juvenile wood, heartwood and sapwood in a conifer stem. Juvenile wood (i.e. corewood,crown wood) forms a cylinder around the pith. In Norway spruce, it is ca. 10 growth rings wide.Heartwood formation increases density of juvenile wood in large stems due to the synthesis ofadditional resins and extractives. The ascent of sap takes place in the sapwood.

Fig

. 4.8

Var

iatio

n of

bas

ic d

ensi

ty f

rom

the

pith

to th

e ba

rk a

t bre

ast h

eigh

t in

240

stem

s of

Nor

way

spr

uce

from

48

diff

eren

t sta

nds

in s

outh

ern

Finl

and

(lat

. 60–

64°)

.B

asic

den

sity

was

det

erm

ined

by

the

wat

er d

ispl

acem

ent

met

hod

appl

ied

to 2

–3cm

wid

e an

d th

ick

woo

d sa

mpl

es s

awn

from

the

pit

h to

the

bar

k. T

he a

vera

geba

sic

dens

ity

of th

e di

sks

at b

reas

t hei

ght w

as 3

80kg

m−3

. The

re w

as n

o si

gnif

ican

t dif

fere

nce

betw

een

sout

h an

d no

rth

side

s of

the

stem

.


each disk, a strip ca. 30 mm wide was cut through the pith. These strips werefurther cut into 30-mm wide samples in the radial direction in order to studythe variation of density from the pith to the bark. The average basic density atbreast height was 380 kg m−3, which corresponds to earlier reports (Hakkila,1966, 1968). The variation of basic density from the pith to the bark wassurprisingly large even though samples with any defect (e.g. compressionwood) were omitted.

Figure 4.9 shows a model of basic density variation from the pith to the barkat various heights in the stem of Norway spruce. The basic density is slightlyhigher close to the pith than further out. Similar horizontal variation of densityhas also been observed in P. glauca (Taylor et al., 1982). The decrease ofdensity and ring width from ring 3 towards the outer rings has been observedto occur near the pith in P. abies (Olesen, 1977; Lewark, 1982; Petty etal., 1990),P. sitchensis (Jeffers, 1959) and P. glauca (Taylor etal., 1982). Danborg (1994)used the minimum wood density (D5%) as the demarcation between corewoodand mature wood. The minimum was reached at the rings 8–10 from the pith.The high density near the pith in P. mariana is mainly due to a higher earlywooddensity and a higher latewood proportion. The subsequent decrease in ringdensity was found to be due to a decrease in both earlywood density and late-wood proportion (Koubaa et al., 2002). This variation of density with tree agemay cause problems with dimensional stability. To avoid this, the juvenilephase should be kept short with dense stands and slow initial growth Juvenile

Fig. 4.9 Variation of basic density from the pith to the bark in Norway spruce. Basic density alsovaried along the stem. The lowest density occurred at 3–6 m height in the stem.


wood and mature wood should not be mixed when comparing effects ofenvironmental factors on growth rate and wood density.

4.4.3 Between sites

Differences between site fertility and geographic location (temperature sum)are the major sources of the variation between different stands even when thegenetic variation is reduced, i.e. by using clonal material (Megraw, 1985). Humanimpact on forests is also considerable. Man-made differences like forest manage-ment including thinning, pruning and fertilisation increase or decrease variationbetween stems.

The variations between individual stems, and between juvenile and maturewood are the major sources of variation in Norway spruce. Figure 4.10 shows thetotal variation of basic density of 240 stems of Norway spruce from 48 stands,divided into three components: between-stands, between-stems and within-tree(samples from the pith to the bark and at the different heights in the stem). Corres-ponding variance components of the unexplained variation of the three selectedcriteria (or explanatory variables) having different fixed parts are also shown inFig. 4.10. The fixed variable of the first model, position in the stem, describes dif-ferences in basic density as a function of distance from the pith to the bark andfrom the stem base to the tree top. The fixed part of the second model, growth ring

Fig. 4.10 The total variation of basic density of Norway spruce is divided into three components: between-stands, between-stems and within-tree (samples from the pith to the bark and at the different heights in thestem). Also shown are the three components and the corresponding variance components of the unexplainedvariation of the three models having different fixed part (position in the stem, growth ring width andGRW + position). The variation between stems and within stems was the major source of variation. GRWcould be used to explain differences between stands and between stems. However, a significant amount ofvariation remained unexplained. Altogether 240 stems from 48 stands.


width (GRW), describes the effect of GRW on wood density. In the third model,GRW+position, i.e. the effect of both stem position and growth ring width isincluded in the fixed part. The variation between stems and within stems was themajor source of variation. GRW could be used to explain differences betweenstands and between stems. However, a significant amount of variation remainedunexplained. Wilhelmsson and co-workers (2002) explained 50% of the totalvariation in the basic density of Norway spruce by using diameter, the number ofgrowth rings and the temperature sum as fixed explanatory variables. Randomamong-tree variance accounted for most of the remaining variation of basicdensity and latewood.

4.5 Is there a correlation between density and growth rate?

Effective timber production has been one of the main topics of research. Thetrend in silviculture in recent decades has been towards enhanced growth andshorter rotations. However, intensive timber production may cause changes inthe anatomical and technical properties of the wood, thus reducing the suit-ability of the wood for many products. As a result, concern has arisen aboutchanges in wood properties caused by intensified silvicultural regimes (Ballard& Long, 1988; Barbour & Kellogg, 1990).

In general, intensive forest management, i.e. regular thinning and fertilisa-tion, results in rapid growth but low density, and thus lower mechanicalstrength of wood. GRW is a widely used measure of overall tree health, grow-ing conditions and even wood quality. The effect of growth rate on wood prop-erties and especially on wood density has been studied intensively. Zobel andvan Buijtenen (1989) and Zobel and Jett (1995) have summarised hundreds ofpapers, and other reviews cover several studies of the effect of growth rate onwood density (Spurr & Hsiung, 1954; Megraw, 1985). The correlation betweengrowth rate and density is of great importance. If increased growth rate results inlow-density wood, it also means poor wood quality which limits the suitabilityof raw material for high quality products. On the other hand, it may also bepossible to select trees with both high growth rate and high wood density forbreeding.

For example, Zobel and van Buijtenen (1989) have listed 59 referencesabout the relationship between growth rate and wood density in hard pines(e.g. Pinus taeda, P. radiata, P. elliottii, P. sylvestris, P. contorta). Of thesestudies, 35 showed no relationship between growth rate and wood density,9 showed only a small correlation and 11 showed a significant reductionin wood density. Only four studies showed a higher density for the fastestgrowing trees. However, it is commonly accepted that rapid growth results inlow-density wood. All combinations were also found among other coniferousspecies (18 species listed) in addition to hard pines. For example, the genus


Picea showed usually a negative correlation (Zobel & van Buijetenen, 1989).Zobel and Jett (1995) listed 38 more studies and separated the effect of growthrate on wood density into four categories:

1. Most of the conifers with dense wood, especially hard pines, show littleor no meaningful relationship.

2. A negative relationship between growth rate and wood density has beenreported in several genera such as spruce (Picea spp.) and fir (Abies spp.).

3. Contradictory reports have been published about the effect of growthrate on Douglas fir (Pseudotsuga menziesii).

4. There are contradictory reports about the relationship between growthrate and wood properties in the hardwoods. Usually, the diffuse-poroushardwoods (i.e. Populus) show little or no relationship between growthrate and wood density.

Zhang (1994, 1995) has studied a number of softwoods and hardwoods andalso divided them into four categories:

1. Softwoods with gradual transition from earlywood to latewood. Piceaspp. belong to this group that shows a significant decrease of density withincreasing growth rate. This will be supported by the literature reviewwhich follows.

2. Softwoods with abrupt transition from earlywood to latewood (i.e. Larixspp. and southern pines or hard pines). Growth rate has little influence ondensity. This has been also shown for loblolly pine (Pinus taeda) byMegraw (1985).

3. Ring-porous species like oak. Growth rate has little influence ondensity, and mechanical properties may increase with increasing growthrate.

4. Diffuse-porous species, i.e. Betula spp., Populus spp. Growth rate haslittle influence on density and mechanical properties. This has beenalso shown for silver birch and downy birch (Betula pendula andB. pubescens) by Hakkila (1966).

Growth rate has a negative correlation with basic density in Norway spruce.This has been demonstrated by numerous Nordic and European researchers(Klem, 1952, 1972, 1974; Nylinder, 1953; Seibt, 1963; Hakkila, 1966; Hakkila& Uusvaara, 1968; Saikku, 1975a,b; Olesen, 1976, 1977; Velling, 1976, 1980;Madsen et al., 1978; Petty et al., 1990; Johansson, 1993; Lindström, 1996;Dutilleul et al., 1998; Herman et al., 1998; Flæte & Kucera, 1999; Pape, 1999;Wilhelmsson et al., 2002). The correlation between GRW and basic density isnon-linear (Fig. 4.11), and the decrease of basic density is faster when the ringwidth decreases to 2–3mm and slower in wider ring widths. Similar trends havebeen reported by many authors (Olesen, 1982; Flæte & Kucera, 1999). Linearcorrelation between ring width and basic density has also been reported (e.g.


Hakkila, 1966) but such models explain much less of the variation comparedto non-linear, logarithmic models. Flæte and Kucera (1999) explained 40% ofthe variation of basic density by a non-linear regression whereas linear modelsexplained less than 10% of the variation. They studied 90 stems from six prov-enances planted in southern Norway (59° N) and also found that the relation-ship between GRW and basic density was slightly different for differentprovenances.

It is very important to establish the difference between juvenile and maturewood when studying the effect of growth rate on basic density. Wide growth ringsand low density are associated with juvenile wood close to the pith and narrowgrowth rings and high density are typical for mature wood. Thus, a negativecorrelation between ring width and density is obvious if both juvenile and maturewood are included in the analysis. Figure 4.11 shows the correlation betweenbasic density and GRW in Norway spruce both for juvenile and mature woodand also at two heights in the stem (breast height and at 8 m). The trend is quitesimilar for both juvenile and mature wood, except that mature wood shows ahigher density level than juvenile wood. The basic density of juvenile woodslightly increases from breast height to the stem height of 8 m. Figure 4.12shows the variation of basic density from the pith to the bark in a fast-grownNorway spruce clone. The stems were cutting-derived clones of a fast-grownmother tree, which was a hybrid between a tree from Pieksämäki, centralFinland (62°18′ N, 27°10′ E, alt. 100m a.s.l.) and a tree from Schilbach, Germany(50°25′ N, 12°30′ E, alt. 600 m a.s.l.). The trees were growing on fertile

Fig. 4.11 Correlation between growth ring width and basic density in Norway spruce both in juvenilewood and in mature wood and at two heights in the stem (breast height 1.3 m and 8 m).


abandoned farmland in Nurmijärvi, southern Finland (60°30′ N, 24°42′ E, alt.100 m a.s.l.) and the volume growth was almost threefold compared to treesgrown in normal conditions. In contrast to the normal trend, the basic densitydecreases from the pith to the bark and is below 300 kg m−3 in the outer rings(Fig. 4.12). The average ring width was 5.9 mm and the average basic densityat breast height, 310 kg m−3, corresponds to the basic density predicted by themodel in Fig. 4.11. Such a low density results in almost 30% lower stiffness andmeans that almost 40% more wood volume is needed for pulping compared toslow-grown wood.

Madsen and co-workers (1978) studied the variation of basic density ina thinning experiment of a 32-year-old stand of Norway spruce in Denmark. Theyexcluded juvenile wood and showed that there was a very similar trend of decreas-ing basic density with increasing ring width in mature wood from ca. 500kgm−3

at ring width of 0.5mm to 300kgm−3 at ring width of 6mm. Olesen (1977, 1982)has reported a similar relationship. The relationship between basic density andring width (basic density level), developed by Olesen (1976) is given by:

(4.3)

where R = basic density, RW = ring width, and a, b and c are coefficients.

Fig. 4.12 Variation of basic density from the pith to the bark at breast height in 15 fast-grown stems ofNorway spruce in southern Finland. Basic density was determined using the water displacementmethod applied to 2–3 cm wide and thick wood samples sawn from the pith to the bark. The averagebasic density at breast height was 310 kg m−3 and the average growth ring width was 5.9 mm.

Ra b+

RW c+( )-----------------------=


The basic density level has been applied by many authors for Norwayspruce (e.g. Madsen et al., 1978; Moltesen et al., 1985; Danborg, 1994; Pape,1999). Basic density level has not been found to be influenced by thinning grade(Madsen et al., 1978; Moltesen et al., 1985; Johansson, 1993; Pape, 1999).However, the block which had the highest site index had also the lowest basicdensity level (Moltesen et al., 1985). A significant correlation between the claycontent of the soil and the basic density level has been reported (Madsen et al.,1978). The authors concluded that the basic density level is negativelycorrelated with the water availability for the tree. However, it may be alsorelated to nutrient availability.

Pape (1999) studied the effect of five different thinning regimes on the basicdensity of Norway spruce in southern Sweden (56–59° N). The age of the stemswas 39–52 years and the first thinning was made when the trees were 23–27-years-old. After a heavy thinning which removed 70% of the stand basal area, theaverage basic density of the growth rings following the thinning decreased by8%. Ring width increased by 41%. A light thinning which removed 20% ofthe basal area had no direct effect on basic density or ring width. A selectiveremoval of dominant trees, thinning from above, resulted in an overall increase inthe basic density of trees in the stand due to the lower growth rates of theremaining trees compared with those removed. However, subsequent thinningshave a strong influence on the wood formed by trees released from theirdominating competitors (Pape, 1999).

According to Larson (1969), manipulation of stand density is the mostpowerful method of regulating wood yield and wood density. As an extremecase, open-grown trees have a long crown and short, clear bole, whereas stand-grown trees have a short crown and long, clear bole. Open-grown trees willhave a high proportion of juvenile wood with low density in the upper crownand a low percentage of latewood in the lower bole. In contrast, stand-growntrees will have a restricted zone of juvenile wood in the upper crown and a highpercentage of latewood and thus dense wood in the lower bole. A thinned pineor spruce never regains its lost lower branches and never reverts to the form ofan open-grown tree. A tree with a long, clear bole, or a mature tree, will continueto form high-quality wood in spite of a relative increase in earlywood productionfollowing a heavy thinning (Larson, 1969).

Zhang and co-workers (Zhang & Morgenstern, 1995; Zhang et al., 1996;Zhang, 1998) studied intensively the variation of growth rate and wood densityin 40 black spruce (Picea mariana) families grown in New Brunswick, Canadaand 86 provenances in Mont-Laurier (46°36′ N, 75°48′ W, alt. 300 m a.s.l.;Koubaa et al., 2000). They found a negative relationship between wood densityand growth rate, but the relationship varied with genotype and environment(Zhang et al., 1996). Correlations between ring density and most intra-ringcharacteristics (i.e. earlywood and latewood width and density) weakened withincreasing age. This implies that fast growth of black spruce may have a less


negative effect on wood density in older trees (Zhang, 1998; Koubaa etal., 2000).Average wood density, earlywood density and latewood density were foundto be under a strong genetic control in black spruce (Zhang & Morgenstern,1995). Basic density has been reported to be under high degree of geneticcontrol, and a highly heritable trait in many softwood and hardwood species(Megraw, 1985; Zobel & Jett, 1995; Rozenberg & Cahalan, 1997).

Extensive research on the effect of spacing on growth rate and wood pro-perties has been carried out on Sitka spruce (Picea sitchensis). Brazier (1970)studied 15 Sitka spruce trees from four plots, which were planted at differentspacings and thinned to maintain the differences in growth rates. The treeswere 31 years old when examined. He showed a decreasing linear relationshipbetween density and ring width in mature wood. He also concluded that therewas an increase in earlywood without a corresponding increase in latewood,and thus a greater proportion of earlywood. Also, the average earlywooddensity was reduced (Brazier, 1970). Petty and co-workers (1990) studied thevariation of basic density in a 48-year-old Sitka and Norway spruce plantedat two spacings. Altogether, they investigated 12 Sitka and 20 Norway sprucetrees. Both species showed a pattern of radial variation of density: high nearthe pith, falling to a minimum and then a gradual increase. Density decreasedwith increasing GRW, strongly in Sitka but weakly in Norway spruce.Mitchell and Denne (1997) showed that on sites having high growth rate,density of Sitka spruce wood decreased more rapidly across the juvenile wood,down to a lower minimum value, than on sites with a slower growth rate. Theystudied a total of 24 trees from seven stands at three sites in mid and northWales.

Simpson and Denne (1997) studied the effect of original spacings andcrown dimensions in a 52-year-old unthinned spacing trial of Sitka spruce.They found that the correlation between ring width and density increased upthe tree. The density of the outer rings of the trees originally at wider spacing wassignificantly higher than that of trees originally at narrower spacing, suggestingthe former were under greater competition at the time of felling. They came to theimportant conclusion that it is crucial to consider the history of spacing aroundthe individual sample trees when quantifying the influence of silviculturalmanagement on tree growth and wood density. Kärkkäinen (1984) observedthat the prediction of basic density with the help of growth rate and some othertree characteristics could be improved if the social status of the tree is takeninto account. He concluded that dominant trees had a higher basic density thanwould be expected on their growth rate alone.

Hakkila (1968) made an intensive study on the variation of basic density ofNorway spruce and Scots pine (Pinus sylvestris) in Finland. The material camefrom 85 sites, and 100 trees (increment corings) from each site. Average diameterof the stems was 12cm. The basic density of Scots pine pulpwood reached itshighest values between 64 and 66° N and decreased in the northern part (>66° N).


Wilhelmsson and co-workers (2002) investigated 252 stems of Norway spruceoriginating from 42 stands and 120 Scots pine stems originating from 20 standsin order to model the variation of wood properties. The stands covered awide variation in climate (56.6–65.8° N, 12.5–21.5° E, alt. 60–440 m a.s.l.).They could explain 59% of the variation of basic density in Scots pine by usingdiameter, the number of annual rings (=growth rate) and temperature sum as fixedexplanatory variables. Figure 4.13 shows the variation of basic density and late-wood percentage in a thinned and in an unthinned 100-year-old Scots pine stemaccording to simulation. Scots pine shows a gradual increase of basic density from

Height (m) Basic density (kg m–3)

latewood (%)

DBH 17.9 cm DBH 20.8 cm

DBH 17.9 cm DBH 20.8 cm

22.7

20.4

18.1

15.9

13.6

11.3

9.1

6.8

4.5

2.31.3

Height (m)

22.7

20.4

18.1

15.9

13.6

11.3

9.1

6.8

4.5

2.31.3

Height (m)

23.6

21.2

18.9

16.5

14.2

11.8

9.4

7.1

4.7

2.41.3

23.6

21.2

18.9

16.5

14.2

11.8

9.4

7.1

4.7

2.41.3

325–350350–375375–400400–425425–450450–475

0–55–10

10–1515–2020–25

Height (m)

Fig. 4.13 Variation of basic density and latewood percentage in a Scots pine stem. Examples ofsimulated distributions of radial growth (every 5th ring), basic density (above) and latewoodpercentage (below) of 100-year-old Scots pine stems in unthinned (left) and thinned stand (right),respectively. Courtesy of Heli Peltola and Veli-Pekka Ikonen.


the pith to the bark and a gradual decrease of basic density from the stem base tothe tree top (Fig. 4.1). Basic density is much more variable within Norway sprucestem (Figs 4.1 and 4.9).

Wodzicki (2001) investigated the effect of thinning in a 33-year-old stand ofScots pine in central Poland. Effects of thinning were associated with increase indaily rate of tracheid formation from the cambial meristematic zone, because theinitiation or termination and seasonal length of cambial activity, as well as theradial diameter or cell wall thickness in earlywood and latewood were not signifi-cantly affected. Thus, thinning affects wood structure considerably only by asingle effect upon the efficiency of cambial cell division.

4.5.1 Effect of fertilisation on growth rate and wood density

Fertilisation is a silvicultural treatment that can be used to control tree growthand wood density. The purpose of fertilisation is to improve the availability ofnutrients for forest trees and increase wood production, especially in poor sites.In Scandinavian countries and in Finland, nitrogen has been found to be themost effective nutrient in increasing tree growth (e.g. Viro, 1972; Albrektsonetal., 1977; Tamm, 1991). Also, increased nitrogen deposition may be one causeof increased tree growth observed throughout Europe. The effects of nitrogenfertilisation on tree growth and wood density have been reported to be ca. 30%in Finland (Gustavsen & Lipas, 1975; Puro, 1977).

The effects of nitrogen fertilisation on the growth and wood density ofNorway spruce and Scots pine (Pinus sylvestris) are nowadays reasonably wellunderstood (e.g. Klem, 1972, 1974; Viro, 1972; Saikku, 1975a,b; Kukkola &Saramäki, 1983; Kenk & Fischer, 1988; Evers & Gussone, 1991; Lindström,1996). As the growth rate increases, earlywood width increases with a lesspronounced or no increase in latewood width. The earlywood of Norwayspruce usually has a lower density than that of latewood, and hence faster treegrowth reduces wood density.

Mäkinen and co-workers (2002a,b; Figs 4.14 and 4.15) studied the effectof nutrient optimisation on growth rate and basic density of Norway spruce.The study was performed in a nutrient optimisation experiment at Flakaliden,in northern Sweden (64°07′ N; 19°27′ E; alt. 310 m a.s.l.). The principal aimof the experiment was to demonstrate the potential yield of Norway spruceunder given climatic conditions and non-limiting soil water, by optimisingthe nutritional status of the stand, avoiding at the same time the leaching ofnutrients to the groundwater (Linder & Flower-Ellis, 1992; Linder, 1995).The annual dose of nitrogen was initially 100 kg N ha−1, the other nutrientsbeing supplied in fixed proportion to N. The N dose was reduced by 25% in1990.

The general trend in the wood density of spruce stems is a decrease fromthe pith outwards, reaching a minimum value around rings 10–20, and an


Fig. 4.14 Mean ring width in 12 control trees and 12 irrigated and fertilised trees of Norway spruce.Fertilisation and irrigation started in 1987 (vertical line) and the radial growth increased threefold.Redrawn from Mäkinen et al. (2002a).

Fig. 4.15 Variation of density from the pith to the bark in 12 control trees and 12 fertilised andirrigated trees. Unextracted weight density was determined at 12% moisture content (ρ12). Fertilisationand irrigation started in 1987 (vertical line) and during the fertilisation period, the mean ring density atbreast height decreased by 23%. Redrawn from Mäkinen et al. (2002a).


increase again towards the bark (e.g. Olesen, 1977; Danborg, 1994; Mitchell& Denne, 1997; Pape, 1999). In contrast, Mäkinen and co-workers (2002a)reported that the average ring density, as well as the earlywood density, ofthe control trees followed a continuous decreasing trend. However, late-wood density showed no apparent trend from the pith to the bark. It isprobable that the time of the minimum wood density occurs at approxi-mately the same time as crown closure, i.e. the onset of inter-treecompetition reduces ring width and increases wood density. As a result ofthe slow growth rate caused by the northern location and high altitude(64°07′ N, alt. 310 m a.s.l.), the stand sampled in this study was only justreaching the crown closure stage.

Fertilisation decreased the average wood density of the sample trees (Fig.4.15). The decrease was partly caused by a shift in the relative widths of theearlywood and latewood zones (Fig. 4.16). This is of course entirely asexpected, because almost all the previous studies on Norway spruce growthhave found that increasing radial growth is mainly caused by increasingearlywood width (e.g. Seibt, 1963; Hakkila, 1966; Klem, 1972; Evers &Gussone, 1991).

However, the decrease in wood density caused by heavy fertilisation wasnot entirely related to the relative proportions of earlywood and latewood(Mäkinen et al., 2002a). The improved nutrient status also decreased the abso-lute wood density throughout the whole annual ring, i.e. the fertilisation hadan effect on the anatomical structure of the wood, in addition to the change inthe ratio of earlywood to latewood. Nutrient optimisation thus results in alarger quantity of thin-walled cells (Fig. 4.16) that may be detrimental for themechanical and chemical wood-processing industries. Fertilisation alsoincreased the needle biomass threefold (Saranpää et al., 2002), which meansthat the current-year needles were the major metabolic sink. The short grow-ing season at the northern latitudes (138 days) limits latewood formation.Seasonal onset of latewood formation coincides with the cessation of heightgrowth by the terminal shoot and new needle maturity. The increase in wooddensity will occur in the future when the crown development diminishes. Thisagrees with Larson’s (1969) conclusions. Fertilisation influences crowndevelopment by increasing the photosynthesising surface and, possibly, byincreasing the photosynthetic efficiency of the foliage. The increase in crowndevelopment promotes earlywood formation in the stem. Larson (1969) madefurther two general statements: (1) Heavy fertilisation in young stands lowerswood quality because it increases crown and branch size and delays naturalpruning and the transition to mature wood; (2) Fertilisation is much moreeffective in older stands and those approaching maturity. The crowns aregenerally closed and the base of the living crown is high in the stem. Thus,fertilisation will generally result in increased growth of both earlywood andlatewood.


Fig. 4.16 Difference in latewood proportion in control trees (above) and fertilised trees (below). Fast-growing fertilised trees had hardly any latewood according to Mork’s definition.


4.6 Conclusions

1. The difference in basic density between juvenile and mature wood, i.e. changefrom the pith to the bark is the major source of variation within conifer trees.

2. In mature wood of Norway spruce, a negative, non-linear correlationexists between GRW and basic density. However, ring width or latewoodpercentage explains only ca. 40% of the variation in basic density. The cor-relation between GRW and density may be negative (softwoods with gradualtransition from earlywood to latewood), positive (ring-porous hardwoods) orinsignificant (southern pines, diffuse-porous hardwoods).

3. Thinning and fertilisation increase growth rate, which in many tree speciesresult in a decrease in wood density and strength. However, the effect of thin-ning depends on thinning degree and on the removal of dominant trees.

4. Heavy thinning and fertilisation have less effect in older stands than inyoung stands with crown not yet closed.

5. Basic density is under strong genetic control. This results in a largebetween-tree genetic variation in basic density.

6. The large genetic variation makes it possible to select trees with high growthrate and moderately high density for breeding.

7. Due to the large genetic variation and seasonal climatic changes in northernlong-rotation forestry, it is difficult to model and predict accurately wood den-sity with the help of stand and tree characteristics, even if management history(thinnings) is known.

References

Andersson, S., Serimaa, R., Paakkari, T., Saranpää, P. & Pesonen, E. (2003) The crystallinity of woodand the size of cellulose crystallites in Norway spruce (Picea abies (L.) Karst.). Journal of WoodScience (in press).

Anttonen, S., Manninen, A.-M., Saranpää, P., Kainulainen, P., Linder, S. & Vapaavuori, E. (2002) Effectsof long-term nutrient optimisation on stem wood chemistry in Picea abies. Trees, 16, 386–394.

Albrektson, A., Aronson, A. & Tamm, C.O. (1977) The effect of forest fertilization on primary productionand nutrient cycling in the forest ecosystem. Silva Fennica, 11, 233–239.

Ballard, L.A. & Long, J.N. (1988) Influence of stand density on log quality of lodgepole pine. CanadianJournal of Forestry Research, 18, 911–916.

Barbour, R.J. & Kellogg, R.M. (1990) Forest management and end-product quality, a Canadianperspective. Canadian Journal of Forestry Research, 20, 405–414.

Beall, F.C. (1972) Density of hemicelluloses and their relationship to wood substance density. Woodand Fiber Science, 4, 114–117.

Beets, P.N., Gilchrist, K. & Jeffreys, M.P. (2001) Wood density of radiata pine, effect of nitrogensupply. Forest Ecology and Management, 145, 173–180.

Bendtsen, B.A. (1978) Properties of wood improved and intensively managed trees. Forest ProductsJournal, 28, 61–72.


Boutelje, J.B. (1968) Juvenile wood, with particular reference to northern spruce. Svensk Papper Stidning,71, 581–585.

Brazier, J.D. (1970) Timber improvement. The effect of vigour on young-growth Sitka spruce. Forestry,43, 135–150.

Cullity, B.D. (1978) Elements of X-Ray Diffraction, Addison-Wesley publishing company (2nd edn)Addison-Wesley Publishing Company Inc., Reading Massachusetts, London.

Danborg, F. (1994) Density variations and demarcation of the juvenile wood in Norway spruce. DanishForest and Landscape Research Institute, Lyngby, Denmark, Forskningsserien, 10, 1–78.

de Kort, I., Loeffen, V. & Baas, P. (1991) Ring width, density and wood anatomy of Douglas fir withdifferent crown vitality. IAWA Bulletin, 12, 453–465.

Denne, M.P. (1989) Definition of latewood according to Mork (1928). IAWA Bulletin, 10, 59–62. Diaz-Váz, J.E., Echols, R. & Knigge, W. (1975) Comparative investigation of the variation of tracheid

dimensions and x-ray densitogrammes within the annual rings of two softwoods. Forstwissens-chaftliches Centralblatt, 94, 161–175 (in German, summary in English).

Dutilleul, P., Herman, M. & Avella-Shaw, T. (1998) Growth rate effects on correlations among ringwidth, wood density and mean tracheid length in Norway spruce (Picea abies). Canadian Journalof Forestry Research, 28, 56–68.

Echols, R.M. (1973) Uniformity of wood density assessed from X-rays of increment cores. WoodScience and Technology, 7, 34–44.

Evans, R., Kibbelwhite, R.P. & Stinger, S. (1997) Karft pulp fibre property prediction from woodproperties of eleven radiata pine clones. Appita Journal, 50, 25–33.

Evers, F.H. & Gussone, H.-A. (1991) Ernährung und Düngung der Fichte, in Die Fichte (ed.H. Schmidt-Vogt), Paul Parey, Hamburg, pp. 524–600 (in German).

Finkenstadt, V.L. & Millane, R.P. (1998) Crystal structure of Valonia cellulose 1ß, Macromolecules,31, 7776–7783.

Flæte, P.-O. & Kucera, B. (1999) Virkesegenskaper til mellomeuropeiske of norske granprovennienserplantet i Ostfold. Rapport fra Skogforskningen, NISK, 1/99, 1–30 (in Norwegian).

Gustavsen, H.G. & Lipas, E. (1975) Effect of nitrogen dosage on fertilizer response. Folia Forestalia, 246,1–20 (in Finnish, summary in English).

Haapanen, M., Velling, P. & Annala, M.-L. (1997) Progeny trial estimates of genetic parameters forgrowth and quality traits in Scots pine. Silva Fennica, 31, 3–12.

Hakkila, P. (1966) Investigations on the basic density of Finnish pine, spruce and birch wood. Communi-cationes Instituti Forestalis Fenniae, 61, 1–98.

Hakkila, P. (1968) Geographical variation of some properties of pine and spruce pulpwood in Finland.Communicationes Instituti Forestalis Fenniae, 66, 1–60.

Hakkila, P. (1998) Structure and properties of wood and woody biomass, in Forest Resources andSustainable Management (ed. S. Kellomäki), Fapet Oy, Helsinki, pp. 117–185.

Hakkila, P. & Uusvaara, O. (1968) On the basic density variation of plantation-grown Norway spruce.Communicationes Instituti Forestalis Fenniae, 66, 1–22.

Hannrup, B., Danell, Ö., Ekberg, I. & Moëll, M. (2001) Relationship between wood density and tracheiddimensions in Pinus sylvestris L. Wood Fiber Science, 33, 173–181.

Harris, J.M. (1969) The use of beta rays determining wood properties, Part 2, measuring earlywood andlatewood. New Zealand Journal of Science, 12, 409–418.

Harris, J.M. & Polge, H. (1967) A comparison of x-ray and beta radiation techniques for measuringwood density. Journal of the Institute of Wood Science, 19, 34–42.

Herman, M., Dutilleul, P. & Avella-Shaw, T. (1998) Growth rate effects on temporal trajectories ofring width, wood density and mean tracheid length in Norway spruce (Picea abies (L.) Karst.).Wood Fiber Science, 30, 6–17.

Hoang, M. & McKimmy, M.D. (1988) Direct scanning X-ray densitometry of thin wood sections. ForestProducts Journal, 38, 23–26.

Jeffers, J.W. (1959) Regression models of variation in specific gravity in four provenances of Sitkaspruce. Journal of the Institute of Wood Science, 4, 44–59.


Johansson, K. (1993) Influence of initial spacing and tree class on the basic density of Picea abies.Scandinavian Journal of Forestry Research, 8, 18–27.

Kärkkäinen, M. (1984) Effect of tree social status on basic density of Norway spruce. Silva Fennica,18, 115–120.

Kellogg, R.M. & Wangaard, F.F. (1969) Variation in the cell-wall density of wood. Wood Fiber Science,1, 180–204.

Kellogg, R.M., Sastry, C.B.R. & Wellwood, R.W. (1975) Relationship between cell-wall compositionand cell-wall density. Wood and Fiber Science, 7, 170–177.

Kenk, G. & Fischer, H. (1988) Evidence from nitrogen fertilization in the forests of Germany. Environ-mental Pollution, 54, 199–218.

Klem, G.G. (1952) The influence of spacing on spruce quality. Det Norske Skogforsøksvesen, 40, 473–506. Klem, G.S. (1972) The influence of fertilization of spruce (Picea abies (L.) Karst.) and pine (Pinus

sylvestris L.) on summerwood percent, specific gravity and extractive content. Det NorskeSkogforsøksvesen, Meddelelser, 122, 283–324 (in Norwegian, summary in English).

Klem, G.S. (1974) Properties of wood from fertilized forests. Norsk Treteknisk Institutt, Meddelelse,51, 1–59 (in Norwegian, summary in English).

Koubaa, A., Zhang, S.Y., Isabel, N., Beaulieu, J. & Bousquet, J. (2000) Phenotypic correlations betweenjuvenile-mature wood density and growth in black spruce. Wood and Fiber Science, 32, 61–71.

Koubaa, A., Zhang, S.Y. & Makni, S. (2002) Defining the transition from earlywood to latewood inblack spruce based on intra-ring wood density profiles from X-ray densitometry. Annals of ForestScience, 59, 511–518.

Kukkola, M. & Saramäki, J. (1983) Growth response in repeatedly fertilized pine and spruce stands onmineral soils. Communicationes Instituti Forestalis Fenniae, 114, 1–55.

Larson, P.R. (1969) Wood formation and the concept of wood quality. Bulletin of the Yale School ofForestry, 74, 1–54.

Larson, P.R., Kretschmann, D.E., Clark III, A. & Isebrands, J.G. (2001) Formation and propertiesof juvenile wood in southern pines. A synopsis. USDA, Forest Products Laboratory, GeneralTechnical Report, FPL-GTR-129, 1–42.

Lewark, S. (1982) Studies of wood characteristics of young Norway spruce from clonal experiments.Forstarchiv, 53, 14–21 (in German, summary in English).

Linder, S. (1995) Foliar analysis for detecting and correcting nutrient imbalances in Norway spruce.Ecological Bulletins, 44, 178–190.

Linder, S. & Flower-Ellis, J.G.K. (1992) Environmental and physiological constraints to forest yield,in Responses of Forest Ecosystems to Environmental Changes (eds A. Teller, P. Mathy &J.N.R. Jeffers) Elsevier Applied Sciences, London, pp. 149–164.

Lindström, H. (1996) Basic density in Norway spruce. Part I. A literature review. Wood and FiberScience, 28, 15–27.

Madsen, T.L., Moltesen, P. & Olesen, P.O. (1978) The influence of thinning degree on basic density,production of dry matter, branch thickness and number of branches of Norway spruce. Det For-stlige Forsøgsvaesen Danmark, 36, 181–204.

Mäkinen, H., Saranpää, P. & Linder, S. (2002a) Wood density variation of Norway spruce (Picea abies(L.) Karst.) in relation to nutrient optimisation and fibre dimensions. Canadian Journal ofForestry Research, 32, 185–194.

Mäkinen, H., Saranpää, P. & Linder, S. (2002b) Effect of growth rate on fibre characteristics in Norwayspruce (Picea abies (L.) Karst.). Holzforschung, 56, 449–460.

Megraw, R.A. (1985) Wood Quality Factors in Loblolly Pine. The Influence of Tree Age, Position inTree, and Cultural Practise on Wood Specific Gravity, Fiber Length, and Fibril Angle, TappiPress, Atlanta.

Mitchell, M.D. & Denne, M.P. (1997) Variation in density of Picea sitchensis in relation to within-treetrends in tracheid diameter and wall thickness. Forestry, 70, 47–60.

Moltesen, P., Madsen, T.L. & Olesen, P.O. (1985) The influence of spacing on the production of drymatter and wood quality of Norway spruce. Det Forstlige Forsøgsvaesen Danmark, 40, 53–76.


Mork, E. (1928) Qualität des Fichtenholzes unterbesonderer Rücksichtnahme auf Schleif- und Papierholz.Der Papier-Fabrikant, 26, 741–747.

Nylinder, P. (1953) Variations in basic density in planted spruce. Meddelanden från Statens Skogforskn-ingsinititut, 43, 1–42 (in Swedish with English summary).

Okstad, T. & Kårstad, H. (1985) Mekaniske egenskpaer hos små, feilfrie prover av granvirke (Picea abiesL. Karst.) fra Nord-Norge. Meddelelser fra Norsk Institutt for Skogforskning, 38, 1–47 (in Norwegian).

Olesen, P.O. (1971) The water displacement method. Forest Tree Improvement, 3, 3–23. Olesen, P.O. (1976) The interrelation between basic density and ring width of Norway spruce. Det For-

stlige Forsøgsvaesen Danmark, 34, 339–359. Olesen, P.O. (1977) The variation of the basic density level and tracheid width within the juvenile and

mature wood of Norway spruce. Forest Tree Improvement, 12, 1–21. Olesen, P.O. (1982) The effect of cyclophysis on tracheid width and basic density in Norway spruce.

Forest Tree Improvement, 15, 1–80. Panshin, A.J. & de Zeeuw, C. (1980) Textbook of wood technology. Structure, Identification, Properties, and

Uses of the Commercial Woods of the United States and Canada, 4th edn, McGraw-Hill, New York. Pape, R. (1999) Influence of thinning and tree diameter class on the development of basic density and

annual ring width in Picea abies. Scandinavian Journal of Forestry Research, 14, 27–37. Pearson, R.G. & Gilmore, R.C. (1971) Characterization of the strength of juvenile wood of loblolly

pine (Pinus taeda L.). Forest Products Journal, 21, 23–30. Petty, J.A., MacMillan, D.C. & Steward, C.M. (1990) Variation of density and growth ring width in

stems of Sitka and Norway spruce. Forestry, 63, 39–49. Polge, H. & Nicholls, J.W.P. (1972) Quantitative radiography and the densitometric analysis of wood.

Wood Science, 5, 51–59. Puro, T. (1977) Results of the second fertilization with nitrogen. Folia Forestalia, 304, 1–15 (in Finnish,

summary in English). Quirk, J.T. (1984) Shrinkage and related properties of Douglas-fir cell walls. Wood and Fiber Science,

16, 115–133. Rozenberg, P. & Cahalan, C. (1997) Spruce wood quality: genetic aspects (a review). Silvae Genetica,

46, 270–279. Saikku, O. (1975a) The effect of fertilization on the basic density of Scots pine (Pinus sylvestris L.).

A densitometric study on the X-ray chart curves of wood. Communicationes Instituti ForestalisFenniae, 85, 1–49.

Saikku, O. (1975b) The effect of nitrogen fertilization on the basic density of Scots pine (Pinus sylves-tris L.), Norway spruce (Picea abies) and common birch (Betula verrucosa). CommunicationesInstituti Forestalis Fenniae, 85, 1–24 (in Finnish, summary in English).

Saranpää, P. (1994) Basic density, longitudinal shrinkage and tracheid length of juvenile wood of Piceaabies Karst. Scandinavian Journal of Forestry Research, 9, 68–74.

Saranpää, P., Pesonen, E., Sarén, M., et al. (2000) Variation of the properties of tracheids in Norwayspruce (Picea abies (L.) Karst.), in Cell and Molecular Biology of Wood Formation (edsR.A. Savidge, J.R. Barnett & R. Napier), BIOS Scientific Publishers Limited, Oxford, pp. 337–345.

Saranpää, P. & Repola, J. (2001) Strength of Norway spruce from both mixed stands and monocultures,in Wood Quality of Norway Spruce Grown in Mixture with Birch and in Monoculture (edsA. Bergstedt, B. Kucera, M. Nylinder, P. Saranpää & E. Ståhl), Finnish Forest Research Institute,Vantaa, Research Reports 822, pp. 33–39.

Saranpää, P., Vapaavuori, E. & Peltola, H. (2002) Effect of forest management on wood quality,in Final report, Finnish Forest Cluster Research Programme WOOD WISDOM, Report 3 (ed.L. Paavilainen), Academy of Finland, Helsinki, pp. 295–301.

Sauvala, K. (1979) Use of compression method in the species-specific calibration of X-ray densito-meter measurements. Wood Microdensitometry Bulletin, 1, 4–8.

Seeling, U. (1999) Einfluss von Richtgewebe (“Druckholz”) auf Festigkeit und Elastizität des Fichtenholzes.Holz als Roh- und Werkstoff, 57, 81–91.

Seibt, G. (1963) Der Einfluß der Düngung auf das Holzgewicht von Kiefer, Japanlärche und Fichte.Aus dem Walde, 6, 51–82 (in German).


Simpson, H.L. & Denne, M.P. (1997) Variation of ring width and specific gravity within trees fromunthinned Sitka spruce spacing trial in Clocaenog, North Wales. Forestry, 70, 31–45.

Sjöström, E. (1993) Wood Chemistry. Fundamentals and Applications, Academic Press, London. Skaar, C. (1972) Water in Wood, Syracuse University Press, Syracuse. Spurr, S. & Hsiung, W.Y. (1954) Growth rate and specific gravity in conifers. Journal of Forestry, 52,

191–200. Stamm, A.J. (1964) Wood and Cellulose Science, Ronald Press, New York. Stamm, A.J. (1969) Correlation of structural variations of lignins with their specific gravities. Tappi,

52, 1498–1502. Sugiyama, J., Voung, R. & Chanzy, H. (1991) Electron diffraction study on the two crystalline phases

occuring in native cellulose from an algal cell wall. Macromolecules, 24, 4168–4175. Tamm, C.O. (1991) Nitrogen in terrestrial ecosystems – questions of productivity, vegetational

changes, and ecosystem stability. Ecological Studies, 81, Springer-Verlag, Berlin. Taylor, F.W., Wang, E.I., Yanchuk, A. & Micko, M.M. (1982) Specific gravity and tracheid length

variation of white spruce in Alberta. Canadian Journal of Forestry Research, 12, 561–566. Thomas, R.J. (1984) The characteristics of juvenile wood, in Proceedings of the Symposium on Utilisa-

tion Changes in Wood Research in the Southern U.S., North Carolina State University, Raleigh,North Carolina, pp. 40–52.

Tsoumis, G. & Passialis, C. (1977) Effect of growth rate and abnormal growth on wood substance andcell wall density. Wood Science and Technology, 11, 33–38.

Velling, P. (1976) The wood basic density variation of pine and spruce provenances. Folia Forestalia,257, 1–32 (in Finnish, English summary).

Velling, P. (1980) Variation in the density of wood of different Scots pine (Pinus sylvestris L.) andNorway spruce (Picea abies (L.) Karst.), origins in the provenance tests. Silva Fennica, 14, 45–51.

Viro, P.J. (1972) Die Walddüngung auf finnischen Mineralböden. Folia Forestalia, 138, 1–19 (in German). Walker, J.C.F. (1993) Primary Wood Processing, Principles and Practice, Chapman & Hall, London, 595,

pp. 74–77. Weatherwax, R.C. & Tarkow, H. (1968) Density of wood substance, importance of penetration and

adsorption compression of the displacement fluid. Forest Products Journal, 18, 44–46. Wilhelmsson, L., Arlinger, J., Spångberg, K., et al. (2002) Models for predicting wood properties in

stems of Picea abies and Pinus sylvestris in Sweden. Scandinavian Journal of Forestry Research,17, 330–350.

Wimmer, R. (1995) Intra-annual cellular characteristics and their implications for modeling softwooddensity. Wood and Fiber Science, 27, 413–420.

Wodzicki, T.J. (2001) Natural factors affecting wood structure. Wood Science and Technology, 35, 5–26. Zhang, S.Y. (1994) Mechanical properties in relation specific gravity in 342 Chinese woods. Wood and

Fiber Science, 26, 512–526. Zhang, S.Y. (1995) Effect of growth rate on wood specific gravity and selected mechanical properties

in individual species from distinct wood categories. Wood Science and Technology, 29, 451–465. Zhang, S.Y. (1998) Effect of age on the variation, correlations and inheritance of selected wood charac-

teristics in black spruce (Picea mariana). Wood Science and Technology, 32, 197–204. Zhang, S.Y. & Morgenstern, E.K. (1995) Genetic variation and inheritance of wood density in black

spruce (Picea mariana) and its relationship with growth, implications for tree breeding. WoodScience and Technology, 30, 63–75.

Zhang, S.Y., Simpson, D. & Morgenstern, E.K. (1996) Variation in the relationship of wood densitywith growth in 40 black spruce (Picea mariana) families grown in New Brunswick. Wood andFiber Science, 28, 91–99.

Zobel, B.J. & van Buijtenen, J.P. (1989) Wood Variation, Its Causes and Control, Springer-Verlag,Berlin.

Zobel, B.J. & Jett, J.B. (1995) Genetics of Wood Production, Springer-Verlag, Berlin. Zobel, B.J. & Sprague, J.R. (1998) Juvenile Wood in Forest Trees, Springer-Verlag, Berlin.

5 Reaction wood John R. Barnett and George Jeronimidis

5.1 Introduction

One of the important attributes of trees is their ability to maintain a verticalalignment of the main stem and a fixed pattern of branch angles in the face ofenvironmental factors such as gravity, wind, sloping site and competition.Since these factors would normally be expected to favour an increase in thetilting of the stem or further drooping of a branch as its mass increases withgrowth, it is clear that the tree is actively countering them. Unlike young primarystems and shoots, which are still undergoing elongation growth and can thereforemaintain or change their orientation by differential longitudinal growth, woodystems have ceased elongation growth and must correct their orientation bybending the existing structure (Wardrop, 1964). They do this by producingmodified wood known as reaction wood. The anatomy, structure and physicalproperties of reaction wood are well adapted to provide the required biome-chanical function. However, they have detrimental effects on the quality ofwood for commercial utilisation.

A fascinating aspect of reaction wood is that gymnosperms and angiospermshave generally adopted different strategies for coping with the mechanics ofactive change of curvature and bending stresses. In leaning gymnosperm stems,reaction wood is formed on the lower side of the stem, which is under compres-sive force. It is therefore called compression wood, and by mechanisms discussedbelow, it may eventually restore the stem to a vertical alignment (Fig. 5.1).Compression wood is also formed on the lower side of branches of gymnospermsand resists the tendency of the branch to bend downwards under its own weight.According to Timell (1983), compression wood evolved about 300 millionyears ago, and is now found in the Ginkgoales, the Coniferales and the Taxales,but not in the Cycadales and the Gnetales.

In leaning stems of angiosperm trees, the situation is the opposite, withreaction wood being formed on the upper side of the stem. Since this side of aleaning stem is in tension, the wood is known as tension wood. Branches inthese trees similarly produce tension wood on the upper side to maintain theirangle of growth. However, unlike compression wood, which has been reportedas present in all coniferous species examined (Westing, 1965, 1968; Timell, 1969),tension wood has been reported as being absent from the branches of manyangiosperm trees (Höster & Liese, 1966) which nevertheless maintain their

REACTION WOOD 119

branch architecture (Fisher & Stevenson, 1981). Tomlinson (2001) recentlyreported that while reaction tissues were absent from the xylem tissues ofGnetum gnomen, gelatinous (tension wood) fibres were present in the cortex onthe upper side of leaning stems.

In both angiosperms and gymnosperms, the wood formed on the oppositeside of the stem or branch to the reaction wood is known as opposite wood,while that on either side and lying between the reaction wood and the oppositewood is referred to as lateral wood. In comparison with wood production in avertically growing stem with almost perfectly circular growth rings (hereafterreferred to as normal wood), compression wood and tension wood are usuallyproduced in larger quantities, while opposite wood is produced in smallerquantities giving the stem a cam-shaped cross section with pronounced eccen-tricity with respect to the pith (Fig. 5.2).

There are reported instances of the occurrence of compression wood insome angiosperms in which the wood type is regarded as primitive because themain axial elements are tracheids (Höster & Liese, 1966; Yoshizawa et al.,1993; Baillères et al., 1997). In contrast, gelatinous (tension wood) fibres havealso been reported in coniferous wood (Jacquiot & Trenard, 1974).

Fig. 5.1 A Pinus sylvestris tree which was blown over but remained alive with its roots partly intact.Compression wood formed subsequently has restored the upper part of the trunk to an alignmentmatching that of other trees by which it is surrounded. Note that these trees lean inwards towards theslope of the hill, a characteristic of trees growing on inclined surfaces.


Reaction wood has a profound effect on the usefulness of timber (see below),but the economic importance of coniferous trees means that compression woodhas been subjected to considerably more research than tension wood. Themonograph by Timell (1986) provides good evidence of this since his three-volume work on compression wood is 2150 pages long. Nothing has been writtenon this scale about tension wood although a recent doctoral thesis goes some wayin filling the gap (Clair, 2001).

Reaction wood formation is also associated with rapid wood formation intrees. Vertically growing conifers with apparently excellent stem form mayproduce complete rings of compression wood, while a spiral pattern has beendescribed within the annual rings of Abies concolor and Pinus taeda (Telewski,

Fig. 5.2 End view of a log of Pinus sylvestris showing the cam-shaped cross section resulting fromcompression wood formation. The growth rings containing compression wood (top left) are muchwider than the rings of opposite wood (bottom right).

REACTION WOOD 121

1988). The latter were suggested to be a result of circumnutational growthmovements.

Fast growing conifers are also more susceptible to compression wood forma-tion in response to gravitational stress (Pillow & Luxford, 1937). The tendencyof fast-growing trees to produce reaction wood in stems of good form haspotentially serious economic consequences for growers. The current emphasison fast growth, short-rotation forestry means that much of the timber producedmay have reduced value owing to the presence of reaction wood. This adds tothe problems caused by the presence of a high proportion of juvenile wood ofhigh microfibril angle in trees harvested at an early age (discussed in Chapter 2).

5.2 Early studies of reaction wood formation

Timell (1980) records that the first scientific descriptions of compression woodwere made by Sanio in 1860, closely followed by Schacht in 1862. Earlyexperiments involved investigating its location in bent stems and twigs when itwas invariably found on the lower side of the bend (Ewart & Mason-Jones,1906). Experiments involving saplings bent into vertical loops were carriedout by Jaccard (1938). In the case of gymnosperms, compression wood wasformed on the lower side of the loops at the top and bottom, while tensionwood in angiosperms was formed on the upper sides regardless of whether thatpart of the loop was in tension or compression. Interestingly, although the loopedpart of the stem in each case must be in tension around its entire outer circum-ference, and in compression around its inner circumference, reaction woodwas not recorded as forming in the vertically oriented parts of the loops. Whenthe loops were cut in half across their horizontal axis, compressive or tensileforces were released which demonstrated that compression wood exerts a push,while tension wood exerts a pull on its side of the stem (Fig. 5.3).

These observations, which demonstrated that reaction wood was formed inparticular locations in the stem apparently regardless of whether the appliedstress was tensile or compressive, led early workers to attribute its formation tothe effect of gravity (Ewart & Mason-Jones, 1906). It could, however, be arguedthat all cells in a tree are subject to the same force of gravity regardless of theirlocation in the structure, or the orientation of the stem or branch of which theyare a part. Tilting or harnessing of stems at an angle to the vertical is a techniquewhich has been used on numerous occasions for induction of reaction woodfor study.

In the case of angiosperm branches, it has been found that bending themvertically out of their normal alignment causes the formation of tension wood onthe lower side in branches bent upwards, and on the upper side in branches bentdownwards (Wardrop, 1964). Similarly, compression wood in gymnospermsforms on the upper side of branches bent upwards from their normal alignment,


and on the lower side of branches bent downwards (Wilson et al., 1989). Itis clear that in this case, reaction wood forms in an attempt to restore apre-ordained branch angle.

5.3 Induction of reaction wood formation

5.3.1 The role of auxin

The discovery of auxin and its effects on plant growth offered an explanationfor many physiological processes in plants including reaction wood formation.Following the observation by Wershing and Bailey (1942) that externalapplication of 3-indoleacetic acid (IAA) could induce compression wood forma-tion, it was hypothesised that changes in auxin levels as a result of basipetalauxin flow were responsible for the formation of both tension and compressionwood. Accumulation of auxin on the lower side of a leaning gymnosperm stem

A B

C D

Fig. 5.3 Location of compression wood (A) and tension wood (B) formation in stems bent into loops.The reaction wood is shown as the shaded regions on the lower sides of the loop in A, and the uppersides in B. Cutting the loops, as shown in C and D, releases the stress and the cut ends move in thedirections indicated by the arrows. After Jaccard (1938).

REACTION WOOD 123

or branch was proposed to lead to compression wood formation, while reducedlevels on the upper side of a leaning angiosperm stem or branch would lead totension wood formation. Nedesaný (1958) applied auxin to the upper side ofleaning stems, which were producing tension wood and found that this inhibitedtension wood formation suggesting that tension wood formation might be aresponse to reduced auxin levels.

The ability to test these hypotheses was, until recently, limited by the difficultyin measuring the real amount of auxin, and its location, in plant tissues. Eventhose experiments in which auxin-containing lanolin was applied to decapitatedstems, for example, could be questioned on the grounds that there was little hardevidence that auxin moved from the lanolin into the tissues or was not metabo-lised once it was there. Lachaud (1987) used tritiated IAA and demonstratedthat in stems bent into loops while still attached to the plant, the label movedtowards the lower sides of the looped stem and extreme tension wood wasformed. When the loop was detached from the plant, there was no directionaltransport of label, and less extreme tension wood was formed on the upper sideof the loops.

Wilson etal. (1989) used gas chromatography–mass spectroscopy to measurethe concentration of auxin at different locations in branches of Douglas fir thathad been bent upwards or downwards from their normal orientation (see above).Compression wood formed on the concave side of the new bend in the branch,i.e. the upper side of branches bent upwards and the lower side of branchesbent downwards. However, IAA concentrations were not significantly differentbetween the upper and lower sides of branches bent upwards. In branches bentdownwards, more IAA was found on the upper side, even though compressionwood formed on the lower side.

Sundberg et al. (1994) applied a ring of IAA transport inhibitors to one-year-old Scots pine trees and using gas chromatography–mass spectroscopyfound a decrease in IAA levels above the ring compared with controls. Despitethis decrease, an increase in compression wood formation occurred in thisregion. Sundberg and co-workers also used a highly sensitive mass spectrometrytechnique to measure endogenous auxin with relatively high resolution in thevarious layers of cambium and its differentiating derivatives in tilted pinestems. The trees formed compression wood tracheids but cambial growth wasnot stimulated during the experimental period. No difference in auxin distributionwas seen in tilted trees from that in controls, supporting the view that undernatural conditions, tracheids with compression wood characteristics are notinduced by elevated auxin concentrations.

Sundberg and co-workers (personal communication) have studied endo-genous auxin using the same technique during tension wood formation in hybridaspen. The auxin content was found to be higher on the upper side where tensionwood was being formed, contrary to the expectation that levels would be loweras proposed by the auxin hypothesis of reaction wood formation. This is, however,


in agreement with the role of auxin in stimulating xylem cell production, sincemany more cells are usually produced on the tension wood side of a stem orbranch. Mellerowicz et al. (2001) suggested that the distribution of auxinacross the cambial zone is more important than its absolute concentration.

A further interesting observation by Sundberg (personal communication)resulted from application of auxin-containing lanolin to decapitated stems. It wasfound that auxin levels in the extraxylary tissues were raised compared tothose in lanolin-only controls, but were still lower than those in intact controltrees. This observation raises wider and interesting questions about the resultsobtained in similar experiments and reported in the literature, where effectsobserved after application of auxin in this way were attributed to raised levelsof auxin.

5.3.2 The role of ethylene

Little and Eklund (1999) found that tilting seedlings of Abies balsameainduced compression wood formation accompanied by increased ethyleneproduction on the lower side of the stem. Application of NPA increased ethyleneproduction and the formation of compression wood tracheids above the pointof application, but inhibited tracheid production and compression wood for-mation below the point of application. From this, the authors deduced thatethylene production is directly related to compression wood formation.

5.3.3 The role of gibberellins

GA3 has been shown to induce tension wood formation on the upper sides ofweeping branches (normally without tension wood) of the weeping form of Prunusspachiana (Baba et al., 1995). Consequently the weeping habit was lost andthe branches grew upwards in the normal manner. This clearly implies a role forgrowth regulators in the formation of tension wood, since stress alone clearlydoes not result in tension wood formation in branches of weeping varietiesof trees.

5.3.4 The role of stress

Boyd (1977) revisited the experiments which had purported to show that auxinconcentration was the key factor in inducing reaction wood. He argued that allthe reported observations could be explained in terms of growth stresses. Theyarise as a consequence of water loss and consequent dimensional changesduring the wood maturation process (discussed in Chapter 6). These stressesare in equilibrium at all points in the stem when it is growing vertically. Lateraldisplacement from the vertical results in additional bending which requires anew set of self-equilibrating stresses to maintain or correct orientation. Similarly,the effect of gravity on a growing branch means that bending stresses due to

REACTION WOOD 125

self-weight must be superposed to the growth stresses, which develop withinthe branch between the upper and lower sides in order to maintain orientationof the limb in space.

Bending stress has been shown to lead to cortical microtubule re-orientationin maize coleoptiles (Zandomeni & Schopfer, 1994) suggesting that mechanicalstress perception involves microtubules associated with the plasma membrane.This might account for the observed differences in microtubule orientationbetween normal and reaction wood cells (Nobushi & Fujita, 1972; Fujita et al.,1974). The microtubule orientation usually parallels that of the cellulosemicrofibrils in the concurrently forming secondary-wall lamellae in reactionwood (e.g. Prodhan et al., 1995; Furosawa et al., 1998), although whetherthe relationship is causal has been disputed (Emons et al., 1992; Barnett et al.,1997).

Kwon et al. (2001) demonstrated that plants of Douglas fir bent at 45°formed compression wood whether under conditions of microgravity or normalgravity. This provides strong evidence that compression wood formation is astress response, rather than a consequence of auxin redistribution.

5.4 Structure and formation of reaction wood

In addition to their location on opposite sides of the stem or branch, the structuralproperties of the wood might also be described as opposite in several ways. Itis a remarkable fact that coniferous trees and angiosperm trees have evolvedentirely different strategies for dealing with the same problem. The structure ofreaction wood is a consequence of deviations from the normal pattern of celldifferentiation of xylem mother cells produced by divisions of the vascularcambium.

5.4.1 Compression wood

Compression wood formation is characterised by changes in the rates ofpericlinal division in the vascular cambium giving rise to growth rings whichare wider than normal on the compression wood side of the stem or branch,and narrower on the opposite side. The more rapid growth on the compressionwood side means that more anticlinal pseudotransverse divisions are requiredto maintain a continuous ring of cambium (Wardrop & Dadswell, 1950). Asa result, the mean length of compression wood tracheids is less than that of thetracheids of opposite wood. The appearance of cells with forked or otherwiseabnormal tips is a feature of compression wood and indicates that the cellsencounter resistance to growth (Münch, 1940; Wardrop & Davis, 1964).Yoshizawa et al. (1985, 1987) explained this as being caused by restrictions onintrusive growth resulting from the increased cell-division activity on thecompression wood side.


In most forms of compression wood, tracheids are more rounded in trans-verse section, and have intercellular spaces, normally absent in normal wood(Fig. 5.4) which may be small in the case of mild compression wood (Fig. 5.5),or large in severe cases, with the cells having a rounded profile in transversesection (Fig. 5.6). Normal wood has no such spaces present between tracheids.Wardrop and Davis (1964) found that these spaces developed schigenouslyduring the phase of cell enlargement and before secondary wall formation

Fig. 5.4 Normal wood of Pinus radiata D. Don. The tracheids fit closely together and no intercellularspaces are present.

Fig. 5.5 Mild compression wood in Pinus radiata. The cells are more rounded than in the normalwood and have intercellular spaces.

REACTION WOOD 127

occurred, a view confirmed by Takabe et al. (1992). They suggested thatabnormally high turgor in the enlarging cells resulting in the cells becomingrounded in section caused the formation of the spaces.

The secondary wall of compression wood tracheids is markedly differentfrom that of normal wood tracheids. Instead of the normal three secondarywall layers (S1, S2 and S3 – see Chapter 2), there are usually two layers only,the S1 and the S2. It is the S2 layer in which the differences are most pro-nounced. Here the microfibril angle is much higher than in normal wood, often45° or more. There is also a greater proportion of lignin present. The wall con-tains splits or checks (Figs 5.6 and 5.7), although it is uncertain whether theseare normal or a feature resulting from the release of stresses during samplingof the wood. Singh et al. (1998) described radial striations in the S2 layer ofmild compression wood which Singh and Donaldson (1999) attributed to alter-nating regions of high and low lignin concentration. They suggested that thesemight be the precursors of the helical checks. The checks enable easy measure-ment of the microfibril angle of the S2 layer in compression wood cells as theyfollow the principal orientation of this layer.

The high microfibril angle and the increased proportion of lignin make thewood stronger in compression and enable the wood to resist the effect of grav-itational bending forces on the tissue. In stems, the active mechanism forinducing the curvature needed to re-establish vertical orientation is associatedwith the greater axial expansion of compression wood cells against older nor-mal wood cells during maturation. This is similar to the change of curvature ofa bi-metallic strip with temperature. The higher cellulose microfibril angle inthe S2 cell wall layers of compression wood is primarily responsible for this.

Fig. 5.6 Severe compression wood in Pinus radiata. The S3 wall layer is absent and there are largesplits present in the S2 layer.


The identification of a callose-like β-D-1,3 linked glucan (laricinan) in extractsfrom compression wood (Hoffmann & Timell, 1970, 1972) led to the suggestionthat this compound was present in the helical cavities in the S2 layer of thewall and might be involved in generating the forces needed for reorientation ofstems and branches. This would occur by swelling of a water–laricinan complex(Brodski, 1972). Boyd (1978) subsequently produced convincing argumentsagainst a role for this substance in the reorientation mechanism.

Attempts to find a molecular basis for compression wood formation in Pinuspinaster have resulted in the detection of up-regulated proteins including,among others, 1-aminocyclopropane-1-carboxylate oxidase (an ethylene-formingenzyme), caffeic O-methyltransferase and caffeoyl CoA-O-methyltransferase(enzymes involved in lignification) (Plomion et al., 2000). A differentiallyexpressed laccase between developing normal and compression wood has beendetected in Picea sitchesis (McDougall, 2000).

5.4.2 Tension wood

The anatomy of tension wood was the subject of an extensive review byWardrop (1964) and the reader is referred to this for a summary of early work.It is probably true to say that little that is fundamentally new has been added tothis aspect of tension wood since then. Among the major features he records,

Fig. 5.7 View of the inner wall of a compression wood tracheid of Pinus radiata showing the splits inthe S2. The cell’s long axis runs from top left to bottom right of the photograph. The orientation of thesplits mirrors that of the cellulose microfibrils in the S2 layer showing that the microfibril angle is large.

REACTION WOOD 129

are the fact that the fibres of tension wood have been variously described bydifferent authors working on different species as being longer, the same lengthor shorter than normal wood fibres. He suggested that this might reflect theposition in the stem from which samples were taken, and whether samples ofopposite wood came from the same growth ring as the tension wood fibresbeing measured, and at the same level in the stem. Alternatively, the resultscould be explained by growth rate since tension wood formation is usually, butnot always, marked by an increase in cell production by the cambium on theupper side of the leaning stem or branch. In those cases in which stem eccen-tricity was the greatest, reflecting faster growth on the tension wood side of thestem, fibres would be expected to be shorter.

In contrast with compression wood tracheids, in which the tips have beendescribed as distorted (see above), tension wood fibres have fewer bifurcationsthan opposite wood fibres (Wardrop, 1964).

As fibres are the main load-bearing elements in the stem, they are structuredto make them strong in tension and capable of resisting gravitational forces onleaning stems and branches and maintaining or restoring normal orientation.Typically, the fibre wall is made up of an S1 layer and an S2 layer whosethickness varies from very much thinner than in normal wood fibres, to beingalmost complete, with the so-called gelatinous layer deposited inside insteadof an S3 layer (Faruya et al., 1970). Occasionally, the presence of an S3-typelayer has been reported inside the gelatinous layer (e.g. Côté et al., 1969). Thegelatinous layer is composed largely of hydrated cellulose (Norberg & Meier,1966) whose microfibrils are more loosely bound together than in normalwood fibres (Côté et al., 1969) and are arranged in concentric lamellae(Casperson, 1961a,b). Compared with normal wood (Fig. 5.8), it may berelatively thin, or may fill almost the entire cell lumen (Figs 5.9–5.11). Thename derives from the fact that it has a glistening gelatinous appearance infreshly cut wood. It is easily detached from the wall by mechanical disturbanceduring preparation of material for microscopy. Cellulose microfibrils withinthis layer are oriented almost parallel to the cell’s long axis, endowing the cellwith strength and high Young’s modulus in tension.

Examination of leaning stems of Betula pendula containing tension woodusing ultra-violet illumination has shown that vessel walls and ray cells remainfluorescent, while fibre walls have a much-reduced fluorescence (Fig. 5.12).This confirms the low level of lignification in the fibre walls compared withopposite and normal wood. Lignin is known to be almost completely absentfrom the gelatinous layer. However, Bentum et al. (1969) found that the con-centration of lignin is higher in the S1 and S2 layers of tension wood fibresthan in normal wood fibres, with the result that the same total amount of ligninis present as in normal wood fibres. They also observed that the extent oflignification of vessel element walls does not appear to differ between normaland tension wood.


Recent work on fast-growing hybrid aspen (Populus tremula× tremuloides)has shown that small pockets of tension wood fibres may be distributed amongnormal fibres in vertical stems. In angiosperms, the mechanism for inducingcurvature changes in stems and branches to correct or maintain orientation issimilar, in principle, to that in gymnosperms. During maturation, tension wood

Fig. 5.8 Transverse section of normal wood of Fagus sylvatica L.

Fig. 5.9 Transverse section of tension wood in Fagus sylvativa L. Note the increase in proportion offibres to vessels and the distortion in vessel shape.

REACTION WOOD 131

shrinks axially more than the normal wood tissue it is bonded to, generating thestrains and stresses which result in curvature change. In recent years, severalpublications have appeared on the relationship between anatomy, ultra-structure,mechanical properties and biomechanical function of tension wood (Sassus,1988; Clair & Thibaut, 2001; Clair etal., 2001; Coutand etal., 2003).

Fig. 5.10 Tension wood of Fagus sylvatica L. Note that the gelatinous wall of the fibre almost entirelyobscures the cell lumen.

Fig. 5.11 Scanning electron micrograph showing the transverse surface of tension wood fibres inFagus sylvatica in which the cell lumen is much reduced.


In general, the proportion of vessels in tension wood is reduced (Ollinmaa,1955), and the vessels appear distorted in section (Fig. 5.9). It may be that theyare non-functional with regard to water transport as has been suggested forcompression wood tracheids (Spicer & Gartner, 1998).

5.4.3 Opposite and lateral wood

Timell (1973) found that the structural features of opposite wood in conifersdiffered from normal wood in the opposite sense to the way compressionwood differed from normal wood. Wilson (1981) suggested that there was acontinuum of characters from those of opposite wood, through normal (orlateral wood) to compression wood in a leaning stem, reflecting the distributionof stress around the circumference.

5.5 Reaction wood and wood quality

Owing to their modified anatomy (cell shape and length, wall thickness),chemistry (lignification) and ultra-structure (microfibrillar angle in cell walls)

Fig. 5.12 Fluorescence micrograph of a transverse section through a leaning stem of silver birch(Betula pendula). Areas of tension wood (lower left, second and outer growth rings) have low levels ofautofluorescence indicating the absence of lignin. Vessel walls and rays have retained their autofluorescentcapability.

REACTION WOOD 133

with respect to normal wood, reaction woods have different physical andmechanical properties. These properties are not necessarily worse than thoseof normal wood and, indeed, some are better: the compressive strength ofcompression wood is higher in normal wood and, similarly, the tensile strengthof tension wood is greater than that of normal wood. However, these biome-chanical optimisations of strength, which may be desirable, are counteractedby changes in other properties which are detrimental to quality. Compressionwood, for example, having been optimised to improve compressive strength,results in a wood which is very brittle and dangerous in structural applicationsof timber, even if present in only limited amounts, particularly as it is alsodifficult to detect (Illston et al., 1979). Tension wood has a higher tensilestrength and Young’s modulus than normal wood, even on a specific basis.It also has a higher fracture toughness and impact resistance, as demonstratedby the difficulty in sawing it and by the very fibrous appearance of machinedsurfaces (USDA, 1974; Sassus, 1988; Huang & Jeronimidis, 2002; Coutandet al., 2003). Mechanically, at least, tension wood is a better wood thannormal.

The main problem associated with quality and utilisation of wood andtimber containing reaction tissue is the fact that their shrinkage characteristicsare different from those of adjacent normal wood. Figure 5.13 shows the extensiveaxial shrinkage of the G layer on drying, for example. This in itself would notbe a problem if the log were made entirely of compression or tension wood.However, since reaction wood is typically localised on one side of the trunkand is often found only in a proportion of the total number of annual rings, itleads to differential shrinkage effects during drying. These manifest them-selves as warping, twisting, bending and cracking of logs, planks, machinedparts and veneers (see Chapter 6).

Fig. 5.13 Axial shrinkage on drying of G layer in Populus cv I4551 (scale bar=10 μm). From Clair (2001).


The ability to form reaction wood is biologically essential for the tree, help-ing to maintain vertical main stems, and in most trees, controlling branchingangles and canopy structure. It is known that the tendency for a particular treeto form reaction wood is under genetic control, but it is clear that it would beimpractical to breed or engineer trees unable to form reaction wood. Selectionof trees with good stem form, or the use of molecular markers to identify treeswith good stem form and without the formation of excessive amounts of tensionwood may offer one approach to the problem. More important, however, may beunderstanding and eliminating the apparently unnecessary formation of reactionwood in fast-growing trees.

References

Baba, K., Adachi, K., Take, T., Yokoyama, T., Itoh, T. & Nakamura, T. (1995) Induction of tensionwood in GA3-treated branches of the weeping type of Japanese cherry (Prunus spachiana). PlantCell Physiology, 36, 983–988.

Baillères, H., Castan, M., Monties, B., Pollet, B. & Lapierre, C. (1997) Lignin structure in Buxussempervirens reaction wood. Phytochemistry, 44, 35–39.

Barnett, J.R., Chaffey, N.J. & Barlow, P.W. (1997) Cortical microtubules and microfibril angle. InMicrofibril Angle in Wood (ed. B.G. Butterfield), University of Canterbury, New Zealand,pp. 253–371.

Bentum, A.L.K., Côté, W.A., Day, A.C. & Timell, T.E. (1969) Distribution of lignin in normal andtension wood. Wood Science and Technology, 3, 218–231.

Boyd, J.D. (1977) Basic causes of differentiation of tension wood and compression wood. AustralianForestry Research, 7, 121–143.

Boyd, J.D. (1978) Significance of laricinan in compression wood tracheids. Wood Science andTechnology, 12, 25–35.

Brodski, P. (1972) Callose in compression wood tracheids. Acta. Soc. Bot. Pol., 41, 321–327. Casperson, G. (1961a) Über die Bildung von Zellwänden bei Laubholzern. 2. Mitt. Der zeitliche Ablauf

der Sekundärwandbildung. Zeitsachruft für Botanik, 49, 289–306. Casperson, G. (1961b) Licht und elektonmikroskopische Untersuchungen über den zeitlichen Ablauf der

Zellwandbildung bei Laubhölzern. Berichte der Deutsche Botanische geselleschaft, 74, 271–273. Clair, B. (2001) Etude des proprietes mecaniques et du retrait au sechage du bois a l’ echelle de la paroi

cellulaire: essai de comprehension du comportement macroscopique paradoxal du bois de tensiona couche gelatineuse. Doctorate Thesis, Universite de Montpellier II.

Clair, B., Ramonda, M., Gril, J. & Thibaut, B. (2001) Longitudinal shrinkage in gelatinous layer: fromobservation to modelling. First International Conference of the European Society for WoodMechanics. Lausanne, 503–509.

Clair, B. & Thibaut, B. (2001) Shrinkage of the gelatinous layer of poplar and beech tension wood.JAWA J, 22 (2), 121–131.

Côté, W.A., Day, A.C. & Timell, T.E. (1969) A contribution to the ultrastructure of tension woodfibres. Wood Science and Technology, 3, 257–271.

Coutand, C., Jeronimidis, G., Chanson, B. & Loup, C. (2003) Comparison of mechanical properties oftension and normal wood in Populus. Wood Science and Technology.

Emons, A.M.C., Derksen, J. & Sassen, M.M.A. (1992) Do microtubules orient plant cell wall microfi-brils? Physiologia Plantarum, 84, 486–493.

Ewart, A.C.J. & Mason-Jones, A.G. (1906) Formation of red wood in conifers. Annals of Botany, 20,201–204.

REACTION WOOD 135

Faruya, N., Takahashi, S. & Miazaki, H. (1970) The chemical composition of the gelatinous layer fromthe tension wood of Populus euroamericana. Journal Japanese Wood Research Society, 20, 26–30.

Fisher, J.B. & Stevenson, J.W. (1981) Occurrence of reaction wood in branches of dicotyledons and itsrole in tree architecture. Botanical Gazette, 142, 82–95.

Fujita, M., Sakai, H. & Harada, H. (1974) Electron microscopy of microtubules and cellulose microfibrilsin secondary wall formation of poplar tension wood. Mokuzai Gakkaishi, 20, 147–156.

Furosawa, O., Funada, R., Murakami, Y. & Ohtani, J. (1998) Arrangement of cortical microtubules incompression wood tracheids of Taxus cuspidate visualized by confocal laser microscopy. Journalof Wood Science, 44, 230–233.

Hoffmann, G.C. & Timell, T.E. (1970) Isolation of a β 1-3 glucan (laticinan) from compression woodof Larix laricinia. Wood Science and Technology, 4, 159–162.

Hoffmann, G.C. & Timell, T.E. (1972) Polysaccharides in compression wood of tamarack (Larix laricina).1. Isolation and characterization of laricinan, an acidic glucan. Svensk Papperstidning, 75, 135–141.

Höster, H.R. & Liese, W. (1966) Über das Vorkommen von Reaktionsgewebe in Wurzeln und Ästender Dikotyledonen. Holzforschung, 20, 80–90.

Huang, X. & Jeronimidis, G. (2002) Comparison of fracture toughness between reaction wood and normalwood from transgenic Poplar plants with modified lignification, in Proceedings of the FirstConference of the European Society for Wood Mechanics (ed. P. Navi), EPFL, Lausanne,pp. 277–284.

Illston, J.M., Dinwoodie, J.M. & Smith, A.A. (1979) Concrete, Timber and Metals – The Nature andBehaviour of Structural Materials. Van Nostrand Reinhold. New York.

Jaccard, P. (1938) Exzentrishces Dickenwachstum und anatomisch-histologisches Differenzierung desHolzes. Berichtes der Schwiezes Botanisches Geselleschaft Zürichi, 48, 491–537.

Jacquiot, C. & Trenard, J. (1974) Note sur la présence de trachéides à parois gélatineuses dans des boisrésineux. Holzforschung, 28, 73–76.

Kwon, M., Bedgar, D.L., Piastuch, W., Davin, L.B. & Lewis, N.G. (2001) Induced compression woodformation in Douglas fir (Pseudotsuga menziesii) in microgravity. Phytochemistry, 57, 847–857.

Lachaud, S. (1987) Xylogenese chez les dicotyledons arborescentes V. Formation du bois de tensionet transport de l’acide indole acétique tritie chez le hêtre. Canadian Journal of Botany, 65, 1253–1258.

Little, C.H.A. & Eklund, L. (1999) Ethylene in relation to compression wood formation in Abiesbalsamea shoots. Trees – Structure and Function, 13, 173–177.

McDougall, G.J. (2000) A comparison of proteins from the developing xylem of compression andnon-compression wood of branches of Sitka spruce (Picea sitchensis) reveals a differentiallyexpressed laccase. Journal of Experimental Botany, 51, 1395–1401.

Mellerowicz, E.J., Baucher, M., Sundberg, B. & Boerjan, W. (2001) Unravelling cell wall formation inthe woody dicot stem. Plant Molecular Biology, 47, 239–274.

Münch, E. (1940) Weitere Untersuchungen über Druckholz und Zugholz. Flora, Jena 34, 45–57. Nedesaný, V. (1958) Effect of β-indoleacetic acid on the formation of reaction wood. Phyton, 11, 117–127. Nobushi, T. & Fujita, M. (1972) Cytological structure of differentiating tension wood fibres of Populus

euamericana. Mokuzai Gakkaishi, 18, 137–144. Norberg, P.H. & Meier, H. (1966) Physical and chemical properties of the gelatinous layer in tension

wood fibres of aspen (Populus tremula L.). Holzforschung, 20, 174–178. Ollinmaa, P.J. (1955) On the anatomic structure and properties of the tension wood in birch. Acta

Forest Fenn, 64, 1–263. Pillow, M.Y. & Luxford, R.F. (1937) Structure, occurrence and properties of compression wood. USDA

Technical Bulletin, no. 546, p. 32. Plomion, C., Pionneau, C., Brach, J., Costa, P. & Bailleres, H. (2000) Compression wood responsive

proteins in developing xylem of maritime pine (Pinus pinaster Ait.). Plant Physiology, 123, 959–969. Prodhan, A.K.M.A., Funada, R., Ohtani, J., Abe, H. & Fukazawa, K. (1995) Orientation of microfibrils

and microtubules in developing tension wood fibers of Japanese ash (Fraxinus mandshurica Var.Japonica). Planta, 196, 577–585.


Sassus, F. (1988) Deformations de maturation et proprietes du bois de tension chel le hetre et lepeuplier: measures et modeles. Doctorate Thesis. ENGREF, Paris.

Singh, A.P. & Donaldson, L.A. (1999) Ultrastructure of tracheid cell walls in radiata pine (Pinus radiata)mild compression wood. Canadian Journal of Botany, 77, 32–40.

Singh, A.P., Sell, J., Schmitt, U., Zimmermann, T. & Dawson, B. (1998) Radial striation of the S2 layerin mild compression wood tracheids of Pinus radiata. Holzforschung, 52, 563–566.

Spicer, R. & Gartner, B.L. (1998) Hydraulic properties of Douglas fir (Pseudotsuga menziesii)branches and branch halves with reference to compression wood. Tree Physiology, 18, 777–784.

Sundberg, B., Tuominen, H. & Little, C.H.A. (1994) Effect of the indole-3-acetic acid (IAA) transportinhibitors N-1-naphthylphthalamic acid and morphactin on endogenous IAA dynamics in relation tocompression wood formation in 1-year-old Pinus sylvestris shoots. Plant Physiology, 106, 469–476.

Takabe, K., Miyauchi, T. & Fukazawa, K. (1992) Cell wall formation of compression wood in Todo fir(Abies saccharensis). 1. Deposition of polysaccharides. IAWA Bulletin, 13, 283–296.

Telewski, F.W. (1988) Intra-annual spiral compression wood: a record of low frequency gravitropiccircumnutational movement in trees. IAWA Journal, 9, 269–274.

Timell, T.E. (1969) The chemical composition of tension wood. Svensk Papperstidning, 72, 173–181. Timell, T.E. (1973) Studies on the opposite wood in conifers. Part II. Histology and Ultrastructure.

Wood Science and Technology, 7, 79–91. Timell, T.E. (1980) Karl Gustav Sanio and the first scientific description of compression wood. IAWA

Bulletin, 1, 147–153. Timell, T.E. (1983) Origin and evolution of compression wood. Holzforschung, 37, 1–10. Timell, T.E. (1986) Compression wood in gymnosperms. Vols I–III. Springer. Berlin, Heidelberg, New York.Tomlinson, P.B. (2001) Reaction tissues in Gnetum Gnomen: a preliminary report. IAWA Journal, 22, 401–413.United States Department of Agriculture (1974) Wood Handbook: Wood as an Engineering Material,

Forest Products Laboratory Handbook No. 72, US Government Printing Office, Washington, DC.Wardrop, A.B. (1964) The Reaction Anatomy of Arborescent Angiosperms. In The Formation of Wood

in Forest Trees (ed. M.H. Zimmermann). Academic Press, New York, pp. 405–456. Wardrop, A.B. & Dadswell, H.E. (1950) The nature of reaction wood II. Cell wall organisation of

compression wood tracheids. Australian Journal of Scientific Research, B3, 1–13. Wardrop, A.B. & Davis, G.W. (1964) The nature of reaction wood. VIII. The structure and differentia-

tion of compression wood. Australian Journal of Botany, 12, 24–38. Wershing, H.T. & Bailey, I.W. (1942) Seedlings as experimental material in the study of “redwood” in

conifers. Journal of Forestry, 40, 411–414. Westing, A.H. (1965) Formation and function of compression wood in gymnosperms. Botanical

reviews, 31, 381–480. Westing, A.H. (1968) Formation and function of compression wood in gymnosperms. II. Botanical

reviews, 34, 51–78. Wilson, B.F. (1981) The development of growth strains and stresses in reaction wood. In Xylem Cell

Development (ed. J.R. Barnett), Castle House Publications, Tunbridge Wells, Kent, pp. 275–290. Wilson, B.F., Ching-Te, C. & Zaerr, J.B. (1989) Distribution of endogenous indole-3-acetic acid and com-

pression wood formation in reoriented branches of Douglas-fir. Plant Physiology, 91, 338–344. Yoshizawa, N., Matsumoto, S. & Idei, T. (1985) Morphological features of tracheid tips associated

with compression wood formation in Larix leptolepis Gord. IAWA Journal, 6, 245–253. Yoshizawa, N., Kiyomiya, M. & Idei, T. (1987) Variations in tracheid length and morphological

changes in tracheid tips associated with the development of compression wood. Wood Scienceand Technology, 21, 1–10.

Yoshizawa, N., Watanabe, N., Yokota, S. & Idei, T. (1993) Distribution of guaiacyl and syringyllignins in normal and compression wood of Buxus microphylla var. insularis Nakai. IAWAJournal, 14, 139–151.

Zandomeni, K. & Schopfer, P. (1994) Microsensory microtubule reorientation in the epidermis ofmaize coleoptiles subjected to bending stress. Protoplasma, 182, 96–101.

6 Growth stresses Bernard Thibaut and Joseph Gril

6.1 Origin of growth stresses

The term growth stresses refers to residual, self-equilibrated stresses within aportion of trunk or branch (designated in this chapter by the common term ofstem), resulting at any given time from the whole growth history of that element.Growth stresses can result from external or internal actions and can express passiveor active mechanical actions of the tree.

6.1.1 Geometric and mass growth: support stress

Due to the action of gravity, each portion of stem must support the weightlocated above it (i.e. including buds and leaves). During tree growth, both the amountof supported weight and the supporting geometry (section, moment of inertia) arecontinually and simultaneously changing. Within a given transverse section,the stress distribution can be evaluated using two basic assumptions (a) a newwood layer contributes to the support of parts above it only after it has beenproduced by cambial activity; (b) the resultant of axial stress acting on a givensection is at all times equal to the axial load and bending moments resultingfrom the gravitational action of upper parts. A step-by-step calculation makes itpossible to compute the stress distribution from pith to bark through the sectionof a vertical or inclined stem, provided that the history of gravitational loading,geometry and material properties is fully known (Fournier etal., 1991a). It shouldbe emphasised that the stress distributions resulting from such progressivegrowth differ strongly from that which would be observed if the gravitationalfield was suddenly applied on an existing stress-free structure (Fig. 6.1).

6.1.2 Cell differentiation: maturation stress

In the first stages of cell development (division, elongation), the cell wall is verysoft and cannot support high stress levels. During the maturation stage, the cellwall of fibres or tracheids thickens and hardens with the deposition of cellulosicmicrofibrils embedded in a hemicellulose and lignin matrix. The polymerisationprocess that leads to higher rigidity also induces small strains, usually axialshrinkage and transverse expansion of the fibre; their order of magnitudeis typically 1/1000 to 1/100, to be compared to initial elongations exceedinga value of 1 relative to the initial length of the cambial cells. Most of these


maturation strains are restrained by the rigid solid wood already present insidethe stem. The new maturation layer sticks on this rigid core, just like a wet paperglued on a wall and, hence, the axial maturation shrinkage leads to a longitudinaltensile stress at the end of the process; similarly, the restraint of the transverseexpansion leads to a compressive tangential stress.

This initial loading process of the material can be expressed as: ε = Sσ0 + α,where ε is the deformation of the newly formed wood during the maturationprocess; σ0, the maturation stress acting on the wood at completion of theprocess; S, the compliance tensor and α the maturation strain.1 The restraint inthe tangential plane at the stem periphery is expressed by the tangential andlongitudinal components of the strain tensor nearly equating zero (εT ≈ εL ≈ 0)and, assuming no pressure from the bark, we have a zero radial stress: σ 0

R ≈ 0.Hence, the relationship between the maturation strain and the correspondingmaturation stress is:

; (6.1)

where EL and ET are Young’s moduli in the longitudinal and tangential dir-ections, respectively, and νTL and νLT are the Poisson’s ratios.

1 This is, however, a simplified view of the problem, where the material is assumed to behave elasticallyand the maturation strain to be of the same nature as a recoverable thermal deformation. The maturationinvolves a lot of biochemical reactions, and occurs by the progressive deposition of cell-wall sub-layerstogether with the cell loading, so that a considerable amount of the deformation induced by maturationis absorbed into irrecoverable strain. The maturation strain mentioned in the text must be understood asthe recoverable part of the total strain.

D

F

FabricationLo

adin

g

D

F

Fabrication +loading

F

D

F(t )

D(t )

D

M(t )

M

D(t )

x

x y

y

(a)

(b)

σ

σ

σ

σ

Fig. 6.1 Longitudinal stress profile along the diameter of a column in compression or a beam inbending: (a) column or beam loaded after being fabricated; (b) growing column or beam loaded duringfabrication (new growth in grey over old growth in white).

αT

σT0 vTLσL

0–ET

---------------------------–= αL

σL0

vLTσT0–

EL

---------------------------σL

0

EL

------≈–=

GROWTH STRESSES 139

Thus, the opposite of the maturation stress is roughly equal, in the longitud-inal direction, to the product of the maturation strain and the Young’s modulus(Fig. 6.2). If the values of both Young’s modulus and maturation strain areknown, one can again calculate, by the same step-by-step approach as above,the internal stress distribution in a stem (Archer, 1986; Kübler, 1987; Fournieret al., 1991b).

6.1.3 Growth stresses

The stress distribution in a stem is the sum of both support and maturation stressfields (Fig. 6.3). It should be noted that: (a) contrary to maturation stresses,support stresses are strongly dependent on the position of the stem portion

Peripheralgrowth stress

Maturationstrain

Fig. 6.2 Maturation strain and maturation stress.

Long

itudi

nal s

tres

s

Relative radial position from pith to bark (r /R )

+ =

20 20 20

– 80 – 80

– 40

1 10 1 10

1 10

Maturation stress support stress total growth stress =+

Fig. 6.3 Growth stresses resulting from the superposition of maturation stress and support stress.


within the whole plant, as for instance, the load supported by bottom parts ishigher than that of upper parts; (b) the longitudinal component of supportstresses is generally compressive, in vertical trunks at least, whereas for matur-ation stresses it changes from tension near the bark to compression near thepith; (c) the initial tensile maturation stress generated in outer rings is usuallyone order of magnitude higher than the highest possible compressive supportstress in a vertical trunk, so that the total growth stress distribution is veryclose to that of the maturation stress alone. For that reason, growth stresses areoften considered as residual stresses since they would not be very much modifiedby removing all external action acting on the structure.

6.1.4 Role of growth stresses

One advantage of growth stresses for tree stems is to improve their strength orflexibility. Wood is strongest in axial tension or transverse compression, andweakest in axial compression and transverse tension. The stress field pre-existingin a standing stem protects the external layers against the effect of bending thatwould induce axial compression together with transverse elongation on one side.

But the most obvious advantage of growth stresses is the control of stemshape and orientation through the fine tuning of growth stress intensity asso-ciated with wood production. A high level of tensile stress systematicallyproduced on one side of a stem, for instance, will produce a bending momentable to counteract the effect of gravity, reorientate towards a light openingor fulfil any biomechanical requirement of the plant. Without such amechanism, no tree could maintain straight stems or non-weeping branches(Wilson & Archer, 1979; Fournier et al., 1994b; Thibaut et al., 2001; Almeraset al., 2002).

6.1.5 General models of growth stresses

Provided sound hypotheses are made for the geometrical and mass growth of thetree, peripheral distribution of maturation strains and constitutive equations ofthe material, it is possible to calculate, at each stage of the tree development, thegrowth stress distribution. The best-known model is that proposed as early as1959 by Kübler in the case of a vertical and axisymmetric trunk made of a homo-genous and transversally isotropic material subjected to a constant maturationstress and no gravitational field (see in Kübler, 1987):

σL (r;R) = [1 + 2 ln (r/R)]

σR (r;R) = [ln (r/R)] (6.2)

σT (r;R) = [1 + ln (r/R)]

σL0

σT0

σT0

GROWTH STRESSES 141

where L, R, T refer to the longitudinal, radial and tangential directions,respectively.

Numerous improvements have been proposed, for instance by Archer (1986),who considered the orthotropic case, and by Fournier’s semi-analyticalcomputations based on the decomposition of the peripheral maturation stresses.

6.2 Measurement of growth stresses

As a general rule, it is not possible to measure directly the stresses in a materialbut only their consequences. The most widely used method consists ofmeasuring the strains resulting from their local suppression by various cuttingtechniques (sawing, drilling).

6.2.1 In situ peripheral measurement

These measurements are performed on the standing tree (sometimes on thetrunk after felling, provided there were no severe cracks), after local debark-ing. Two main techniques can be used for measuring the longitudinal locked-in strains: (i) strain gauges glued directly on tree surface with isocyanate glue;or (ii) transducers measuring the distance between two pins hammered into thewood surface. The gauges necessitate a battery and an electronic device meas-uring the variation of their electrical resistance, directly proportional to thestrain; if necessary, the tangential strain can be obtained as well, at the samelocation. To reach a sufficient degree of precision (1 μm), the transducers mustbe left unmoved during cutting operations. These can also be classified in twogroups: (i) groove sawing perpendicular to fibre directions, usually to a depthof few millimetres on both sides of the gauge or transducer (and in case oftangential measurement also along the fibres on tangential sides); (ii) holedrilling, either two holes on both sides of the gauge or transducer, or one singlehole with gauges or transducer pins on both sides (Nicholson, 1971; Saurat &Guéneau, 1976; Archer, 1986; Fournier etal., 1994a; Loup etal., 2001; Yoshida& Okuyama, 2002), as shown in Fig. 6.4.

In all cases, the effect of the cutting operation is measured either by a localstrain, or a local variation of distance. Usually, strains are measured in percentageor microstrains, με (1 με = 10−6 m/m), and variations of distance in μm.The term growth stress indicator or GSI may be used for these values meas-ured directly in the field. Mechanical analysis of this operation has shown that,with a very good approximation, GSI is proportional to the longitudinalmaturation strain of recently formed wood layers; the coefficient relating GSIand peripheral growth stress can also be computed, provided that the elasticmoduli of green wood are known.


6.2.2 Measurement of residual stresses in logs

For technological purpose, it is of great importance to know the residualstresses inside a log after felling and crosscutting. This can be achieved bysuccessive cutting operations associated with many measurements of strain orrelative distance variations at chosen positions, permitting backward calcula-tions to estimate the released stresses able to produce them. Sophisticatedexperiments such as those developed by Jacobs (1945) or Chardin (1982) (Fig. 6.5)gave good agreement with general models; they confirmed the common occur-rence of non-axisymmetric distributions.

Although for most tree biomechanics applications related to growth, move-ment and stability of stems, it is sufficient to only estimate the longitudinalcomponent of growth stresses, for industrial applications, the knowledge of thetransverse components can be important. In situ measurements of tangentialgrowth strains at stem periphery have been done using the above-describedmethods, using 2-directional strain gauges or transducers, but the final precisionis less than that of longitudinal measurements. A good evaluation of the globallevel of transverse locked-in strains can be obtained using disks a few centime-tres thick subjected to a V-shape cutting (Jullien & Gril, 1996 and Fig. 6.6).

6.2.3 Main results for normal maturation strain

A great number of in situ measurements have been performed on standingtrees using the different methods (Saurat & Guéneau, 1976; Chardin, 1982;Ferrand, 1982; Fournier et al., 1994a; Yamamoto et al., 1995) either on soft-woods (Pinus, Cryptomeria, . . .), temperate hardwoods (Fagus, Castanea,

Ø20 mm

T

L

Hole

Frame

Pin

Fig. 6.4 The single-hole method for measurement of longitudinal residual strain at stem periphery.

GROWTH STRESSES 143

E

N

S

W

Strain (10–6)

Distance from pith (mm)

Fig. 6.5 Measurement of growth stress in the volume of a log of Wallaba (Eperua falacata Aubl.) afterChardin (1982). Left: longitudinal strips are sawn successively, and their shortening is recorded. Topright: radial profile of apparent residual strains. Bottom right: reconstitution of the initial growth straindistribution taking into account the stress redistribution at each extraction.

Fig. 6.6 Measurement of transverse locked-in strains in a disk. Left: the sector extraction provokes theclosing of the opening, related to the global level of locked-in strain. Right: in addition to the globalclosing, a number of local deformation can be measured between pins nailed on the disk.


Populus, . . .) or tropical hardwoods (Eucalyptus hybrids, Eperua, . . .). Oftenfour to eight measurements around the trunk at breast height were made tocheck the degree of dissymmetry of the peripheral distribution, and in someinstances, the effect of height has been investigated. The hole drilling methoddeveloped by CIRAD, for instance, has been used in routine measurements onaround 1000 trees, yielding a GSI databank of about 10 000 measurements thatgives some insight into the distribution of their values and factors influencinggrowth stress levels in the trunks of standing trees.

Maturation strains can be calculated by multiplying the CIRAD GSI expressedin micrometres by a typical factor of −12 × 10−6 (or −12 if the strain isexpressed in με). The multiplying factor actually ranges from −9 × 10−6 forlow-density woods with high degree of anisotropy to −15 × 10−6 for the lessanisotropic, very dense woods. Examination of histograms of this GSI valueshows, for any tree population (either a given species, or a sub-population ofthat species) a normal GSI level of around +30 to +70μm (or −0.04 to −0.08%),corresponding to a longitudinal tensile stress at the periphery of around+5MPa for poplar or pine but up to +25MPa for Wallaba from French Guyana.Interestingly, the normal tensile maturation strain does not vary markedlybetween species of very different density. On the contrary, as the longitudinalYoung’s modulus is more or less proportional to wood density, the same generaltrend is expected for the maturation stress.

Within a given species or group of hybrids, the normal level of maturationstress was found to depend on many factors: genotype in eucalypt or poplarhybrids; growth rate in eucalypt or chestnut, where very slow growth was associ-ated with lower strains (Gérard, 1994; Gril etal., 1995; Loup etal., 2001); treemorphology (which itself depends on sylviculture) in beech where a high ratioof tree height to trunk diameter at breast height was associated to higher strains(Loup etal., 2001). Tree age is a complex factor because of combined influenceof tree history and morphology. For young fast-growing hardwoods such aseucalypt, poplar or chestnut, constant levels of normal strain were observed,whereas in young softwoods, the normal-strain pattern seemed to consist ofcompressive values in the juvenile stage, shifting to usual tensile values in theadult stage.

6.2.4 Growth stresses and reaction wood

In hardwoods, peripheral growth strains can fall to nearly zero, or very seldomto slightly positive values like +0.01% for chestnut or beech, corresponding toa compressive growth stress of −1 to −2 MPa. But, they can reach very highnegative values, down to −0.4% corresponding to a tensile stress of +35 MPa(poplar) or +120 MPa (wallaba). These high values appear in the tail of histo-grams such as the one shown in Fig. 6.7. For softwoods, the lowest values

GROWTH STRESSES 145

rarely exceed −0.1% (+15 MPa for Pine) but can rise to +0.2%, correspondingto a compression stress of −20 MPa for Pine.

Highest and lowest growth strain/stress values are usually observed in thesame trees of each species. In such cases of highly dissymmetrical distribu-tions, typically a peak of very high tensile (in hardwood) or compressive (insoftwood) stress is observed on the upper (hardwoods) or lower (softwoods) sideof the slightly or markedly leaning stem. This is known to relate to reactionwood formation; there is a clear correspondence between reaction wood occur-rence, in the anatomical sense, and reaction values of the growth stress:compression wood produces a compressive stress, and tension wood an unusuallyhigh tensile stress (Trénard & Guéneau, 1975; Fournier-Djimbi et al., 1997;Yamamoto et al., 2002). The rigidity of tension wood can be higher than thatof normal wood, but often not by a large amount. Compression wood, by contrast,is generally much less rigid, typically by a factor 2 or 3, even though it is oftendenser. Thus, from the biomechanical point of view, tension wood is a moreefficient way to create dissymmetrical stress for a given amount of woody matter.This explains why growth eccentricity is much more common in softwoods, whereit is required to compensate for the lower rigidity of the compression woodside, than in hardwoods.

The association between dissymmetrical stress distributions and reactionwood formation legitimises the concept of reaction stress as applied to suchphenomena. Hence a complementary and even more universal definition ofreaction wood in a wood species: one whose differentiation generates a reac-tion stress, or, more precisely, a reaction maturation strain. Reaction woodformation and consecutive dissymmetrical growth stress serve mechanicalpurposes such as counteracting a creeping tendency of inclined stems, allowinga rapid change of stem orientation to restore verticality (negative-gravitropism),search for light (phototropism), or achieve a programmed architectural sequencesuch as plagiotropism of secondary axes, organisation of forks or restoration ofa main axis after the death of terminal bud. Thus, border trees accidentallyinclined by wind or snow action, or partly dominated trees, typically exhibit a

Fagus sylvatica10 stands 545 trees

0

1500

0 300GSI (microns)

Num

ber

of d

ata

Tension wood

Normalwood

Fig. 6.7 Histograms of growth stress indicator.


marked peak of reaction stress. However, it should be emphasised that steminclination does not imply necessarily the production of reaction wood: coppiceshoots, although individually inclined and supporting a dissymmetric crown,do not contain reaction wood, provided that the whole stump remains equili-brated (Fig. 6.8).

E

S

N

1

2

34

5

0 360°0

400

Peripheral position

0 360°0

400

0 360°0

400

Str

ain

(μm

)

Str

ain

(μm

)

Str

ain

(μm

)

Str

ain

(μm

)S

trai

n (μ

m)

Str

ain

(μm

)

0 360°0

400

0 360°0

400

0 360°0

400

1

2

3

4

5

6

W 6

Fig. 6.8 Per-tree profile of surface longitudinal maturation strain measured at eight points at stemperiphery (from 0° to 315°), on six shoots from the same stump in a chestnut coppice. Shoots 3, 4 and5 were situated in the shadow of tall neighbouring oaks, so that they searched light and producedreaction stress although they were closer to verticality. By contrast, shoots 1, 2 and 6, althoughnormally inclined in the stump, expressed no reaction stress.

GROWTH STRESSES 147

6.3 Consequence of growth stresses for quality

The timber industry has long known the problems created by growth stresses,and most of the studies on the subject originated in response to the qualityissues raised (Guéneau, 1973; Kübler, 1987; Baillères et al., 1996).

6.3.1 Log-end cracks

While a log, destined to be processed by industry, was still a part of the trunk ofa tree, it was subjected to a growth stress field characterised typically by highlongitudinal tension near the bark changing gradually to high compression nearthe pith (Fig. 6.3), and transverse tension near the pith falling gradually toperipheral zero radial stress and tangential compression. Consequently, thisvolume of matter contained initially a considerable initial amount of stored elasticenergy (W i), roughly proportional to the product of longitudinal Young’s modulus(EL), square of longitudinal maturation strain (αL) and log volume (V):2

W i∝ELαL2 V (6.3)

Timber processing will modify this initial stress distribution, and sometimesprovokes a brutal release of the stored elastic energy. This is true for all typesof cutting operation (felling, crosscutting, sawing) as well as some heatingoperations such as log steaming.

After tree felling and crosscutting, the remaining log is almost free fromexternal loads except for the negligible effect of self-weight; it is, however,subjected to considerable self-equilibrating stresses. This residual stress fieldis generally close to the growth stress field, pre-existing in the standing tree (aswas explained above with maturation stress dominating support stress invertical stems), except for both log ends where the creation of a free surfaceexacerbates the transverse tension, pulls the heart and compresses the periphery.In many cases, small cracks are initiated near the pith as a result of thesetransverse tensions. The crack may then propagate into more or less numerousand far-reaching radial cracks, depending on the energetic balance betweenelastic energy release resulting from stress redistribution allowed by free surfacecreation, and the amount of energy required for matter separation. This canbe expressed in a simple way using the Griffith theory for brittle elasticmaterials:

dWe + GcdA < 0 or G = −dWe/dA > Gc (6.4)

2 This simplified equation can be obtained assuming constant maturation strain, constant rigidity and noviscoelastic relaxation, so that the pattern of stress distribution remains unchanged during stem growth,as well as neglecting the contribution of transverse components of the stress.


where dWe is the variation of elastic energy stored in the whole log for a givenvariation of crack area, dA; G, the elastic energy release rate and Gc, the materialtoughness.

The elastic energy release rate, G, is a structural quantity depending on thewood behaviour, log geometry and crack shape and extension; it typicallyreaches a maximum at the beginning of crack propagation and a minimumwhere it approaches the periphery. Gc, on the contrary, is an intrinsic woodproperty that only depends on the local mode and direction of crack propagation.In some tree populations, where the toughness corresponding to tangentialdelamination is unusually low (e.g. chestnut), ringshakes may also develop asan alternative to radial cracks.

Mechanical analysis, using numerical finite-element methods and confirmedby numerous field observations, has shown the following results for the crackingrisk at log end (as quantified, for instance, by the ratio of maximum cracklength to log radius): (i) the risk is considerable in hardwoods containing largeamounts of tension wood characterised by much higher locked-in strain andslightly more rigidity; (ii) it is generally small in softwoods – compressionwood in particular does not present additional problems because of its lowerrigidity and its negative growth stress which are not much higher in absolutevalue than that of normal wood; (iii) the risk tends to increase with log diameter,as can be predicted by the fact that the crack surface is proportional to thediameter, while the amount of initially stored energy (as shown by Equation 6.3)varies like its square (Jullien & Gril, 2002).

The viscoelastic nature of wood, not taken into account in the above ana-lysis, has two contradictory effects. On the one hand, it produces additionaldissipation that tends to stabilise the propagation. Indeed, cracks running alongthe whole log are rather rare – although they do exist, see Fig. 6.9. On the otherhand, a proportion of the growth strains that have been locked-in as viscouselastic strains are gradually recovered as a delayed consequence of stress

Fig. 6.9 Log-end cracks in a beech log.

GROWTH STRESSES 149

removal: viscoelastic locked-in strains are converted into elastic locked-instrains, so that the amount of elastic energy actually stored in the log structuregradually increases, and cracks develop further. This phenomenon is dramati-cally accelerated by log heating or steaming, because viscoelasticity is athermally activated process: owing to the so-called hygrothermal recovery ofgrowth strains, log heating above the glass transition of lignin (60–70°C)induces a considerable development of end cracks (Kübler, 1987; Gril &Thibaut, 1994; Thibaut et al., 1995).

6.3.2 Lumber distortion

Cutting operations are another means of creating free surfaces in the structure,but in a controlled way. They also result in partial stress release and deforma-tion. Lumber distortion (or warp) is a serious problem for sawmills, some of itoccurring before any drying, and can be explained as an expression of growthstress release; in particular, spring (or crook) and bow during sawing resultfrom the gradient of longitudinal residual strains in the living stem. After thefirst flat-sawn pieces have been cut, not only do they bow, but the remaininglog tends also to bend, resulting in irregular thickness of the next pieces cut(Fig. 6.10). All these deformations can be predicted numerically using data ongeometry, material rigidity and initial growth stress distribution. Sawing

Fig. 6.10 Lumber distortion during sawing.


operations do not allow full release of growth stresses. This can be easily verifiedwhen boards are re-cut and express further distortion, or by the spontaneoussplitting of quarter-sawn pieces containing the pith or close to it, a phenomenonof the same nature as the log-end splitting described previously.

Log-end cracks or board distortion after sawing can be considered as thedirect consequences of growth stresses. The correlation between growth stresslevel and crack extension or board deformation is always highly significant forany population of trees with a large range of growth stress levels. But, therelationship itself depends on parameters not necessarily related to growthstresses, so that caution is necessary to avoid misinterpretation of results. Asan example, the occurrence of ringshakes in chestnut, instead of radial splitting,is governed by factors completely independent of the growth stress level (Fontiet al., 2002). The log-end cracks are probably the worst consequence of growthstresses because of the danger they represent for wood cutters who can bekilled during the harvesting of big trees, and because they strongly affect therecovery and productivity of all primary conversion operations (sawing, rotaryand slicing veneer cutting). The distortion of boards or remaining log portionscan be spectacular in sawmills, and even dangerous in some cases; essentiallythese phenomena result in poor recovery. A good knowledge of the residualstress distribution, using computer assistance, could be useful to improve andoptimise the sawing patterns.

6.3.3 Reaction wood

In addition to their direct effect on wood processing, reaction stresses influenceindirectly wood quality through the existence of reaction wood. Reactionwood has properties differing from normal wood, that will be studied specifi-cally in Chapter 5 and only briefly evoked here.

Although, compressive reaction stresses in softwoods do not cause consid-erable log-end cracks, compression wood as such is generally the source ofmany drawbacks in the later stages of processing, essentially because of its veryparticular properties compared to normal wood (Timell, 1986): mainly a higherspecific gravity (up to 20% more) and lower MOE (down to 3–5 times lessthan normal) associated to a higher longitudinal shrinkage (up to 10 or 15 timesmore), lower transverse or volumetric shrinkage (down to 20% less), higherlignin and lower cellulose contents. As a result, boards containing compressionwood exhibit considerable problems during drying. Moreover, the usually largergrowth rings (up to 2 or 3 times more) may result in a very marked eccentricitycomplicating cutting operations.

On the contrary, tension wood which is almost always associated with moreconsiderable growth-stress induced problems, does not always cause troubledue to its different behaviour from normal wood although a higher longitudinalshrinkage has been often observed (up to 5 times more than normal wood).

GROWTH STRESSES 151

Figure 6.11 shows the variation of wood properties around the periphery of aninclined hardwood trunk. In some species, for instance in poplar and eucalypt, thelongitudinal shrinkage and the fuzzy grain are serious drawbacks; it is also aproblem in relation to particular end-uses, like the green vein of cherry tensionwood in furniture (Polge, 1984; Castéra et al., 1994; Baillères et al., 1995;Coutand et al., 2003).

6.4 Prediction and treatment

Methods to predict either growth stresses or their consequences are often askedfor by foresters or the timber industry. The demand may concern standing treesor felled stems in the forest, logs from the log yard or timber stock in woodfactories. Basically, a mechanical stress cannot be observed or directly measured;all methods should proceed from indirect relationships, which in the case ofgrowth stress can be classified into three main groups.

6.4.1 Tree and log morphology

From the known effect of growth stresses, and particularly that of reactionstresses, valuable information can be obtained through the acquisition ofmorphological data using various tools (drawing, digital picture with imageanalysis, distance measured by magnetic or laser beam, length, curvature ordiameter measured manually, etc.) at essentially two levels. First, the generalshape and position of the tree within the plot can be analysed based, for a givenspecies, on known correlation of growth stress level and reaction wood occurrence

Eperua falcata

0

1

0 360Peripheral position (°)

Val

ue r

elat

ive

to m

ax

GSIaLE/daTaR

Fig. 6.11 Variation of material properties around the periphery of an inclined hardwood. The tension woodzone is located around 180°; GSI = growth stress indicator according to CIRAD single-hole method; aL,aR, aT = total shrinkage in L, R, T directions; E/d = specific Young’s modulus in air-dry conditions.


with the dissymmetry of crown and shape, slenderness, crown amplitudeand tree vigour (Fig. 6.12). Second, the existence and position of reactionwood zones can be inferred from the observation of tree morphology (leaning,curvature, evidence of verticality restoration or phototropism, tree architec-ture) considered as resulting from an assumed history of maturation strains.However, there is a non-negligible risk of misinterpretation; for instance, an

Straight

Bending

curved at base one great

curve several

curves

mean

total

symmetry

84 118

40

100 121

4

133 158

11

69 96

10

94 101

10

91 119

75

asymmetry

104 138

34

113 128

18

102 129

6

85 102

13

109 140

19

104 131

90

large crown

eccentricity

158 206

2

68 90

8

202 215

1

0

120 141

1

98 124

12 means (large crowned trees)

nb of trees

95 129

76

99 117

30

127 152

18

78 99

23

105 127

30

98 125

177 96 137

50

123 142

6

94 117

7

114 144

17

119 147

24

106 139

104 107 141

43

134 159

15

125 145

9

137 165

24

116 155

15

120 151

106

eccentricity

86 109

9

122 146

10

105 122

10

130 191

3

115 132

7

109 132

39 means (narrow crowned trees)

nb of trees

100 136

102

128 151

31

109 128

26

128 159

44

117 147

46

112 143

249

max-minmaximum

tree nb

Total

98 133

178

114 134

61

116 138

44

111 138

67

112 139

76

107 136

426

crownshape

trunkshape

Means(all trees)

narrowcrown

asymmetry

symmetry

Fig. 6.12 Tree shape and growth stress level. Beech from ten European stands, including a largeproportion of risky stems. GSI in μm, single-hole CIRAD method.

GROWTH STRESSES 153

apparently vertical stem can be in the final phase of restoration of verticalityafter accidental leaning, and contains a large amount of reaction wood on oneside. A good knowledge of the growth patterns of the studied species isrequired for conducting such analysis.

6.4.2 Consequences of cutting operations

This type of semi-destructive method can be used for evaluating the stresslevel in trees to be felled shortly afterwards or that were just harvested and cutinto logs, or for the establishment of relationships between growth stresses andgrowth parameters such as genotype, soil, climate, sylviculture.

For standing trees, the groove or hole drilling methods as described abovecan be used. From long experience, it is known that the mean level of growthstress in a tree is highly correlated to the maximum level in the reacting sector,which in turn is well correlated to the consequences of growth stresses andreaction wood. It is generally possible to localise the reacting sector in a trunkby observing its inclination, which is nearly always existing and visible. Verygood correlations have been found between direction of stem inclination andthat of reaction stress, with compression wood on the lower side in softwood andtension wood on the upper part in hardwoods. However, poor correlationswere obtained between stem inclination and growth stress levels, so that theactual measurement of surface growth strain is absolutely necessary – but asubstantial economy of data acquisition can be gained by limiting measure-ments to the suspected portions of the stem, provided the results have beencalibrated for a given species and stand type.

For felled logs, observation of end cracks is highly instructive. It can bequantified in various ways, like the total or mean crack length, the length of thelongest crack, crack opening. A high amplitude of crack development isalways the signature of much stored elastic energy linked to a high level ofgrowth stresses. In the case of species susceptible to ringshake (e.g. chestnut),the early observation at log end is a good indication of later risk (Fonti et al.,2002). More or less rapid observation of cracks at both ends of logs can beused – after proper calibration – to separate risky logs from others beforeindustrial processing. At a later stage, the cracks can be used for choosing thebest sawing pattern, in order to limit the future drawbacks resulting from growthstress and/or reaction wood occurrence.

6.4.3 Observation of reaction wood

In some species (softwoods, chestnut, poplar, . . .), reaction wood zones can bemore or less identified by visual inspection, either on increment cores or logsection after crosscutting, through colour, reaction to light, ring width, colourreaction to simply handled chemicals, etc. In the case of extraction from thestanding tree, it is recommended to choose by visual assessment, as explained


above, the most risky side of the trunk. According to species, large sectors oftypical reaction wood are always linked to reaction stresses and to forthcomingproblems in industry, that can be addressed by log sorting or proper sawingpattern.

6.5 Conclusions

Whatever the progress in the knowledge and measurement technology, growthstresses and their consequences involve a complex chain of relations which arevery much species-dependent, and where strong human experience and ingenuityremain the main key to success.

There is no way to change reaction wood into normal wood, once it hasbeen produced; however, genetic control through selection or more advancedtechnologies is possible in order to reduce its occurrence or its most negativefeatures, once foresters or tree breeders have been made aware of itsimportance.

Techniques are available to ameliorate the direct effects of growth stresses.If the cracks are prevented from occurring, the restraint imposed to the mater-ial is compensated by a fall of the residual stresses. This phenomenon knownas stress relaxation originates in the material viscoelasticity; it is greatly ampli-fied by temperature increase, so that log heating, known to cause the worst oflog-end cracks (see end of Section 3.1), can also serve to reduce their impact;irreversible chemical changes might contribute to their effectiveness. The mostobvious method for preventing the actual crack propagation consists in ham-mering S-shaped fasteners over starting cracks or in girdling the log ends. Ifthe cracked ends are removed after stress relaxation has been allowed, theremaining portion of the log will exhibit much less cracking risk.

References

Almeras, T., Gril, J. & Costes, E. (2002) Bending of apricot-tree branches under the weight of axillaryproductions: test of a mechanical model with experimental data. Trees – Structure and Function,16, 5–15.

Archer, R.R. (1986) Growth stresses and strains in trees, in Springer Series in Wood Science (ed.E. Timell), Springer-Verlag, Berlin, Heidelberg, New York, Tokyo.

Baillères, H., Chanson, B., Fournier, M., Tollier, M.T. & Monties, B. (1995) Structure, compositionchimique et retraits de maturation du bois chez les clones d’eucalyptus. Annales des SciencesForestières, 52, 157–172.

Baillères, H., Fournier-Djimbi, M., Gérard, J. & Thibaut, B. (1996) Growth stresses, end splitting andsawing distorsions in Eucalypts from plantation, in International Conference on Wood Mechanics(ed. S. Aicher), FMPA-Otto-Graf-Institut, Stuttgart, pp. 483–494.

Castéra, P., Nepveu, G., Mahé, F. & Valentin, G. (1994) A study on growth stresses, tension wooddistribution and other related defects in poplar (Populus euramericana cv. I214): end splits,specific gravity and pulp yield. Annales des Sciences Forestières, 51, 301–313.

GROWTH STRESSES 155

Chardin, A. (1982) Les contraintes de croissance du Wapa. Etude en vue de déterminer les condi-tions technologiques d’utilisation du Wapa, essence principale de Guyane. Final report of con-tract No 78-2-34-0535, French Ministry of Industry and Research, Centre Technique ForestierTropical.

Coutand, C., Jeronimidis, G., Chanson, B. & Loup, C. (2003) Comparison of mechanical properties oftension and normal wood in Populus, to be published in Wood Science and Technology.

Ferrand, J.C. (1982) Etude des contraintes de croissance. Deuxième partie variabilité en forêt descontraintes de croissances du hêtre (Fagus sylvatica L.). Annales des Sciences Forestières, 51,329–338.

Fonti, P., Macchioni, M. & Thibaut, B. (2002) Ring shake in chestnut (Castanea sativa Mill.): State ofthe art. Annals of Forest Science, 59, 129–140.

Fournier, M., Chanson, B., Guitard, D. & Thibaut, B. (1991a) Mécanique de l’arbre sur pied:modélisation d’une structure en croissance soumise à des chargements permanents, évolutifs.Partie 1: application à l’analyse des contraintes de support. Annales des Sciences Forestières,48, 513–526.

Fournier, M., Chanson, B., Thibaut, B. & Guitard, D. (1991b) Mécanique de l’arbre sur pied: modéli-sation d’une structure en croissance soumise à des chargements permanents, évolutifs. Partie 2:application à l’analyse tridimensionnelle des contraintes de maturation. Annales des SciencesForestières, 48, 527–546.

Fournier, M., Chanson, B., Thibaut, B. & Guitard, D. (1994a) Mesure des déformations résiduelles decroissance à la surface des arbres, en relation avec leur morphologie. Observation sur différentesespèces. Annales des Sciences Forestières, 51, 249–266.

Fournier, M., Moulia, B. & Thibaut, B. (1994b) Growth, prestresses and regulation of movements inliving trees, in Evolution of Natural Structures, Principles, Strategies and Models in Architectureand Nature, 3rd International Symposium SFB 230 (eds Universities of Stuttgart and Tübingen),Mitteilungen des Sonderforschungsbereich 230, Heft 9, Stuttgart, pp. 157–162.

Fournier-Djimbi, M., Sassus, F., Combes, J.G. & Grzeskowiak, V. (1997) Growth strains and reactionwood, in 2nd International Conference on Plant Biomechanics, Centre for Biomimetics, TheUniversity of Reading, Reading, pp. 231–236.

Gérard, J. (1994) Contraintes de croissance et valorisation du bois des Eucalyptus de plantation auCongo, Doctoral dissertation, University of Bordeaux 1, France, 160pp.

Gril, J. & Thibaut, B. (1994) Tree mechanics and wood mechanics: relating hygrothermal recovery ofgreen wood to the maturation process. Annales des Sciences Forestières, 51, 329–338.

Gril, J., Thibaut, B., Macchioni, N., Pividori, M. and Cielo, P. (1995) Growth stresses related problemsin using chestnut wood from coppices, in International Conference in Forest Products Research(eds G. Becker & M. Hecker), Göttingen, pp. 43–46.

Guéneau, P. (1973) Contraintes de croissance, Bois et Forêts des Tropiques, Cahier scientifique no 3,pp. 1–52.

Jacobs, M.R. (1945) The growth stresses of woody stems, Bulletin Commonwealth Forestry Bureau 28,pp. 1–67.

Jullien, D. & Gril, J. (1996) Mesure des déformations bloquées dans un disque de bois vert. Méthode dela fermeture. Annales des Sciences Forestières, 53, 955–966.

Jullien, D. & Gril, J. (2002) Modelling crack propagation due to growth stress release in round wood, inProceedings of EMMC6 Conference on Non Linear Mechanics of Anisotropic Materials, Liège,pp. 203–210.

Kübler, H. (1987) Growth stresses in trees and related wood properties. Forest Products Abstracts, 10,61–119.

Loup, C., Thibaut, B., Gril, J. & Jullien, D. (2001) Stresses in beech in Contribution to the final reportof European FAIR project CT98-3606.

Nicholson, J.E. (1971) A rapid method for estimating longitudinal growth stresses in logs. WoodScience and Technology, 5, 40–48.


Polge, H. (1984) Essai de caractérisation de la veine verte du Merisier. Annales des SciencesForestières, 41, 45–58.

Saurat, J. & Guéneau, P. (1976) Growth stresses in beech. Wood Science and Technology, 10, 111–123. Thibaut, B., Fournier, M. & Jullien, D. (1995) Contraintes de croissance, recouvrance différée à l’étuvage

et fissuration des grumes: cas du Châtaignier. Forêt Méditerranéenne, XVI, 85–92. Thibaut, B., Gril, J. & Fournier, M. (2001) Mechanics of wood and trees: some new highlights for an

old story. Comptes-rendus de l’Académie des Sciences, Paris, Série 2b, 329, 701–716. Timell, T.E. (1986) Compression Wood in Gymnosperms, Springer-Verlag, Berlin, Heidelberg, New

York, Tokyo. Trénard, Y. & Guéneau, P. (1975) Relation entre contraintes de croissance longitudinales et bois de

tension dans le hêtre ( fagus sylvatica L.). Holzforschung, 29, 217–223. Yamamoto, H., Okuyama, T. & Yoshida, M. (1995) Generation process of growth stresses in cell walls

IV: analysis of growth stress generation using a cell model having three layers (S1, S2 and I + P).Mokuzai Gakkaishi, 41, 1–8.

Yamamoto, H., Kojima, Y., Okuyama, T., Abasolo, W.P. & Gril, J. (2002) Origin of the biomechanicalproperties of wood related to the fine structure of the multi-layered cell wall. Journal ofBiomechanical Engineering, 124, 432–440.

Yoshida, M. & Okuyama, T. (2002) Techniques for measuring growth stress on the xylem surface usingstrain and dial gauges. Holzforschung, 56, 461–467.

Wilson, B.F. & Archer, R.R. (1979) Tree design: some biological solutions to mechanical problems.Bioscience, 29, 293–298.

7 Wood quality for pulp and paper Denilson Da Silva Perez and Thierry Fauchon

7.1 Introduction

7.1.1 Why wood?

The necessity of finding an adequate medium for the written word led humansto invent papermaking. Before the introduction of paper, history records theuse of several mediums for writing purposes, e.g. soft stone blocks, earthen-ware tablets, and parchment and papyrus, obtained respectively from animal andplant sources. Some rudiments of papermaking can be found in the preparationof papyrus, since this material was obtained by beating and organising thin layersof plant stems to make a single sheet. However, one of the most important charac-teristic of papermaking, the defibering, was completely absent in the process.

The invention of paper is attributed to Ts’ai Lun in China around 105 AD, whoproduced sheets from a suspension of bamboo fibres. The process was thendeveloped in China for a long time before reaching the south of Europe at theend of the 14th century. At that time, paper was obtained exclusively from cottonand linen rags. The raw material supply became scarce with the arrival of modernprinting techniques – illustrated, for example, by the printing of Gutenberg’sBible, and the consequent increase in the demand for paper. Moreover, thecontinuous development of papermaking machines, evidenced by the subsequentpatents attributed to Robert, Fourdrinier and Dickinson in 1798, 1803 and 1809,respectively, aggravated the situation. A method to mechanically transformwood into paper was invented by Keller in 1844 and since then, wooden rawmaterial has become the most important source of fibres for papermaking.

A few years later, the first manufacture of pulp from wood, using the sodaprocess was developed in England. Two patents attributed, at the end of the19th century to Benjamin Tilghman in America (1867) and Carl Dahl in Germany(1884) for sulphite and Kraft processes respectively, were the basis of thedevelopment of modern wood pulping for the paper industry.

The most important requirement for a fibre to be useful for papermaking isits conformability, i.e. the capability of using it to form a mat and then auniform sheet under pressure (Smook, 1997). In fact, fibres from almost anyvascular plant can be used for such purposes. However, a high yield of fibres isneeded, and only economically beneficial sources of fibre are used. This is whywood, as the most abundant fibre source in nature, is by far the most importantraw material for papermaking fibres.


For pulp and paper applications, the most important differences betweenhardwoods and softwoods lie in their fibre characteristics and the paper propertiesresulting from them. The most important softwood species are pines (radiata,white, jack, ponderosa, Caribbean, maritime, loblolly, among others), spruces(Norway, white, Sitka, Engelmann), firs (balsam, silver, Douglas) and hemlock.Important hardwood species include aspen, birch, beech, oak, chestnut anddifferent species of Eucalyptus. Pulpwood, worldwide, is derived from about30% hardwood and 70% softwood. However, since pulp manufacture dependson the wood supply available, production is site- and species-dependent. Thisexplains why Brazil’s production is composed of almost 100% hardwood pulpwhereas Scandinavian countries produce mainly softwood pulps.

7.1.2 Wood versus non-wood fibres

Non-wood fibres represent less than 10% of the resource used for producing pulpin the world. They can be separated into two different groups, those available asagricultural residues like cereal straws, sugar cane bagasse or oil seed rape, andthose grown specifically for fibre production, such as sisal, cotton linters, manilla,hemp, flax, jute, bamboo, reeds and kenaf. Although in developed countries,such fibres account for less than 1% of total pulp production, in developingcountries they represent 56% of the fibre supply, reaching 90–100% in countriessuch as China and Pakistan (Moore, 1996).

The characteristics of wood and non-wood fibre sources for pulp are summar-ised in Table 7.1. The fibres from agricultural residues are shorter and thickerthan wood fibres, although some annual plants possess significantly longerfibres than those from wood. Chemical composition also differs slightlyamong softwood, hardwood and annual plants. Besides small variationsobserved for major macromolecular components, non-wood sources have ahigher ash (mainly silica) content (Table 7.2).

There are numerous advantages in using wood for producing pulp. Themost important relate to the fact that it is a renewable resource which is

Table 7.1 Fibre dimensions of plants used for pulp production

Length (mm) Width (μm) Ratio (L/W)

Softwood (tracheids) 1.4–6 20–50 80–200 Hardwood (fibres and vessels) 0.2–1.6 10–300 2–100 Straws (rice, wheat, rye) 0.5–1.5 8–15 50–120 Sugar cane bagasse 1.2–2 15–25 80–130 Cotton 25–65 18–30 1000–4000 Flax (linen) 10–36 12–20 1000–2000 Ramie 100–150 35–50 2000–4500 Sisal 2.5–3 18–25 100–160 Bamboo 2–3.5 12–18 110–300

WOOD QUALITY FOR PULP AND PAPER 159

ecologically friendly. The possibility of harvesting all year round is a veryimportant point when compared to non-wood sources, which are usually har-vested at a particular time of year. The higher density of wood is economicallybeneficial in relation to transport and storage of chips, and permits the designof larger and economically competitive mills. High ash and silica contents innon-wood plants make chemical recovery difficult, especially in the Kraft pro-cess. Although wood is far from being a homogeneous material, the uniformityof its fibre dimensions and composition is much higher than that observed forannual plants. In the case of flax, for example, only 20% of bast fibre is usedfor pulp purposes. Whole sugar cane bagasse cannot be used since the pithcontains too much silica, increasing the consumption of alkali during pulping.In both cases, separation of fibres from other tissues is difficult and results inyield loss. On the other hand, except for agricultural residues, annual plant fibrecosts are much higher than wood fibre costs because of their use in high added-value products. In Table 7.3, a comparison of fibre costs is shown based oncotton ranked as 100 (Chisholm, 1994). Finally, the quality of paper obtainedfrom wood, especially long fibre-rich softwood species, is considerably higherwith respect to both physical and optical properties than non-wood fibres atcomparable cost, i.e. agricultural residues.

Wood will undoubtedly be the major source of virgin fibres for a long timeto come except in countries where there is no forest resource. Nevertheless, thereis a growing interest in the potential use of agricultural residues and annual

Table 7.2 Gross chemical composition of woody and non-woody sources of pulp (Moore, 1996; Alén,2000)

* Includes silica.

Component (%) Hardwood Softwood Annual plants

Cellulose 42–49 41–46 30–45 Hemicelluloses 23–26 25–32 20–35 Lignin 20–26 26–31 3–24 Extractives 3–8 10–25 5–15 Ash* 0.2–0.8 0.2–0.4 3–18 Silica <0.1 <0.1 2–12

Table 7.3 Comparison of fibre costs (Chisholm, 1994)

Relative price

Cotton 100 Flax (linen) 67 Abaca 50 Jute 22 Sisal 21 Wood 6 Straws 4


plants as part of the pulp and paper raw material supply. Such considerationsform part of the strategy to expand the activities of the agro-industry, currentlyalmost uniquely focused on food production, into a potential supplier for integratedsectors such as paper, textiles, energy and chemistry. Therefore, non-wood fibres,to be used in mixture with wood pulp, could become an important componentof paper raw material market in the future.

7.2 From wood to paper

7.2.1 Wood as a raw material

The number of tree species widely distributed on the earth is around 1000 forsoftwoods and 30 000–35 000 for hardwoods. However, the wood of only asmall proportion of these species is currently used commercially (Alén, 2000).The number of species of commercial importance depends on different param-eters such as availability, diversity, and technical interest and feasibility. Thus,the number of commercially important wood species in the world is around 120.

Since wood has a unique structural and chemical organisation, it is importantto understand both its macroscopic and chemical composition in order toestablish wood–pulp fibre relationships.

Wood tissue is not constituted of a single type of cell, but of a variety ofthem in different anatomical arrangements according to, among other factors, thebotanical class (hardwood or softwood), species, growing conditions and physio-logical function. Softwoods mainly comprise tracheids or prosenchyma cells,ray parenchyma and ray tracheids. Hardwoods are structurally more complexand are composed of fibres, tracheids, vessel elements, ray and longitudinalparenchyma. The characteristics of the cell types are summarised in Table 7.4.

A cross section of a tree reveals the existence of an inner part, usually ofdead cells, the heartwood (Fig. 7.1). The physiologically active wood aroundthe heartwood is the sapwood. This region grows every year as the tree producesa new layer annually. The variation in climatic conditions during the year

Table 7.4 Characteristics of the most important cells in softwoods and hardwoods (Alén, 2000)

Cell type Length (mm) Width (μm) Amount (%)

Softwood Tracheid (fibre) 1.4–6.0 20–50 90 Ray tracheid <5 Ray parenchyma 0.01–0.16 2–50 <10 Epithelial parenchyma <1

Hardwood Fibres 0.4–1.6 10–40 55 Vessel element 0.2–0.6 10–300 30 Longitudinal parenchyma

<0.1 <30 <5

Ray parenchyma 15

} }

} }


generates cells growing at different rates, resulting in different earlywood andlatewood proportions. Earlywood cells are shorter and thin walled, while late-wood cells are longer and thick walled. The difference between the two woodtypes leads to the appearance of the annual growth rings, most notably in woodfrom temperate zones. In wood from tropical regions, the difference betweenthe tissues is less pronounced.

Seasonal changes is not the only factor influencing the shape of cells. Externalforces (wind for example) which force the tree stem or branch out of its normalequilibrium position induce an accelerated growth on either the lower or upperside of the stem or branch. The wood formed due to this reaction of the tree iscalled reaction wood, more specifically compression wood for softwoods,because it is formed on the underside of the stem or the branch, and tensionwood for hardwoods, formed on the upper side of inclined stems or branches.Compression wood is reddish-brown, characterised by high density, high hardnessand low moisture-holding capacity (Smook, 1997). An example of compressionwood formed in an inclined maritime pine (Pinus pinaster) is shown in Fig. 7.2.

Tension wood owes its properties largely to a special wall layer, the gelatinouslayer, composed of almost pure and highly crystalline cellulose. Reaction woodis discussed in greater detail in Chapter 5.

7.2.2 Wood–pulping process interactions

As wood properties are highly variable, especially between species, pulpabilityof different wood species will differ from one another. In order to obtain apulp, it is necessary to break down the wood structure to separate the fibres. Themain methods for achieving this separation are either mechanical or chemical,

Fig. 7.1 Earlywood/latewood and heartwood/sapwood structures in Pinus pinaster.


or a combination of both. The most important characteristics of the mechanical,chemical and hybrid processes are summarised in Table 7.5.

Mechanical pulps are obtained by subjecting wood material to forces sufficientto break the internal cell organisation, especially between fibres (Sundholm,1999a,b; Kure, 1997). This type of pulping process has changed considerablyin the last 40 years. Until the 1980s, most pulp was produced by the basic stonegroundwood (SGW) process, in which wood logs were ground against a largeabrasive stone or a wheel made up of stone segments. Nowadays, mostmechanical pulp is produced by pressurised refining equipment which is fedwith pre-steamed chips. The refiner is composed of two opposing grooved steeldiscs, where one or both discs rotate to break down wood chips introducedbetween them by a screw-feed mechanism.

Fig. 7.2 Formation of compression wood in Pinus pinaster.


In the chemical process, fibre separation is obtained through delignification,especially of the middle lamella that bonds the fibres together. Fibre liberationis achieved when enough lignin is dissolved to allow fibres to be separated fromthe wood matrix without, or with only a small amount of, mechanical treatment(Smook, 1997). The ultimate goal is the production of lignin-free fibres, usuallyachieved in a second step called bleaching, since the selectivity of lignin removalin relation to polysaccharides becomes too low at the end of the pulping process.The elimination of lignin gives fibres a higher flexibility and better papermakingcharacteristics, therefore better physical and optical properties. Chemical pro-cesses are by far the most important pulping processes in the western world. Twoprocesses are used for this purpose, the Kraft process and the sulphite process.The liquor used in the Kraft process is composed of sodium hydroxide (NaOH)and sodium sulphide (Na2S). The sulphite process is carried out with a liquorcomprising sulphurous acid and a base (Ca, Mg, Na or NH3). Both free andbase-conjugated SO2 coexist in the liquor. The ratio depends on the pH and onthe solubility of the conjugated base. Thus, the traditional Ca-process operates atpH 1.5 to avoid precipitation, whereas processes based on soluble bases (Mg,Na and NH3) are carried out at pH between 4.0 and 5.0. Cooking temperaturesfor both Kraft and sulphite processes are between 160 and 180°C depending onthe wood species to be cooked, except for calcium sulphide, where the drasticpH conditions limit the maximum temperature, which is 140°C.

A combination of sulphite chemical and mechanical pulping is used in theso-called neutral sulphite semichemical (NSSC) process. In this process, a chem-ical pre-delignification takes place followed by a mechanical defibering to pro-duce a particular type of pulp, mainly used for the internal reinforcement ofcorrugated papers.

Table 7.5 Main characteristics of mechanical, semichemical and chemical pulping processes (Smook,1997; Salmén et al., 1999)

Mechanical Semichemical Chemical

Pulping agent Mechanical energy Hybrid Chemicals and heat Wood Softwood (mostly) Hardwood (mostly) Both Temperature (°C) Room temperature–100 160–180 120–180 Pulp yield (%) 85–95 55–85 40–55 Relative strength

Hardwood 3 6–7 7–8 Softwood 5–7 9–10

Fibres characteristics Short, weak, unstable Intermediate (unique properties)

Long, strong, stable

Processes Groundstone pulp Neutral sulphite Kraft Refiner mechanical pulp High-yield kraft Sulphite Thermomechanical pulp High-yield sulphite Soda Main application Newspaper Corrugated papers Printing paper


An interesting approach to pulping which has been developed recentlyinvolves producing chemical pulps by delignification with an organic solvent.Such processes are called organosolv pulping. The most advanced technology wasAlcell® technology using ethanol–water as liquor for cooking. However, milloperations ceased just a few years after the beginning of commercial activity foreconomic reasons.

The major difference between mechanical and chemical pulps is the natureof wood components present in the pulp. In mechanical pulps, almost the totalityof the wood components is present in the final pulp. This pulp is primarily usedfor newspapers, coated printing grade paper and SC (super calendered) grades.A major problem is the presence of lignin among the pulp components that reactsunder the action of heat or light, leading to a rapid brightness reversion. Opticaland physical properties of chemical pulps are better than those obtained formechanical pulps, but the pulp yield is lower due to the removal of lignin andsome polysaccharides.

7.2.3 Wood–pulp fibre relationships

Although wood and paper are basically constituted of the same type of fibres,they are organised in very different ways. In wood, the cells are arranged in apreferential way and are bonded together by a thin lignin-rich layer called middlelamella. In paper, fibres are randomly organised and bonded by strong physico-chemical interactions at the contact points between the fibres. In fact, beforebecoming suitable for papermaking purposes, the fibres must be mechanicallytreated (by refining or beating processes) in order to increase the contact pointsbetween them (Sundholm, 1999a,b).

The differences observed between hardwoods and softwoods directly influencethe quality and therefore the end-use of the pulps. The two most importantaspects are fibre dimensions, especially length, and cell wall thickness. Fibrelength is directly connected with interfibre bonding and is proportional to tearstrength. This is valid for both softwoods and hardwoods.

The effect of cell wall thickness is more complex. Thin-walled cells contributeto burst/tensile strength since they are more flexible and also because theycollapse during sheet formation generating a more compact network in paper.On the other hand, thicker-walled cells contribute favourably to tear strength,breaking length, bulk and absorbance properties, but are less conformable thanthin fibres. This analysis is reasonably true for softwoods. For hardwoods,the relationships are less well established because of the presence of verydimensionally different types of cells such as fibres, vessels and parenchymacells; the parenchyma cells being responsible for the high fines contents ofhardwood pulps.

Since the cells found in contrasting wood tissues such as latewood/earlywood,sapwood/heartwood, juvenile/mature wood or normal/reaction woods are


different, it is normal to expect different behaviours in papermaking pro-perties. Pulp yield, physical and optical properties, and chemical or energyconsumption are tremendously affected, as shown in Table 7.6 for the Kraftcooking of normal and compression wood of maritime pine (Da Silva Perez &Chantre, 2002).

7.3 Resource management and biological decay

Since wood species and freshness (see below) are among the main factors whichinfluence pulping processes and end-product quality, origin and managementof the forest resource have to be taken into account as one of several links in thechain of variability.

7.3.1 Origin, supply of resources and pulp production

According to a 1995 FAO forecast, global paper consumption will be about440 million tonnes by 2010. The growth of demand for paper will be higherin Asia, whereas the demand growth will be comparatively low in NorthAmerica. The development in Asia is expected to spread from Japan to the restof Asia. Particularly high consumption growth is expected in China.

There are several reasons for changes in paper furnish composition. Firstlyeconomic:

• To reduce furnish cost by using lower cost fibre (hardwood instead ofsoftwood, mechanical instead of chemical pulp, recycled instead of virginfibres). The hardwood/softwood ratio in pulp production is expected toincrease from slightly below 30/70 in 1994 to 33/67 in 2010.

• To reduce furnish cost by using less fibre (higher filler content). • To reduce furnish cost by using locally grown non-wood fibres espe-

cially in fibre-deficit regions such as the Asia-Pacific region.

Table 7.6 Comparison between properties of pulps obtained from normal and compression wood(maritime pine)

Normal wood Compression wood

Pulp yield (%) 49.2 43.7 Residual lignin (%) 8.2 15.1 Burst index (kPa m2 g−1) 4.82 3.76 Tear index (mN m2 g−1) 4.62 3.62 Breaking length (km) 9.76 6.84 Young’s modulus (Mpa) 275.0 215.0TEA (J/g−1) 1.90 1.52


Secondly, there are reasons concerning process changes and quality demandevolution which include:

• Changing process requirements, e.g. higher paper machine speed or moredemanding converting operations which require a stronger paper toensure better runnability.

• Improvement of pulp quality as reinforcement components allows inclusionof a greater proportion of weaker and less expensive components.

• Changing paper quality requirements, mostly towards higher quality andbetter performance characteristics.

• Increasing market demands for paper with special characteristics and forthe fibre raw material used.

Thirdly, changes are required to meet environmental concerns relating to legis-lation and other demands for preservation of natural resources or minimisationof detrimental emissions from pulp and/or paper manufacture.

The share of pulpwood in total fibre use is expected to grow from 11 to14% by 2010 as a result of increasing demand for reconstituted products.Competition for pulpwood is likely to intensify in a number of regions as a res-ult of increasing production capacity for reconstituted panels. The increasinginterdependence of mechanical wood processing and pulp and paper manufac-ture will not only intensify competition but also open up opportunities forcollaboration. Sawmills will become a more important source of raw materialfor pulp manufacture.

Natural softwood resources exist mainly in the northern hemisphere in theboreal coniferous zone in North America, the Nordic countries and Russia. Mostof the world’s untapped softwood resources are found in the Asian part of Russia.Significantly, increased harvesting from these areas is, however, at presentdifficult owing to a lack of infrastructure and other problems. Tropical andtemperate plantations in the southern hemisphere (Brazil, Chile, New Zealand,Australia and South Africa) provide a complementary resource of increasingimportance.

About three quarters of the world’s closed hardwood forests are in thedeveloping countries, and to a significant extent in the southern hemisphere.The world’s largest untapped natural hardwood resources are in Latin America,especially in Brazil. The availability of trees from tropical forests will, however,be very restricted for conservation reasons. The wide range of sources of woodfor pulp will inevitably lead to high diversity in paper characteristics.

7.3.2 Biological decay

Some experts in meteorology predict major climatic changes in the near future(Houghton et al., 1996; Forbes et al., 1997). For example, an increase in thefrequency of hurricanes in western Europe is forecast. Wind-thrown trees are


rapidly subjected to fungal and insect attack (Peek & Liese, 1974; Verkasalo,1993; Bucking etal., 1997). Also, biological decay develops in standing treesthrough wounds or pruned surfaces within a range of moisture content between20 and 60%. This has to be taken into consideration for supply managementfor pulp mills in crisis situations (Dillner, 1972; Oliver-Villanueva & Sachsse,1992; Kucera & Katuscak, 1993; Eisenbarth, 1995). Figure 7.3 summarises therisk associated with a range of moisture contents in wood.

Before lignivorous fungal attack takes place, and from the beginning of spring,blue stain rapidly appears in the sapwood of softwoods and some hardwoodspecies (beech, poplar, ash, etc.), although it has no effect on chemical pulping.Then, depending on season, time, species and storage conditions, and generallya few months after felling, stains caused by Stereum purpureum appear in lessdurable species like beech. Decay which degrades the wood at greater depthsalso occurs, partially altering wood quality. This results in pulp yield loss anddecrease of pulp strength properties. However, some fungi preferentially attacklignin (white rot), whereas others attack cellulose (brown rot) (Nilsson, 1972).The former is commonly named fibrous rot (Ungulina on birch for instance)and the latter cubical rot (bracket fungi on all species). Finally, some fungalspecies attack lignin and cellulose simultaneously. This natural diversity offungi and the variability of wind-thrown trees lead to a high variability of pulpproperties, depending on species, meteorological conditions, type of damage andstand. In the long term, all logs stored in the forest will be degraded in quality.The more slowly the wood is dried, the more effective is the decay phenomenon,particularly in wet and enclosed conditions. It is worst of all when the logs arein contact with the ground. Both pulp yield and pulp strength properties areaffected by the degradation of polysaccharides (Bergman, 1972). A decreaseof six points in average pulp yield between fresh wood and decayed wood hasbeen found to be caused by different cubical rots in five clones of aspen in

Time evolution of fungi decays

Blue stain

Brown rot

Other fungi

40 60 80 100 120Woodmoisturecontent

Fig. 7.3 Effect of moisture content on fungal attack of timber.


Ontario (Hunt et al., 1978). Attack in the early stages has no significant effecton strength properties, whereas in the later stages, a drop of 20–30% in burstindex and of 9–13% in breaking length is found (Fig. 7.4).

Janin (1983) observed a two-point loss in pulp yield between fresh beechand decayed beech, and a decrease of 22% in breaking length. The phenomenonis less noticeable in more durable species like hornbeam and even less so in verydurable species like oak or chestnut. Unfortunately, durable hardwood specieshave a lower pulp yield than the vulnerable species such as beech, birch andpoplar. Moltesen and Bang (1973) analysed the decay of beech roundwood lefton the roadside in Denmark for three years after windthrow in 1967. Theyobserved a 15% density loss for shadow-stored wood, and 19% loss for sun-stored wood. On average, these losses were more pronounced in the peripheralzone of logs (−25%) than in the central part (−7%). These losses of woodenmaterial were very good predictors of degradation of pulp strength properties.According to Moltesen and Bang (1973) and Viala (1963), hemicelluloseswere less degraded than lignin and cellulose in beech. In laboratory conditions(Viala, 1963), the decrease in cellulose content was severe (−22%) in beechand greater in Norway spruce (−42%). In Norway spruce, the cubical fungus(Fomes) degrades specifically the cellulosic compounds, leading to a verysignificant decrease in pulp yield and pulp viscosity and an increase in alkaliconsumption (Lonnberg & Varhimo, 1981).

6.1

5.7

5.3

4.9

4.5

4.1

3.70 3 6 9 12 15 18

Months of storage

Bur

st in

dex

at 2

5°S

R (

kPa

m2 g

–1)

Fig. 7.4 Variation of burst index over 18 months of storage of wood prior to pulping.


Other Swedish authors (Henningson, Whilhemssen, quoted by Moltesen)found a significant decrease in beech pulp yield only after one or two years ofstorage (−1%) and a real degradation after three years of storage (−4 to −5%).

High material losses in partially decayed logs occur also during transporta-tion, handling, debarking, chipping and screening. Losses during debarkingcan reach 1–5% when using decayed wood, and the same level of losses occursin chippers (Pulkki, 1991). The pulping value of roundwood stored roadsidefor three years was found to be half the pulping value of fresh roundwood(Moltesen & Bang, 1973).

Other detailed studies have confirmed the detrimental effects of rust onsouthern pines (Zobel & van Buijtenen, 1989). Table 7.7 shows the effects onpulping of the presence of 29% of rusted trees in a living stand of loblolly pine.

Fungal attack on living trees leads to an abnormally high production ofresins in response to the attack. This explains the gain due to increase in turpen-tine and tall-oil production. In windthrown trees, the resin content decreasesduring storage and there is an oxidation of resins. This means a decrease intall-oil and turpentine production in addition to a decrease in pulp yield andstrength properties resulting in a dramatic financial loss ($0.55/1000lbs oven-driedwood). In this case the authors considered that a decrease of 5–10% in pulpstrength has no economic effect. However, specifications for strength propertiesmade by the customer (Kraft liner, corrugating grades, etc.) mean that the decreaseof physical properties of the pulp will be compensated for by an increase in theratio of virgin/recycled fibres with a resultant financial loss for the mill.

There are two main methods for avoiding such degradation of wood andpulp quality: dry storage and wet storage.

• Dry storage is not suitable for wood to be used for mechanical pulping wherehigh moisture content is a key factor for pulp quality (Whitman, 1957). Forchemical pulping, the main problems arise with the production of oversizedchips and impregnation of dry chips by cooking liquor. In addition, morefines are generated during chipping, increasing the wood yard losses.

Table 7.7 Modifications of pulping caused by rust (Zobel & van Buijtenen, 1989)

Modifications induced by fungus % Gain($/1000 lbs o.d. wood)

Loss ($/1000 lbs o.d. wood)

Decrease of pulp yield 0.9 – 0.51 Increase of alkali consumption 0.8 – 0.04 Decrease of pulp strength properties 5–10 – – Increase of tall-oils production 25 0.23 – Increase of turpentine production 40 0.25 –

Total 0.48 0.55 Net loss 0.07


• Wet storage, under sprinklers, is more expensive, but the stored woodretains the quality of fresh wood (Moltesen et al., 1974; Watson et al.,1992). Another interesting point is the reduction of extractives content,responsible for pitch in the pulp and paper process equipments (Ekman& Hafizoglu, 1993; Borga et al., 1996).

However, some disadvantages of the wet storage method must be taken intoaccount:

• Debarkability of logs and brightness and bleachability of mechanicalpulps are strongly affected the during storage.

• Modification of extractive compounds during wet storage makes the woodunsuitable for sulphite pulping.

• The kinetic of delignification of watered wood may differ a great deal fromthat of fresh wood due to bacterial attack on cell walls (Gross etal., 1996).

• There is an important decrease in bark content, used normally as bio fuelby pulp mills.

Hence, depending on durability of species and storage conditions and time, thepulp properties of roundwood can vary greatly. Storage must be considered asan emergency solution, as in Europe after the hurricane of December 1999.The elapsed time from felling to delivery must be used as a freshness criterionin meeting the requirements of pulpwood users (Chantre, 2000).

7.3.3 Mill specifications and quality control measurement

Because of the variability of wood and its influence on both pulping processesand end-product quality, mills impose specifications on the quality of their woodsupply. Mills specify wood quality to enable them to increase yield, reduceoperating costs, optimise wood consumption and increase profitability of bothpulp mills and suppliers (Stevenson, 2001).

The specifications set out by the pulp mills were initially based on predictedinteractions between raw material quality and end-product quality with respectto resource availability close to the mill. Greater process stability is a key factorfor major potential gain, and the variability of supplies is considered as the mainsource of instability.

Depending on the type of process (mechanical or chemical pulping), differentparameters acquire different degrees of importance. In mechanical pulping,strict specifications are applied for freshness and moisture content, which areimportant for chips steaming and defibering efficiency, whereas in chemicalpulping the chip size distribution is the most important factor, affecting chipimpregnation and cooking efficiency. An example of influence of chip thickness


heterogeneity is presented in Fig. 7.5. The more heterogeneous the mixture ofchips, the lower is the Kraft pulp yield.

Raw material quality can be defined as follows: fitness for use, customerexpectations met by the product, continuous improvement, value added andreduced variability. Figure 7.6 represents the traditional view of quality control,whereas what might be considered the correct view is presented in Fig. 7.7.

Each mill has to define its own quality parameters depending on the processand the end-product. The specifications for the raw material have to be made

12 3

4

5

5

6

8 10

0 10 15

9

7

41

40

39

38

37

36

Standard deviation

Pu

lp y

ield

(%

)Influence of chip size heterogeneity on pulp

yield

Fig. 7.5 Example of pulp yield loss due to the increase of chip thickness heterogeneity in batchlaboratory digesters.

TargetLowerspecification

Upperspecification

NotOK

NotOKOK OK

Fig. 7.6 The traditional view of quality specifications.


according to this definition. Specifications concern both roundwood and sawmillchips, but the control methods and devices are different for the two types ofmaterial. The main parameters affecting process runnability and/or end-productquality that are controlled at the mill gate are:

• For sawmill chips: moisture content, chip size distribution, bulk densityand bark content (eventually other contaminants).

• For roundwoods: species, freshness (colouration, stains, etc.) and moisturecontent.

For roundwood, quality control is mainly visual. Samples for moisture contentmeasurements are taken either directly in logs on the trucks, or on the conveyorbelt after the chipper. For sawmill chips, automatic control devices have beendeveloped and are available on the market for the measurement of moisturecontent, chip size distribution, bark content and bulk density. Nevertheless,those devices are not standardised yet and represent high investment; semi-manual measurements are most commonly practised in mills.

Sawmill chip samples are taken in order to control the compliance of a loadto the specifications laid down by the mill. Representative sampling is the keyfactor for an accurate measurement. It can be done directly in the trucks, duringunloading or on the conveyor belt. In any case, manual or automatic samplingis possible. Automatic sampling is safer, more representative and faster but farmore expensive. The main interest of automatic sampling is the possibility ofanalysing more samples which pre-supposes investment in automatic qualitycontrol devices or in human resource in order to measure all the samplesprovided by the sampler.

Although raw material quality control is of major importance due to variability(species, supplier, season, etc.), most of the pulp mills still use old-fashioned

Target

Lowerspecification

Upperspecification

NotOK

NotOK

OK OK

Fig. 7.7 The correct view of quality specifications.


rules for quality measurement. One of the biggest potential gains for the mill isa better knowledge of the supply allowing optimised woodyard management.

7.4 Wood-quality variability and its consequences for pulp and paper quality

7.4.1 Wood species and mixtures

Wood species is probably the most significant parameter affecting the relation-ship between the pulping process and pulp quality. Variations in chemicalcomposition or fibre characteristics are observed not only between species, butalso within a species and a tree. Moreover, the wood supply of a given mill isdependent on its geographic location and therefore with the species producedin its neighbourhood. Local contexts are also important for understanding thewood supply chain. Thus in Brazil or Chile, the mills operate almost exclusivelywith eucalypt and loblolly pine respectively, in most cases cultivated in standsowned by the mill. In some European countries (France, for example), themills use a mixture of softwood or hardwood according to the wood availabilityin forests managed by either the state or private owners. This generates situ-ations where the mill operates with a wood mix containing up to a dozen ofdifferent species. Because of their difference in terms of anatomy, chemicalcomposition, density, age, diameter, etc., the species do not react in the same wayunder mechanical or chemical pulping.

Wood species used in mechanical pulping processes have to conform tosome specific characteristics such as the absence of colour, since the pulps arenot fully bleached, low density for reduced energy consumption and minimumresin content since resin contributes to pitch deposits and affects paper quality.The favoured species is Norway spruce (Picea abies), because it conforms to allthe aspects mentioned here (Tyrvaïnen, 1995; Heikkureinen etal., 1999). Severalother softwood species are used, including spruces, pines, firs and hemlock(Sahlberg, 1998). The advantages of spruce in relation to other species are clearlyseen from the data in Table 7.8, where the properties of pulps made from different

Table 7.8 Characteristics of thermomechanical pulps obtained from different softwoods (Varhimo &Tuovinen, 1999)

Spruce Jack pine Loblolly pine Caribbean pine

Energy required (MWh t−1) 1.92 2.20 2.14 2.56 Fibre length (mm) 1.29 1.00 1.19 1.32 Tensile index (Nm g−1) 40.8 34.9 31.9 0.50 Tear index (mN m2 g−1) 9.4 8.5 9.2 10.2 Scattering coefficient (m2 kg−1) 63.5 59.4 49.5 53.6 Brightness (%) 58.0 47.8 53.3 54.1


softwood species through groundwood and thermomechanical pulping processare compared.

Hardwoods are less competitive for mechanical pulps because of their poorstrength properties, despite having good light scattering and good sheet surfaceproperties, as shown in Table 7.9 (Marton et al., 1979; Varhimo & Tuovinen,1999). Several factors have contributed to an increase in interest in hardwoodsincluding chemical pre-treatment to obtain a compromise between strength andoptical properties (Varhimo & Tuovinen, 1999), the refining at higher energy orthe use of young trees (less than 35 years old) to avoid heartwood.

Spruce-type softwoods are the most useful species for semichemical andchemical sulphite-based processes. The reasons are not the same as for themechanical process but, rather, the fact that wood chips impregnate readily.Heartwood, especially from pines and firs, is to be avoided because the phenolicsubstances contained in heartwood tend to condense with lignin and cause highlevels of screening rejects and high lignin content in pulps (Varhimo &Tuovinen, 1999). However, the sapwood of these species, as sawmill chips, canbe pulped without major problems. High resin content is also undesirable sinceit inhibits penetration of the cooking liquor into the chips and also because itgenerates pitch problems.

Despite their low strength, hardwood bisulphite pulps have found somespecific markets such as dissolving pulp or as an improving agent for opacityin fine papers. The most important species for this are aspen, poplar, beech,birch and maple.

The Kraft pulping process is less demanding than the mechanical and sulphiteprocesses in terms of wood species. Most lignocellulosic materials can be pulpedby this process if the conditions of liquor composition, temperature and cookingtime are correctly chosen. This explains its adoption worldwide and justifies itsposition as the most important process in the pulp industry.

In contrast with the mechanical process, in the Kraft process, the separationof fibres is brought about slowly (if compared with contact time of wood chipswith refiner or groundstone) by the action of chemicals which remove ligninfrom the wood matrix. However, the strong alkaline conditions involved alsolead to a partial removal of polysaccharides, especially hemicelluloses. On theother hand, lignin is the macromolecular component that varies the most between

Table 7.9 Characteristics of thermomechanical pulps obtained from various hardwood species(Varhimo & Tuovinen, 1999)

Aspen Poplar Eucalypt Birch Oak Maple

Energy required (MWh t−1) 2.3 2.6 2.3 3.3 3.1 4.0 Tensile index (Nm g−1) 25.5 27.8 12.6 21.2 10.2 20.0 Tear index (mN m2 g−1) 2.1 2.3 1.4 2.5 1.7 1.6 Scattering coefficient (m2 kg−1) 77.0 87.9 57.6 67.0 48.3 66.5 Brightness (%) 60 55 35 39 31 46


species, especially between hardwoods and softwoods. Therefore, since therate of delignification governs the pulp production rate, significant differencesare expected to occur whether hardwoods or softwoods are pulped. Delignifi-cation usually occurs in three distinct phases, called initial, bulk and residualdelignification. The typical behaviour of softwood and hardwood speciesunder Kraft cooking conditions is presented in Fig. 7.8. The H-factor, appearingin the x-axis of this figure is a concept developed to group time and tempera-ture variations in a single variable (Vroom, 1957). When polysaccharide removalbecomes significant at the end of the residual delignification stage, cooking isstopped. The residual lignin is removed by bleaching sequences in which lignin-specific chemicals are used.

The kappa number, a test widely used in pulp and paper industry, is posi-tively correlated with residual lignin in the pulp. Figures 7.9 and 7.10 show thepulp yield–kappa number relationships for several species and in particular thedifferences that exist between hardwood and softwood species in terms ofresidual lignin.

As a result of the intrinsic species variability and the effect of the delignifi-cation process, the fibre morphology and physical properties of the paper obtainedfrom different species are quite different. Tables 7.10 and 7.11 show thevariations observed in pulp fibre morphology and paper physical properties ofdifferent softwoods and hardwoods. The global differences between these twogroups in terms of paper strength properties are clearly shown in Fig. 7.11.

Res

idu

al li

gn

in (

%)

Hardwoods

Softwoods

Residual delignification

Bulk delignification

Initial delignification

H factor

0 500 1000 1500 2000

Fig. 7.8 Different delignification phases during Kraft pulping of softwood and hardwood.


Because of the variation between species and the wood-supply problemsdiscussed above, a mill can rarely operate using a single species. In mostcases, wood feedstock is composed of two or more species. Recent experimentshave shown that specific interactions exist during Kraft cooking of mixedspecies, which result in values different from those expected from the simple

Fig. 7.9 Kraft delignification behaviour, expressed as pulp yield and kappa number, of severalsoftwood species.

0 10 20 30 40 50 60 70

Kappa number

Spruce

Pin noir

Douglas fir

Radiata pine

Scot pine

Loblolly

Aleppo pine

Maritime pine

Pul

p yi

eld

(%)

56

54

52

50

48

46

44

42

40

Table 7.10 Morphological and physical properties of Kraft pulps obtained from softwood species

Wood Length(mm)

Width (μm)

Bulk (cm3 g−1)

Burst index (kPa m2 g−1)

Tear index (mN m2 g−1)

Breaking length (m)

Spruce 1.8 33.8 1.23 5.5 10.2 8900 Aleppo pine 1.8 34 1.32 4.3 10.6 7200 Scot pine 1.8 32.1 1.30 4.8 10.8 7700 Black pine 1.9 33.5 1.41 4.7 12.1 7400 Douglas fir 1.9 32.3 1.43 4.7 13.9 7800 Loblolly pine 2.0 34.1 1.20 5.3 8.0 8300 Radiata pine 2.0 34.1 1.30 5.7 10.0 10000 Maritime pine 1.9 27.5 1.32 5.6 9.9 9600


weighted addition of properties (Medina et al., 2001). Two different mixturesof hardwood were studied: (1) poplar, beech and oak and (2) poplar, beechand chestnut. The modelling for pulp yield at kappa number 18 and fibrelength, based on Scheffé’s triangle methodology is presented in Fig. 7.12(Scheffé, 1958, 1963).

Birch

Beech

Eucalypt (E. gunnii )

Chestnut

Poplar

Eucalypt (E. urograndis)

Oak

Aspen

Hornbeam

Kappa number

Pul

p yi

eld

(%)

60

56

52

48

44

4010 20 30 40 50 60 70 80 90

Fig. 7.10 Kraft delignification behaviour, expressed as pulp yield and kappa number, of severalhardwood species.

Table 7.11 Morphological and physical properties of Kraft pulps obtained from hardwood species

Wood Length(mm)

Width(μm)

Bulk (cm3 g−1)

Burst index(kPa m2 g−1)

Tear index (mN m2 g−1)

Breaking length (m)

Birch 1.33 33.5 1.25 3.8 8.5 7900 Chestnut 1.20 34.3 1.21 4.0 7.7 8100 Oak 1.07 38.5 1.46 3.1 7.6 5700 Beech 1.08 36.5 1.40 3.3 6.9 6600 Poplar 1.42 30.3 1.51 3.8 7.8 7600 Hornbeam 1.41 36.3 1.36 3.5 9.6 6700 Aspen 1.27 31.7 1.24 2.8 6.7 7200 Eucalypt

(E. gunnii) 1.23 37.6 1.18 4.3 6.1 8200

Eucalypt (E. urograndis)

1.34 33.9 1.31 4.0 9.4 8200


8.0

7.0

6.0

5.0

1.0 2.0 3.0 4.0 5.0 6.0

Burst index (kPa m2 g–1)

Bre

akin

g le

ngth

(km

)

Hardwoods

Softwoods

10.0

9.0

Fig. 7.11 Comparison of strength properties of softwood and hardwood pulps.

POPLAR

POPLARPOPLAR

POPLAR

Pulp yield

Fibre length

CHESTNUT

CHESTNUT BEECH

BEECH BEECH

BEECH

OAK

OAK

45464748495051525354plus

45464748495051525354plus

1.151.161.171.181.191.21.211.221.231.24plus

1.151.161.171.181.191.21.211.221.231.24plus

Fig. 7.12 Modelling of Kraft pulping of mixed hardwood species.


Pulp yield models are a perfect example of the effect of specific interactions.The use of either oak or chestnut as a third component of the mix with beech orpoplar gives different results, despite the similar morphological properties ofthe two added species. The pulp yield model for the oak-containing mixture isquadratic, whereas the chestnut one is linear. It is evident from these resultsthat chestnut has a more detrimental effect on pulp yield behaviour of the mixthan oak. The same tendency is observed for fibre length results, but the finalmodel is linear for both hardwood mixtures.

The understanding of such interactions is essential if solutions are to be foundfor some important problems of the pulp and paper industry such as pulpingyield or paper strength properties. A better understanding of the differencesbetween species as well as the interactions between species when in a mixtureis of fundamental importance in providing the mills with some practical rules forwood supply and woodyard management.

7.4.2 Within- and among-tree property variation

Because fibres form the basic structure of various paper grades, fibre propertiesin the original wood raw material can be expected to have a substantial effecton sheet properties. Fibre properties vary not only between wood species, butalso within and among trees of the same species. A tree contains several differ-ent types of fibre. Variations in the type and distribution of these fibres in thewood raw material can lead to significant variations in pulp quality. Within-treevariability can be very large. There exists at least as much variability in woodcharacteristics within a single tree as among trees growing on the same site orbetween trees growing on different sites.

Other within-tree variations influencing the pulping process and/or pulp andpaper properties include: wood density, chemical composition, moisture content,juvenile wood, heartwood and sapwood.

There are several patterns of variability within trees that are of importance.The first is the within-ring difference, the second the changes from the centreof the tree to outside and the third the differences associated with differentheights within the tree.

7.4.2.1 Within-ring variation Wood variation within an annual ring is caused by the coexistence of earlywoodand latewood (Wang & Braaten, 1997). The main difference between earlywoodand latewood is specific gravity, particularly for species like Douglas fir. Notonly does specific gravity vary but also greater differences in chemical compos-ition were found between earlywood and latewood within an annual ring thanbetween heartwood and sapwood of the same Douglas fir tree (Andrews,1986). Gladstone etal. (1970) reported that latewood of Pinus taeda hadhigher holocellulose, alpha-cellulose and glucan contents than earlywood, and


latewood gave 2–7% higher pulp yield from a given weight of dry wood thanearlywood. Figure 7.13 shows the differences between earlywood and latewoodfibre morphology.

The primary impacts on thermomechanical pulps are seen in sheet structureand fibre composition: high latewood content leads to surface disturbances andpoor sheet structure, and a high content of latewood in wood chips produceslonger fibres in pulp (Corson, 1991, 1997). For chemical pulps, latewoodcontent variation is of major importance for pulp yield mainly because it affectsphysical properties.

There are two sources of high wood basic density: a high content of nor-mal latewood fibres or a high content of compression wood fibres. The effectof pure compression wood on the mechanical pulp has not been previouslystudied or published. However, it is known that the glass transition tem-perature is slightly higher for compression wood compared with the othertypes of fibres.

7.4.2.2 Radial trends The second pattern of within-tree variability mostly depends upon the presenceof juvenile wood and its location. The differences between juvenile and maturewood is particularly important for softwood, but less important for hardwoodspecies. Juvenile wood is characterised by shorter cells with larger lumendiameters, thinner cell walls, larger microfibril angles, more compressionwood, lower specific gravity, higher lignin content, lower cellulose content,lower strength, and larger longitudinal and lesser transverse dimensionalchanges than mature wood.

Fig. 7.13 Transverse view of fibres from maritime pine in juvenile wood (left) and mature (late)wood(right) as seen in an electron microscope. Source: Tembec R&D Tartas.


Pulp yields from juvenile wood are lower than from mature wood on a dryweight basis and even more so on a green weight or volume basis. In the Kraftprocess, tall-oils and turpentine yields are three to five times higher in oldmature wood than in juvenile wood.

As shown in Figs 7.14a–d, paper produced from pulp prepared from juvenilewood (thinning logs) exhibits higher tensile strength, sheet smoothness, burststrength, fold endurance and apparent density but lower tear and opacity thanpaper from mature wood (sawmill chips).

Increasing the juvenile wood fibre content also improves the printability andISO brightness. For this reason, juvenile wood is excellent for thermomechanicalpulp, chemico-thermomechanical pulp or pressure groundwood. It is useful fornewsprint, some tissues, high quality printing and writing paper, among others.

A high basic wood density means that the cell walls are thick and stiff which mayresult in fibre cutting during mechanical pulping and therefore a pulp with low tearindex. It has also been shown that wood with high basic density has higher energyconsumption to a given freeness value than wood with low basic density. An excep-tion is juvenile wood which has a higher energy consumption when refined to agiven tensile index compared with mature wood. The reason for this is not known.

The optical properties on the other hand are better for pulp made of juvenilewood. Generally, mature wood gives pulps with better strength properties thanjuvenile wood.

For mechanical pulping, the relationship between wood fibre and pulp prop-erties is more difficult to determine than for Kraft pulping. To obtain a goodpulp quality, the wood chips have to be properly impregnated with waterbefore refining. A high content of water extractives or a high content of corewood (dry wood with closed pores) might impede the penetration of water orchemicals. This is why the resulting pulps may have properties poorer thanthose, which may have been predicted from the wood fibre properties.

7.4.2.3 Variations from the base to the top Variations from the base to the top of the tree are mainly due to juvenile woodcontent and moisture content (heartwood/sapwood ratio) variations. Butt logsproduce pulps with the highest and the top logs with the lowest long-fibre content.On the other hand, pulp fines content are less sensitive to wood variations.Irrespective of whether the comparison is made as a function of freeness, sheetdensity, or energy consumption, the tensile index of the butt log pulps is greater thanthat of the second and top log pulps. Although the differences are even more pro-nounced, the same result is obtained in the comparison of tear index. The order ofthe different samples is switched when a comparison of optical properties is made.

Log position has a more pronounced effect on the inverse relationshipbetween the strength and optical properties of TMP than on wood basic density.Log position is an indication of wood age. As both wood density and fibreproperties are positively correlated to wood age, fibre properties must havea major influence on pulp properties.


Fig. 7.14a–d Comparison between thinning logs (juvenile wood) and sawmill chips (mature wood) forstrength properties and porosity.

Sawmill chips Sawmill chips Sawmill chips Thinning logs

1600

2000

2400

2800

3200


12 13 14 15 16 17 18 19 20

12 13 14 15 16 17 18 19 20

12 13 14 15 16 17 18 19 20

(m)

0

1

2

3

4

5

6

10000

2000

3000

4000

5000

6000

7000

8000

9000


°SR

Breaking length

Burst index

(a)

(b)

(c)

(kP

a m

2 g–1

)P

oros

ity

°SR

°SR


7.4.3 Forestry practice, site index and growth conditions

The increasing demand for wood products means that the forest industry willrely increasingly on fast-growing trees and short rotations (Zobel, 1984). Anythingthat changes the growth pattern of a tree can affect wood properties and conse-quently pulp and paper quality. Wood properties which may be affected bychanges in growth patterns are mainly juvenile wood content and the earlywood/latewood ratio. Their influence on pulp and paper quality has already beendiscussed. The specific gravity is another parameter influencing the pulp andpaper quality and it is related to forestry practices, site index, growth condi-tions and wood species (Zobel, 1984).

For mechanical pulp, the influence of basic density is expected to be minor.The areas of expected primary impacts on thermomechanical pulp are energyconsumption and strength properties: constant variations in basic density causenon-uniformity in pulp quality, especially in pulp strength, and elevatedenergy consumption (Bergander & Salmén, 1998).

A research programme on maritime pine (Chantre et al., 2000) has shownthe possibility of using stand parameters as predictors of pulp strength properties.Predictive models for assessment of fibre morphology and densitometricparameters from forest parameters and, from these, pulp strength propertiesmay be used to predict the pulp potential of a stand. The influence of growthrate on chemical pulp physical properties is clear (Fig. 7.15), by affecting, forexample, the ratio of latewood to earlywood. From this, it is possible to modelthe influence of growth rate on paper-strength properties, particularly burstand tear.


Tear index20

18

16

14

12

10

8

(d)

12 13 14 15 16 17 18 19 20

mN

m2 g

–1

°SR

Fig. 7.14a–d (continued).


7.5 The future

Breeding programmes are becoming increasingly important in forestry plans,for example, in South America, where specific clones selected on their pulppotential are planted in stands owned by the pulp mills and managed in short-rotation coppice. In other countries, breeding and selection based on pulppotential are emerging. This is the case for loblolly pine in USA, radiata pinein New Zealand, Norway spruce in Scandinavia or maritime pine in France.The benefits of integrating pulp traits in breeding programmes can be tremen-dous. Both pulp yield and pulp strength properties can be significantly increased.However, since trees are in almost all the cases (except Brazil already cited)planted for sawmilling, furniture or structural uses, there is competition betweenbenefits of a breeding programme focused on pulp traits or wood properties.

References

Alén, R. (2000) Structure and chemical composition of wood, in Forest Products Chemistry (ed. P. Stenius),FAPET/TAPPI, Helsinki, 11–57.

Andrews, E.K. (1986) Impact of fiber morphology and chemical composition on the kraft process andsubsequent handsheet properties. R&D Conf TAPPI Conf., Raleigh, North Caroline, Proceedings.pp. 111–116.

Bergander, A. & Salmén, L. (1998) Wood and its properties – a challenge for the improvement ofmechanical pulping. MWP/STFI Symposium, Stockholm.

Circumference at breast height/age

Burst indexC

ircum

fere

nce

of th

e pu

lp lo

g

201.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

6.75

6.50

6.25

6.00

5.755.505.255.00

100

80

60

40

Fig. 7.15 Chart giving the burst index as a function of the tree growth rate and the log girth for Frenchmaritime pine (Pinus pinaster) thinning logs.


Bergman (1972) Wood substances losses in some chip storage experiments. IUFRO SymposiumProtection of Wood in Storage, Norway and Sweden, XVII, 1–28.

Borga, P., Elowson, T. & Liukko, K. (1996) Environmental loads from water-sprinkled softwoodtimber: 2. Influence of tree species and water characteristics on wastewater discharges. Environmentaland Toxicological Chemistry, 15, 1445–1454.

Bucking, M., Eisenbarth, E., Jochum, M. & Padberg, A. (1997) Investigations on the live storage ofwindthrown timber of the species Norway spruce, Scots pine, Douglas fir and oak. Mitteilungen-aus-der-Forstlichen-Versuchsanstalt-Rheinland-Pfalz. No. 42–97, p. 148.

Chantre, G. (2000) Qualité des bois après la tempête. Recommandations pour une conservation optimaledes chablis, bibliographical report AFOCEL, unpublished.

Chantre, G., Bouvet, A., Sens, D., Robin, E. & Bongrand, O. (2000) Modelling Kraft fibre morphologyand paper properties from forest data: the example of maritime pine thinning logs. Proceedings ofTAPPI Pulping Conference/Process and Product Quality. Boston.

Chisholm, C.J. (1994) Towards a UK Research Strategy for Alternative Crops. Silsoe Research Institute,371pp.

Corson, S.R. (1991) Wood characteristics influence pine TMP quality. Tappi Journal, 74 N °11. Corson, S.R. (1997) Tree and fibre selection for optimal TMP quality. PAPRO New Zealand, Report

B211, March 1997. Da Silva Perez, D. & Chantre, G. (2002) Variabilité naturelle des fibres du pin maritime, dans le champ

des fibres végétales. Colloque ARBORA, Bordeaux. Dillner, B. (1972) Outside stored chips as raw material for pulping. IUFRO Symposium: Protection of

Wood in Storage, Norway and Sweden, XIII, 1–9. Eisenbarth, E. (1995) Storing windthrown alive. AFZ Allgemeine-Forst-Zeitschrift, 50, 82–84. Ekman, R. & Hafizoglu, H. (1993) Changes in spruce wood extractives due to log storage in water.

Seventh International Symposium on Wood and Pulping Chemistry, Beijing, China, 3, 92–96. Forbes, D.L., Shaw, J. & Taylor, R.B. (1997) Climate change impacts in the coastal zone of Atlantic

Canada, in Climate Impacts and Adaption, Canada Country Study, Atlantic Regional Report.Environment Canada, Bedford, NS.

Gladstone, W.T., Barefoot, A.C. & Zobel, B.J. (1970) Kraft pulping of earlywood and latewood fromloblolly pine. For. Prod. J., 20, 17–24.

Gross, M., Metzler, B. & Schumacher, P. (1996) Honey-fungus attack on windthrown wood storedunder water sprinklers. AFZ-der-Wald, 51, 329–332.

Heikkureinen, A., Lucander, M., Sirvio, J. & Varhimo, A. (1999) Effect of spruce wood and fibreproperties on pulp quality under varying defibration conditions. 1999 International MechanicalPulping Conference, Houston, Texas, pp. 11–34.

Houghton, J.T., Meira Filho, L.G., Callander, B.A., Harris, N., Kattenberg, A. & Maskell, K. (1996)Changements climatiques 1995. Contribution du GTI au Deuxième Rapport d’évaluation duGroupe d’experts intergouvernemental sur l’évolution du climat; GIEC.

Hunt, K., Basham, J.T. & Kemperman, J.A. (1978) Kraft pulping evaluation of decayed tremblingaspen (Populus tremuloïdes) from Ontario. Canadian Journal of Forest Research, 8, 181–187.

Janin, G. (1983) Microcuissons, microclassage microraffinage, mesure automatique de la longueur desfibres. Doctoral Thesis INPG 5th May 1983, 236pp. + annexes.

Kucera, L.J. & Katuscak, S. (1993) Conditions of windthrown Norway spruce roundwood after oneyear’s storage at Lukmanier’s canton. Schweizerische-Zeitschrift-fur-Forstwesen, 144, 873–892.

Kure, K.-A. (1997) The alteration of the wood fibres in refining. 1997 International Mechanical PulpingConference, Stockholm, Sweden. pp. 137–150.

Lonnberg, B. & Varhimo, A. (1981) Effect of Fomes annosus on wood and pulping properties ofspruce. Paperi-ja-Puu, 63, 269–272.

Marton, R., Goff, S., Brown, A.F. & Granzow, S. (1979) Tappi, 62 (1), 49. May, W.D. (1998) The Miles and May Model – a presentation. MWP/STFI Symposium, Stockholm. Medina, S., Deleuze-Brézins, C. & Chantre, G. (2001) Characterisation and competitiveness of the

French wood supply for the Kraft pulp manufacturers. Mixed hardwood Kraft pulp: modellingand optimising. TAPPI International Pulping Conference, Seattle.


Moltesen, P. & Bang, L. (1973) Storage losses in beech pulp wood stored at forest roadside for threeyears. IUFRO Symposium Protection of Wood in Storage, Norway and Sweden, IX, 1–16.

Moltesen, P., Dalgas, K.F. & Herlow, M. (1974) Storing windthrown spruce under sprinklers. Dansk-Skogforenings-Tidsskrift, 59, 253–295.

Moore, G. (1996) Nonwood Fibre Applications in Papermaking, Pira International, Surrey. Nilsson, T. (1972) Microorganisms in chip piles. IUFRO symposium Protection of Wood in Storage,

Norway and Sweden, XIV, 1–20. Oliver-Villanueva, J.V. & Sachsse, H. (1992) Keeping windthrown beech alive. Forst und Holz, 47,

227–232. Pape, R. (1999) Effects of thinning regime on the wood properties and stem quality of Picea Abies.

Scandinavian Journal of Forest Research, 14. Peek, R.D. & Liese, W. (1974) Erste Erfahrungen mit der Beregnung von Sturmholz in Niedersachsen.

Forst und Holzwirt, 29, 261–263. Pulkki, R. (1991) Synthèse de la littérature traitant des effets de la qualité du bois sur la fabrication des

pâtes et papiers. Fiche technique FERIC FT, 171, p. 8. Sahlberg, U. (1998) Variations of wood and fibre properties and their influence on the properties of the

mechanical pulp. RTH, Stockholm. Salmén, L., Lucander, M., Härkönen, E. & Sundholm, J. (1999) Fundamentals of mechanical pulping,

in Mechanical Pulping (ed. J. Sundholm), FAPET/TAPPI, Helsinki, 34–65. Scheffé, H. (1958) Experiments with mixtures. Journal of the Royal Statistical Society, 25, 344–350. Scheffé, H. (1963) The simplex-centroïd design for experiments with mixtures. Journal of the Royal

Statistical Society, 25, 235–263. Smook, G.A. (1997) Handbook for Pulp & Paper Technologists, 3rd edition, Angus Wilde Publications,

Vancouver and Bellingham. Stevenson, R. (2001) Chip Production and Quality. A One Day Tutorial. TAPPI Pulping Conference,

Seattle. Sundholm, J. (1999a) What is mechanical pulping?, in Mechanical Pulping (ed. J. Sundholm), FAPET/

TAPPI, Helsinki. Sundholm, J. (1999b) Mechanical pulping. Papermaking Science and Technology, Book 5. TAPPI

Press 1999, 427pp.Tyrvaïnen, J. (1995) Wood and Fiber Properties of Norway Spruce and its Suitability for Thermo-

mechanical Pulping. Acta Forestalia Fennica, 249. Varhimo, A. & Tuovinen, O. (1999) Raw materials, in Mechanical Pulping (ed. J. Sundholm), FAPET/

TAPPI, Helsinki.Verkasalo, E. (1993) Deterioration of birch timber during prolonged storage in the forest and its effect

on yield, quality and value of rotary cut veneer. Folia Forestalia, 806, p. 31.Viala, D. (1963) Contribution à l’étude des modifications anatomiques et chimiques des bois attaqués

par des Stereum. Annales de l’Ecole Nationale des Eaux et Forêts, XX, 373–399. Vroom, K.E. (1957) The H-factor/the means of expressing cooking time and temperatures as single

variable. Pulp and Paper Magazine of Canada, 58, 228–231. Wang, X. & Braaten, K.R. (1997) Growth rings and spruce pulpwood sorting. Nordic Pulp and Paper

Journal, 12 N°3. Watson, W.F., Koger, J.L., Dubois, M.R. & Belli, M.L. (1992) A comparison of storage under sprinklers

versus open storage on the quality and yield of chips from southern hardwood. TAPPI PulpingConference, 1–5 November 1992, pp. 473–475.

Whitman, F.A. (1957) The effect of pulpwood aging on groundwood brightness. Tappi Journal, 40,20–24.

Zobel, B. (1984) Wood where will it come from where will it go? School For. Res., 1–12. Zobel, B.J. & van Buijtenen, J.P. (1989) Wood Variation. Its Causes and Control. Springer-Verlag,

Springer Series in Wood Science, 363pp.

8 The mechanical properties of wood Audrey Zink-Sharp

8.1 Introduction

Humans have utilized the unique and first-rate mechanical properties of woodsince tools were first crafted. From primitive times to today, wood remainsthe construction material of choice. The enduring use of wood in structurescan be traced to many characteristics, including ease of fabrication andconversion, favorable strength to weight ratio, impact resistance, dimensionalstability, extreme versatility, low conversion energy and sustained availability.The natural beauty and warmth of wood is unmatched in other architecturalmaterials.

The mechanical properties of wood make it a material that can be used toconstruct a variety of structures, ranging from conventional residential buildingsto modern large-scale structures like domes, bridges, or industrial complexes.Worldwide, more buildings are constructed with wood than with any otherstructural material. Most of these are wood-framed single-family homes, butmany larger multi-family and commercial buildings also use wood framing.An example of multi-story apartment complexes constructed with roof trussesmade of structural lumber and oriented strandboard sheathing is seen in Fig. 8.1.Other wooden building systems utilize pole building construction, and glue-laminated (glulam) beams and arches. Wood is also commonly used in com-bination with other construction materials such as concrete, brick, and masonry.It is fire-resistant in large cross sections, light weight and durable, and pro-duces aesthetically pleasing structures. Wood is the only construction materialthat is truly a renewable resource.

Wood is available for structural applications in many forms. The mostobvious is sawn lumber, which is wood that has been manufactured by simplycutting it directly from a log. Other structural materials such as glulam beamsand arches start as lumber and then undergo additional processing. Manyother wood-based products are available for structural applications includingstructural composite lumber, structural-use panels such as plywood andoriented strandboard, and manufactured components such as trusses, wood Ijoists, and box beams. A better understanding of the mechanical propertiesof wood, lumber, plywood, and the other engineered wood products will enableresearchers, engineers, and scientists to utilize the full range of possibilitieswith wood materials.


8.2 Advantages and disadvantages of wood as a structural material

Structural applications are those in which strength is usually the primarycriterion for adequacy of a material. The term strength has many meanings incommon usage, but in the setting of structural materials, strength is defined asthe ability to carry or resist applied loads or forces. The strength of wood determinesits mechanical performance and is an important factor in drying, machining,bending, gluing, and fastening (Hoadley, 1992).

Structural uses of wood products include floor joists and rafters in wood-frame homes, heavy timber construction, beams and stringers, power linetransmission poles, plywood roof sheathing and subflooring, glulam beamsand decking in commercial buildings, particleboard flooring in mobilehomes, steps and rails of wooden ladders, sailboat masts, and frames ofupholstered furniture (Haygreen & Bowyer, 1996). When wood is used instructural applications, it offers great challenges as well as opportunities tothe user and the designer. Wood is widely available in a variety of shapes andsizes and offers a greater mixture of unique characteristics than other structuralmaterials.

Fig. 8.1 Multi-story apartments buildings constructed with timber trusses in the roof and orientedstrandboard sheathing.

THE MECHANICAL PROPERTIES OF WOOD 189

8.2.1 Advantages

8.2.1.1 High strength and flexural rigidity in spite of light weight Wood is unusually strong for its weight. For example, it takes almost 267 N ofload to cause failure of a Douglas-fir (Pseudotsuga menziesii) wood beamof span length 40 cm and 1.27 by 1.27 cm in cross section. The wood beamitself only weighs about 120 g. Picture this as a bucket holding just a little over36 liters of water dangling from a very small, light-weight rod of wood. Thefavorable strength-to-weight ratio of wood is traceable to the nature of cellwall material and its distribution as a system of thin-walled tubes (Panshin &de Zeeuw, 1980). Because the cell wall constituents are distributed in the formof thin tubes surrounding a hollow lumen, the flexural rigidity is enhancedover solid rods. As a result, wood exhibits good rigidity in comparison to solidstructural materials. This feature makes wood well suited for use in situationsthat require elastic stability such as long beams and columns or stress skinconstruction. However, wood suffers in comparison with metals for uses thatrequire high shear and compression resistance because the distribution of thintubes that enhances rigidity in bending reduces the shear and compressionefficiency (Panshin & de Zeeuw, 1980).

8.2.1.2 Available and renewable resource Wood is an unusual structural material because its supply can be renewed byforest regeneration. Forest management techniques which take into considerationrapidly changing wood technology can ensure a continuing supply of structuraltimber. However, production and consumption of wood materials continue toincrease at a somewhat staggering rate. For example, the weight of wood usedevery year in the United States is roughly equal to the weight of all metals,plastics, and Portland cement combined. Fortunately, modern forest managementpractices have kept pace with the high demand for wood materials, and forestsworldwide are generally increasing in area coverage or are stable in size (Sedjo& Lyon, 1990). Additionally, new areas for management are becoming availablein the far eastern region of Russia, northern Europe, and several parts of SouthAmerica. Improvements in reforestation, material properties, processing, andconversion are extending the supply of structural timber.

8.2.1.3 Requires less energy to process into structural material Wood has a substantial advantage over other materials in terms of energyconsumption per unit of finished products (National Research Council, 1976).Even with the significant advancements in technology since 1976, recent studieshave confirmed the advantages of wood (Buchanan, 1991; Marcea & Lau,1992; Meil, 1993). Wood-frame construction was found to require far less energythan steel, aluminum, concrete block, or brick when the energy required to buildwall systems for residential homes was examined. Energy use associated with


raw-material gathering (harvesting or mining), transportation, manufacturing,and building construction was considered in the evaluation. Wood constructionon a concrete foundation was found to require only 35% as much energy as steelconstruction on a concrete foundation. Office and industrial construction usingtimber required only 55% as much energy as steel construction and 66–72% asmuch as concrete construction. Wood-frame construction of residential buildingswas found to require only 42% as much energy as a brick-clad, steel-frameddwelling. Even though all types of construction, including wood frame, involveenvironmental impacts, Meil (1993) discovered that environmental impactssuch as air emissions, water-borne effluents, and solid wastes, from woodconstruction are minuscule compared to those of steel. When recycled steel isused, the differences are smaller, but wood retains a significant advantage.

8.2.1.4 Ease of fabrication and conversion Simple hand and power tools can be used to manufacture products in the factoryas well as on the construction site. This plus low cost is largely what keepswood-frame construction fully competitive with other materials and methods offabrication (Loferski, 1997). Wood and wood-based composites can be joinedwith simple connectors and adhesives, which allows on-site assembly fabricationof differing shapes and almost unlimited dimensions. Large wood trusses and trusssystems, laminated beams and arches, and stress-skin panels have made woodconstruction competitive in building and represent an opportunity to engineer highquality/low-cost products from a diminishing supply of high quality lumber.

8.2.1.5 Dimensionally stable and durable if used correctly When wood is used in conditions that are not favorable to wood-degradingagents, it is quite durable and stable. Wood will endure continuously if keptdry, and if it is used in situations where accessibility of biological organisms isminimized. In addition, wood can be treated with preservatives that repelinsects and poison the food supply. Preservative-treated wood that is also keptdry will provide desirable structural behavior indefinitely.

8.2.1.6 Low electrical, thermal, and acoustical conductivity These properties are unrelated to strength but nonetheless are advantageous ina structural material. Other structural materials do not naturally provide goodelectrical, thermal, or acoustical insulation, and extra materials must beincluded in the structure at an additional cost.

8.2.2 Disadvantages

8.2.2.1 Variability Because wood is produced in a biological environment and the tree is subjectto varying growth conditions, it is a highly variable material. Coefficients of


variation (COV) range from 10% for relative density to 34% for work to maximumload in static bending. Within any species of wood, there is considerable variationin material properties, and even from location to location within individualpieces of wood. In the past, this uncertainity was managed with overdesign andmaterial redundancy. Wood’s variability has even been considered by some tobe of such magnitude that alternative materials were used. Fortunately, advancesin mathematical analyses and engineering techniques have produced morereliable wood structures using composites which are less variable than previoussolid wood materials.

8.2.2.2 Natural built-in defects Material characteristics such as knots, spiral grain, reaction wood, moisturepockets, rotted areas, stains, etc. are all natural features of wood but they all serveto reduce the value and strength of wood in some manner. Modern structuralcomposites have minimized and randomized the influences of these naturallyoccurring defects, but the dimensional instability of wood in service remainsa definite problem.

8.2.2.3 Dimensional instability Wood is an anisotropic, hygroscopic substance that has a chemical affinity forwater vapor. Confounding the problem is the capillary structure of wood thatmakes it an excellent sponge when in contact with liquid water. As a result,a piece of wood will change dimensions when moisture is absorbed or des-orbed due to changes in environmental conditions or when it comes in contactwith water. The dimensional changes are not equal in the three structuraldirections and the result is warp, splits, and checks, all of which can result inunacceptable quality or behavior if of sufficient magnitude.

8.2.2.4 Susceptibility to biological attack Since wood is biodegradable, it is reduced to its chemical constituents throughattack by biological organisms such as fungi, bacteria, and certain insects.Degradation due to biological attack can severely reduce the mechanicalproperties of wood and cause structural instability. Fortunately, there arepreservative chemicals that can be applied which significantly reduce thesusceptibility of wood to biological degradation.

8.2.2.5 Anisotropy Wood is an anisotropic material in its cellular organization, and this results inanisotropic physical properties as well. Properties parallel to the grain differsignificantly from those in the transverse direction. In fact, wood is consideredan orthotropic material because it exhibits different properties in three mutuallyperpendicular directions or axes. Mechanical properties are a good example ofthe anisotropy of wood – values for strength and stiffness along the grain


versus across the grain may vary by a factor of 20 or even more. In addition,transverse properties radially and tangentially also differ. Obviously, it ispossible to engineer and design with wood while lacking a working knowledgeof anisotropy or ignoring its influence but inefficient usage, material misuse,and even disasters can result.

8.2.2.6 Combustibility To many, this is clearly an attractive feature but where structural applicationsare concerned, the combustibility of wood can be very much a disadvantage ifoverlooked when used in light frame and timber construction.

8.3 Importance of density

There is probably no other single physical property that has as significant animpact on the mechanical properties of wood as density. Almost all mechan-ical properties of wood are closely connected to density, some more so thanothers. In fact, density is most likely the best single predictor of mechanicalproperties of clear, straight-grained defect-free wood. Properties such as elas-ticity in bending and maximum crushing strength parallel to the grain increaselinearly, others increase through a power function, and some are only slightlyaffected. Dependence of material properties on density is certainly well known,but no other structural material relies so heavily on natural-growth characteris-tics to impart density. In wood, density is a reflection of the amount of cellwall material present per unit area, and this in turn is a function of the type andsize of cells and the amount of latewood versus earlywood in the piece underconsideration, among other factors. Additionally, relative percentages of latewoodand earlywood are functions of growth conditions and tree genetics. As a result,density varies within individual pieces and across wood types.

In structural applications, density is commonly calculated as weight per unitvolume, not mass per unit volume. Weight density, or weight of wood per unitvolume, is usually calculated assuming that both the weight and the volume aretaken at the same moisture content. However, density is sometimes calculated andreported in ways that mix weight and volume at different moisture contents.Consequently, knowledge of the basis for calculation is required. Density isexpressed as pounds per cubic foot, grams per cubic centimeter, or kilogramsper cubic meter using the following equation:

(8.1)

Dry, solid cell wall substance has been demonstrated to have an averagedensity of 1.53 g cm−3 (Kellogg & Wangaard, 1969). Little variation exists in

Weight densityweight of wood with moisturevolume of wood with moisture--------------------------------------------------------------------------=


this value for different kinds of wood. However, due to the manner of calculation,density of wood bears a direct relationship to the moisture content. Whenmoisture is absorbed into the cell wall, the wall increases in volume with theincreasing moisture content. This lessens or reduces cell wall density belowthat for solid wood. Minimum values for weight density occur at the oven-drycondition and maximum when wood is fully saturated. Moisture content atwhich the density was calculated should be included whenever density isconsidered. Additionally, it is just good practice to calculate density usingweight and volume at the same moisture content to avoid confusion andinvalid comparisons.

Because there is very little difference in the density of cell wall substance,density is in effect a measure of porosity, i.e. proportion of void volume. Mechan-ical properties are directly linked to relative proportion of solid cell wallsubstance – the inverse of void volume. Consider a piece of white oak (Quercusspp.) with a dry density of 620 kg m−3 and a piece of balsa (Ochroma pyrami-dale) with a dry density of 180 kg m−3. The oak has a void volume of about38% and about 62% solid cell wall material and the balsa, 82% void, 18% solid.The direct relation to mechanical properties is immediately obvious when wecompare the oak and the balsa: the oak which has a 62% solid cell wall willcertainly withstand a higher crushing load than the balsa with 18% cell wallvolume.

In some sectors of wood science, the terms relative density and specific gravityare used interchangeably with density. However, they are distinctly different eventhough they refer to the same material characteristic. In general terms, specificgravity (or relative density) is the ratio of density of a material to density ofwater. In wood science, specific gravity is always calculated using the oven-dryweight (ODW); however, the sample volume used in the calculation can be atany given moisture content (X% MC). The following equation is used.

(8.2)

where X% MC, the moisture content at which the volume was measured; ODW,oven-dry weight of the piece (weight at 0% MC).

There is some difference in specific gravity values at moisture levels belowthe fiber saturation point due to small changes in volume with changes in moisturecontent, but the difference is small compared to the change in density withmoisture. This is because when density is calculated, the weight used in theequation is the weight with moisture, not the oven-dry weight, and both theweight and the volume change with moisture content (see Equation 8.1). Owingto their relevance to properties and quality, this is an important distinction tounderstand when evaluating density and specific gravity values.

Specific gravityX% MC

ODW/volumeX%MC

density of water -----------------------------------------------=


8.4 Mechanical properties important for structural applications

Since wood is composed of long, slender, hollow tubes aligned parallel to theaxis of the tree trunk, it may be thought of as a bundle of tubes (see Fig. 8.2)aligned in three mutually perpendicular directions. The tubes are quite strongwhen forces are applied parallel to their long axis, but will collapse easilywhen the load is perpendicular (a bit like pushing on a rope). This explainswhy the properties of wood are highly dependent on the direction in whichthey are measured. Mechanical property values along the three material axesmay vary by a factor of 20 or even more, and therefore wood is a material thatis extremely anisotropic.

A thorough discussion of the mechanical properties that characterizematerials exceeds the scope of this discussion. Consequently, only the mostfundamental and commonly used will be included here. Mechanical proper-ties are often divided into two categories: elastic and strength properties.Elastic properties are those that indicate stiffness of a material, whilestrength reflects the load-carrying capacity or resistance to forces. Theproperty most critical to any application is determined by the type of loadingthat component will be expected to withstand. For example, the strength of along column is partially controlled by resistance to bending, i.e. the modulusof elasticity, but on the other hand, a short column depends on compressionstrength parallel to the grain.

8.4.1 Elastic properties

Elastic properties are a measure of a material’s stiffness and include modulusof rigidity and modulus of elasticity (MOE or simply E). Modulus of elasticitycan be either a measure of axial stiffness (sometimes called Young’s modulus)

LR

T

Fig. 8.2 Highly idealized model of wood cells showing the tubular structure.


or a measure of stiffness in bending. Axial MOE measures a material’sresistance to elongation or shortening under axial tension or compression load-ing. MOE in bending is a measure of bending or beam stiffness. Wood is amaterial that exhibits elastic properties within a certain range of loading, butpast that range, additional increments of stress result in increasingly largerincrements of strain as illustrated in Fig. 8.3.

The proportionality constant that relates stress to strain is called MOE. Inother words, MOE is the ratio of the amount a material will deflect in propor-tion to an applied load within the elastic range. The higher the MOE value, thestiffer the wood and the lower the deformation under a given load. In axialloading, the limit of proportionality between stress (load) and strain (deforma-tion) usually occurs at about one-half to two-thirds the maximum stress(Hoadley, 1992).

As mentioned, wood is an orthotropic material that exhibits different MOEvalues (EL, ER, and ET) when load is applied longitudinally (L), radially (R), ortangentially (T). Since EL is the stiffness parallel to the grain, it exhibits thehighest magnitude. Values for ER and ET are often given as ratios with EL.Ratios of ER and ET to EL range from a low of 0.015 in balsa (Ochroma spp.) toa high of 0.197 in black cherry (Prunus serotina). In addition to varying withinand between species, stiffness is influenced by moisture content and density.Increasing moisture content up to the fiber saturation point results in decreasedstiffness. The opposite is true for density: increased density results in increasedstiffness.

X

Str

ess

(or

load

)

Strain (or deflection)

Proportionallimit

Slope isstress/strainor MOE

Fig. 8.3 Plot illustrating the relationship between stress and strain in a typical compression parallel-to-grain test.


Since wood is an elastic material up to a point, deformations imposed belowthe proportional limit are completely recoverable upon removal of the load.Past the proportionality limit, permanent deformations occur that are notrecovered when the load is removed. Examples of non-recoverable deforma-tions are the hammer missing the nail with the result that the wood ispermanently dented, a diving board that remains slightly bent after the diverhas bounced away, hygrothermal stresses that occur during drying perman-ently warp a board, and furniture parts that are permanently deformed intocurved patterns during steam bending. Proper design of wood structuresensures that wood members used in construction will be stressed only withinthe elastic limit.

Modulus of rigidity (G), also called the shear modulus, is a measure ofstiffness in resistance to shear forces. Shear forces attempt to make one part ofa material slide past the material adjacent to it. Wood exhibits three G con-stants, GLR, GLT, GRT in the LR, LT, and RT planes, respectively. For example,GLT is the modulus of rigidity based on shear strain in the LT plane and shearstresses in the LR and RT planes. Shear moduli are always lower than axial andbending MOE. Ratios of G to MOE range from a low of 0.037 in balsa(Ochroma spp.) for GLT/EL to a high of 0.220 in northern white cedar (Thujaoccidentalis) for GLR/EL. As with MOE, modulus of rigidity varies within andacross species, and with moisture content and density.

8.4.2 Strength properties

These include compression, tension, and shear strength parallel and perpendicularto the grain, bending strength (MOR – modulus of rupture), toughness, resilience,side hardness, and work to maximum load. The properties that are most com-monly measured and used in design are MOR, maximum stress in compressionparallel to the grain, compressive stress perpendicular to the grain, and shearstrength parallel to the grain. Toughness, resilience, side hardness, and work tomaximum load are less common strength properties and usually do not playa significant role in wood structures.

MOR is a measure of the maximum load-carrying capacity in bending, i.e. theload a beam will carry. Compression parallel to the grain reflects maximumstress sustained by a compression load applied parallel to the grain in a shortspecimen (slenderness ratio (l/d) >11), i.e. the load a short post or column willcarry. Compression stress perpendicular to the grain is not a maximum strengthvalue but it is reported as stress at the proportional limit in a specimen loadedperpendicular to the grain. During compression perpendicular to the grain, woodcells crush upon themselves, densify, and continue to carry load indefinitely. Asa result of the densification, there is no clearly defined maximum stress for theproperty. Connections between wood members in buildings and at the supportsfor a beam rely on compression strength perpendicular to the grain.


Wood is quite strong when stressed along the grain. Something as small as ashort column of tamarack (Larix laricina) with a 2.54 cm2 cross section willwithstand almost 50 kPa when stressed in compression parallel to the grain. Itis weaker when loaded perpendicular to the grain or in shear. Compressionperpendicular to the grain is imposed by heavy objects resting on tables or floors,heavy columns resting on a wooden deck or bookshelves on floors, and incases where swelling is restrained such as dowels in joints. Wood exhibitsrelatively low values for compression perpendicular to the grain, and thisproperty can be a limiting strength. Tables of commonly used strength andelastic properties are found in many engineering handbooks, for exampleTables 4-1 through 4-7 in Wood Handbook: Wood as an Engineering Material(FPS, 1999).

Strength properties are very closely linked to density. In fact, where gradingrules for structural timber are not highly developed, grading is based primarilyon density alone. Some properties increase with density more rapidly thanothers, for example, maximum crushing strength parallel to the grain increaseslinearly while side hardness increases through a power function.

Other factors affecting the strength of clear wood include moisturecontent, time, temperature, and exposure to chemicals. Of these factors,moisture content has the most deleterious influence and is the most meaning-ful in practice. As wood dries below the fiber saturation point, most strengthproperties increase. An exponential relation that involves the properties ofgreen wood and wood at 12% MC can be used to estimate the strength valueat any given moisture content. The moisture–strength relationship variesamong species, so estimates cannot be used to generalize all predictions.Changes in moisture content do not influence the properties equally.Compression perpendicular to the grain is the most influenced followedby compression parallel to the grain, MOR, with shear strength the leastinfluenced. The strength of wood does not change significantly over timeunless it is subjected to damaging factors such as decay, high temperatures,extreme moisture fluctuations, or powerful chemicals. If continuous loadingis encountered, there may be some loss of strength and an increase in thedeformation. The influence of time on deformation is more fully described inthe next section. As with most materials, the mechanical properties of wooddecrease when it is heated and increase when it is cooled. However, as longas temperatures do not exceed 100°C, there is little permanent strength loss.Exposure of wood to extreme chemicals such as severe alkaline or acidicconditions may reduce strength, but wood is actually more resistant thansteel to acidic conditions. Strength losses can result from hydrolysis ofcellulose, oxidation by oxidizing agents, or delignification by alkalis.Chemical deterioration can occur near iron or steel fasteners due to chemicalinteractions with moisture that create iron salts which promote localizedweakening of the wood.


8.5 Creep effects on deformation and fracture

Structural components are frequently subjected to constant loading overextended periods. The time-dependent deformation caused by these sustainedloads is called creep. Some structural materials are prone to creep and othersare not. Moreover, some materials will creep under certain conditions and notunder others.

Structural timber and composites are examples of materials that undergocreep. In these materials, creep is thought to occur due to slippage in therelative position of the long-chain molecules in the cell wall. The bonding sitesshift and slide with respect to adjacent molecules when stress is applied overextended periods. Creep is influenced by the amount of water in the cell wall.The more water present in the cell wall, the more easily this slippage will occur.Ordinary climatic variations in temperature and humidity will cause creep toincrease. An increase of about 28°C in temperature can cause a two- to threefoldincrease in creep (FPS, 1999). Green wood may creep four to six times the initialdeformation as it dries under load. Creep behavior is approximately the samefor the various commercial structural timbers but generally higher for thewood-based composites (Gnanaharan & Haygreen, 1979). Creep is greater underhigher stresses than lower ones, and is more likely to occur when the stress isapplied across the grain.

Common examples of creep are gradual sagging of the header beam overa garage door after many years of loading, and drooping of a shelf heavilyloaded with books. A number of factors influence the magnitude of creep,including composite and wood type, geometry, duration of stress, imposedstress level, and the surrounding environmental conditions. Laboratory researchhas discovered that strength of a wood member decreases about 8% for eachtenfold increase in load duration. Thus, strength is approximately linearlyrelated to the logarithm of time (Schniewind, 1982). In situations where loadswill be imposed on solid timber, glue-laminated timber, or plywood for periodsof load from 6 months to 10 years, maximum design loads are reduced by 30%and for permanent loads (all those beyond 10 years), an additional 10% reductionis imposed (CEN, 1995).

Creep rupture (fracture) occurs because of sufficiently high stress overextended periods. In situations where wood must continuously carry a load fora long period, the load required to produce fracture is much lower than that fora short-term loading situation. For example, a wood beam under continuousaction of bending stress for 10 years may carry 70% or less of the load requiredto produce fracture in the same specimen loaded in a standard bending strengthtest of only a few minutes duration (CEN, 1995). For the creep rupture processto occur, a wood specimen must undergo substantial deformation prior tofracture. However, the deformation at fracture is approximately the same forcreep rupture as for standard short-duration laboratory strength tests. Changes


in environmental conditions hasten the rate of creep and shorten the time tocreep rupture. This influence is greatest for small wood specimens and largecyclic changes in temperature and humidity. Fortunately, modern design practicesand most in-use situations minimize circumstances conducive to excessivecreep in wood materials.

8.6 Defects affecting mechanical properties

When mechanical properties of wood are evaluated, many naturally occurringfeatures such as knots and sloping grain are considered defects. Other defectssuch as splits and checks occur during sawing, sorting, drying, and shipping.Improper use or installation and changing environmental conditions such ashumidity, temperature and time also influence mechanical properties of wood.This section will cover only the most commonly encountered natural andprocessing defects.

8.6.1 Naturally occurring defects: knots and sloping grain

8.6.1.1 Knots Knots are branch bases embedded into the tree trunk. Because all trees havebranch bases included in the trunk, knots are the most commonly encountereddefect. Cell structure in branches is significantly different from that of trunkcells, and alignment of longitudinal cells in branches is at an angle to cells inthe tree trunk. In addition, longitudinal cells in the trunk must deviate aroundthe branch base which causes localized grain alterations as seen in Fig. 8.4.

Knots can adversely influence mechanical properties because of the hetero-geneity they introduce, and create stress concentrations due to interruption ofthe continuous, parallel arrangement of the trunk cells. Frequently, checksoccur around the knots during the drying process.

Degree of strength reduction varies from significant to minimal dependingon the size and quantity of knots. A few very small, round pin knots havenegligible impact, but numerous, large, loose spike knots can actually cause apiece to fall apart when dried. A conservative approach to predicting strengthreduction is to consider knots as empty holes, and then total strength isassumed to come from what solid wood remains. Influence of grain deviationis more difficult to visualize but is equally significant because it compoundsstrength reductions over that of empty holes. For example, a grain slope of2.54 cm in 15.14 cm (1″ in 6″) has been found to reduce bending strength by60% in the region of the knot. Knot location and load type must also be con-sidered. Knots on the top or bottom edge of a beam are more critical than thesame knots located near the center. Tensile strength is more affected by knotsthan compression strength. A 2.54-cm knot on the edge of a 5 × 20 cm beamwill reduce the strength by 23% but the same 2.54-cm knot at the center


reduces the strength by only 12%. Examples of the amount of strength reduc-tion from knots are shown in Table 8.1. Influence of knots is incorporated intograding rules and strength classes such as those outlined in British StandardsBS EN 384:1995, BS EN 518:1995, BS EN 519:1995, BS EN 1912:1998 (BSI,2001), and for North America – National Design Specifications for WoodConstruction (AF & PA, 1995).

8.6.1.2 Sloping grain

Straight-grained wood products are, of course, the goal during fabrication, butin reality grain deviations usually occur to some extent in every piece. Devi-ations in grain from parallel to the longitudinal axis occur naturally in the treefrom knots, spiral, or interlocked grain, and they can result also from failure tosaw boards parallel to the bark.

The terms slope of grain or sloping grain are used to describe deviation oflongitudinal cells from a line that is parallel to the long edge of a piece of

Fig. 8.4 (A) Edge knot.


lumber or wood product. An example of sloping grain in lumber used in anoutdoor bench is shown in Fig. 8.5.

As seen in the figure, surface splits in the wood are parallel to the grain, whichis not parallel to the long edge of the board. Slope of grain is expressed as theratio of unit deviation across the grain to the corresponding distance alongthe grain. For example, sloping grain of 2.54 in 25.4 (1 in 10 or 1:10) would bethe length (25.4 cm) through which a 2.54-cm deviation in the grain occurs.Localized deviations around knots are disregarded in the general grain slopemeasurement of an entire piece of structural lumber.

Sloping grain is best detected by observing orientation of resin streaks,hardwood vessels, surface checks, mineral stains, and other features that tendto be oriented with the longitudinal cells. Orientation of growth rings can berelied on only with true radial surfaces, and rays only with true tangential surfaces.The most reliable method is to split a short scrap and measure deviation fromthe long axis of the piece.

Sloping grain always reduces the strength of wood. This is because loads placedparallel to the axis of the piece are no longer applied parallel to the longitudinaldirection of the cells and are not resisted by the superior parallel-to-the-grain

Fig. 8.4 (continued) (B) Grain deviation around a circular knot.


Table 8.1 Strength reduction in lumber resulting from knots and slope of grain(ASTM Standard D245; ASTM 1987)

Width of beam (cm)

Size of knot (cm) 9 14 19 24

Strength reduction from knots in the center of the wide face of a beam (%) 2.54 25 16 12 10 5.08 51 33 24 20 7.62 – 50 37 30

Strength reduction from knots on one edge of the wide face of a beam (%) 2.54 43 30 23 18 5.08 81 55 43 35 7.62 – 79 63 50

Slope of grain (cm) In bending In compression

Strength reduction in bending and compression parallel to the grain resulting from sloping grain (%) 2.54 in 15.24 (1″ in 6″) 60 44 2.54 in 20.32 (1″ in 8″) 47 34 2.54 in 25.4 (1″ in 10″) 39 26 2.54 in 38.1 (1″ in 15″) 24 0 2.54 in 50.8 (1″ in 20″) 0 0

Fig. 8.5 Illustration of sloping grain (arrow indicates grain direction).


strength of wood in tension or compression. The consequence of sloping graindepends on the extent to which the weaker perpendicular-to-the-grain strengthcomponent is required to act. Limitations on slope of grain are incorporatedinto all grades of structural timber. Tension strength is more severely affectedby sloping grain than either bending or compression. As seen in Table 8.1,sloping grain no worse than 1:15 has a minimal influence in compression loading,but a 24% strength reduction is taken for the same 1:15 slope when loaded inbending. In many standards, there are especially strict limitations in place forcritical structural members such as ladder rails (Hoadley, 1992).

8.6.2 Processing defects: checks and splits

8.6.2.1 Checks Checks are material separations caused by non-uniform dimensional changessuch as moisture content decrease during the drying process. A diagram ofend checks and a split is shown in Fig. 8.6. Because more shrinkage occurstangentially to the growth rings than perpendicularly (radially), radial checksare created perpendicular to the growth rings. Radial checks appear on theends of boards and are usually called end checks. Longitudinal checks (alsocalled surface checks) can occur during drying because the exterior dries andshrinks before the wood at the inner core, and non-uniform shrinkage createsstresses that exceed tension strength. Checks and splits are also affected bythe presence of high residual stresses. Longitudinal drying checks are morecommon in thick materials such as heavy timbers, and reductions in shearstrength are taken into account in lumber grading rules and design values.Figure 8.7 illustrates examples of longitudinal checks (Fig. 8.7A) and endchecks (Fig. 8.7B). Checks can noticeably reduce stiffness and load-carryingcapacity. Compression parallel to the grain tends to be influenced to a greaterdegree than tension strength. When a severely checked column of wood isloaded in compression, it will tend to act as several individual columns eachof which will have a greater tendency to buckle due to their higher slender-ness ratio (Bodig & Jayne, 1982).

8.6.2.2 Splits Splits are complete separation of the wood cells across the entire thickness of apiece. A split in the end of a board is illustrated in Fig. 8.8. Splits and checksthat are sufficiently large in relation to the cross section can reduce strengthsignificantly and induce failure at lower loads than expected. Extension of pre-existing splits and checks can even result in failure in tension perpendicular tothe grain or shear parallel to the grain at much lower stress levels than wouldcause failure of intact wood. Checks and splits may detract from the appear-ance of structural timber and possibly limit their usage in some applications. Inaddition, they enhance admittance of moisture and biological degraders such


as insects and decay fungi. Location of splits and checks is also an importantconsideration. Splits located midspan in bending members are more critical tobending stress and deflection, but splits and checks on the ends of the beamhave minimal influence. Separations due to splits and checks must be relativelylarge and located in a critical zone before performance will be significantlyaffected. However, lumber grading rules do set limits on these types of defects.

8.7 Problems with mechanical joints

While it is possible to construct simple mechanical joints in wood with simpletools and fasteners, even simple mechanical joints can create very complexproblems. Adequately addressing mechanical connections in wood and theaccompanying problems alone could fill many chapters; consequently this

Fig. 8.6 Diagram of various types of checks and split in timber.

SplitCompleteseparationof woodcells

Longitudinalsurface checksAlong grain

End checksPerpendicularto annual rings


section is limited to only the most common and focuses primarily on thematerial issues rather than engineering design issues.

Structural lumber and timber can be connected with a variety of mechanicalfasteners such as nails, dowels, staples, bolts, lag screws, and punched metalplate fasteners (also known as toothed metal plates and metal plate connectors),among others. It has been estimated that about 75000 fasteners (mostly nails)are used in the average wood-frame residence (Hoadley, 1992). Complex

Fig. 8.7 (A) Longitudinal surface checks, and (B) end checks.


mathematical analyses and centuries of experience have gone into current con-struction practices that ensure sound mechanical connections; however, poorlymade wood joints occur with considerable regularity.

Many of the problems encountered with mechanical joints can be traced tofactors inherent in the wood and its sensitivity to environmental conditions suchas temperature and humidity. Fasteners have undergone many design analysesand iterations, and the modern versions are generally not the source of problems.However, misuse or lack of understanding of modern fasteners and woodproperties is frequently encountered. For example, research on nails has shownthat for many woods, a spear point and slim taper results in the greatest holdingcapacity, but the wood cells tend to separate. So, if this type of nail is used incertain woods, it will cause splitting. In contrast, a blunt point has fewer tendenciesto split but has reduced holding power. Therefore, a diamond point nail shouldbe used as a compromise between highest holding power and least splitting.

It is usually assumed that predrilling nail holes is unnecessary. Lack of pre-drilling often leads to splitting with certain combinations of nail and wood types.However, it is well known that the best holding power with nails can be achievedwith prebored holes. Another common assumption is that all connections madeby nailing into end grain will be inferior. Predrilling holes can minimize low

Fig. 8.8 End split (at arrow) extending into length of piece.


holding capacity when nails are driven into end grain. Unfortunately, thistime-consuming practice is usually omitted in routine construction.

Other problems encountered with nails and fasteners occur over time andmany are due to changing environmental conditions. When extreme moisturevariation causes dimensional changes in wood, nail-holding capacity can bereduced by as much as 80%. Nails can even become loose and pop out (i.e. nailpopping or nail withdrawal) of the wood member if the moisture variations arecyclic in nature and subject the connection to routine dimensional changes.Specialty nails such as resin-coated nails and annularly or spiral-threaded nailshave been developed to minimize the reduction with time and the tendency towithdraw due to cyclic moisture variations.

Perhaps it goes without saying that a fastener must be strong enough totransmit the applied load without bending, but fastener yielding does occurregularly. An example is a long, slender nail that crushes the wood near thesurface and bends. It then becomes loaded in withdrawal and works its way outof the hole.

Nails are usually designed for connections where the loads transmitted arerelatively small and other types of fasteners (e.g. bolts) are used for larger loads.However, recent trends in Europe and New Zealand use nails in connectionssubject to relatively large loads. In this case, a large number of nails are requiredand special construction practices are necessary. Adequately connecting woodwith any mechanical fastener requires the right combination of moisture content,material, diameter, length, and common sense.

Correctly connecting with screws also requires a combination of factors, butpilot holes and wood density play significant roles in connection success. Thekey is preboring pilot holes whose size mesh the threaded portion of the screwand the wood density. Pilot holes for the threaded portion should be about 70%of root diameter (threaded portion) for low-density woods and about 90%for high-density woods (Hoadley, 1992). As with nails, the holding powerincreases linearly with diameter and length, but exponentially with density.Additional problems are encountered when joints are designed so that screwsmust be driven into end grain. In this situation, screws have only about 75%as much holding power as those driven into side grain and the holding poweris less predictable.

There is an endless array of factors that influence the performance of a boltedwood connection. A partial list includes species, moisture content, number andthickness of the members, type of side member material, direction of load, boltdiameter, bolt yield stress, and spacing, end and edge distance of bolts. Inaddition, multiple-bolted connection behavior depends on load distributionamong the bolts. Prior to 1983, design for strength of a single-bolted woodconnection was based mainly on the research conducted by Trayer in 1932.

Recommendations made by Trayer (1932) were based on an empirical fitof experimental data. Trayer’s work only considered a few base cases of


joints, and these data were extrapolated over the years to apply to a widerange of geometries and species. In the late 1970s and early 1980s, Europeanresearchers developed a new analytical joint strength model now known asthe European Yield Theory. The yield theory ‘represents a change to an engin-eering mechanics approach to the design of wood connections. It replaces theold empirical method of predicting the capacity of many common woodfasteners’ (Breyer, 1993). The yield theory uses connection geometry andmaterial properties to evaluate the strength of two- or three-member dowel-typeconnections. Equations are provided to predict the yield load that correspondsto various failure modes.

Practical considerations must not be overlooked when evaluating possibleproblems with bolted connections. For example, washers of adequate size mustbe used. Washer size needs to be sufficient so that bearing stress in the wooddoes not exceed allowable compressive strength perpendicular to the grain.Something as simple as using a washer that is too small in a connection loadedin tension, tightening the nut too tight, or wood shrinkage away from theconnector can cause localized catastrophic failures.

Appropriately sized predrilled holes are important to bolted connectionintegrity. Holes that are too small will require excessive driving of the bolt andsplitting of the wood may occur. Splits in the wood can greatly reduce shearcapacity of a bolted connection, plus they have a tendency to enlarge oncecreated. Holes that are too large will allow non-uniform bearing stresses thatresult in stress concentrations. Even the manner in which the hole is drilled hasa role. A smooth surface develops a higher load capacity and less deformationthan a rough surface. Dull drill bits, improper drill speeds, and too rapid a feedrate will create rough surfaces and potential connection problems. Fabricationinaccuracies such as improper hole sizes or hole quality have been shown toaccount for many joint inadequacies.

Something as obscure as microscopic anatomy of wood used in a connectioncan have an impact on failure modes (Zink-Sharp et al., 1999). This researchfound that uniform anatomy features tended to produce a uniform and predictablefailure mode but variable anatomy such as that found in ring porous hardwoods,would produce non-uniform failure modes and bolt displacement.

Decreasing the number or size of fasteners below what is recommended byeither a product manufacturer or the engineering design seriously reduces theconnection strength. Disregarding spacing recommendations and insufficientpenetration of fasteners into the connecting member almost always results ininadequate joints. Defects such as knots, sloping grain, splits, and low-densitywood in the regions to be joined will also contribute to in-service problems.

Assuming base conditions of normal load duration (10 years), dry wood,normal temperature range, and adequate spacing provisions, equations for designof timber structures and connections can be found in various codes and manuals,for example, National Design Specifications for Wood Construction (AF &


PA, 1995) and EUROCODE 5 (CEN, 1995). However, it is crucial to rememberthat the equations, tables, and formulas all refer to clear wood specimens (noor minimal defects) that will not undergo extreme moisture conditions or cycles.Once a design value for strength parameters is established, this nominalstrength is subject to other modification factors that take into account numberof fasteners, moisture content, extended duration of load, and other factorswhere appropriate.

Punched metal plate fasteners (commonly called truss plates) and steelplates are common methods for joining prefabricated rafter and joist trusssystems. These connectors transmit loads through the metal teeth, plugs, ornails. A plate is placed on the front and the back of the lumber in each connectionto ensure a balanced joint. The plate acts as an array of short nails attached to acommon head and transfers forces between members in the plane of the plate.Neither the plate nor the teeth are intended to be subjected to withdrawalforces. Numerous designs for truss and steel plates are available. Connectorplates are usually made from galvanized sheet steel or, in special situations,from stainless steel. Installation of the plates requires a hydraulic press or otherheavy equipment. Lateral resistance of plate-type connections is a function ofplate and teeth characteristics and wood factors such as density, moisturecontent, and duration of load. However, it has been found that relative density(specific gravity) has a very significant influence on lateral resistance andembedment of these types of fasteners (Via etal., 1999; among numerous others).Eccentric loads, misalignment of truss plates, shallow embedments, improperstorage and installation on site can produce situations that result in partial toothwithdrawal and connection failures.

It might seem at first that the list of problems that could be encounteredwith mechanical connections made with wood is rather long, and the designfairly complicated. Fortunately, the practice has been accomplished successfullyfor centuries, and modern engineering analyses, composite materials, and acomplete understanding of material properties can only lead to improvements.

References

American Forest and Paper Association (AF & PA) (1995) National Design Specifications for WoodConstruction. Washington, DC.

American Society for Testing and Materials (ASTM) (1987) Standard methods for establishingstructural grades and related allowable properties for visually graded lumber. ASTM D245,Philadelphia, Pennsylvania.

Bodig, J. & Jayne, B.A. (1982) Mechanics of Wood and Wood Composites. Van Nostrand ReinholdCompany, New York.

Breyer, D.E. (1993) Design of Wood Structures. 3rd edn, McGraw-Hill Inc., New York. British Standards Institute (BSI) (2001) BS EN 384:1995 Structural Timber – Determination of Charac-

teristic Values of Mechanical Properties and Density, BS EN 518:1995 Structural Timber –


Grading – Requirements for Visual Strength Grading Standards, BS EN 519:1995 StructuralTimber – Timber – Requirements for Machine Strength Grading – Timber and Grading Machines,BS EN 1912:1998 Structural timber – Determination of Characteristic Values of MechanicalProperties and Density, Berkshire, UK.

Buchanan, A. (1991) Building materials and the greenhouse effect. New Zealand Journal of TimberConstruction, 7, 6–10.

European Committee for Standardization (CEN) (1995) Eurocode 5 – Design of timber structures –Part 1-1: General rules and rules for buildings. Brussels, Germany.

Forest Products Society (FPS) (1999) Wood Handbook: Wood as an Engineering Material. Madison,Wisconsin.

Gnanaharan, R. & Haygreen, J.G. (1979) Comparison of creep behavior of waferboard and that of solidwood. Wood and Fiber, 11, 155–170.

Haygreen, J.G. & Bowyer, J.L. (1996) Forest Products and Wood Products, 3rd edn, Iowa State UniversityPress, Ames, Iowa.

Hoadley, R.B. (1992) Understanding Wood. Taunton Press, Newtown, Connecticut. Kellogg, R.M. & Wangaard, F.F. (1969) Variation in the cell wall density of wood. Wood and Fiber,

1, 180–204. Loferski, J. (1997) Long term performance and durability of engineered wood products, in

Engineered Wood Products (ed. S. Smulski), PFS Research Foundation, Madison, Wisconsin,pp. 193–222.

Marcea, R.L. & Lau, K.K. (1992) Carbon dioxide implications of building materials. Journal of ForestEngineering, 3, 37–43.

Meil, J.K. (1993) Environmental measures as substitution criteria for wood and nonwood buildingproducts, in The Globalization of Wood: Supply, Processes, Products, and Markets, Forest ProductsSociety Proceedings 7319, 50–53.

National Research Council (1976) Renewable Resources for Industrial Materials. National Academyof Sciences, Washington, DC.

Panshin, A.J. & de Zeeuw, C. (1980) Textbook of Wood Technology. 4th edn, McGraw-Hill Inc.,New York.

Schniewind, A.P. (1982) Mechanical behavior and properties of wood. In Wood as a Structural Material(eds A. Dietz, E. Schaffer & D. Gromala), The Pennsylvania State University, University Park,Pennsylvania, pp. 55–94.

Sedjo, R.A. & Lyon, K.S. (1990) The long term adequacy of world timber supply, in Resources for theFuture, Washington, DC.

Trayer, G.W. (1932) The Bearing Strength of Wood Under Bolts. Technical Bulletin No. 332, USDepartment of Agriculture, Washington, DC.

Via, B.V., Zink-Sharp, A., Woeste, F.E. & Dolan, J.D. (1999) Relationship between tooth withdrawalstrength and specific gravity. Forest Products Journal, 49, 56–63.

Zink-Sharp, A., Stelmokas, J.W. & Gu, H.-M. (1999) Effects of wood anatomy on the mechanicalbehavior of single bolted connections. Wood and Fiber Science, 31, 249–263.

Index

Note: Page numbers in italics refer to figures and tables

Abies 30, 103 Abies balsamea 2, 124 Abies concolor 120 abietic acid 73

absorbance 164 Acer macrophyllum 77 Acer negundo 77 Acer saccharum 15, 16 O-acetyl groups 64 O-acetyl-4-O-methylglucuronoxylans 63 O-acetylgalactoglucomannans 63 acetylglucuronoxylan 77 acoustic properties 38 acoustic technology 43 acoustical insulation 190 adhesives 82, 190 Agathis australis 47 agricultural residues 158, 159 agro-industry 160 air-dry density 90 air emissions 190 Alectryon excelsus 33 alkaloids 76, 82 allergenic action 82 Alnus incana 88 alpha-cellulose 179 American chestnut 9 amino acids 76 1-aminocylcopropane-1-carboxylic acid (ACC)

19, 128 β-amyrin 73, 74 anatomy 53, 111, 173, 208 Angiospermae 30 angiosperms 118, 119, 121, 125, 130 anisotropy 12, 144, 191, 192, 194 annual plants 158, 159

fibre costs 159 annual ring 3, 7, 15, 96, 108, 120, 133, 161, 179 anomalous wall bars 20 anti-microbial activity 81 anticlinal pseudotransverse divisions 125 ants 12

ant cavity 13 apical meristems 7 apoptosis 6

Arabidopsis thaliana 26 arabinofuranose 61 L-arabinofuranosyl 64 α-L-arabinofuranosyl units 63 arabinogalactan 60, 64, 77, 91 arabino-4-O-methylglucuronoxylans 63 arabinoglucuronoxylan 76 arabinopyranose 61 3-O-β-L-arabinopyranosyl-L-arabinofuranosyl 64 arabinose 64 Aralidium pinnatifidium 45 Araucaria sp. 93 arches 187, 190 aromatic and aliphatic hydroxyls 67 artificial woods 18 ash 77, 158, 159, 167 aspen 158, 167, 174 aspirated pit 47 atomic force microscope 35 aurones 74 auxin 3, 6, 16, 18, 19, 122, 123, 125 average wood density 107, 111 axial 137, 196

compression 140 expansion 127 maturation shrinkage 138 parenchyma 6, 31, 33, 34, 35, 49 parenchyma cells 30 resin canals 30 shrinkage 133 stiffness 37, 38 tension 140, 195 tracheids 49

bacteria 191 bacterial attack 170 balsa 14, 193, 195, 196 balsam 158 bamboo 158

fibres 157 bark 137, 138, 140, 147, 170, 172 basic density 87, 88, 90, 91, 93, 96, 98–109,

113, 183 basipetal transport 6 bast fibre 159

212 INDEX

beam 22, 188, 189, 196, 198, 199 box 187 stiffness 195

bearing stress 208 beating 164 beech 68, 144, 158, 167, 168, 174, 177, 179

pulp yield 169 roundwood 168

beetles 12 Beilschmiedia tawa 44 bending 118, 121, 124, 133, 140, 188, 194–6, 202, 204

MOE 196 moment 137, 140 strength 81, 87, 198, 199 strength MOR 197 stress 118, 124, 125, 198, 204

benzyl alcohol 67 benzyl ester 68 berberine 76 Betula alleghaniensis 3 Betula papyrifera 78 Betula pendula 88, 103, 129 Betula pubescens 88, 103 Betula spp. 103 biflavonoids 76 bi-metallic strip 127 bio fuel 170 biochemistry 1, 3, 5, 25 biodiversity 20 biological attack 191 biological decay 165, 166, 167 biological degradation 191 biological degraders 203 biological organisms 190, 191 biomass 8, 11 biomechanical function 131 biomechanics 142 biophysics 1, 25 biosphere 23 birch 93, 158, 167, 168, 174 biseriate rays 30 black cherry 195 black spruce 94, 96, 106, 107 bleachability 170 bleaching 18, 23, 82, 163, 175 blunt point 206 board 22

deformation 150 distortion 150

bole 7, 9, 13, 21 bolt 205, 207, 208

connection 208 yield stress 207

bordered pit 2–4, 6, 79 membrane 47 number 8

boreal coniferous zone 166

bow 149 bowed 14 bracket fungi 167 branch 6, 7, 118, 119, 121, 123–5, 128–30, 161

angle 11, 122, 134 architecture 119 breaking length 164, 168 diameter 11 form 11 gymnosperms 118 number 11 self-prunability 11

breeding 23, 113, 184 brick 187, 189 bridges 187 brightness 164, 170

reversion 82 brittle 18, 133

elastic 147 brooming 44 brown rot 167 buckle 203 building construction 190 buildings 187 burst 168, 183

strength 181 tensile strength 164

Buxus 19

caffeic O-methyltransferase 128 caffeoyl CoA-O-methyltransferase 128 calcium sulphide 163 Callitris glauca 36 Callitroid thickenings 36 cambial zone 1, 2, 124 cambium 1, 2, 6, 7, 9, 14, 18, 19, 26, 76, 79, 97, 123,

125, 129 cambial activity 95, 109, 137 cambial age 7 cambial cell division 32, 93, 109 cambial cells 7, 19, 137 cambial derivatives 5, 18, 19 cambial growth 1, 6–9, 11, 25, 123 cambial meristematic zone 109 cambial region 7, 9, 20

canopy structure 134 cants 12, 14 caoutchouc 74 carbohydrate 72, 76, 79 carbon content 11 carbon dioxide 1 carcinogenic activity 82 3-carene 78 Carya ovata 15 Castanea 142 Castanea dentata 10 catastrophic failures 208

INDEX 213

catechin 75, 76 cavitation 46 Cedrus libani 33 cell 53, 121, 124–7, 129, 160, 161, 164, 167, 180,

199–201 autolysis 47 biology 1, 3, 9, 25 corners 4 death 6 development 137 diameter 95, 97 differentiation 32, 125, 137 division 3, 18, 125 expansion 37 length 97 lumen 12 maceration 35 organisation 162 stiffness 38 structure 53, 199

cell-free biological system 5 cell wall 34, 37, 53, 54, 60, 64, 72, 79, 81, 82,

87, 132, 137, 170, 180, 189, 192, 193, 198

area 92 checking 38 density 193 deposition 46 loosening 3 organization 71 softening 14 structural components 55 substance 192, 193 thickness 8, 37, 38, 92, 97, 109, 164

cellobiose 55, 88 cellulose 5, 8, 11, 19, 54–60, 64, 68–71,

77, 80–3, 87–9, 125, 167, 168, 180, 197

angle 127 content 150 crystal lattice 88 crystalline 57, 161 crystalline lattice 59 crystallite orientation 70 density 88 fibres 60 flexibility 70 microfibril 5, 35–7, 47, 58, 59, 129 synthase 5

α-cellulose 77, 87 cellulose:lignin ratio 11 cellulosic compounds 168 cellulosic fibre 23 cereal straws 158 chalcones 74 Chamaecyparis 30 checking 13, 14

checks 12, 127, 191, 199, 203–5 chemical 165, 174, 197

composition 53–5, 83, 158, 160, 173, 179 deterioration 197 engineering 24 organisation 160 process 163 properties of lignin 68 pulp 164, 165, 180, 183 pulping 167, 169, 170, 173 recovery 159 substances 6 sulphite-based 174 variability 55

chemistry 12, 160 chestnut 144, 148, 150, 153, 158, 168,

177, 179 chipper 169, 172 chipping 169 chips 22, 159, 162, 169, 170–2, 174, 180, 181

size 170, 172 steaming 170

circumnutational growth 121 clay content of the soil 106 clear or knot-free wood 9 clear wood specimens 209 climate 23, 25, 153

change 1, 18, 22, 113, 166 conditions 109, 160

clones 21, 167, 184 cloning 22 coated printing grade paper and SC 164 collapse 164 coloration 172 colour 80, 82, 153, 173 colour reaction 153 columns 189, 194, 196, 197, 203 combustibility 192 communication 4 communities 23 competition 118 compliance tensor 138 composites 191, 198, 209 compound light microscope 8 compound middle lamella 2, 4, 36 compression 18, 121, 127, 133, 140, 161, 189, 195–7,

202, 203 efficiency 189 loading 203 strength 194, 196, 199

compression wood 3, 16, 18, 40–2, 80, 83, 98, 100, 118–28, 132, 133, 145, 148, 150, 153, 161, 165, 180

cells 127 fibres 180 hemicelluloses 80 tracheids 36, 123–5, 127, 129, 132

214 INDEX

compressional load 34 compressive 121, 140, 144, 145

force 118 growth stress 144 reaction stresses 150 strength 81, 133, 208 stress 12, 145, 196 support stress 140 tangential stress 138

concrete 187, 190 block 189 construction 190

conduction 31 confocal laser microscope 35 confocal microscopy 38 conformability 157 conformable 164 conifer 26

cell walls 87 needles 19 stems 18 tracheids 2

Coniferales 118 Coniferophyta 24 conifers 1, 6, 7, 16, 24, 77, 92, 120, 121,

125, 132 coniferyl alcohol 65–7 coniine 82 connection 196, 204, 206–9

failures 209 strength 208

connectors 190, 209 conservation 166 constant loading 198 constitutive equations 140 construction 87, 187, 190 consumption 170, 189 contact points 164 content 193, 209 control 172 controlled pollination 25 controlled-pollination crosses 22 conversion 38, 187, 189 conversion operations 150 cooking efficiency 170 cooking liquor 169, 174 corewood 38, 40–2, 97, 98,

100 corewood density 38 corrugated papers 163 corrugating grades 169 cortical microtubule 5, 125 cost 159 cotton 157, 159 cotton linters 158 p-coumaryl alcohol 64, 65 covalent bonding 7

crack 141, 147–9, 153, 154 extension 150 length 148, 153 propagation 148, 154 surface 148

cracked ends 154 cracking 133, 148, 154 creep 198

rupture fracture 198, 189 creeping 145 crook 149 crosscutting 142, 147, 153 crowding 16 crown 9, 25, 34, 152

size 16 wood 97, 98

crushing strength 197 cryofixation 1 Cryptomeria 142 crystalline lattice 60 crystallinity 58, 60, 83 crystallites 60 cubical fungus 168 cubical rots 167 Cupressus 30 curvature 118, 127, 130, 131, 151, 152 cut 153 cutting 141, 142, 147, 149, 150, 153 Cycadales 118 cytokinin 18

Dacrycarpus dacrydioides 31 damage 167 de-oxy-hexoses 61 debarkability 170 debarking 141, 169 decay 6, 9, 12, 167, 168, 197

fungi 204 microbes 12

decayed beech 168 decayed wood 167, 169 decking 188 defects 12, 191, 199, 204, 208, 209 defence 24, 72 defibering 157, 163, 170 deformation 149, 197, 198 degradation 167, 168 degree of polymerization 55, 60, 63, 64 dehydrogenation 66 delamination 148 delignification 83, 163, 164, 170,

175, 197delignification rates 83 dendrochronology 7 dendroclimatology 25 dense woods 144 densification 196

INDEX 215

density 14, 16, 40, 81, 100, 101, 103, 107, 111, 144, 159, 168, 173, 179, 180, 192, 193, 195–7, 207, 209

density 191, 193, 209 determination 90 dry 19 profile 92 variation 91–3, 94

design loads 198 developing countries 166 developing leaf 19, 21 developing xylem 19 developmental biology 25 developmental zones 1 diameter 9, 173, 207

growth 15, 16 diamond point nail 206 differentiating xylem cells 46 differentiation 5, 18, 145 diffraction pattern 40 diffuse porous 34 diffuse-porous hardwoods 113 digallic acid 76 digital picture 151 β-1 dilignol 65 dilignol radical 67 dilignols 65, 67 dimensional changes 124, 180, 191, 207 dimensional instability 12, 16, 191 dimensional stability 83, 100 disaccharides 76 discoloration 82 disks 142 disoriented cells 6 dissolving pulp 174 dissymmetry 144, 152 distorted grain 81 distortion of boards 150 disymmetrical stress 145 diterpenic acids 72, 81 diterpenoids 72–4 diurnal and phenological changes 6 dominant trees 113 dormancy 14 Douglas fir (Pseudotsuga menziesii) 93, 94, 96, 123,

125, 158, 179, 189 dowels 197, 205, 208 downy birch 103 drilling 141 dry weight 181 drying 14, 82, 133, 149, 150, 188, 196, 199, 203drying checks 203 durability 18, 168, 187, 190 duration of load 209 duration of stress 198

earlywood 2, 3, 7, 32–4, 93–6, 100, 103, 107, 109–11, 113, 161, 179, 180, 183, 192

cells 38, 161 density 94, 107, 110 lumen diameter 97 vessels 14

earlywood/latewood 183 eastern white pine 9 eccentric loads 209 eccentricity 119, 150 ecophysiological fitness 22 ellagic acid 75, 76 elastic 194, 196

energy 147, 149 energy release rate 147, 148 limit 196 locked-in strains 149 moduli 141 properties 194, 195, 197 stability 189

elasticity 192 electrical resistance 141 electron microscopy 35 elementary fibrils 58 elongation 137, 195

growth 118 emissions 166 end checks 203, 204 end cracks 153 end grain 206, 207 end-product quality 170 end-use properties 53 energy 8, 23, 147, 160, 165, 173, 187, 189, 190

consumption 183 Engelmann 158 engineering 21

material 6 properties 22

Entelea arborescens 34, 35 environment 11, 53, 106 environmental 9, 166, 190, 198

conditions 78, 191, 199, 206, 207 factors 55, 101 pollution 23

Eperua 144 ester bonds 64 esters 76 ether

groups 67 linkages 68 solubles 77

ethylene 19, 124, 128EtOH-benzene solubles 77 Eucalypts 78, 144, 151, 173 Eucalyptus 77, 79, 158

botryoides 78 camaldulensis 78, 83 deglupta 83 delagatensis 32

216 INDEX

Eucalyptus (Continued)diversicolor 77 globulus 78, 83, 88 maculata 77 nitens 83 regnans 78 tereticornis 77, 78, 83

extension growth 8 extractive components 71 extractive compounds 170 extractive content 8, 91, 97 extractives 11, 38, 53–5, 71, 72, 74, 76, 77–82, 91, 98, 170 extreme events 18 extreme moisture 197 extrinsic environment 7

Fagus 142 failure 189, 200, 203, 208 false rings 33, 93 farnesene 73 fast-grown trees 9, 25 fast growth 121, 134, 144, 183 fast-rotation plantation 41 fasteners 154, 204–9

metal plate 209 fats 76 fatty acids 76 fatty reserve molecules 72 feed rate 208 feedstock 176 felled stems 151 felling 142, 147, 167, 170 fencing 24 fertilisation 109–11, 113 fibre 1, 2, 6, 8, 12, 18–21, 31–6, 43, 46, 78–81, 98, 119,

129, 137, 141, 157–9, 160–6, 173, 174, 179, 180 collapse 83 costs 159 elongation 6 length 8, 164, 177, 179 length/width/cell-wall thickness 11 morphology 175, 183 properties 87, 97 saturation point 89, 90, 193, 195, 197 strength 83 wall 129

fibre:vessel ratios 14 fibrillation 18, 83 fibrous rot 167 filler 165 fine papers 18, 174 fines 6, 164, 169 finishing 80, 81 finite-element methods 148 firewood 24 firs 158, 173, 174 flavan-3,4-diol 76 flavan-3-ol 76

flavane 76 flavanes 74 flavanones 74 flavones 74 flavonoids 74–6, 81 flax 158, 159 flexibility 140, 163 flexural rigidity 189 fluorescence 129 fold endurance 181 Fomes 168 forest 24

decline 25 ecosystem 23, 24 habitat 24 management 9, 22, 24, 102, 189 manager 23 regeneration 189

forestry 1 research 26

forks 145 foxtailed pine 18 fracture 70, 198

morphology 70 toughness 133

frame 192 Fraxinus americana 15 free surfaces 147, 149 frost cracking 12, 13 fructose 76 fucose 61 fungal 24, 167, 169 fungi 74, 167, 191 furnish cost 165 fusiform cambial cells 1, 6, 30 fusiform cells 2

G or gelatinous wall layer 43, 80, 83, 129, 133, 161 G type lignin 65, 77, 78, 83 GA3 124 galactan 80 galactoglucomannan 64, 76 α-D-galactopyranosyl 63, 64, 80 galactose 61, 63, 64 D-galactosyl units 80 galacturonic acid 61 β-D-galacturonic acid 80 D-galacturonic acid residues 80 gallic acid 75, 76 galvanized sheet steel 209 gas chromatography–mass spectroscopy 123 gelatinous fibres 44 gene expression 7, 22 gene manipulation 22 gene pools 22 genetic 9

component 11 control 97, 107, 113, 134, 154

INDEX 217

diversity 23 engineering 12, 18 factors 55 modification 22, 23 theory 22 variation 78, 101, 113

genetics 25, 53 genome 11 genotype 7, 11, 106, 144, 153 geographic location 101 germplasm 25 gibberellin 6, 18, 19, 124 Ginkgoales 118 girdling 154 girth 9 glass transition 149

temperature 180 GS-lignin 65 glucan 179 glucomannan 5, 60, 63, 77 β-D-glucopyranose 55 β-D-glucopyranosyl units 63 glucose 55, 59, 61, 76 D-glucuronic 80 Glucuronic acid 61 D-glucuronic acid 64 glue-laminated (glulam) beams 187, 188 glue-laminated timber 198 gluing 80–2, 188 glycerol 76 glycosides 74, 76 glycosidic bond 55, 60 α-(1-2) glycosidic bonds 63 β-(1-3) glycosidic bonds 64 β-(1-4) glycosidic bonds 55, 63 Gnetales 118 Gnetum gnomen 119 grading 197

rules 200 grain 13, 191, 192, 194–201, 203, 208

angle 11 deviation 199, 200 slope 199, 201 strength 202

gravitational bending 127 field 140 stress 121

gravity 8, 80, 118, 121, 124, 125, 137, 140, 193

green vein 151 volume 90, 91 weight 181 wood 141, 197, 198

greenhouse gases 23 Griffith theory 147 groove sawing 141 groundstone 174 groundwood 174

growing conditions 160 space 15, 16

growth 11, 118, 125, 129, 137, 142, 153, 161, 183, 190, 192

eccentricity 145 episodes 9 history 137 patterns 153 rate 93, 101–3, 107, 109, 113, 129, 144, 183 regulators 124 ring width 101 strain 144, 148, 149, 153 strain/stress 145

growth ring 31, 32, 34, 53, 93, 95, 106, 119, 120, 125, 129, 140, 150, 201, 203

width 96, 102–5, 113 growth stress 124, 125, 137, 139, 140–2, 144, 145, 147,

148, 150, 151, 153, 154 distribution 149 indicator 141 level 150 release 149

guaiacyl 65, 77 gums 79 guttapercha 74 G×E interactions 11 gymnosperm 118–21, 121, 129, 130

stems 118 Gymnospermae 30

hard pines 103 hardness 161 hardwood 1, 6, 8, 9, 11, 14, 30–2, 46, 53, 61,

63, 65, 72, 74, 76–80, 83, 87–9, 107, 113, 142, 145, 148, 151, 153, 158, 160, 164, 165, 166, 173–5, 177, 179, 180

bisulphite pulps 174 fibres 71 hemicelluloses 63 lignin 68, 70 pulp 158, 165 species 167, 175 vessels 201 xylans 63

harvesting 24, 25, 150 H-bonds 56, 60, 64 health of forest ecosystems 18 heart 147 heartrot 13 heartwood 11, 24, 25, 38, 53, 71, 72, 74, 78–81, 91, 97,

98, 160, 174, 179 formation 5, 91, 98

heat 164 heating 147 Hedycarya arborea 45 height growth 11, 16, 111 Heimerliodendron brunonianum 33

218 INDEX

helical checks 127 helical thickening 36, 37 hemicellulose 19, 37, 54, 55, 58–61, 63, 64, 68, 69, 71,

76, 77, 79, 80, 87, 88, 137, 168, 174 density 89

hemiterpenoids 72 hemlock 158, 173 hemp 158 heredity 93 hexoses 61 H-factor 175 high compression 147 high density 161 high value woods 25 holding capacity 206, 207 holding power 206, 207 hole drilling 141, 144, 153 holocellulose 77, 179 holocoenotic linkages 24 holocoenotic principle 23 hornbeam 168 hot water solubles 77 human impact 101 humidity 198, 199, 206 hybrid aspen 123, 130 hybrids 144 hydrated cellulose 129 hydraulic press 209 hydrolysis 197 p-hydroxy-cinnamyl alcohols 64 hydroxyl groups 60, 64 p-hydroxyphenyl 18, 65 hygrothermal recovery 149 hygrothermal stresses 196

image processing techniques 35 impregnation 170 inclined stems 145 increment cores 153 3-indoleacetic acid (IAA) 122 industrial chemicals 23 industrial processing 35 industrial revolution 24 inorganic components 54 insects 11, 74, 190, 191, 204

attack 167 inter-tracheary pit membranes 48 inter-vessel pit membranes 47 inter-vessel pit-pairs 46, 48 intercellular spaces 126 interfibre bonding 164 interlocked grain 200 intermolecular H-bonds 57 internal failure 14 internal stress 139 inter-tracheary pits 46, 49 intramolecular H-bonds 57

intrinsic environment 7 intrusive growth 43 iodine crystals 38 iodine staining 39 iroko 78 iron 197 irrigation 14, 110 isocyanate glue 141 isoflavones 74 isoprene 72 isoprenoid polymers 74 Itrax 92

joint strength 208 joints 197, 204, 206–8, 209 joist truss 209 joists 187, 188 jute 158 juvenile wood 53, 79, 94, 96–8, 104–7, 113, 121, 144,

164, 179–81, 183 and mature wood 79 characteristics 41 core 2, 16, 25 sapwood 25 stock 1

kappa number 175, 177 kauri (Agathis) 34 kenaf 158 kiln drying 82 kino 79 Klason lignin 77 Knightia excelsa 45, 47 knots 11, 53, 82, 191, 199–201, 208 knotwood 79, 81, 83 Kraft liner 169 Kraft pulp yield 171 Kraft pulping 83, 159, 163, 165, 174–6, 181

laccase 128 ladder rails 203 lag screws 205 laminated beams 190 lanolin 123, 124 larch 64, 82, 96 laricinan 128 Larix 30, 31, 96, 103 Larix sibirica 88, 91, 96 laser beam 151 lateral roots 7 lateral wood 119, 132 latewood 2, 3, 20, 21, 32–4, 46, 94–8, 100, 103, 106,

107, 109, 111, 113, 161, 179, 180, 183, 192 cells 38 density 94, 107, 110 fibres 180 formation 93, 95, 111

INDEX 219

percentage 92, 93, 108 proportion 97, 112 to earlywood ratio 14, 165 tracheids 39, 93, 97

leaning stem 119, 122, 123, 129, 132, 145 leaves 9, 19 legislation 166 leucocyanidin 76 libriform fibres 34, 44 life cycle 23 light absorption 83 light scattering 174 lignan 74, 75, 81, 82 lignification 2, 4, 5, 24, 128, 129, 132 lignin 3, 5, 7, 8, 11, 12, 19, 23, 24, 35, 54, 55, 64,

65, 68, 69, 71, 76, 77, 78, 80–3, 87–9, 127, 129, 137, 149, 150, 163, 164, 167, 168, 174, 175, 180

lignin-free fibres 163 lignivorous fungi 167 limonene 72, 73, 78 linen rags 157 β-O-4 linked dilignols 65 Liriodendron tulipifera 15 load 191

capacity 208 duration 198

loading 18 loblolly pine 91, 103, 158, 169, 173, 184 local strain 141 locked-in strains 141, 142, 148, 149 log 8, 12, 14, 22, 133, 142, 147–51, 153, 154, 162, 167,

168, 172, 187 diameter 148 end cracks 147, 150, 154 end splitting 150 heating 149, 154 length and diameter 9 steaming 147

long-chain alkanols 76 long-chain fatty acids 76 long-chain molecules 198 long-rotation forestry 113 longitudinal checks 203 longitudinal compression 69

maturation strain 141 parenchyma 160 residual strains 149 shrinkage 37, 41, 79, 80, 98, 150, 151 strength 70 tensile strength 79 tensile stress 138, 144 tension 69, 147 Young’s modulus 144

low-density wood 14, 87, 144, 207, 208 low moisture-holding capacity 161 lower cost fibre 165

lower pulp yield 168 lower rigidity 145 lumber 22, 187, 201, 209

dimensions 9 distortion 149 grading 203, 204

lytic enzymes 4

machinability 38 machining 188 macromolecules 54, 158 Magnoliophyta 24, 26 maize coleoptiles 125 management 189 manilla 158 β-D-mannopyranosyl 63 mannose 61, 63 maple 174 margo 4, 46–8 maritime pine 165, 183, 184 mass density 87 mass spectrometry 123 maturation 127, 130, 138

stage 137 strain 138–40, 142, 144, 147, 152

maturation stress 138–41, 144, 147 mature wood 25, 79, 94, 96–8, 100, 101, 104, 105, 107,

113, 180, 181 maximum height 9 maximum stress 196 mechanical failure 14 mechanical properties 37, 103, 187 mechanical pulp 162–4, 174, 180, 183 mechanical pulping 169, 170, 173 mechanical stress 78

perception 125 mechanical tests 90 merchantable diameter 12 metabolism 6

in the cambium 9 metal teeth 209 meteorological conditions 167 methoxyl groups 67 methoxyl substitution 65 2-methyl-1,3-butadiene 72 4-O-methyl-α-D-glucopyranosyluronic acid 63 methylglucuronic acid 61 Metrosideros umbellata 35 microcavities in the cell wall 89 microdomains 6, 22 microfibril 4, 37, 46, 57, 58, 68, 125, 129, 137

angle 8, 37–42, 44, 45, 70, 80, 83, 98, 121, 127, 132, 180

variation 40 orientation 5 polymerization 5 structure 70

220 INDEX

microfibrillar strands 4 microgravity 125 microorganisms 11 microtubules 125 middle lamella 35, 36, 68, 78, 163 mill 159, 164, 167, 169–73, 176, 179 milling 24 mineral stains 201 minimum wood density 111 model species 26 modelling 12 modulus of elasticity (MOE) 90, 150, 194–6 modulus of rigidity 194, 196 modulus of rupture MOR 196, 197 moisture 15, 191, 193, 197, 203, 207, 209

conditions 209 content 89, 90, 167, 169, 170, 172, 179, 191–3,

195–7, 203, 207, 209 strength 197 stress 93 variations 207

molecular markers 134 moment of inertia 137 monosaccharides 61, 76 monoterpenes 72, 78, 81 monoterpenoids 72–4 MOR 196 morphogenesis 7 mother cells 125 multiple-bolted connection 207 multiple perforation plates 48 multiseriate rays 31, 48, 49 Myrtaceae 34

nail 205, 206, 207, 209 nail-holding capacity 207 natural durability 71, 81 natural-growth 192 natural selection 22 needle biomass 111 negative-gravitropism 145 newspapers 164 night frosts 93 nitrogen fertilisation 109 non-autolyzed fibres 6 non-recoverable deformations 196 non-weeping branches 140 non-wood fibre 158–60, 165 normal-strain 144 normal wood 119, 126, 127, 129, 131–3, 145,

148, 150, 154 cells 127 fibres 129, 130 tracheids 127

northern white cedar 196 Norway spruce 87, 89, 92–101, 103–10, 113, 168,

173, 184

nutrient 9 availability 106 optimisation 109, 111 status 111

oak 103, 158, 168, 177, 179, 193 obeche 78 Ochroma lagopus 14 Ochroma spp. 195, 196 oil seed rape 158 old growth 15, 24 oleananes 74 oligolignols 67 opacity 174, 181 open-grown trees 106 operating costs 170 opposite wood 119, 125, 129, 132 optical properties 174 organosolv pulping 164 oriented strandboard 187 outerwood 38, 40

density 38 oven-dried wood 89, 169 oven-dry weight 90, 193 oxidation 169, 197 oxidizing agents 197

paint and varnishes 82 paleobotanical record 22 panels 187 paper 157–60, 164–6, 170, 175, 179, 181

furnish 165 machine speed 166 physical properties 175 quality 166, 173, 183 raw material supply 160 strength 175, 179, 183 yield 38

papermaking 18, 23, 24, 157, 163–5 papyrus 157 parchment 157 parenchyma cells 34, 35, 71, 79, 164 partially decayed logs 169 particleboard flooring 188 Paulownia spp. 14 pectic compounds 35 pectin 77 pentosans 77 pentoses 61 perforation plates 6, 48 performance in use 87 periclinal division 1, 125 permanent deformations 196 permanent loads 198 permeability 76, 82 petiolar traces 20 pH 163

INDEX 221

pharmaceuticals 82 phenol 74 phenolic 82, 174

extractives 74, 75, 81 phenotype 7, 10, 22 phenotypic plasticity 11, 14 phenoxy radicals 65 phenyl glycosidic linkages 68 phenylpropane 64 phenylpropanoid units 67–70, 74 phloem 1, 26 photoperiod 6 photosynthate 9 phototropism 145, 152 physiological age 40 phytohormones 5, 6, 18 Picea 31, 103

abies 88, 92–4, 100, 173 glauca 2, 10, 100 mariana 94, 96, 100, 106 rubens 13 sitchensis 89, 100, 107, 128

pilot holes 207 pin knots 199 Pinaceae 4, 74 pine 40, 82, 144, 145, 158, 173, 174 β-pinene 72 α-pinene 72, 78 pinoresinol 74 Pinosylvin 75 Pinus 26, 31, 142

caribea 10 contorta 2, 13, 102 elliottii 78, 102 nigra 45 pinaster 128, 161 radiata 16, 36, 47, 102 resinosa 2 strobus 10 sylvestris 78, 88, 94, 97, 102, 109 taeda 91, 102, 103, 120, 179

pit 45, 46 apertures 36, 38, 39 aspiration 46 border 4, 47 membrane 46–9, 79, 82 pair 45, 46

pith 12, 15, 119, 137, 140, 147, 150, 159 plagiotropism 145 planks 12, 14, 133 plant evolution 24 plant growth 122 plantation 23, 166

stock 11 plasma membrane 4, 125 plastic-waste wood structural members 18 plastics 189

plate connectors 205 plate-type connections 209 plugs 209 plywood 187, 188, 198 pockets 191 poison hemlock 82 Poisson’s ratios 138 polarized light microscopy 35 pole building 187 polyflavonoids 76 polymer/wood composites 80 polyoses 76 polyphenolic 79 polysaccharide 4, 7, 60, 163, 164, 167, 174, 175 polyterpenoids 72, 74 ponderosa pine 158 poor sites 109 poplar 78, 93, 144, 151, 153, 167, 168, 174,

177, 179 hybrids 144

Populus 26, 44, 47, 103, 144 balsamea 19 balsamifera 20 tremula × tremuloides 130

pore content 15 porosity 193 Portland cement 189 potential yield 109 pre-delignification 163 prebored holes 206 preboring pilot holes 207 predrilled holes 208 predrilling 206 preservation 80, 166 preservative 190, 191

treatment 82 pressurised refining 162 primary cell wall 3, 35–7, 68

expansion 2, 3 primary growth 7–9 proanthocyanidins 76 process 157, 162–4

changes 166 requirements 166 runnability 172 stability 170

processing 166, 189 defects 199, 203

productivity 25, 150 profitability 170 progressive growth 137 prosenchyma cells 160 protecting compounds 78 proteins 76 protoplasmic autolysis 5, 6 provenance 11, 25 pruned surfaces 167

222 INDEX

Prunus serotina 195 Prunus spachiana 124 pseudo-transverse divisions 30 Pseudostuga menziesii 30, 31, 93–6,

103 Pseudowintera colorata 49 pulp 9, 157–62, 164–7, 169, 170, 173–5, 179–81,

183, 184 and paper 21, 23, 24, 54, 87

industries 18 brightness 83 fibre 164

morphology 175 mills 170, 172, 184 properties 167 quality 166, 169, 173, 179, 183 screen rejects 83 strength 167, 168, 183, 184 traits 184 viscosity 168 wettability 82 yield 53, 82, 83, 164, 165, 167–9, 175, 177, 179,

180, 181, 184 pulping 18, 82, 83, 105, 157, 159, 161, 163–5,

169, 179 liquors 82 process 162, 170, 173 yields 83

pulpwood 158, 166, 170 quality 87

punched metal plate fasteners 205 pyranose ring 55

quality 118, 132, 133, 147, 159, 164–7, 170, 171, 172, 173, 191, 193

assessment 8, 9 control 170–2 measurement 173

Quercus alba 17 Quercus robur 33 quinine 76 quinone methide 65

radial checks 203 cracks 147, 148 diameter 93, 109 expansion 18, 97 growth 110 shrinkage 41 splitting 150 stress 138, 147 striations 127 tension 70 walls 3

radiata pine 39, 40, 41, 158, 184 rafters 188, 209 rainfall 25 raw material quality 170

ray 1, 2, 6, 160, 201 cells 129 parenchyma 30, 71, 78, 160 tracheids 2, 31, 160

reacting sector 153 reaction maturation strain 145 reaction stress 145, 146, 150, 151, 153, 154 reaction tissues 119 reaction to light 153 reaction wood 11, 25, 53, 79–81, 118, 119, 120–5,

132–4, 144–6, 150, 151–4, 161, 191 recombinant DNA 22 reconstituted panels 166 recycled steel 190 redwood 9, 78 reeds 158 refining 164, 174 reforestation 189 reinforcement components 166 relative humidity 90 release 150 released stresses 142 residual lignin 175 residual stress 140, 142, 147, 154, 203 residual stress distribution 150 resilience 196 resin 72, 78, 98, 169, 173, 174

canals 72 streaks 201

resin-coated nails 207 resource management 23 restoration of verticality 153 reticulate perforation plates 48 rhamnose 61, 63 rigidity 137, 145, 148, 149, 189 ring boundary 32, 33 ring number 34 ring porous 14, 34, 103, 113, 208 ring width 15, 34, 96, 100, 105, 106, 107, 110, 153 ringshake 148, 150, 153 rock elm (Ulmus thomasii) 16 roguing 22 roof trusses 187 root 6, 8, 19 rotted areas 191 roundwood 169, 170, 172 rubber tree 74 rust 169

S/G ratio 77, 78S1 layer 37, 68, 127, 129 S2 layer 3, 37, 38, 43, 69, 70, 127–9 S3 layer 37, 43, 69, 127, 129 S-lignins 83 S2 microfibrils 18, 44 S type lignin 78 sagging 198 sailboat masts 188

INDEX 223

sandalwood 72 sanding 82 α- and β-santalol 72 saponins 74 sapwood 5, 6, 11, 24, 38, 53, 72, 76, 78, 79, 81, 98,

160, 167, 174, 179 sapwood/heartwood 164 sawing 133, 141, 147, 149, 150, 154, 199

pattern 153 sawmill 149, 150, 166

chips 172 sawmilling 184 scalariform perforation plate 45, 48 scanning microdensitometer 92 Scheffé’s triangle methodology 177 Scots pine (Pinus sylvestris) 91, 94, 97, 107–9, 123 screened-pulp rejects 82 screening 169, 174 seasonal climatic changes 93 secondary growth 1, 8, 9, 26 secondary metabolites 71 secondary-wall 5, 35–7, 46, 78, 125–7

deposition 19, 37 formation 2, 18, 19 lamellae 3 layers 127 thickening 97

secondary xylem 26 selection 23, 25, 154, 184 self-equilibrating stresses 124, 147 self-sizing 82 self-weight 125 SEM 3 septate axial parenchyma 6 Sequioa 30 Sequoiadendron giganteum 10 sesquiterpenoids 72, 73 shake 12, 13 shear 189, 197, 203, 208

forces 196 modulus 196 slip 7 strain 196 strength 196, 197, 203 stresses 196

sheet 157, 164, 179, 180 smoothness 181 surface 174

shoots 118 short rotation 121, 183

coppice 184 short-term loading 198 shrinkage 42, 81, 87, 133, 137, 203, 208 Siberian larch 91, 96 side grain 207 side hardness 196, 197 silica 158, 159 silver birch 103

silviculture 14, 22 Silviscan 40, 92 sinapyl alcohol 65 sinks 1 sisal 158 site 78

fertility 101 index 106

Sitka spruce 107, 158 β-sitosterol 73, 74 slenderness 152

ratio 196, 203 slicing veneer cutting 150 sloping grain 199–202, 208 sloping site 118 slow-grown conifers 9 sodium hydroxide 163 sodium sulphide 163 softwood 30, 31, 33, 34, 46, 53, 61, 63, 72,

74, 77–83, 87–9, 93, 107, 113, 142, 144, 145, 148, 150, 153, 158–61, 164–7, 173–5, 180

cell wall 89 hemicellulose 64, 89 lignin 67, 68, 69 pulps 158 resin 74 species 175 tracheids 71

soil 7, 153 nutrient levels 25 water 109

solid timber 54 solid wastes 190 solubility of lignin 82 solvent extraction 71 southern pines 103, 113, 169 spacing 107 spear point 206 specific gravity 15–17, 90, 150, 179, 180, 183,

193, 209 spike knots 199 spiral grain 6, 25, 191 splits 127, 191, 199, 201–4, 206, 208 splitting 150, 206, 208 spruce 68, 82, 93, 97, 158, 173

lignin 89 tracheids 68

spruce-type softwoods 174 stability 142, 187 stand density 106 starch 79 steam bending 196 steaming 149 steel 189, 190, 197

construction 190 fasteners 197 plates 209

224 INDEX

stem 6, 118–21, 123–5, 128–30, 134, 137–40, 142, 147, 149, 153, 161

and branch axes 7 breakage 18 dimensions 9eccentricity 129 fibres 129 form 134 inclination 146, 153 orientation 145 shape 140 straightness 11 symmetry 12

Stereum purpureum 167 sterol 73 stiffness 37, 41, 43, 87, 81, 105, 191, 194, 195, 196, 203 stilbenoid 74, 75

pinosylvin 78 stone groundwood 162

stored energy 147, 148, 153 straight-grain 192, 200 strain 131, 137, 141, 142, 144, 195

gauges 141, 142 tensor 138

strength 22, 38, 83, 87, 113, 129, 133, 140, 168, 169, 174, 180, 183, 187, 188, 190–2, 194, 196–203, 207, 208, 209

classes 200 parameters 209 properties 81 reduction 199, 200, 202

stress 7, 121, 124, 127, 131, 132, 137, 139, 141, 144, 145, 148, 151, 153, 195, 196, 198, 203

concentrations 199, 208 distribution 147 field 140 gradients 14 redistribution 147 relaxation 154 release 149 response 125 skin construction 189 skin panels 190

stringers 188 structural support 8 structural testing 22 structural timber 87, 187, 189, 197, 198, 201–3, 205 strychnine 76 sub-cellular environment 7 sucrose 76 sugar cane bagasse 158, 159 sugar maple 16 sugars 76, 79 sulphite 157, 163 sulphite process 163, 174 sulphite pulping 170 sulphurous acid 163

sunstored 168 super races 23 super trees 22 support stresses 139, 140, 147 surface checks 201, 203 surface properties 83 sustainable forestry 23 swelling 81, 87, 128, 197 sylviculture 144, 153 syringaresinol 74, 75 syringyl lignin 65, 77

tall-oil 169, 181 tamarack (Larix laricina) 197 tangential 140, 141

compression 147 growth strains 142 measurement 141 plane 138 shrinkage 41, 81 strain 141 walls 3

T-angle 39 tannin phenolic units 75 tannins 74, 76, 81, 82 taper 7, 9, 11, 12, 13 Taxales 118 Taxodium 30 Taxus baccata 36 tear 181, 183

strength 164 temperate hardwoods 34 temperate-zone conifers 14 temperate zones 6 temperature 25, 154, 197, 198, 199, 206

range 208 sum 108

tensile 121, 144, 145 and bending strength 81 maturation strain 144 maturation stress 140 strength 12, 60, 133, 140, 145, 181, 199, 202, 203

tension wood 12, 44, 80, 81, 118–24, 128–34, 140, 145, 148, 150, 153, 161, 196, 203, 208

terminal bud 145 terpenes 82 terpenoid 72, 81 tertiary wall layers 37 tetrahydrofuran 74 tetraterpenoid 73 textiles 160 thermomechanical pulp 183 thermomechanical pulping 174 thinning 16, 113

grade 106 regimes 106

INDEX 225

Thuja 30 Thuja occidentalis 13, 196 tilting 121–4 timber 42, 81, 120, 121, 133, 147, 151, 188–90, 198,

205, 208 construction 192 design 21

time-dependent deformation 198 toothed metal plates 205 torus 4, 46–8 toughness 148, 196 trace-cambium junction 20, 21 tracheary cells 46 tracheary elements 6, 20 tracheid 2, 6, 18, 19, 30, 31, 33, 34, 36,

41–3, 46, 48, 49, 79, 96, 119, 123–6, 137, 160

cell wall 35, 37 differentiation 19 differentiation factor 19 length 41 lumens 93

trans- or inter-wall failures 35 transducer 141, 142 transition zone 96 transport of water 30 transportation 190 transverse 142, 150

compression 140 elongation 140 expansion 137 shrinkage 41 tension 140, 147

traumatic resin and gum pockets 82 traumatic resin canals 78 traumatic tissues 53, 78 tree 8, 118, 120, 121

age 93 breeders 154 genetics 192 growth 1, 18, 109 improvement 11, 12, 16, 18, 22–5 physiology 34 rings 25 taper 34

tricyclic diterpenic acids 74 tripartite secondary wall 2 triterpenes 82 triterpenoids 72, 74 tritiated IAA 123 tropical forests 166 tropical hardwoods 76, 144 tropical woods 72, 74, 78 tropical zones 6 trunk form 34 truss plates 209 truss systems 190

trusses 187 Tsuga 30, 31 turgor 127 turgor pressure 3 turpentine 169, 181

Ungulina 167 uniseriate rays 30–2 unit cell 88 unstable wood 25 up-regulated proteins 128 uronic acids 61

van der Waals’ forces 57 variability 165, 167, 170, 172, 173, 175, 179 variation 8, 21, 22, 179

in density 92 of wood properties 108 taraplas 12, 18

vascular cambium 125 veneers 133 vertical alignment 118 vertical orientation 127 verticality 152 vessel 31, 32, 34, 35, 43, 45, 46, 48, 78–80, 132, 160, 164

diameter 14 element 31, 44, 45, 48

walls 129 members 6, 19, 21

vessel-less angiosperms 48, 49 vine 8 virgin fibres 159 virgin/recycled fibres 169 viscoelasticity 148, 149, 154 viscous elastic strains 148 void volume 87, 89, 193 volumetric shrinkage 150 V-shape cutting 142

wall 193 hydrolysis 48 thickness 93

Wallaba 144 warp 42, 149, 196 warty deposition 35–7 warty layer 36 waste fibres 18 water 14, 191, 198

availability 6, 106 -borne effluents 190 conduction 31, 46 displacement method 90 flow 46, 47 movement 47 -potential gradients 7 potential relations 9 transport 132

226 INDEX

waxes 76 weeping branches 124 weight density 90, 95, 192, 193 wettability of wood 82 white oaks 17 white rot 167 white spruce 9 wild-type populations 11 wind 118, 145, 161

forces 80 -inflicted stress 41 throw 168 -thrown trees 167

Winteraceae 48 wood 1, 6, 7, 8, 9, 11–13, 15, 16, 18, 19, 23, 25, 26,

161, 197 anatomy 30 cells 196, 206 chemistry 53 density 8, 12, 15, 18, 34, 38, 87, 97 durability 53 engineering 24 fibres 129 formation 1, 2, 19, 24–6 frame 190 framing 187 identification 34 joints 206 machining 81 matrix 163, 174 maturation 124 processing 76 production 109, 140 properties 4, 25, 40, 53, 97, 107 pulp 160 quality 1, 33, 34, 38, 41, 43, 87, 150

quantity 9 science 24, 25 specific gravity 11, 14 stiffness 33, 34, 70 supply 1 trusses 190 utilization 54, 80 yield 106

wood-based composites 190, 198 wood-frame construction 190 wood-pulp fibre 160 wooden ladders 188 work to maximum load 191, 196 wound wood 53 wounding 79 wounds 167

X-ray analysis 92 X-ray densitometric analysis 91, 92, 95 X-ray diffraction 38, 39 X-ray diffractometry 40 xylans 60, 63, 64 xylem 119, 125 xylogenesis 5 xyloglucan 5 xylophagous organisms 81 β-D-xylopyranosyl units 63 xylose 61, 63

yield 157, 159, 170, 179 theory 208

young trees 174 Young’s modulus 129, 133, 138, 139, 147, 194

zonation 1

wood quality and its biological basiseeb.lu.lv/ftp/grozs/augufiziologijas/augu_resursu... · wood...

Documents